Our research addresses engineering and materials issues in fuel cells and electrolyzers, particularly mass transport within and stability of reactor electrodes. We focus on non-precious metal catalysts based on transition metals and redox enzymes; these materials have lower costs compared to precious metals, but are challenging in terms of overall activity and stability, and often are implemented at high loadings that lead to transport limitations. Below are brief descriptions of current projects.
- Electrostatic Channelling in Enzyme Cascades
- Electrochemical Hydrogenation (ECH) Flow Reactor Modeling
- Calcium ion Transport in Battery Cathodes
- Electrocatalytic hydrogenation of biomass
- Quantitative nanoscale scanning electrochemical microscopy
Electrostatic Channelling in Enzyme Cascades
We study electrostatic channeling catalytic cascades using computational techniques ranging from molecular dynamics (MD) to kinetic monte carlo (KMC) to continuum models such as finite element (FE) and finite difference (FD). The enzyme systems we have studied include a complex of hexokinase (HK) and glucose-6-phosphate dehydrogenase (G6PDH) and the more complex surface between malate dehydrogenase (MDH) and citrate synthase (CS) – of the Tricarboxylic Acid or Krebs Cycle.
We have described channelling on the surface of the MDH-CS complex using a Markov state model that uses a network of 500 nodes, as shown below. Analysis of such a network using KMC can be computationally demanding. Instead, we take a finite difference approach by treating the network as a system of differential equations that describe transport on the enzyme surface and in the bulk between catalytic active sites where reactions occur.
We can use the finite difference model to predict experimental lag times, which are determined by time dependent measurements in which the product concentration, CP, is plotted vs. time. Transporth via desorption into the bulk results in a delayed onset of reaction at the second active site (here citrate synthase) which is measured as a lag time, tl. Perfect channeling leads to zero lag time.
In fact, direct use of the network model does not allow prediction of experimental results. Instead, we need to modify the network to convert the desorbed intermediate nodes (in purple) to a single bulk concentration. We can then consider adsorption equilibrium between the surface nodes and the bulk, like so:
In so doing, and using an adsorption coefficient, Kads = 10−5, we are able to closely match the model lag times to experiment, as shown below. Based on these predictions, we can employ the model to make further design predictions, offering a way to connect enzyme structure to electrostatic channeling efficiency.
| Source | Recombinant Lag Time τl / s |
Mutant Lag Time τl /s |
|---|---|---|
| Experiment | 0.03 ± 0.01 | 0.88 ± 0.06 |
| Model | 0.44 ± 2e-9 | 2.1 ± 4e-6 |
References
-
Y. Xie, S.D. Minteer, S. Banta, S. Calabrese Barton, “Markov State Study of Electrostatic Channeling within the Tricarboxylic Acid Cycle Supercomplex”, ACS Nanosci. Au (2022). doi:10/gqkvtq
-
Y. Xie and S. Calabrese Barton, “Infrequent Metadynamics Study of Rare-event Electrostatic Channeling”, Phys. Chem. Chem. Phys. (2021). doi:10/gkgmtf
-
Y. Liu, D. P. Hickey, S. D. Minteer, A. Dickson and S. Calabrese Barton, “Markov-State Transition Path Analysis of Electrostatic Channeling”, The Journal of Physical Chemistry C, 123, 15284–15292 (2019). doi:10/ghfprh
-
Y. Liu, I. Matanovic, D.P. Hickey, S.D. Minteer, P. Atanassov, S. Calabrese Barton, “Cascade Kinetics of an Artificial Metabolon by Molecular Dynamics and Kinetic Monte Carlo”, ACS Catalysis, 8, 7719–7726 (2018). doi:10/gd7xgn
-
Y. Liu, D. P. Hickey, J.-Y. Guo, E. Earl, S. Abdellaoui, R. D. Milton, M. S. Sigman, S. D. Minteer and S. Calabrese Barton, “Substrate Channeling in an Artificial Metabolon: A Molecular Dynamics Blueprint for an Experimental Peptide Bridge”, ACS Catalysis, 7, 2486–2493 (2017). doi:10/f94shk
Electrochemical Hydrogenation (ECH) Flow Reactor Modeling
While there are many studies investigating mechanistic impacts of electrode/catalyst and chemical environment on electrochemical hydrogenation (ECH) of bio-derived compounds, very little has been done to scale these phenomena in electrochemical flow reactors. To address this need, we investigated kinetic impacts on reactor design by comparing an adsorption ECH model for benzaldehyde hydrogenation, recently proposed by Sanyal et al. (2021), against a first order kinetic scheme. The Sanyal model assumes that the rate of benzaldehyde hydrogenation is the sum of single-site and benzaldehyde and phenol coupled two-site adsorption kinetics. This results in the following rate law:
$$r_{\text{TOF}} = a \cdot r_{ECH} = \frac{{a \cdot k}_{1}{K_{b}C}_{b}}{(1 + K_{b}C_{b} + K_{p}C_{p})} + \frac{a \cdot k_{2}{K_{b}K_{p}C}_{b}C_{p}}{\left( 1 + K_{b}C_{b} + K_{p}C_{p} \right)^{2}}$$
where TOF (h^-1) is the turnover frequency, a is the reactor volume per mole of active catalyst, and rECH is the volumetric rate law. The first term on the right side represents the single-site reaction, and the second term the dual-site reaction. As can be seen, the Sanyal model differs substantially from a first order kinetic model, which will affect predictions of reactor performance.
