EUTECNIQ Sciences

See every page on one mapArrow keys move through results · Enter opens · Esc closes

Ab initio reactivity EUTECNIQ computation · DFT Methods peer-reviewed

Our DFT calculations rank plant phenols by how easily they give up a hydrogen atom, a screening descriptor, not a health effect

The O–H bond dissociation enthalpy (BDE), the energy to remove the phenolic hydrogen as an atom, computed for plant phenols at the B3LYP/def2-TZVP//def2-SVP level, with its limits.

5of 14 queued phenols complete, parent and radical (5 O–H sites)
3.95kcal/mol: our phenol BDE (82.75) below the gas-phase reference, 86.7 ± 0.7
9.9kcal/mol: catechol's O–H weaker than phenol's here (hydrogen-bonded radical; B3LYP overstates it)
3.3hours of archived wall time for all five, on two CPU threads
Molecules

Five plant phenols optimized by density-functional theory so far, each shown as the geometry the calculation found

The computed phenols, each at its B3LYP/def2-SVP energy minimumEUTECNIQ computation

Completed

Eugenol

Formula
C₁₀H₁₂O₂
Molar mass
164.20 g/mol
PubChem CID
3314
Found in
Clove (Syzygium aromaticum)
O–H BDE
80.5 kcal/mol
ΔBDE vs phenol
−2.3 kcal/mol
HOMO
−5.63 eV
Drag or use the arrow keys to rotate. Parent molecules in the gas phase; the radicals and their spin densities are in the spin-density section.Geometries: B3LYP/def2-SVP optimization with PySCF 2.14.0 and pyberny 0.7.0 from RDKit starting structures (src/modules/abinitio/compute/bde.py); bond orders drawn from the same SMILES. Identity: PubChem, retrieved October 24, 2026. Gas-phase phenol reference: Mulder et al. 2005.
Data table: the completed phenols
The completed parent phenols: identity (PubChem) and computed O–H BDE (B3LYP/def2-TZVP//def2-SVP, kcal/mol)
MoleculeFormulaPubChem CIDMolar mass (g/mol)Found inAtomsBDEΔBDE
PhenolC₆H₆O99694.11Reference compound1382.750.00
GuaiacolC₇H₈O₂460124.14Wood smoke, guaiacum resin1781.52−1.23
CatecholC₆H₆O₂289110.11Core of many plant polyphenols1472.86−9.89
VanillinC₈H₈O₃1183152.15Vanilla planifolia pods1983.710.96
EugenolC₁₀H₁₂O₂3314164.20Clove (Syzygium aromaticum)2480.46−2.29

Status. EUTECNIQ's own calculations, run in October 2026 with open-source software; no laboratory work is involved. At this update 5 of 14 molecules (5 O–H sites) are complete. Unfinished molecules are listed under the results and will be added when their calculations finish.

  • A computed BDE≠a measured BDE
  • A gas-phase value≠the value in an oil, in water or in a cell
  • Fast hydrogen donation in a test tube≠an antioxidant benefit in people
Explanation, details and sources

Solvent screening with COSMO-SAC from our earlier DFT work is on Quantum chemistry. How in vitro, animal and human evidence differ is on Evidence and pharmacokinetics.

What is computed Established theory

A phenol stops a radical chain by handing over its O–H hydrogen, and the weaker that bond, the larger the driving force

