AI-Guided Discovery of Novel Ionic Liquid Solvents for Industrial CO2 Capture

TL;DR

The paper tackles the challenge of finding refinery-appropriate CO$_2$ capture solvents with lower regeneration energy and reduced corrosion by leveraging an AI-driven, modular five-stage pipeline. It combines candidate IL generation, a GNN-based predictor for temperature- and pressure-dependent CO$_2$ solubility and viscosity, Van’t Hoff thermodynamic post-processing to derive working capacity and regeneration energy, Pareto-front optimization, and synthesis-feasibility filtering to identify practical IL candidates. The study screens over 400k ILs and highlights 36 that are synthesis-feasible and offer favorable trade-offs, including potential 5–10% OPEX savings and up to 10% CAPEX reductions, while revealing chemical trends across cation/anion families. This approach provides a scalable, interpretable framework to accelerate industrial deployment of IL solvents for CO$_2$ capture and guides subsequent experimental validation and pilot testing.

Abstract

We present an AI-driven approach to discover compounds with optimal properties for CO2 capture from flue gas-refinery emissions' primary source. Focusing on ionic liquids (ILs) as alternatives to traditional amine-based solvents, we successfully identify new IL candidates with high working capacity, manageable viscosity, favorable regeneration energy, and viable synthetic routes. Our approach follows a five-stage pipeline. First, we generate IL candidates by pairing available cation and anion molecules, then predict temperature- and pressure-dependent CO2 solubility and viscosity using a GNN-based molecular property prediction model. Next, we convert solubility to working capacity and regeneration energy via Van't Hoff modeling, and then find the best set of candidates using Pareto optimization, before finally filtering those based on feasible synthesis routes. We identify 36 feasible candidates that could enable 5-10% OPEX savings and up to 10% CAPEX reductions through lower regeneration energy requirements and reduced corrosivity-offering a novel carbon-capture strategy for refineries moving forward.

Figures (15)

• Figure 1: Overview of AI-driven IL candidate screening framework
• Figure 2: Schematic of absorption–desorption process for CO$_2$ capture in industrial scrubbing systems. Absorption occurs at low temperature and pressure, while regeneration requires elevated thermal conditions to release CO$_2$ and recycle solvent
• Figure 3: Distributions of predicted CO$_2$ solubility and natural log viscosity properties for generated ionic liquids compared to those from raw training dataset. Plots show property distributions at absorption conditions ($T=313.15 \text{K}, P=1 \text{bar}$). Pareto front distribution represent isolated top performers and a subset of generated.
• Figure 4: Pairwise model agreement for CO$_2$ solubility. Hexbin plots (model $j$ vs. $i$) across all IL$\times$T with $y{=}x$; panels report $r$, RMSE, bias ($y_j{-}y_i$), and LoA half–width ($\pm 1.96\,\sigma_d$)
• Figure 5: Pairwise model agreement for $\ln(\eta)$. Hexbin plots (model $j$ vs. $i$) across all IL$\times$T with $y{=}x$; panels report $r$, RMSE, bias ($y_j{-}y_i$), and LoA half–width ($\pm 1.96\,\sigma_d$)