In this model we assumed a constant density and velocity field and that ECH of benzaldehyde occurred at the cathode indicated by the blue boundary (see schematic above). In this model both convective and diffusive transport were considered. As a result, the solutions at varying mass diffusion related Damköhler numbers (Da = rate law dependent), Péclet numbers (Pe = $\frac{{(u}_{z}/l)}{{(D}_{b}/l^{2})}$), and aspect ratios (AR = l/w) were quantified.
The dimensionless solutions to this flow reactor system were obtained using FEnics, a finite element coding package in python. To compare reactor performance of differing kinetic models, we integrated the flux across the outlet boundary to determine the overall conversion of benzaldehyde. An example of possible contour plots for both the first order and Sanyal model can be found below. At the same value of Da and Pe, the first order model almost always had a much higher conversion than the Sanyal model. This implies that while designing a reactor for the ECH of benzaldehyde, we will have to generally utilize much higher and lower Da and Pe values, respectively than if the kinetics followed a first order scheme. This is a consequence of the kinetics of a first order model not being slowed down by adsorption of the reactants and subsequent desorption of the products.
Calcium ion Transport in Battery Cathodes
The transition towards renewable energy sources has increased the demand for advanced energy storage technologies with advantages over conventional lithium ion, particularly for stationary applications such as electric load leveling. Aqueous Calcium-ion batteries are a promising alternative secondary battery system, offering advantages in terms of resource abundance, environmental friendliness, and high volumetric/gravimetric capacity. Among the various cathode materials explored for calcium-ion batteries, Prussian Blue Analogues (PBAs) have showed significant potential due to their unique structural properties, as shown in figure a below. The open framework of PBAs supports fast ion transport and high-rate capabilities.
Previous research has mainly concentrated on analyzing the kinetics of calcium ions in PBAs electrodes using methods like Cyclic Voltammetry (CV) and Electrochemical Impedance Spectroscopy (EIS). However, these studies have not yet provided a comprehensive understanding of the calcium ion transport mechanisms in these materials. This gap in knowledge hinders the optimization of PBAs for calcium-ion batteries. Our work focuses on analyzing Ca ion transport in PBAs by employing various electrochemical characterization techniques. We synthesized CuHCF through co-precipitation method and assembled calcium aqueous battery, as shown in figure b and c above. A combination of CV, GITT, and EIS was used to investigate the chemical diffusion coefficients of calcium ions within the CuHCF framework. CV measurement revealed a diffusion coefficient of 1.26 × 10-6 cm2 s-1 for calcium ions, determined using the Randles-Sevcik equation from the slope of dI/dν1/2.The EIS measurements were conducted at the the end of each GITT cycle, as depicted in figure b below, to facilitate a direct comparison between the diffusion coefficients derived from both methodologies. Figure c below showed the comparison between the diffusion coefficients calculated through GITT and EIS. GITT analysis yielded coefficients ranging from 10-9 and 10-7 cm2 s-1, indicating rapid diffusion of calcium ions within the CuHCF framework. A notable decrease in the diffusion coefficients was observed during the discharge process, wherein Ca ions are intercalated into the CuHCF frameworks. This decrease may be attributed to the obstruction caused by already occupied sites, which impedes the further diffusion of Ca ions into the structure. In contrast, the EIS-derived diffusion coefficients remained relatively constant across varying voltages, with values primarily in the order of 10-9 cm2 s-1. After investigating the fundamental electrochemical properties and diffusion coefficients of the material through experimental electrochemical techniques, to further elucidate the transport mechanism of calcium ions within the CuHCF framework, we are employing semi-empirical computational tools, such as GFN-xTB, to provide a deeper understanding of the underlying electrochemical processes.