The O–H bond strength is the cost side of the hydrogen-transfer balanceEstablished theory
Hydrogen atom transfer as a thermochemical cycle: breaking the phenol O–H bond costs BDE(ArO–H), forming R–H returns BDE(R–H), and the reaction enthalpy is their difference; proton-coupled electron transfer and sequential proton loss then electron transfer reach the same products by other pathsEnthalpy (schematic levels, not to scale)ArO–H + R•ArO• + H• + R•ArO• + R–H+ BDE(ArO–H)bond broken− BDE(R–H)bond formedΔHHAT​ = BDE(ArO–H) − BDE(R–H)Negative (downhill) when R–H is the stronger bond,as drawn herePhenol, ArO–H, gives phenoxyl, ArO•OHPhenolOPhenoxyl radical− H•HATOne step: the hydrogen moves as an atomPCETProton and electron transfer coupled in one stepSPLETProton lost first, then an electron; needs an ionizing solventWhich path dominates depends on solvent and reactants(Litwinienko and Ingold 2007).Hydrogen atom transfer as a thermochemical cycle, and three routes from phenol to phenoxylEnthalpy (schematic, not to scale)ArO–H + R•ArO• + H• + R•ArO• + R–H+ BDE(ArO–H)− BDE(R–H)ΔHHAT​ = BDE(ArO–H) − BDE(R–H)Negative (downhill) when R–H is the stronger bond.OHPhenolOPhenoxyl radicalHATHydrogen moves as an atomPCETProton and electron coupledSPLETProton first, then electron; ionizing solvent
In hydrogen atom transfer (HAT), a radical R• takes the hydrogen of a phenol ArOH; the reaction enthalpy is the bond broken minus the bond formed. Schematic enthalpy levels, not to scale. Phenols can also react by proton-coupled electron transfer (PCET) and, in ionizing solvents, by sequential proton loss then electron transfer (SPLET).Litwinienko G, Ingold KU. Acc Chem Res. 2007;40(3):222–230. Structures: PubChem CID 996 (phenol) and 123143 (phenoxy radical).
Explanation, details and sources
Hydrogen atom transfer and its enthalpyArO–H + R• → ArO• + R–H
ΔHHAT = BDE(ArO–H) − BDE(R–H)
Bond dissociation enthalpy at 298.15 K, gas phaseBDE = H(ArO•) + H(H•) − H(ArOH)
H = Eelec + Hcorr(298.15 K), H(H•) = E(H) + 52RT
Relative to phenol, and anchored to experimentΔBDE = BDE(ArO–H) − BDE(PhO–H)
BDEanchored = 86.7 kcal/mol + ΔBDE

Absolute BDEs from B3LYP are known to run low, so substituent effects are read as ΔBDE, where much of the method error cancels. The anchored value adds ΔBDE to the recommended gas-phase BDE of phenol, 86.7 ± 0.7 kcal/mol (Mulder et al. 2005). One kcal/mol is 4.184 kJ/mol.

Litwinienko G, Ingold KU. Acc Chem Res. 2007;40(3):222–230. Mulder P, et al. J Phys Chem A. 2005;109(11):2647–2655. Wright JS, Johnson ER, DiLabio GA. J Am Chem Soc. 2001;123(6):1173–1183 (DFT procedure for phenolic BDEs; concludes HAT dominates in most cases studied).

Results EUTECNIQ computation

Ring substituents shift the O–H bond strength by several kcal/mol, which is enough to change which phenol can regenerate which radical

O–H bond dissociation enthalpy relative to phenol, B3LYP/def2-TZVP//def2-SVP, gas phase, 298.15 KEUTECNIQ computation
Weakest O–H in the moleculeOther O–H sites
Our computation with PySCF 2.14.0 (src/modules/abinitio/compute/bde.py; make_data.py). Thermal corrections from harmonic frequencies for phenol and guaiacol; the mean correction of those for every other molecule (see Methods).
Show the table (5 rows): Computed O–H bond dissociation enthalpies and descriptors for every completed site
Computed O–H bond dissociation enthalpies and descriptors for every completed site
Molecule and siteBDE (kcal/mol)ΔBDE vs phenolPhenol-anchored BDESpin on OHOMO (eV)⟨S²⟩Thermal correctionWeakest O–H in molecule
Catechol †72.9−9.976.80.353−5.790.771mean of phenol and guaiacolyes
Eugenol80.5−2.384.40.377−5.630.777mean of phenol and guaiacolyes
Guaiacol81.5−1.285.50.382−5.710.778own frequenciesyes
Phenol82.80.086.70.436−6.160.788own frequenciesyes
Vanillin83.7+1.087.70.337−6.230.780mean of phenol and guaiacolyes

BDE: B3LYP/def2-TZVP electronic energies on B3LYP/def2-SVP geometries plus thermal corrections. Spin on O: Mulliken spin population of the radical oxygen, B3LYP/def2-SVP. HOMO: highest occupied orbital energy of the parent phenol, B3LYP/def2-SVP. ⟨S²⟩: spin expectation value of the radical, def2-TZVP (0.75 for a pure doublet). Anchored = 86.7 + ΔBDE.

