From Biomimicking to Bioinspired Design of Electrocatalysts for CO2 Reduction to C1 Products

Abstract The electrochemical reduction of CO2 (CO2RR) is a promising approach to maintain a carbon cycle balance and produce value‐added chemicals. However, CO2RR technology is far from mature, since the conventional CO2RR electrocatalysts suffer from low activity (leading to currents <10 mA cm−2 in an H‐cell), stability (<120 h), and selectivity. Hence, they cannot meet the requirements for commercial applications (>200 mA cm−2, >8000 h, >90 % selectivity). Significant improvements are possible by taking inspiration from nature, considering biological organisms that efficiently catalyze the CO2 to various products. In this minireview, we present recent examples of enzyme‐inspired and enzyme‐mimicking CO2RR electrocatalysts enabling the production of C1 products with high faradaic efficiency (FE). At present, these designs do not typically follow a methodical approach, but rather focus on isolated features of biological systems. To achieve disruptive change, we advocate a systematic design methodology that leverages fundamental mechanisms associated with desired properties in nature and adapts them to the context of engineering applications.


Introduction 1.Fundamentals of CO 2 Reduction
Electrocatalytic reduction of CO 2 (CO 2 RR) offers a compelling avenue to address growing concerns about our carbon footprint, as it offsets carbon emissions and produces valueadded products. [1]It is a multiple proton and electron transfer reaction (Eq. 1) resulting in the formation of several products and water (Figure 1) where k, n, and m are the reaction coefficients (e.g., k = 1, n = 2, m = 1 when Product = CO; k = 2, n = 12, m = 3, when Product = CH 3 CH 2 OH, etc.). [2]If the reduction of CO 2 involves the transfer of two electrons and protons (n = 2), then the overall reaction is reversible, since the employed electrocatalyst can catalyze the corresponding reactions in both directions.On the contrary, if CO 2 RR occurs via the transfer of more electrons and protons (n > 2), then additional intermediates are formed and the overall reaction becomes irreversible, as the binding energy of each intermediate follows linear scaling relationships. [2]These scaling relationships are due to the presence of similar chemical bonds between adsorbed species and catalytic surfaces imposing a high overpotential for the reduction of CO 2 .
Examples include the scaling relationship between *OCHCH 2 , *OCHCH 3 , and *OCH 2 CH 3 intermediates for the reduction of CO 2 to CH 3 CH 2 OH, *OH and *OCHCH 2 intermediates for the reduction of CO 2 to C 2 H 4 , etc. [2]

