Thermochemistry Lab (beta) | Casey D. Foley, PhD

How to cite

ATcT's table attributes its values to the two method papers [1, 2] and to the network version used [3]:

Cite this tool only for the arithmetic (balancing, unit conversion, channel search); the thermochemistry is ATcT's.

Covariances

Correlated uncertainties use the ATcT covariance matrix of Δ_fH°(298.15 K) for 1874 gas-phase species, fetched from the ATcT API (/api/v1/covariance/species/) on 2026-11-06. The API serves TN 1.220, not 1.228. The page keeps only the pairs that change the variance of a sum or difference of two species by at least 0.1 % (326 497 of 1 732 591), which changes two-body fragmentation uncertainties by 0.04 % (median) and 0.24 % (99th percentile). Each correlation coefficient is stored with enough digits that its rounding changes no pair combination by more than 0.5 %, including strongly correlated pairs such as Br₂ and Br, where 1 − r ≈ 2 × 10⁻⁶ sets the uncertainty of D₀(Br₂). The tool uses the TN 1.220 covariances as served, except that a species whose printed uncertainty changed between TN 1.220 and TN 1.228 has its row and column scaled to the TN 1.228 uncertainty; both values are shown. The file (2.1 MB, about 0.4 MB compressed) is loaded only when a reaction, channel list, diagram or the self-test needs it.

Excited electronic states

106 term values of low-lying states of 10 atoms, 16 diatomics and 4 polyatomics, retrieved on 2026-11-06. Where ATcT lists a state (14 cases, such as O(¹D), O₂(a¹Δ_g), CO(a³Π), ¹CH₂ and the halogen ²P½ levels), the ATcT entry is used instead, so no state is counted twice. States given only as estimates (SO a¹Δ, BrO and IO X₂) and species without an ATcT ground state (P, NS) are left out.

A. Kramida, Yu. Ralchenko, J. Reader and NIST ASD Team (2024). NIST Atomic Spectra Database (ver. 5.12), physics.nist.gov/asd [2026, November 6]. NIST, Gaithersburg, MD. doi:10.18434/T4W30F (atomic levels).
K. P. Huber and G. Herzberg, “Constants of Diatomic Molecules” (data prepared by J. W. Gallagher and R. D. Johnson, III), in NIST Chemistry WebBook, NIST Standard Reference Database Number 69, eds. P. J. Linstrom and W. G. Mallard, NIST, Gaithersburg, MD, doi:10.18434/T4D303 (retrieved November 6, 2026); print source: K. P. Huber and G. Herzberg, Molecular Spectra and Molecular Structure IV. Constants of Diatomic Molecules (Van Nostrand Reinhold, 1979), data through May 1977 (diatomic T₀).
M. E. Jacox, “Vibrational and Electronic Energy Levels of Polyatomic Transient Molecules,” in NIST Chemistry WebBook, NIST Standard Reference Database Number 69, eds. P. J. Linstrom and W. G. Mallard, NIST, Gaithersburg, MD, doi:10.18434/T4D303 (retrieved November 6, 2026) (CH₂, HCO, HO₂, NH₂).

T₀ for molecules is the Huber–Herzberg (0,0) band origin to the ground state where one is listed, otherwise chained through a lower state or computed as Tₑ + G′(0) − G″(0). Huber and Herzberg give some origins from the zero point of the Hill–Van Vleck expression rather than from the lowest level (their notes on OH, SH, CH and CN); these are converted to lowest-level T₀ with the spin-orbit and rotational constants on the same pages (OH A²Σ⁺ +38.18 cm⁻¹, SH A²Σ⁺ +170.23 cm⁻¹, CH A²Δ +28.27 cm⁻¹, CN A²Π −23.06 cm⁻¹). Where the value refers to a multiplet centre or to a hypothetical J = 0 level instead, it is not converted; the spread of the multiplet (|A|ΛS, up to 146 cm⁻¹ for O₂ A′³Δ_u) or the offset (about 1.3 cm⁻¹ for NO) is added to the uncertainty as an allowance. Rows marked “note” say which. Huber and Herzberg stop at 1977.

