A Chemical Space Perspective on Diastereomeric Barriers in Alkylperoxy-to-Hydroperoxyalkyl Isomerization

This study demonstrates that explicit stereochemical treatment is critical for accurately modeling low-temperature hydrocarbon autooxidation, as it reveals significant energy barriers between diastereomeric pathways that are systematically missed by constitutionally collapsed molecular representations.

The Gist

The Big Picture: Why "Left" and "Right" Matter in Fire

Imagine you are trying to predict how a campfire burns. You know the wood, the oxygen, and the heat. But what if the way the wood is twisted or shaped changes how fast it catches fire?

This paper is about a specific, invisible dance that happens inside burning fuel (hydrocarbons) at low temperatures. It's a dance between tiny, unstable particles called radicals. The author, Raghunathan Ramakrishnan, discovered that if you ignore the 3D "handedness" (stereochemistry) of these particles, you might completely miss how the fire behaves.

The Analogy: The Key and the Lock

To understand the problem, imagine a Key (a chemical molecule) trying to fit into a Lock (a reaction pathway).

1. The Old Way (The Flawed Map):
   For years, scientists have been drawing maps of chemical reactions using 2D sketches. They look at the atoms and say, "Okay, this carbon is connected to that oxygen." It's like looking at a flat blueprint of a house.

   • The Problem: A flat blueprint doesn't tell you if a door is on the left or the right side of the hallway. In chemistry, this matters. If a molecule is "twisted" one way, it might fit the lock perfectly. If it's twisted the other way, it might jam.
   • The Result: Scientists were often counting only one path for a reaction, missing the fact that there are actually two different paths (like a left-handed key and a right-handed key) that behave very differently.

2. The New Discovery (The 3D Map):
   Ramakrishnan built a massive, 3D database called SEARS (Stereochemically Expanded Autooxidation Reaction Space). He took 498 different fuel molecules and simulated what happens when they react with oxygen.

   • He didn't just look at the atoms; he looked at how they are arranged in 3D space.
   • He found that for many reactions, there are two distinct "twists" (diastereomers) that lead to the same result, but one is easy and the other is hard.

The "Ephemeral Diastereomers": Ghosts in the Machine

The paper focuses on a very specific, fleeting moment in the reaction called a Transition State.

• The Metaphor: Imagine a hiker trying to cross a mountain range. The "Transition State" is the very top of the mountain pass.
• The Twist: Sometimes, there are two mountain passes right next to each other.
  • Pass A (The Easy Path): The terrain is smooth. The hiker (the reaction) zooms through.
  • Pass B (The Hard Path): The terrain is a cliff with a massive boulder blocking the way. The hiker gets stuck or turns back.
• The Surprise: In the old 2D maps, scientists thought there was only one mountain pass. They didn't realize that for every reaction, there might be two "ghost" passes (diastereomers) that exist only for a split second. Sometimes these two passes are identical (degenerate), but often, one is a highway and the other is a dead end.

What the Data Showed

Ramakrishnan ran thousands of supercomputer simulations (using a method called Density Functional Theory) to map these mountain passes. Here is what he found:

1. The Gap is Huge: For some molecules, the difference between the "Easy Pass" and the "Hard Pass" was tiny (they are almost the same). But for others, the difference was massive—up to 60 kcal/mol.

   • Translation: That's like the difference between walking up a gentle hill and trying to climb a vertical wall. If you ignore the wall, your prediction of how fast the fire spreads will be totally wrong.

2. It's About Crowding: The reason for these huge differences is steric strain (crowding). Imagine trying to squeeze through a doorway while carrying a large sofa.

   • If you turn the sofa one way, it fits.
   • If you turn it the other way, the sofa hits the doorframe and gets stuck.
   • In the chemical world, if the atoms are crowded in a specific 3D arrangement, the reaction gets stuck. If they are arranged differently, it flows.

3. The "Hidden" Channels: Because many computer programs only look at the "flat" version of the molecule, they often miss the "Hard Pass" entirely. They assume the reaction happens at a certain speed, but in reality, the reaction might be much faster (because the "Easy Pass" was ignored) or much slower (because the "Hard Pass" was the only one they found).

Why Should You Care?

This isn't just about abstract chemistry; it's about predicting the future.