Electrocatalyst Design: The Heart of The Challenge
Several noble (Au, Ag) and non-noble (Cu, Ni, Bi, Fe, Mn, etc.) metals, alloys, metal oxides, and transition metal chalcogenides are employed as CO 2 RR electrocatalysts. [3]mong them, copper is the most widely used electrocatalyst, since it can directly reduce CO 2 to hydrocarbons and alcohols, such as methane, ethanol, CO, and formate. [4]owever, Cu based electrocatalysts exhibit poor selectivity and large overpotentials, due to the complexity of the reaction pathway for hydrocarbon production involving several proton and electron transfer steps. [4]oble metal based electrocatalysts (such as Au, Pd, Ag, etc.) have also been used for the reduction of CO 2 to CO and formate. [5]Even though these electrocatalysts exhibit high selectivity towards the production of CO (FE CO > 90 % for Au and Ag based electrocatalysts) [6] and formate (FE formate > 97 % for Pd based electrocatalysts), [7] their high cost and scarcity makes them less attractive for large scale production.
Metal nitrogen doped carbon (MNCs) materials are a promising alternative to these electrocatalysts due to their stability and high selectivity toward CO 2 RR.4a] Further research is needed to improve these materials, as the origin of their activity is not well understood.Thus far, a very low FE CH4 ( � 0.4 %) [8] to produce CH 4 is achieved due to the inability of MNCs to co-adsorb CO* and H* limiting CO protonation. [9]8] From these reviews, it becomes clear that the lack of rational design principles for the development of highly efficient and selective CO 2 RR catalysts impedes disruptive progress in CO 2 RR devices.3a,d,10] Additionally, they should mediate multiple proton and electron transfers to CO 2 without resorting to excessive reducing overpotentials, leading to low energy efficiency. [11]Finally, it is crucial to embrace additional factors affecting their activity and selectivity, such as the local environment around the active sites (concentration of available CO 2 ) and the structure of the catalyst layer, which can influence the transport of reacting species. [12]n terms of local environment, the local pH at the interface between the catalyst and the electrolyte can significantly affect the selectivity of the electrocatalyst, since the overall reaction rate depends on the selected pH value (Eq.2). [13]4a,14] where R is the overall reaction rate, A is the reaction prefactor, q m X* is the coverage of species X*, G 0 a is the activation energy of the process at 0 V vs. standard hydrogen electrode (SHE), DG 0 is the free energy of the process at 0 V vs. SHE, U SHE is the potential vs. SHE, e is the electric elementary charge of an electron, n is the number of protonelectron transfers before the rate-limiting step, β is the transfer coefficient, k B is the Boltzmann constant, and T is the reaction temperature.For n = 0, the reaction rate depends only on U SHE . [13]he shift in pH values (À 2:3k ) creates different overpotentials favoring the formation of different products.A � À 71 mV and � À 357 mV shift in overpotential for C 1 and C 2 products, respectively, between pH = 7 and pH = 13 is reported in the case of Cu electrocatalysts. [13]The observed � 0.36 V shift in overpotential for C 2 products represents a two-orders of magnitude increase in C 2 selectivity compared to C 1 products in that pH range. [13]The C 2 production is limited by the rate of first proton-electron transfer to the O=CÀ C=O intermediate (Figure 1) and CO coverage at low and high overpotentials, respectively. [13]On the contrary, C 1 production is limited by the rate of protonelectron transfer to the CHOH intermediate, resulting in a smaller increase in activity with increasing pH, compared to C 2 formation. [13] similar trend is observed for noble metal based electrocatalysts to produce CO; the selectivity of Ag electrocatalysts (FE CO > 85 %) is enhanced at local pH values greater than the buffer pH ( � 7), [15] while for Au electrocatalysts, CO selectivity does not depend on local pH as it is limited by the CO 2 adsorption step.[16] For MNCs, pH plays an important role in their selectivity as well.High faradaic efficiency (> 80 %) toward CO production is achieved for FeÀ NÀ C electrocatalysts at high pH values (� 7), where the HER is suppressed.[17] Based on the above challenges, it is evident that a novel approach to design CO 2 RR electrocatalysts could drive accelerated innovation.Incremental changes through minor design modifications or alteration of the reaction conditions may not produce the required transformative solutions.An example of such approach is our nature-inspired chemical engineering (NICE) methodology [18] (discussed in more detail in Section 4) which recognizes universal fundamental mechanisms in nature underpinning desired properties (like scalability, efficiency, and resilience), which can be leveraged to achieve similar properties in an applied context.These ubiquitous mechanisms define NICE Themes, of which there are presently four: (T1) hierarchical transport networks; (T2) force balancing and nano-confinement; (T3) dynamic self-organization; (T4) ecosystems, networks, and modularity. For xample, (T1) includes ubiquitous networks that greatly promote reliably scalable performance across a wide range of length scales, as observed in trees, lungs, and the vascular network.Within or across these themes, NICE deploys a design methodology that comprises four stages to bridge nature with technology in applications, namely: nature-inspired concept, design, prototype, and application (Figure 2). Figure 2 presents an example of the application of the NICE methodology to design lung-inspired flow fields for proton exchange membrane fuel cells (PEMFCs).[19] Inspiration is derived from the lung ("Nature") due to its ability to scale-up irrespective of size, providing uniform distribution of oxygen into the blood stream while keeping the thermodynamic losses across its volume at a minimum.This is achieved via its fractal architecture, ensuring that the Péclet number, Pé, is close to 1, and, hence, the convective air flow dominating the upper part of the lung is equal to the diffusive air flow at the lower part of the lung ("Natureinspired concept").A mathematical model is built based on these characteristics of the lung to calculate the optimum number of fractal generations in a flow field to achieve Pé � 1 ("Nature-inspired design"), and then, lung-inspired flow fields are created via 3D printing ("Prototype") exhibiting higher performance ( � 30 % increase in current and power density) and � 75 % lower pressure drop than serpentine flow field based PEMFCs ("Application") minimizing the parasitic power losses.[19] In terms of CO 2 RR electrocatalysts, Nature can be an excellent guide to rational design, as it is full of biological organisms that efficiently catalyze the same reactions as the electrocatalysts and robust structures that are intrinsically scaling.Based on recent literature, we now review opportunities resulting from imitating or taking inspiration from nature to design better catalysts. We fcus on C 1 products, since they are important chemical feedstocks to produce fuels and value-added chemicals.

