Redox-active molecule discovery

Know the energy.
Know the molecule.

Explore redox-active molecules, predict energies, and understand the landscape of your chemistry.

Powered by AIMNet2. Grounded in Chemistry.

2,500,000+Molecules screened, with quantum chemical accuracy
AIMNet2-NSE-CPCMMachine learned interatomic potential powered chemistry
One, UnifiedWorkspace for calculations, visualizations and exploration

About

Project EQUUS (Effective QUest into boUnded chemical Space) is an initiative started by Garcia-Bosch and Isayev labs at the Chemistry department in Carnegie Mellon University. We develop experimentally-validated, machine-learning based methods to study redox-active molecules. We explore the energetics of large chemical spaces to guide molecular design decisions. We are a highly interdisciplinary team trying to understand Chemistry better.

We are interested in the thermochemistry (free energies, bond strengths, reduction potentials, pKa, asynchronicity, frustration, polarities and rate constants) of redox-active molecules involved in Proton-coupled electron transfer (PCET). It underpins catalysis, energy storage and several biological processes.

A landscape of possibilities

Explore the PCET space

See how bond energies vary across molecular structures, then explore the dataset to find candidates for your next reaction.

Explore the dataset

Phase 1 - BDFE

This is the first iteration of Project EQUUS. We have developed an ultrafast screening workflow (ALOE) for finding the BDFE of redox-active organic small molecules, particularly diamines, for their applications in catalysis and redox mediation. Using AIMNet2 and AIMNet2-NSE, we have virtually screened over 0.5 million small molecules for their bond dissociation free energy (BDFE), which is one of the most important parameters in understanding and exploring the chemistry of these molecules. As a result of this work, we have discovered several new patterns in our dataset. We believe this tool will now allow chemists to have an idea of the thermodynamic landscape of their reaction before it even happens!

A subset (0.018%) of molecules studied in this work

What is BDFE?

Bond Dissociation Free Energy (BDFE) is the energy required to homolytically break a bond between two atoms. This work focuses on measuring the (average) BDFE of redox-active molecules involving two N-H, O-H and S-H bonds, vicinal to each other. Most of our work focuses on phenylenediamines, because of several reasons — they are highly relevant as ligands in organometallic reactions and there is a significant gap in information about this family of molecules (compared to C-H, O-H and S-H). This can potentially open new avenues of research.

o-Phenylenediamine · drag to rotate. Side-chains (R) replace the N–H hydrogens; backbone groups (R′) sit on the ring.
Energy XH2 fully reduced •X–H + H• X + 2 H• fully oxidized ΔG°1 ΔG°2 ΔG°1,2 BDFEavg = ½ ΔG°1,2

From structure to insight

Understand your
molecule better

EQUUS Phase 1 Publication

Machine Learning-Accelerated Screening of Hydroquinone Analogs for Proton-Coupled Electron Transfer

1 Department of Chemistry, Carnegie Mellon University
2 Department of Mathematical Sciences, Carnegie Mellon University

Phase 2 - Square Scheme

Bond Dissociation Free Energy (BDFE), although a helpful metric, only paints part of the picture.
Hydrogen atom transfer systems heavily rely on the energetics of proton transfer (H+, pKa) and electron transfer (e−, E°). These parameters provide a more complete picture, particularly into newly discovered phenomena like asynchronicity and frustration. The PCET field has contexualized all these elementary transformations into something called a square scheme. Here is our attempt at zero-shot prediction of all the parameters, for both 1- and 2-hydrogen atom transfer system(s).
Disclaimer: This is a work in progress. The predictions are not entirely validated yet, but we are getting closer every day.

The 1-Hydrogen system

One Square.
4 states. 5 values.

A hydrogen atom transfer from XH is essentailly a proton and an electron transfer. This can happen either as a single concerted step (along the diagonal), or in two separate steps (along the edges of the square).

Moving right removes H+, measured by the pKa value; moving up removes e−, measured by the redox potential (E°) value. Either edge route lands on the same dehydrogenated state X•. Its pKa, E° add up to BDFE following the Bordwell Equation:
BDFE = 1.37 pKa + 23.06 E° + CG (kcal/mol).

4 edges and 1 diagonal give the square its 5 free energy values, and this help us understand the true reaction mechanism.

Scroll to follow the square

PCET single-square scheme
Generic square scheme for a redox-active molecule, XH. Edge labels name the free-energy dependent parameter for the respective step. H•, H⁺ and e⁻ are omitted from the off-diagonal species. PCET square-scheme convention ↗

The 2-Hydrogen system

Four Squares.
9 states. 16 values.

The 2-H atom system noticably more complex, but nature uses it to its advantage. We now have four squares, each connecting unique proton transfers, electron transfers, and their coupled pathway. Moving horizontally denotes proton (H+) transfer; moving vertically denotes electron (e−) transfer.

The diagonal follows two successive (or maybe concerted) net hydrogen-atom transfers: XH2 → •X–H + H• → X + 2 H•.

Each square has its own set of 5 free energy values, totaling up to 16 unique values for 1 redox-active molecule, XH₂.

Scroll to follow the 2-HAT square scheme

PCET double-square scheme
Generic square scheme for a redox-active molecule, XH₂. H•, H⁺ and e⁻ are omitted from state representations. Each edge label shows the free-energy dependent parameter for that step and the loss of the corresponding particle, it does not necessarily mean its formation in a reaction.

More phases coming up, stay tuned