No radical in the current set lacks an available intramolecular hydrogen bond. † Hydrogen-bonded radical conformer from compute/bde_hbond.py, lower in enthalpy than the radical bde.py generated: Catechol. Both conformers are listed in the next table.

Radical conformers of sites with an ortho OH group: in a catechol unit the radical is most stable when the remaining OH hydrogen-bonds to the radical oxygen

Lucarini, Pedulli and Guerra (2004) put the intramolecular hydrogen bond between the hydroxyl proton and the radical oxygen of the catechol semiquinone at 5.6 kcal/mol and the catechol O–H BDE at 81.8 kcal/mol. Our first radical structures sometimes kept the remaining OH pointing away; compute/bde_hbond.py rotates it toward the radical oxygen and re-optimizes. The table lists both, and the results use the lower enthalpy.

Radical conformers of sites with an ortho OH group
Molecule and siteAs generated: H···O• (Å)As generated: BDERotated OH: H···O• (Å)Rotated OH: BDEDifferenceUsed in results
Catechol3.6681.91.9672.9−9.1rotated (hydrogen-bonded)

BDE in kcal/mol at the same level of theory as above. H···O•: shortest distance from an ortho hydroxyl hydrogen to the radical oxygen; about 2 Å indicates an intramolecular hydrogen bond. Lucarini M, Pedulli GF, Guerra M. Chem Eur J. 2004;10(4):933–939 (values from the published abstract).

Queued, not yet computed at this update (9 molecules)
  • Thymol: 0 of 1 site computed
  • Carvacrol: 0 of 1 site computed
  • Sesamol: 0 of 1 site computed
  • PMHC (α-tocopherol model): 0 of 1 site computed
  • Hydroxytyrosol: 0 of 2 sites computed
  • Ferulic acid: 0 of 1 site computed
  • Caffeic acid: 0 of 2 sites computed
  • trans-Resveratrol: 0 of 2 sites computed
  • Quercetin: 0 of 5 sites computed
Explanation, details and sources

Each bar is one phenolic O–H site; negative values are weaker bonds than in phenol, and a molecule's weakest site decides its hydrogen-donating ability. As a check on direction, electron-withdrawing para groups such as CHO raise the O–H BDE in EPR measurements (Brigati et al. 2002), and vanillin comes out above phenol here.

Spin density EUTECNIQ computation

The farther the unpaired electron spreads onto the ring, the more stable the radical and the weaker the O–H bond

BDE (kcal/mol)82.8
ΔBDE vs phenol0.0
Spin on the radical O0.436
Spin on ring and chain carbons0.600
⟨S²⟩ (ideal 0.75)0.788
H···O• distance (Å)none

Phenoxyl radical: the spin is shared between the oxygen and the ortho and para ring carbons.

Phenoxyl radical: Mulliken spin populationsB3LYP/def2-SVP

Drag to rotate. Halo size and color show the spin population on each atom: green positive (excess α spin), rose negative.

Geometry: B3LYP/def2-SVP optimum of the radical. Spin populations: Mulliken analysis of the unrestricted Kohn–Sham density (Mulliken 1955). Mulliken values depend on the basis set; read them as a qualitative map.
Per-atom spin populations of the selected radical
AtomElementSpin population
1O0.436
Explanation, details and sources

Mulliken spin populations show where the unpaired electron sits: in phenoxyl, mostly on the oxygen and the ortho and para ring carbons. Substituents that share the spin, or hydrogen-bond to the radical oxygen, stabilize the radical.

