Multiscale Mathematics for Biomass Conversion to Renewable Hydrogen

Petr Plecháč, Dionisios G. Vlachos, Markos A. Katsoulakis2013 · DOI

TL;DR

This research pioneers multiscale modeling to understand and design complex catalytic processes for converting biomass into renewable hydrogen and chemicals. By integrating quantum mechanics (DFT), kinetic Monte Carlo (KMC), and continuum models, they're bridging the gap between atomic-level understanding and macroscopic reactor behavior, crucial for developing efficient and sustainable energy solutions.

What Problem Does This Paper Address?

The efficient conversion of biomass into valuable products like hydrogen is a cornerstone of renewable energy initiatives. However, these processes involve complex chemical reactions occurring at multiple scales, from the quantum interactions of atoms on catalyst surfaces to the bulk flow and reaction dynamics within a reactor. Traditional modeling approaches often struggle to capture this multiscale complexity, leading to a disconnect between theoretical predictions and experimental observations. Specifically, understanding the "structure sensitivity" of catalysts – how their atomic arrangement influences their performance – has been a persistent challenge. This paper aims to bridge this gap by developing and applying a hierarchical multiscale modeling framework that can accurately simulate these complex systems.

Methods

The core of this work lies in the development and application of a hierarchical multiscale modeling framework. This framework integrates several key computational techniques:

  • Density Functional Theory (DFT): Used for high-fidelity, first-principles calculations to understand the fundamental electronic structure and reaction pathways at the atomic level. This is crucial for determining reaction energies, activation barriers, and adsorption properties.
  • Kinetic Monte Carlo (KMC): A stochastic simulation method that models the time evolution of a system by considering discrete events (e.g., adsorption, desorption, reaction) based on their probabilities. This allows for the simulation of microscopic processes over longer timescales than DFT alone. The paper highlights the use of a Graph-Theoretical KMC approach, which is particularly adept at handling complex reaction networks and catalyst structures.
  • Mean-Field Microkinetic Modeling: Used to simulate reaction kinetics on specific catalytic surfaces, often informed by DFT data.
  • Hybrid Multiscale Simulations: The paper describes efforts to link these methods. For instance, coupling KMC simulations with deterministic Partial Differential Equation (PDE) models to simulate entire reactors, and also linking KMC with DFT calculations.
  • Parallelization: A significant effort is dedicated to parallelizing these models to leverage high-performance computing resources (Petaflop scale) for tackling real-world applications.

Datasets

Not discussed in the paper. The research relies on ab initio calculations (DFT) and existing experimental data for validation, rather than large, pre-existing datasets for training machine learning models.

Key Insights and Ideas

  • Structure Sensitivity is Key: The research strongly suggests that the specific atomic arrangement (structure) of catalytic surfaces, even at the nanoscale, plays a critical role in reaction performance. For steam methane reforming (SMR), the difference between Pt(111) and Pt(211) surfaces significantly impacts activation energies and reaction orders, explaining discrepancies in previous models.
  • Bridging the DFT-KMC Gap: The successful integration of DFT-derived reaction parameters into KMC simulations allows for a more accurate representation of catalytic processes, moving beyond simplified models.
  • Understanding Deactivation Mechanisms: For gold nanoclusters in CO oxidation, the work identified that while some clusters (like Au6) can be poisoned by carbonate formation, others (like Au8 and Au10) show no deactivation under similar conditions, offering insights into catalyst longevity.
  • Predictive Power of Descriptors: The study on biomass conversion to p-xylene demonstrates that site-specific descriptors (like Fukui functions) and proton affinities can quantitatively correlate with activation energies, enabling a priori prediction and screening of furan derivatives for desired reactions.
  • Catalyst Design Implications: The multiscale approach provides a powerful tool for designing catalysts with improved activity, selectivity, and stability by understanding how atomic-level properties translate to macroscopic performance.

Key Results

  • Steam Methane Reforming (SMR): The developed Pt(111) and Pt(211) models showed very good agreement with experimental data for conversion, with the Pt(211) model highlighting a potentially blocked step or terrace dominance in Pt(111) for this reaction. Sensitivity analysis identified methane adsorption and C-O bond formation as critical steps warranting further DFT investigation.
  • CO Oxidation on Gold Nanoclusters: DFT calculations revealed low-barrier reaction pathways for Au8, including a 4-center intermediate. KMC simulations on Au8 and Au10 showed no deactivation, unlike Au6, and matched experimental turnover frequencies, with Au8 being at least twice as active as Au10.
  • Biomass Conversion to p-Xylene: The model predicted that the uncatalyzed cycloaddition is the rate-limiting step for DMF and ethylene conversion to p-xylene on HY zeolites, explaining experimental observations. It also suggested that Lewis acids and alkali-exchanged zeolites (like NaY) can accelerate this reaction, and bifunctional NaHY zeolites could further enhance production rates.

Limitations and Future Work

  • Computational Cost: While parallelization is a focus, multiscale simulations, especially those involving DFT, remain computationally intensive.
  • Model Complexity: Developing accurate KMC models requires detailed knowledge of all relevant elementary steps and their rates, which can be challenging to obtain comprehensively.
  • Experimental Validation: While agreement with experimental data is shown, continuous validation and refinement of models against new experimental results are crucial.
  • Future Work: The paper indicates that the work on SMR is near completion. Future directions likely involve extending the developed multiscale frameworks to other biomass conversion processes, designing new catalysts with enhanced properties, and further optimizing computational efficiency for larger-scale industrial applications. The development of predictive descriptors for catalyst design is also a promising avenue.

Summary updated

Abstract

The overall objective of this project is to develop multiscale models for understanding and eventually designing complex processes for renewables. To the best of our knowledge, our work is the first attempt at modeling complex reacting systems, whose performance relies on underlying multiscale mathematics. Our specific application lies at the heart of biofuels initiatives of DOE and entails modeling of catalytic systems, to enable economic, environmentally benign, and efficient conversion of biomass into either hydrogen or valuable chemicals. Specific goals include: (i) Development of rigorous spatio-temporal coarse-grained kinetic Monte Carlo (KMC) mathematics and simulation for microscopic processes encountered in biomass transformation. (ii) Development of hybrid multiscale simulation that links stochastic simulation to a deterministic partial differential equation (PDE) model for an entire reactor. (iii) Development of hybrid multiscale simulation that links KMC simulation with quantum density functional theory (DFT) calculations. (iv) Development of parallelization of models of (i)-(iii) to take advantage of Petaflop computing and enable real world applications of complex, multiscale models. In this NCE period, we continued addressing these objectives and completed the proposed work. Main initiatives, key results, and activities are outlined.
BibTeX
@misc{plechac2013multiscale,
  author = {Petr Plecháč and Dionisios G. Vlachos and Markos A. Katsoulakis},
  title = {{Multiscale Mathematics for Biomass Conversion to Renewable Hydrogen}},
  year = {2013},
  doi = {10.2172/1091947},
}
Figure