• Combustion Engines: If we want to design cleaner, more efficient engines (or prevent engines from knocking), we need to know exactly how fuel burns. If our computer models miss these "hidden paths," our engines might be inefficient or dangerous.
• Atmospheric Science: These same reactions happen in the air, creating smog or breaking down pollutants. Getting the speed of these reactions wrong means our climate models could be off.
• The Future of AI: The paper suggests that future AI models for chemistry need to be "stereo-aware." They can't just look at a list of ingredients; they need to understand the 3D shape of the molecules to predict how they will react.

The Takeaway

Think of this paper as a warning to all the cartographers of the chemical world: "Don't just draw the roads; draw the hills and the valleys."

By ignoring the 3D twists and turns of molecules, we have been missing half the story. This new dataset (SEARS) gives us a complete, 3D map, showing us that sometimes, the difference between a reaction happening instantly or not happening at all is just a matter of which way the molecule is twisted.

Technical Summary

1. Problem Statement

Low-temperature hydrocarbon autooxidation is a critical process in combustion and atmospheric chemistry, governed by complex radical-chain networks. A key step in this process is the isomerization of alkylperoxy radicals ($ROO^\bullet$) to hydroperoxyalkyl radicals ($^\bullet QOOH$) via intramolecular hydrogen transfer. This step controls access to chain-branching pathways.

The Core Issue:
Current automated reaction mechanism generation and kinetic modeling often rely on "constitutionally collapsed" molecular representations. These models treat molecules based solely on atom connectivity, ignoring stereochemistry. However, the $ROO^\bullet \to ^\bullet QOOH$ transition state (TS) often involves the formation of a transient stereogenic center.

• Ephemeral Diastereomers: A single $ROO^\bullet$ conformer can access the same $^\bullet QOOH$ product through two distinct transition states that are diastereomeric (non-superimposable, non-mirror images).
• Kinetic Consequence: These diastereomeric pathways may have significantly different activation barriers. If ignored, models may either miss kinetically relevant channels (if one barrier is much lower) or fail to account for rate enhancements (if barriers are near-degenerate, effectively doubling the rate).
• Gap: Large-scale chemical space datasets and automated workflows rarely track stereochemical variations explicitly along reactive coordinates, particularly for transient transition states.

2. Methodology

The study introduces a new dataset, SEARS (Stereochemically Expanded Autooxidation Reaction Space), to systematically quantify the impact of stereochemistry on reaction barriers.

Data Generation Workflow:

1. Source Material: Started with the bigQM7$\omega$ dataset (12,880 small molecules).
2. Filtering: Selected 498 non-aromatic, aliphatic C1–C7 hydrocarbons (excluding benzene/toluene and highly fused ring systems).
3. Stereochemical Enumeration:
   • Generated geometrical isomers (E/Z) and diastereomers by inverting tetrahedral centers.
   • Propagated stereochemistry through the reaction sequence: Hydrocarbon $\to$ Alkyl Radical ($R^\bullet$) $\to$ Alkylperoxy Radical ($ROO^\bullet$) $\to$ Hydroperoxyalkyl Radical ($^\bullet QOOH$).
   • Used RDKit for SMILES enumeration and canonicalization, ensuring enantiomers were treated as identical while retaining distinct diastereomers.
4. Transition State (TS) Construction:
   • Constructed cyclic TS scaffolds (4- to 8-membered rings) for intramolecular H-transfer.
   • Used a "dummy atom" strategy: A nitrogen atom ([N]) was inserted in the SMILES string between the peroxyl oxygen and the H-donor carbon to enforce the cyclic topology. This was replaced by Hydrogen in 3D coordinates prior to DFT calculations.
5. Electronic Structure Calculations:
   • Software: ORCA (v6.1.1).
   • Functional/Basis: $\omega$B97M-D4/def2-SV(P) for geometry optimization and frequency analysis; $\omega$B97M-D4/def2-TZVP for single-point energy refinement.
   • Validation:
     • Confirmed TSs via one imaginary frequency.
     • Intrinsic Reaction Coordinate (IRC) analysis was performed to verify that each TS connected the intended $ROO^\bullet$ and $^\bullet QOOH$ minima.
     • Structures failing connectivity checks (e.g., evolving into O-O bond scission) were excluded.