Previous version (TN 1.220)

The species card shows the TN 1.220 values of the 2352 species whose value or uncertainty differs from TN 1.228, from the ATcT API bulk endpoint /api/v1/all/ (downloaded 2026-11-06); 13 species are new in TN 1.228.

Use your own table

Load a CSV with the columns of the cleaned ATcT table (atct_id, formula, state, charge, composition, dHf_0K_kJmol, dHf_298K_kJmol, unc_95_kJmol…), for example a newer ATcT version. It stays in your browser. atct_id, formula and dHf_298K_kJmol are required. Composition and charge are read from their columns, or parsed from the formula where those are missing; rows where neither works are skipped and counted, so the balance check never runs without a composition. The ATcT covariances and the version diff are not applied to a CSV; the excited states are, by ATcT ID.

Load CSV…

How it is calculated

Every result is a sum of ATcT enthalpies of formation, Δ_rH° = Σ ν_i Δ_fH°_i, with ν positive for products. The tool checks that elements and charge balance before it shows a number; electrons (e-) carry Δ_fH° = 0.

Bond dissociation energy D₀(A–B) = Δ_fH°₀(A) + Δ_fH°₀(B) − Δ_fH°₀(AB); D₂₉₈ uses the 298.15 K values.
Photodissociation threshold λ = N_Ahc / D₀ = 119 626.6 nm·kJ mol⁻¹ / D₀. It is the one-photon limit for a molecule in its lowest rovibrational level with no barrier above the products. Parent internal energy lowers the real onset; a barrier (common for molecular channels such as H₂CO → H₂ + CO) raises it.
Available energy at wavelength λ is E_hν − D₀, shared among product translation, rotation, vibration and electronic excitation.
Ionization and appearance energies are 0 K differences: IE = Δ_fH°₀(M⁺) − Δ_fH°₀(M) and AE(A⁺) = Δ_fH°₀(A⁺) + Δ_fH°₀(B) − Δ_fH°₀(AB). ATcT's 298.15 K ion values follow the stationary-electron convention, in which the electron carries no enthalpy at any temperature; to treat free electrons as a thermal gas, add 5/2 RT = 6.197 kJ/mol per electron produced at 298.15 K.
Proton affinity PA(M) = Δ_fH°₂₉₈(M) + Δ_fH°₂₉₈(H⁺) − Δ_fH°₂₉₈(MH⁺).
Excited products. A fragment in an excited electronic state X* has Δ_fH°₀(X*) = Δ_fH°₀(X) + T₀(X*), with T₀ the term value above the lowest level from NIST (atoms: ASD; diatomics: Huber and Herzberg; polyatomics: Jacox). Where ATcT lists the state itself its own value is used. A channel's excited thresholds are D₀ plus the sum of the products' T₀; no 298.15 K values are given for these. The uncertainty of such a state adds the term value's own uncertainty, where the source gives one (Jacox gives ±0.05 to ±10 cm⁻¹), and an allowance where the source refers the state to a level other than the lowest, to the ATcT uncertainty of the ground state.
Electron channels. Dissociative electron attachment AB + e⁻ → A + B⁻ has threshold Δ_fH°₀(A) + Δ_fH°₀(B⁻) − Δ_fH°₀(AB) = D₀(A–B) − EA(B); ion pairs A⁺ + B⁻ have D₀ + IE(A) − EA(B), for photons and for electron impact alike.
Energy diagram. Product levels are the 0 K reaction enthalpies from the reactants; your levels are drawn at the energies you enter, converted from their unit.