Descriptors EUTECNIQ computation

Two cheap descriptors, spin on the radical oxygen and the parent's HOMO energy, do not track bond strength across five sites

Across five sites neither descriptor tracks ΔBDE closely; catechol's hydrogen-bonded radical sits far below bothEUTECNIQ computation
Two scatter plots of the five completed O–H sites: ΔBDE relative to phenol against the parent's HOMO energy and against spin on the radical oxygen. Catechol, with its hydrogen-bonded radical, sits far below the others in both; with five points neither descriptor tracks ΔBDE closely.Against the parent's HOMO energy−10−8−6−4−20+2−6.40−6.20−6.00−5.80−5.60HOMO energy of the parent phenol, eVΔBDE vs phenol, kcal/molPhenol: −6.16; ΔBDE +0.0 kcal/molPhenolGuaiacol: −5.71; ΔBDE −1.2 kcal/molGuaiacolCatechol: −5.79; ΔBDE −9.9 kcal/molCatecholVanillin: −6.23; ΔBDE +1.0 kcal/molVanillinEugenol: −5.63; ΔBDE −2.3 kcal/molEugenolPearson r = −0.49, n = 5Against spin left on the radical oxygen−10−8−6−4−20+20.320.360.400.44Mulliken spin on the radical oxygenΔBDE vs phenol, kcal/molPhenol: 0.44; ΔBDE +0.0 kcal/molPhenolGuaiacol: 0.38; ΔBDE −1.2 kcal/molGuaiacolCatechol: 0.35; ΔBDE −9.9 kcal/molCatecholVanillin: 0.34; ΔBDE +1.0 kcal/molVanillinEugenol: 0.38; ΔBDE −2.3 kcal/molEugenolPearson r = +0.29, n = 5Two scatter plots of the five completed O–H sites: ΔBDE against HOMO energy and against spin on the radical oxygen.Against the parent's HOMO energy−10−8−6−4−20+2−6.4−6.0−5.6HOMO energy, eVΔBDE vs phenol, kcal/molPhenol: −6.2; ΔBDE +0.0 kcal/molPhenolGuaiacol: −5.7; ΔBDE −1.2 kcal/molGuaiacolCatechol: −5.8; ΔBDE −9.9 kcal/molCatecholVanillin: −6.2; ΔBDE +1.0 kcal/molVanillinEugenol: −5.6; ΔBDE −2.3 kcal/molEugenolPearson r = −0.49, n = 5Against spin on the radical oxygen−10−8−6−4−20+20.320.380.44Mulliken spin on the radical OΔBDE vs phenol, kcal/molPhenol: 0.44; ΔBDE +0.0 kcal/molPhenolGuaiacol: 0.38; ΔBDE −1.2 kcal/molGuaiacolCatechol: 0.35; ΔBDE −9.9 kcal/molCatecholVanillin: 0.34; ΔBDE +1.0 kcal/molVanillinEugenol: 0.38; ΔBDE −2.3 kcal/molEugenolPearson r = +0.29, n = 5
Same calculations as the results table: HOMO of the parent (B3LYP/def2-SVP), Mulliken spin on the radical oxygen (B3LYP/def2-SVP) and ΔBDE (B3LYP/def2-TZVP//def2-SVP). Correlations among descriptors computed at one level of theory describe this set of molecules only; they are not a validated prediction model.
Data table: descriptors for every completed site
HOMO energy, spin on the radical oxygen and ΔBDE (EUTECNIQ computation)
PhenolHOMO (eV)Spin on radical OΔBDE (kcal/mol)
Phenol−6.160.4360.0
Guaiacol−5.710.382−1.2
Catechol−5.790.353−9.9
Vanillin−6.230.337+1.0
Eugenol−5.630.377−2.3
Explanation, details and sources

Three numbers from the same calculations: the HOMO energy of the parent (a common proxy for electron donation), the spin left on the radical oxygen, and ΔBDE. Correlations across 5 sites (shown once at least three sites are complete): ΔBDE with spin on O, r = 0.29; ΔBDE with HOMO energy, r = −0.49 (Pearson). With fewer than about ten sites these coefficients are descriptive only.

