Identification of Solid-Electrolyte Interphase Species by Joint Characterization of Li-ion Battery Chemistry by Mass Spectrometry and Electro-Chemical Reaction Networks

This study presents a computational-experimental framework combining high-throughput quantum chemistry, data-driven electro-chemical reaction networks, and advanced mass spectrometry to comprehensively characterize the solid-electrolyte interphase (SEI) in lithium-ion batteries, successfully identifying 28 novel SEI species and elucidating their kinetically feasible formation mechanisms to enable rational battery design.

The Gist

Imagine a lithium-ion battery is like a bustling city. Inside, tiny messengers called Lithium ions zip back and forth between two districts: the Anode (the negative side, usually made of graphite) and the Cathode (the positive side).

For this city to run smoothly for years, the Lithium messengers need a safe, paved road to travel on. But here's the problem: the Anode is like a very aggressive construction site. When the battery is first turned on, the chemicals in the "air" (the liquid electrolyte) crash into the Anode and break apart.

This crash creates a messy pile of debris on the surface of the Anode. In the battery world, this pile is called the SEI (Solid Electrolyte Interphase).

Think of the SEI as a protective shield or a force field.

• If it's good: It's a sturdy, flexible wall that lets the Lithium messengers pass through but stops the messy chemicals from attacking the Anode again. The battery lasts a long time.
• If it's bad: It's a crumbling, leaky wall. The chemicals keep attacking, the mess piles up, the battery gets clogged, and it dies quickly.

The Problem: We Didn't Know What Was in the Shield

For decades, scientists knew this shield existed, but they were like detectives trying to solve a crime scene in the dark. They could see the shape of the debris (using microscopes) or guess the types of materials (using X-rays), but they couldn't read the specific "DNA" of the molecules making up the shield.

Because they didn't know exactly what molecules were in the SEI, they couldn't design better batteries. They were just guessing which ingredients to mix, hoping for the best.

The Solution: A Detective Team with Superpowers

This paper describes a new, super-smart detective team that solved the mystery of the SEI's composition. They combined two powerful tools:

1. The "Crystal Ball" (Super-Computer Simulation):
   Instead of just guessing, the team built a massive digital library of every possible chemical reaction that could happen in a battery. They simulated over 209 million reactions involving 10,000 different molecules.

   • Analogy: Imagine a chef who doesn't just cook one dish, but simulates every possible combination of ingredients in a giant kitchen to predict exactly what flavors will result. They used a "stochastic" method (like rolling dice millions of times) to see which chemical "recipes" were most likely to happen.

2. The "Super-Microscope" (Mass Spectrometry):
   They took real batteries, peeled off the shield (the SEI), and put it under a machine called LDI-FTICR-MS. This machine is so precise it can weigh a molecule and tell you its exact chemical formula, down to the difference of a single electron.

   • Analogy: If the Crystal Ball predicted a specific type of cookie might be baked, the Super-Microscope is the taste-tester that confirms, "Yes, that is exactly a chocolate chip cookie, and here is the proof."

The Big Discovery

By matching the computer's predictions with the real-world measurements, the team achieved two amazing things:

1. They confirmed the old suspects: They found 27 known molecules that scientists had suspected were in the shield. This proved their computer model was working correctly.
2. They found the "Ghost" molecules: They discovered 28 brand-new molecules that no one had ever seen in a battery before!

These new molecules are like secret agents in the battery's defense system. Some of them act like glue (polymers) to make the shield flexible so it doesn't crack when the battery expands and contracts. Others act like reinforced concrete (fluorinated compounds) to make the shield stronger and better at conducting electricity.

Why This Matters

Before this paper, designing a better battery was like trying to build a better house by throwing random bricks at a wall and hoping it stands.

Now, thanks to this work, battery engineers have a blueprint. They know exactly which molecular "bricks" create a strong, flexible, long-lasting shield.

• If they want a battery that lasts longer, they can now engineer the electrolyte to specifically create those "glue" molecules.
• If they want a battery that charges faster, they can target the molecules that conduct electricity better.

The Bottom Line

This research is a giant leap from "guessing" to "knowing." By combining a massive computer simulation with ultra-precise real-world testing, the team has finally cracked the code on what makes a battery's protective shield tick. This paves the way for electric cars that drive further, phones that last all day, and energy storage systems that are safer and more efficient than ever before.

Technical Summary

1. Problem Statement

The Solid Electrolyte Interphase (SEI) is a critical passivation layer in lithium-ion batteries (LIBs) that determines long-term performance, safety, and longevity. Despite decades of research, the SEI remains poorly understood due to its:

• Heterogeneity: It is a complex, multi-phase mixture of organic and inorganic species.
• Dynamic Nature: It evolves continuously during battery cycling.
• Characterization Limitations: Traditional techniques (XPS, SEM, FTIR, ToF-SIMS) provide ensemble-averaged data or surface morphology but struggle to resolve specific molecular formulas, isotopic signatures, and the complex chemical bonding environments of unknown organic-inorganic hybrid species.
• Knowledge Gap: There is a lack of comprehensive molecular-level data linking electrolyte decomposition (both salt and solvent) to specific SEI components, hindering the rational design of advanced electrolytes.

2. Methodology

The authors developed a computational-experimental framework that integrates high-throughput quantum chemistry, data-driven reaction networks, stochastic simulations, and ultra-high-resolution mass spectrometry.