Uncertainty

The ± values are 95 % limits. ATcT values are correlated because they come from one network: when a bond energy was itself measured (H–OH, H–HCO), the parent and the fragment move together and the true uncertainty of the difference is much smaller than the independent-error estimate √Σ(ν_iU_i)². When every species of a reaction is in ATcT's covariance file (1874 gas-phase species), the tool uses √(νᵀCν) with ATcT's covariance matrix C from TN 1.220 (the version the ATcT API serves, covariance of Δ_fH° at 298.15 K). For a species whose printed uncertainty changed in TN 1.228, its row and column are scaled by U_1.228/U_1.220, so that its own uncertainty is the one shown on this page while its correlations stay those of TN 1.220. For D₀(H–HCO), where both H₂CO and HCO changed, that gives ±0.0066 kJ/mol instead of ±0.110 (±0.0080 with the TN 1.220 covariances as served). If the scaling would leave less than 0.1 % of the sum of the terms (strong cancellation, as in X₂ → 2 X), the value as served is used; a result that cancels below the precision of the stored correlations is shown as a bound (“< …”), never as 0. Otherwise the tool falls back to independent errors and says so. The Reaction tab shows all three numbers and which one is used; species whose printed uncertainty changed by more than ×1.5 between versions are flagged.

Typical values

Bond energies of small molecules span about 100 to 1100 kJ/mol (1 to 11 eV), so one-photon thresholds run from the near infrared to the vacuum ultraviolet. All values below are D₀ at 0 K to ground-state products, computed by this tool from ATcT TN 1.228 [3]; λ max = 119 626.57 nm·kJ mol⁻¹ / D₀ (CODATA 2022 [6]).

Reaction	D₀ or IE, kJ/mol	eV	λ max, nm
O₃ → O₂ + O	102.434	1.0617	1167.8
Cl₂ → 2 Cl	239.242	2.4796	500.0
NO₂ → NO + O	300.604	3.1155	398.0
H₂CO → H + HCO	362.804	3.7602	329.7
H₂ → 2 H	432.068	4.4781	276.9
CH₄ → CH₃ + H	432.457	4.4821	276.6
H₂O → H + OH	492.214	5.1014	243.0
O₂ → 2 O	493.686	5.1167	242.3
N₂ → 2 N	941.310	9.7560	127.1
CO → C + O	1072.016	11.1107	111.6
H₂O → H₂O⁺ + e⁻ (adiabatic IE)	1217.398	12.6174	98.3

Real photolysis onsets are often well to the blue of λ max, because the bright or spin-allowed transition leads to excited products: ozone's main ultraviolet channel gives O(¹D) + O₂(a¹Δ_g), not ground-state O + O₂.

Worked example

The default reaction, CH2O -> H + HCO in kJ/mol, uses the TN 1.228 values Δ_fH°₀ = 216.034 (H), 41.384 ± 0.078 (HCO) and −105.386 ± 0.078 (H₂CO) kJ/mol [3].

D₀ = 216.034 + 41.384 − (−105.386) = 362.804 kJ/mol.
In other units: 362.804 / 96.485 332 = 3.7602 eV, and 362.804 / 0.011 962 657 = 30 328.05 cm⁻¹.
Threshold: λ = 119 626.57 / 362.804 = 329.73 nm.
Independent-error uncertainty: √(0.0005² + 0.078² + 0.078²) = ±0.110 kJ/mol (H is printed as ± 0.000, taken as 0.0005). With the ATcT correlations the uncertainty is ±0.0066 kJ/mol, because H₂CO and HCO are tied by the measured bond energy.
At 298.15 K: 217.998 + 41.758 − (−109.226) = 368.982 kJ/mol.

The Reaction tab shows the same numbers, and the Self-test tab checks them against Terentis and Kable's measured 30 328.5 ± 0.5 cm⁻¹, which ATcT uses as an input.