HAT calculator Illustration

Comparing two computed BDEs predicts which way a hydrogen would move between them, with most of the method error cancelling

Reaction enthalpyArO–H + Y• → ArO• + Y–H
ΔH = BDE(ArO–H) − BDE(Y–H)
Donor BDE used (kcal/mol)82.8
Acceptor BDE used (kcal/mol)82.8
ΔH (kcal/mol)0.0

Thermoneutral: the two O–H bonds are equally strong at this level of theory.

Reaction enthalpy for every donor against the selected acceptorModel output
Phenoxyl acceptors use the computed BDEs on both sides, so the B3LYP offset cancels. For an entered Y–H value the donor uses the phenol-anchored BDE (86.7 + ΔBDE), so both numbers sit on an experimental scale. Enthalpy only; entropy, solvent and kinetics are not included.
Method, assumptions and limits: HAT calculator
What it calculates
The gas-phase reaction enthalpy for moving a hydrogen atom from a phenolic O–H site to a radical.
Equations
ΔH = BDE(donor O–H) − BDE(acceptor Y–H). With a phenoxyl acceptor from this set, both BDEs are the computed B3LYP values. With an entered Y–H value, the donor BDE is anchored: 86.7 kcal/mol + ΔBDE(donor).
Variables and units
kcal/mol at 298.15 K; Y–H BDE 60–120 kcal/mol.
Assumptions
Entropy changes of phenol–phenoxyl exchange are small; Lucarini et al. (2004) measured ΔS° = +1.3 ± 0.2 cal mol⁻¹ K⁻¹ for one such exchange. Gas phase; one conformer per species.
Valid range
Hydrogen atom transfer between neutral species. Ionizing solvents can open the proton-loss pathway, which this calculation does not describe.
Does not establish
A rate constant, an antioxidant capacity or any biological effect. Rates also depend on steric hindrance, solvent hydrogen bonding and polar effects (Litwinienko and Ingold 2007).
References
Amorati R, et al. J Org Chem. 2002;67(26):9295–9303. Lucarini M, Pedulli GF, Guerra M. Chem Eur J. 2004;10(4):933–939. Litwinienko and Ingold 2007. Mulder et al. 2005.
Verification
tests/numerics/ref/insights-ab-initio-reactivity.py recomputes every BDE from the cached energies with its own code and checks the table, the viewer and calculator readouts.
Who can hand a hydrogen to whom: reaction enthalpy for every computed phenol (row) against every computed phenoxyl radical (column)Model output
ΔH = BDE(row O–H) − BDE(column O–H), both at B3LYP/def2-TZVP//def2-SVP, so the method's offset cancels. Negative (copper): the transfer releases heat. Enthalpy only; rates, entropy and solvent are not included. Select a column heading to sort.BDEs: the results table above (catechol with its hydrogen-bonded radical). Computed by us (src/modules/comp_abi/compute/make_data.py).
Data table: ΔH for every donor and acceptor (kcal/mol)
ΔH (kcal/mol) for ArO–H (row) + Ar′O• (column) → ArO• + Ar′O–H; negative = exothermic
DonorPhenol•Guaiacol•Catechol•Vanillin•Eugenol•
Phenol0.001.239.89−0.962.29
Guaiacol−1.230.008.66−2.191.06
Catechol−9.89−8.660.00−10.86−7.60
Vanillin0.962.1910.860.003.26
Eugenol−2.29−1.067.60−3.260.00

At this level of theory hydrogen transfer from catechol is exothermic toward every other phenoxyl here, and from vanillin toward none, since its para aldehyde strengthens the O–H.

Explanation, details and sources

A phenol can reduce another phenol's radical when its own O–H bond is weaker. Amorati et al. (2002) state that reversible recycling of α-tocopherol is observed only when the co-antioxidant's BDE is lower than, or close to, that of α-tocopherol. For a radical outside this set, enter the strength of the bond it forms.