Electrocatalytic hydrogenation of biomass
Interest in production of fuels and chemicals from promising renewable resources, such as biomass continue to grow, as the environmental impact of fossil resources becomes more apparent. As an example, furfural is bio-derived chemical which can be converted to furfural alcohol and methyl furan by hydrogenation, which have applications in perfumery, polymer and pharmaceutical industries.
In this work, we aim to convert furfural to furfural alcohol (FAL) and methyl furan (MF) by electrocatalytic hydrogenation (ECH), which is depicted in Figure (a) above. The motivation is to build high surface area metal nanoparticle catalysts for the selective ECH of furfural. The preliminary tests for screening the activity of the metals were performed using bulk metal catalysts. Steady state voltammetry and electrolytic measurements were performed to analyze their performance on the basis of conversion, yield and Faradaic efficiency (Figure b below). The effect of various reaction parameters such as applied potential, electrolyte pH were also studied. Zinc stood out as an efficient metal catalyst in comparison to copper and nickel, achieving highest Faradaic efficiency (FE) for the ECH of furfural in neutral pH electrolytes (Figure b below). However, the yield of desired products was low based on the fractional conversion of furfural, which is possible due to electrodimerization of furfural. Oxidation of zinc was observed during electrolysis, only in the presence of furfural, suggesting that oxidized zinc may play a role in the reaction mechanism.
Zinc nanoparticles can help to increase the yield of FAL and MF while maintaining the same FE due to high surface area to volume ratio, which will aid in increasing the reaction rate for the conversion of intermediate radicals in ECH of furfural to FAL and MF instead of forming dimerized furfural product. Thus, we are working on to make zinc nanoparticle catalysts by electrodeposition and study it’s activity for ECH of furfural to achieve higher selectivity. The catalysts are characterized using SEM , XRD and nitrogen physisorption to determining morphology, crystalline structure, elemental composition, surface area, pore size and volume distributions. Catalyst characterization will help us in elucidating reaction mechanism.
References:
- M. S. Dhawan, G. D. Yadav and S. Calabrese Barton, “Zinc-electrocatalyzed hydrogenation of furfural in near-neutral electrolytes”, Sustainable Energy Fuels, 5, 2972 (2021). [doi:10/gj3gp7](http://doi.org/10/gj3gp7
Quantitative nanoscale scanning electrochemical microscopy
Nanoscale catalysts are used in high surface area systems as well as to increase intermediate transport efficiency. Experimental investigation of the local environment to these sites is crucial but challenging. Our group uses Atomic Force Microscopy based Scanning Electrochemical Microscopy (AFM-SECM) in aqueous systems to quantitatively analyze the local response for parameterization of nanoscale systems, which can vary from the bulk average response.
The in situ concentration profile differs from the in operando profile due to the presence of the tip. The physical alteration of chemical transport, as well as the kinetic activity of the SECM tip electrode, alter the in situ response as compared to the in operando. This can be seen in the animation below, where a conical SECM electrode traverses over an active nanoparticle at constant separation. Quantitative analysis through FEM inverse analysis allows for discussion of the in operando system from the in situ response. Through a combined experimental and simulation approach, our group investigates submicron active sites and their surrounding environment.
Our group investigates transport and kinetic mechanisms relating to cascade channeling. Sensing intermediates allow correlation to concentration profiles surrounding active sites and channeling pathways. Visualization of the environment of active sites increases insights into both transport and kinetic attributes. The correspondence of topographical Pt nanoparticles with their electrochemical response is illustrated below. Insights into transport efficiency of cascade intermediates helps optimize design of catalyst cascades.
References
- A. Mirabal and S. Calabrese Barton, “Numerical Correction of In Situ AFM-SECM Measurements”, Anal. Chem., 93, 12495–12503 (2021). doi:10/gmrwqq
- M. C. O. Monteiro, A. Mirabal, L. Jacobse, K. Doblhoff-Dier, S. Calabrese Barton and M. T. M. Koper, “Time-Resolved Local pH Measurements during CO 2 Reduction Using Scanning Electrochemical Microscopy: Buffering and Tip Effects”, JACS Au, 1, 1915–1924 (2021). doi:10/gm6t96