Dataset Scale:

• Total Species: 5,356 unique species (including radicals and TSs).
• Validated TSs: 2,324 transition states corresponding to 1,162 unique diastereomeric pairs of $ROO^\bullet \to ^\bullet QOOH$ reactions.

3. Key Results

A. Energetic Diversity of Diastereomeric Pairs

• Barrier Splitting ($\Delta\Delta E_{barrier}$): The energy difference between the two diastereomeric TSs for a single reaction varies widely.
  • Near-Degenerate: 32% of pairs have $\Delta\Delta E_{barrier} \le 1$ kcal/mol.
  • Significant Splitting: A substantial tail exists where differences exceed 60 kcal/mol.
  • Distribution: 52% of pairs lie within 3 kcal/mol, but the remaining 48% show significant kinetic differentiation.
• Structural Drivers: Large splittings are governed by steric strain at the carbon bearing the peroxyl group and the rigidity of the ring system.
  • Systems with the peroxyl group on an $sp^3$ carbon (especially secondary) show the largest splittings.
  • Systems with the peroxyl group on an $sp^2$ carbon tend to have smaller splittings due to geometric constraints limiting conformational divergence.
  • Ring Constraints: Cyclic systems exhibit larger splittings than acyclic ones because the cyclic TS motif forces one diastereomer into a high-strain conformation (e.g., eclipsing interactions) while the other remains low-strain.

B. Kinetic Implications

• Rate Enhancement: When barriers are near-degenerate, the total rate constant ($k_{tot}$) is approximately the sum of both pathways, potentially doubling the rate compared to a single-pathway model.
• Kinetic Negligibility: When the splitting is large (e.g., >10–15 kcal/mol), the higher-barrier pathway becomes kinetically irrelevant, and the reaction proceeds almost exclusively via the lower-barrier diastereomer.
• Case Studies: Analysis of 41 exothermic reactions showed:
  • 5-membered TSs: Often exhibited small splittings (near-degenerate).
  • 6-membered TSs: Often exhibited larger splittings, with one pathway significantly favored (e.g., a Case 2 example showed a 18 kcal/mol split, rendering the high-barrier path negligible).

C. Failure Modes

• Approximately 32 TS candidates (16 pairs) failed IRC validation, evolving instead toward O-O bond scission and $^\bullet OH$ formation. These failures were strongly correlated with peroxyl groups attached to $sp^2$ carbons, suggesting these systems favor alternative reaction channels over H-transfer.

4. Significance and Contributions

1. SEARS Dataset: Provides the first large-scale, stereochemically resolved quantum-chemical dataset specifically for $ROO^\bullet \to ^\bullet QOOH$ isomerization, containing 1,162 diastereomeric TS pairs.
2. Validation of Stereochemical Necessity: Demonstrates that constitutionally collapsed models systematically miss reactive channels. Ignoring stereochemistry can lead to:
   • Underestimation of reaction rates (by missing the lower-barrier diastereomer).
   • Overestimation of rates (by failing to account for the kinetic irrelevance of a high-barrier diastereomer).
   • Failure to capture rate enhancements from near-degenerate pathways.
3. Guidance for Automated Mechanism Generation: Argues that future automated reaction discovery tools must explicitly track stereochemistry at the transition state level. The study suggests that "constitution-only" graph formalisms are insufficient for accurate combustion modeling.
4. Chemical Space Insights: Identifies specific structural motifs (e.g., $sp^3$ peroxyl centers in constrained rings) where stereochemical resolution is most critical for predictive accuracy.
5. Benchmark for ML: The dataset serves as a benchmark for machine learning models aiming to predict not just average barrier trends, but the specific patterns of stereochemically induced barrier splitting.

Conclusion

The paper establishes that stereochemistry is not merely a labeling detail but a fundamental determinant of reaction kinetics in low-temperature autooxidation. By explicitly treating diastereomeric transition states, the study reveals a "hidden kinetic dimension" where barrier splittings can range from negligible to massive, fundamentally altering the predicted reactivity of hydrocarbon oxidation. The SEARS framework provides the necessary foundation for developing stereochemistry-aware combustion models and automated mechanism generators.