Comparison Published values

Our absolute BDEs run about 4 kcal/mol below experiment, as B3LYP is known to, so ΔBDE is the usable quantity

Phenol sits a constant offset below the reference values; the hydrogen-bonded catechol radical sits well below that offsetEUTECNIQ computation Published values
Parity plot: this work's B3LYP O–H BDEs against published reference values for phenol and catechol; phenol sits 3.9 kcal/mol below the line of equality, catechol about 9 kcal/mol below7075808590Reference value (kcal/mol)7075808590This work, B3LYP (kcal/mol)equal3.9 kcal/mol belowPhenol: this work 82.8; G3, CBS-APNO, CBS-QB3: 86.4–87.7Catechol: this work 72.9; CCSD, 81.6Catechol: this work 72.9; Earlier DFT, 77.6Phenol: this work 82.8; Recommended gas-phase BDE, 86.7 ± 0.7Catechol: this work 72.9; EPR-derived, 81.8PhenolCatecholRelative to phenol, ΔBDE (kcal/mol)−12−10−8−6−4−20+2CatecholReference estimate about −5 to −6 (CCSD catechol minus composite phenol range)Catechol: ΔBDE -9.9 kcal/mol (this work)GuaiacolGuaiacol: ΔBDE -1.2 kcal/mol (this work)EugenolEugenol: ΔBDE -2.3 kcal/mol (this work)VanillinVanillin: ΔBDE +1.0 kcal/mol (this work)Only catechol has a reference ΔBDE; ours is too negative byseveral kcal/mol, the B3LYP overstabilization of itshydrogen-bonded radical.Experiment-based reference (bar: ± 0.7)High-level theory (band: composite range; CCSD)Earlier DFT, judged in poor agreementThis work, ΔBDEReference ΔBDE estimate
Only values read in the primary papers are used; they come from different phases and methods, and a solution-phase value is not directly a gas-phase value. The catechol reference ΔBDE (about −5 to −6) is our subtraction of two papers' values, so it is approximate. Molecules without a verified reference value appear only in the right panel. PMHC is pending; its reference, 78.2 kcal/mol, is for α-tocopherol in solution.Mulder et al. 2005, J Phys Chem A 109:2647–2655; Lucarini, Pedulli, Guerra 2004, Chem Eur J 10:933–939.

Catechol: ΔBDE −9.9 kcal/mol here; rotating the remaining OH away raises the BDE by 9.1. Read catechol ΔBDE values on this page as too negative by several kcal/mol.

Text version: reference values and their sources
CompoundThis work, B3LYP (kcal/mol)Reference value (kcal/mol)What the reference isSource
Phenol82.886.7 ± 0.7Recommended gas-phase BDE from thermolysis and three corrected liquid-phase methodsMulder et al. 2005
Phenol82.886.4–87.7G3, CBS-APNO and CBS-QB3 composite calculationsMulder et al. 2005
Catechol72.981.8Derived from EPR radical-equilibration data by group additivityLucarini, Pedulli, Guerra 2004
Catechol72.981.6CCSD with diffuse functions on oxygenLucarini, Pedulli, Guerra 2004
Catechol72.977.6Earlier DFT values in the literature, as cited by the same authors, who judged them in poor agreementLucarini, Pedulli, Guerra 2004
Catechol, ΔBDE vs phenol−9.9about −5 to −6Our subtraction: CCSD catechol value (81.6) minus the composite phenol range (86.4–87.7); two papers and two methods, so approximateLucarini, Pedulli, Guerra 2004; Mulder et al. 2005
PMHC (α-tocopherol head group)pending78.2α-Tocopherol itself, EPR radical equilibration in solution; PMHC carries a methyl group where α-tocopherol carries its phytyl chainLucarini, Pedulli, Guerra 2004

Mulder P, Korth HG, Pratt DA, DiLabio GA, Valgimigli L, Pedulli GF, Ingold KU. Critical re-evaluation of the O–H bond dissociation enthalpy in phenol. J Phys Chem A. 2005;109(11):2647–2655. Lucarini M, Pedulli GF, Guerra M. A critical evaluation of the factors determining the effect of intramolecular hydrogen bonding on the O–H bond dissociation enthalpy of catechol and of flavonoid antioxidants. Chem Eur J. 2004;10(4):933–939. Mulder et al. also note that the calibration standard of the EPR equilibration method had to be revised, so older solution values carry that uncertainty. We did not find primary experimental or high-level values for the other molecules that we could verify, so they are shown without a reference.