Assumptions and limits

Only species in ATcT TN 1.228 (3455 entries, about 2700 in the gas phase). Products the network does not contain are not found.
One uncertainty per species as printed in the ATcT table, applied at 0 K and 298.15 K. ATcT prints each value to the decimals of its uncertainty (at most 0.001 kJ/mol), so every printed value also carries up to half a last digit of rounding. Both the independent and the correlated uncertainties include that rounding (½·10^−d per species, combined in quadrature); it is what limits differences where the covariances cancel, such as Br₂ → 2 Br. For independent errors "± 0.000" is taken as 0.0005.
Thresholds are thermochemical: no barriers, no tunnelling, no spin or symmetry restrictions, no hot bands.
Channel search uses gas-phase entries in their listed ground states unless you include ATcT's other states and conformers (for example O(g, singlet) = O(¹D) and CH₂(g, singlet) = ¹CH₂). Ortho and para nuclear-spin entries are left out.
The covariances are from TN 1.220, the version the ATcT API served on 2026-11-06, while the values are TN 1.228; 2352 species changed value or uncertainty between the two. They are for 298.15 K and are applied to 0 K results as well. Pairs too weak to matter were dropped, so a variance can come out slightly negative; it is then shown as an upper bound, not as 0.
Excited states cover the low-lying states of 30 atoms and small molecules only. Ion fragments are always in their ground state. Huber and Herzberg's diatomic data stop in 1977. Where their reference level is not the lowest level and could not be converted (multiplet centres, hypothetical J = 0 levels), the offset, up to 146 cm⁻¹, is carried as an allowance in the uncertainty rather than corrected. Spin, symmetry and Franck–Condon factors decide whether a channel is actually reached; thresholds say only that it is energetically open.
The energy diagram draws your intermediates and transition states as entered; it does not check them.
A plain formula such as C2H4 is matched by composition and charge. When several isomers share it, the tool takes the gas-phase entry ATcT lists first (here ethylene, not ethylidene) and offers the others; check the species name in the result.

References

Primary sources and standards

B. Ruscic et al., “Introduction to Active Thermochemical Tables: Several ‘Key’ Enthalpies of Formation Revisited,” J. Phys. Chem. A 108, 9979–9997 (2004), doi:10.1021/jp047912y (method).
B. Ruscic et al., “Active Thermochemical Tables: Thermochemistry for the 21st Century,” J. Phys. Conf. Ser. 16, 561–570 (2005), doi:10.1088/1742-6596/16/1/078 (method).
B. Ruscic and D. H. Bross, “Active Thermochemical Tables (ATcT) values based on ver. 1.228 of the Thermochemical Network,” Argonne National Laboratory (2026), atct.anl.gov (all values).
B. Ruscic, “Uncertainty Quantification in Thermochemistry, Benchmarking Electronic Structure Computations, and Active Thermochemical Tables,” Int. J. Quantum Chem. 114, 1097–1101 (2014), doi:10.1002/qua.24605 (95 % uncertainties).
B. Ruscic, “Active Thermochemical Tables: Sequential Bond Dissociation Enthalpies of Methane, Ethane, and Methanol and the Related Thermochemistry,” J. Phys. Chem. A 119, 7810–7837 (2015), doi:10.1021/acs.jpca.5b01346 (D₀ and BDE₂₉₈ definitions).
P. J. Mohr, D. B. Newell, B. N. Taylor, and E. Tiesinga, “CODATA recommended values of the fundamental physical constants: 2022,” Rev. Mod. Phys. 97, 025002 (2025), doi:10.1103/RevModPhys.97.025002 (unit factors; all but the hartree are exact).
A. Kramida, Yu. Ralchenko, J. Reader and NIST ASD Team, NIST Atomic Spectra Database (ver. 5.12), NIST (2024), doi:10.18434/T4W30F (atomic term values).
K. P. Huber and G. Herzberg, “Constants of Diatomic Molecules,” and M. E. Jacox, “Vibrational and Electronic Energy Levels of Polyatomic Transient Molecules,” in NIST Chemistry WebBook, SRD 69, doi:10.18434/T4D303 (molecular term values).

Free introductions

Argonne, Active Thermochemical Tables home page (what a thermochemical network is).
Spectroscopic units on this site (cm⁻¹, eV, kJ/mol and wavelength).

Thermochemistry Lab BETA

Diagram

Data source