A. Experimental Characterization

• System: Graphite anodes cycled in a standard carbonate-based electrolyte (1 M LiPF₆ in EC/EMC 30:70 v/v) at 60°C for one formation cycle.
• Technique: Laser Desorption/Ionization Fourier Transform Ion Cyclotron Resonance Mass Spectrometry (LDI-FTICR-MS).
  • Utilized an 18 T superconducting magnet for ultra-high resolution (700,000 at m/z 200).
  • Analyzed in both positive and negative ion modes.
  • Achieved mass accuracy allowing for distinct molecular formula assignment (tolerance < 0.5 ppm) and isotopic fine structure validation.

B. Computational Framework

1. Species Generation:

   • Fragmentation/Recombination: Principal molecules (EC, EMC, LiPF₆, and early decomposition product PF₄OH) were systematically fragmented into smaller units. These fragments were then recombined to generate a diverse pool of candidate species.
   • DFT Calculations: High-throughput Density Functional Theory (using $\omega$B97X-V/def2-TZVPPD) was used to optimize geometries and calculate thermodynamic quantities (Gibbs free energy, enthalpy, entropy) and vibrational frequencies.
   • Dataset: Generated an initial set of 23,625 species (including C-H-O-Li-P-F systems).

2. Electrochemical Reaction Network (eCRN) Construction:

   • Applied filtering criteria to remove chemically unreasonable species, resulting in 11,867 filtered species.
   • Enumerated all stoichiometrically valid reactions, creating a massive network of 208,929,658 reactions. This is the largest eCRN for LIB SEI formation to date and is unique for including coupled salt-solvent decomposition pathways (hybrid species).

3. Stochastic Analysis (kMC):

   • Performed 50,000 parallel Kinetic Monte Carlo (kMC) simulations using the Reaction Network Monte Carlo (RNMC) framework.
   • Simulated under 0.0 V vs. Li/Li⁺ at 298.15 K.
   • Used uniform rate coefficients based on transition state theory to sample the network stochastically without needing pre-defined kinetic data for every step.
   • Identified "network products" based on formation/consumption ratios and pathway accessibility.

4. Validation & Kinetic Refinement:

   • Mass Matching: Computationally predicted species were matched against LDI-FTICR-MS data by generating theoretical isotopic patterns and exact masses.
   • Kinetic Validation: For a subset of jointly identified species, transition states (TS) were explicitly calculated to determine activation barriers ($\Delta G^\ddagger$) and confirm kinetically feasible pathways (barriers < 1 eV).

3. Key Contributions

• Largest eCRN to Date: Constructed a reaction network with >10,000 species and >200 million reactions, explicitly modeling the joint decomposition of salt (LiPF₆) and solvents (EC/EMC).
• Discovery of Novel Species: Identified 28 novel SEI species previously unreported in literature, effectively doubling the known molecular inventory of SEI components in this context.
• Hybrid Species Identification: Successfully predicted and confirmed fluorinated organophosphate-carbonates, bridging the gap between inorganic salt fragments (P, F) and organic solvent fragments (C, H, O).
• Methodological Synergy: Demonstrated a robust workflow where computational predictions guide experimental mass spectrometry interpretation, and experimental data validates computational hypotheses, moving beyond trial-and-error electrolyte design.

4. Key Results

• Recovery of Known Species: The framework successfully recovered 27 known SEI components (e.g., Li₂CO₃, LiF, LEC, LMC, LBDC, gases like H₂ and CO), validating the robustness of the approach even without explicit kinetic data.
• Novel Species Discovery:
  • Identified 28 new molecules including fluorinated organophosphate-carbonates, cyclic/bicyclic alkyl carbonates, oxalates, formates, diols, ethers, and dioxolanes.
  • Experimental Confirmation: Each novel species was confirmed via LDI-FTICR-MS by matching exact mass (to the 4th/5th decimal place) and isotopic fine structure (e.g., $^{6}$Li, $^{13}$C, $^{18}$O patterns).
• Mechanistic Insights:
  • Pathway Analysis: Kinetic refinement of representative pathways revealed activation barriers generally below 1.0 eV (many < 0.5 eV), confirming they are kinetically accessible under SEI formation conditions.
  • Competing Pathways: Identified multiple competing reaction channels (e.g., nucleophilic attack vs. ethyl group transfer) that lead to different products, highlighting the sensitivity of SEI composition to local conditions.
  • Salt-Solvent Coupling: Confirmed mechanisms where salt decomposition products (like POF₃) react directly with solvent-derived intermediates to form hybrid species.
• Functional Implications:
  • Mechanical Stability: Identified species (e.g., vinyl-bearing formates, cyclic carbonates) capable of polymerization, suggesting mechanisms for forming flexible, cross-linked SEI layers that resist cracking.
  • Ionic Conductivity: Highlighted the role of mixed-anion amorphous phases (Li-O-P-F) and fluorinated hybrids in facilitating Li⁺ transport.

5. Significance

• Paradigm Shift: Moves SEI characterization from empirical observation to hypothesis-driven molecular engineering.
• Rational Design: Provides a roadmap for designing electrolyte additives. For example, additives containing vinyl groups can be targeted to generate specific polymerizable formates, or formulations can be tuned to promote salt-solvent coupling for better ionic conductivity.
• Scalability: The framework demonstrates that massive-scale stochastic analysis combined with high-resolution mass spectrometry can unravel complex reactive landscapes that are intractable for traditional experimental or purely theoretical approaches alone.
• Future Outlook: The study lays the groundwork for integrating Machine Learning Interatomic Potentials (MLIPs) to further refine kinetic barriers across the entire network, enabling predictive modeling of battery aging and failure.

In conclusion, this work provides the most comprehensive molecular-level map of SEI chemistry to date, validating a powerful computational-experimental loop that is essential for developing next-generation, high-performance lithium-based energy storage systems.