For catechol, our hydrogen-bonded radical gives ΔBDE = −9.9 kcal/mol, and rotating the remaining OH away raises the BDE by 9.1 kcal/mol. Both are larger in magnitude than the estimates of Lucarini et al. (2004), who derived a 5.6 kcal/mol semiquinone hydrogen bond and a catechol BDE of 81.8 kcal/mol, and who judged earlier DFT values (9.5 and 77.6 kcal/mol) to be in poor agreement. B3LYP here shows the same overstabilization, so catechol ΔBDE values on this page should be read as too negative by several kcal/mol.

Limits Read before use

Six simplifications bound what these numbers can support

Three error sources are sized on this page; solvent, kinetics and health effects lie outside the calculationLimits
Six simplifications ordered from the computed number outward: functional, thermal corrections and conformer error are quantified here; solvent, kinetics and health effects are outside what the calculation coversFrom the computed number outwardSize of the effect on this pageFunctionalB3LYP, no dispersion correctionphenol 3.9 kcal/mol lowThermal correctionsFrequencies for phenol and guaiacol only; their mean used elsewherespread 0.11 kcal/molOne conformerLowest of 24 MMFF94 conformers, optimized oncecatechol OH rotation 9.1 kcal/molGas phaseNo solvent: solvents change rate and mechanismnot modeledOne descriptorRate also needs steric access, polar effects, ionization potential, aciditynot computedNot a health effectNCCIH: evidence does not support antioxidant supplements to prevent cancer orcardiovascular diseaseoutside scopeQuantified hereNot modeledOutside what any BDE can showSix simplifications, from the computed number outward, with the size of each effect1Functional: phenol 3.9 kcal/mol lowB3LYP, no dispersion correction2Thermal corrections: spread 0.11 kcal/molFrequencies for phenol and guaiacol only; theirmean used elsewhere3One conformer: catechol OH rotation 9.1kcal/molLowest of 24 MMFF94 conformers, optimized once4Gas phase: not modeledNo solvent: solvents change rate and mechanism5One descriptor: not computedRate also needs steric access, polar effects,ionization potential, acidity6Not a health effect: outside scopeNCCIH: evidence does not support antioxidantsupplements to prevent cancer or cardiovasculardisease
Sizes from this page's results and methods. Litwinienko and Ingold 2007 (solvent effects). NCCIH, Antioxidant supplements: what you need to know (last updated July 2023).
Text version of the six limits

Gas phase

No solvent was modeled. The hydrogen-bond-accepting and anion-solvating abilities of a solvent strongly change how fast phenols react with radicals, and by which mechanism (Litwinienko and Ingold 2007).

One conformer

Each structure starts from the lowest-energy MMFF94 conformer of 24 and is optimized once. Catechols are sensitive to which OH points at the radical oxygen; the H···O• column of the viewer shows the arrangement each radical reached.

Thermal corrections

Frequencies were computed for phenol and guaiacol only, to keep the run within two shared processor cores. Every other molecule uses their mean thermal contribution, −6.91 kcal/mol; the spread between them, 0.11 kcal/mol, is a rough indication of the error this adds.

Functional

B3LYP without a dispersion correction underestimates the absolute O–H BDE of phenol here by about 4 kcal/mol. Relative values are more reliable than absolute ones.

One descriptor

BDE is a thermodynamic quantity. Antioxidant rate also depends on steric access, polar effects, ionization potential and acidity, and on what the radical does next.

Not a health effect

Radical scavenging measured or computed outside the body is not evidence of benefit. NCCIH states that current evidence does not support antioxidant supplements for preventing cancer or cardiovascular disease.

NCCIH. Antioxidant supplements: what you need to know. Last updated July 2023. https://www.nccih.nih.gov/health/antioxidant-supplements-what-you-need-to-know

Methods Reproducible

Every number on this page can be regenerated from the archived scripts and cached outputs

What the five molecules cost: archived wall time by calculation stepRun record
Parent and radical structures together; harmonic frequencies were run for phenol and guaiacol only, which is why those two bars are longer for their size. Two CPU threads on a cloud machine; times vary with machine load.Cache records in src/modules/abinitio/compute/cache/ (t_opt_s, t_hess_s, t_sp_s).
Data table: wall time per molecule and step
Archived wall time per molecule (parent and radical structures; two CPU threads)
MoleculeStructuresOptimization (min)Frequencies (min)def2-TZVP single point (min)Total (h)
Phenol26.014.31.50.36
Guaiacol212.447.43.51.05
Catechol315.10.04.00.32
Vanillin222.50.05.30.46
Eugenol255.40.08.81.07
Show the table (8 rows)
StepSettingReference
Starting structuresSMILES to 3D with RDKit 2026.09.1: 24 embedded conformers (seed 7), MMFF94 minimization, lowest-energy conformer keptRiniker and Landrum 2015 (embedding); Halgren 1996 (MMFF94)
GeometryB3LYP/def2-SVP, density fitting with the def2-universal-jkfit auxiliary basis, integration grid level 3, pyberny 0.7.0 optimizer; restricted Kohn–Sham for phenols, unrestricted for radicalsBecke 1993; Lee, Yang, Parr 1988; Weigend and Ahlrichs 2005; Weigend 2008
RadicalsThe phenolic H is removed from the optimized parent and the radical is re-optimized; symmetry-equivalent sites are computed oncebde.py
EnergiesB3LYP/def2-TZVP single points on the def2-SVP geometries; H atom at the same levelWeigend and Ahlrichs 2005
Thermal correctionsHarmonic frequencies and ideal-gas, rigid-rotor, harmonic-oscillator enthalpy at 298.15 K for phenol and guaiacol (parents and radicals). Thermal contribution to the BDE, Hcorr(radical) − Hcorr(parent) + 5/2 RT: phenol −6.97, guaiacol −6.86 kcal/mol. Their mean, −6.91 kcal/mol (range 0.11 kcal/mol over 2 sites), is applied to every other moleculePySCF thermo module
Spin and orbitalsMulliken spin populations of the radicals and HOMO energies of the parents at B3LYP/def2-SVP; ⟨S²⟩ reported for both basis setsMulliken 1955
SoftwarePySCF 2.14.0 on Python 3.13; constants: 1 hartree = 627.5094740631 kcal/mol, R = 1.987204259 × 10⁻³ kcal mol⁻¹ K⁻¹Sun et al. 2020 (PySCF)
Filessrc/modules/abinitio/compute/bde.py (calculations), cache/*.json (one record per structure), make_data.py (this page's data and tables)Archived with the site source
Run record: convergence, imaginary frequencies and wall time per structure
StructureSCF converged (SVP, TZVP)Imaginary frequencies⟨S²⟩ SVP / TZVPWall time (s)
Phenol (parent)yes, yes0closed shell486
Phenol, radical OHyes, yes00.791 / 0.788824
Guaiacol (parent)yes, yes0closed shell1,320
Guaiacol, radical OHyes, yes00.780 / 0.7782,477
Catechol (parent)yes, yesnot computedclosed shell241
Catechol, radical OH (either; equivalent)yes, yesnot computed0.781 / 0.779504
Vanillin (parent)yes, yesnot computedclosed shell520
Vanillin, radical OHyes, yesnot computed0.782 / 0.7801,151
Eugenol (parent)yes, yesnot computedclosed shell1,157
Eugenol, radical OHyes, yesnot computed0.779 / 0.7772,698

Need a computational screen of phenolic constituents?

Partners can propose a set of structures and a question. EUTECNIQ would state the method, its expected error and what the result can and cannot support before any calculation.

Discuss a screenCOSMO-SAC screening