Biomimetic Electrocatalysts for CO 2 Reduction to C 1 Products
The most widely used source of inspiration for the design of novel electrocatalysts for CO 2 RR are metalloenzymes, which have exceptional catalytic efficiency and selectivity towards the same reactions occurring in electrochemical CO 2 reduction devices.Biomimetic design is focused on isolated features of biological organisms (such as their chemical structure), whereas bioinspired design considers both the structure and function of metalloenzymes recognizing the different context between the biological example and the technological application.

Enzyme-Mimicking Electrocatalysts
A successful biomimetic design approach is based on the presence of imidazole groups of histidine residues at the active site of carbon monoxide dehydrogenases (CODHs) facilitating proton transfer to and from the active site via the formation of hydrogen bonds with water molecules. [20]s a result, N,N-di(2-picolyl)ethylenediamine (DPEN), a source of imidazole groups, is incorporated into iron porphyrin, one of the most active and selective electrocatalysts for CO 2 to CO conversion in organic solvents, [21] and its properties are evaluated towards CO 2 RR in acetonitrile.The poly-pyridine/amine sites of DPEN form hydrogen bonds with water molecules assisting in proton transfer, as they function as multiple proton relays, and the protonated DPEN units are positively charged, stabilizing the negatively charged CO 2 RR reduction intermediates. [22]ence, the iron porphyrin with incorporated DPEN units is highly active towards CO 2 RR, exhibiting a TOF of � 5 • 10 4 s À 1 (acetonitrile (MeCN) electrolyte solution) for CO 2 to CO conversion with water as the proton source, four times higher than the activity of DPEN-free iron porphyrin. [22]nother efficient biomimetic electrocatalyst is pentlandite (Fe 4.5 Ni 4.5 S 8 ), due to its structural resemblance with CODHs, as it contains FeÀ Ni sites with a bond length of � 2.6 Å, similar to the ones in CODHs ( � 2.8 Å), connected via sulfur atoms. [23]The efficiency of CO 2 RR depends on the concentration of protons (the water content in the chosen solvent to conduct CO 2 RR) at the surface of the electrocatalyst.The selectivity of the electrocatalyst towards CO 2 RR increases as the water content decreases (MeCN electrolyte solution), and an � 87 % FE for CO is observed at low water concentration ( � 24 ppm H 2 O). [23]This preliminary result illustrates the potential of aprotic solvents with low water content as electrolytes for CO 2 RR. [23]

Alveolus-Mimicking Electrocatalysts
The efficiency of CO 2 RR electrocatalysts can also be improved via an increase of the local CO 2 concentration.The ratio of CO 2 to H 2 O molecules in an aqueous solution is � 1 : 1,300 at 1 atm pressure; this CO 2 concentration can be increased by increasing the pressure, [24] but this is only a temporary solution.To design such electrocatalyst with high gas permeability and low water diffusivity, the structure of the mammalian lung is imitated, where the alveoli are enclosed by several epithelial membranes ( � 1 μm thickness) with high gas permeability and low water diffusivity. [25]ence, an artificial alveolus is engineered from a flexible polyethylene (PE) membrane sputtered with a thin layer ( � 20 nm thickness) of gold nanoparticles (Figure 3).The PE membrane is hydrophobic and porous (pore radius � 40-500 nm) making it impermeable to water but allowing gas transport.Gold serves as the catalyst in this study as it is highly efficient towards the production of CO. [26] This Au/ PE composite membrane is rolled into a bilayer structure, and its bottom and top edges are sealed to form a closed pouch-type structure.The activity of this biomimicking electrocatalyst is evaluated in an H-type cell achieving � 92 % Faradaic efficiency of CO production (in a CO 2saturated 0.5 M potassium bicarbonate (KHCO 3 ) electrolyte solution). [26]gure 2. Systematic employment of the NICE methodology for the design and engineering of lung-inspired flow fields for proton exchange membrane fuel cells (PEMFCs).This figure is adapted from ref. [19].

Bioinspired Electrocatalysts for CO 2 Reduction to C 1 Products
Enzyme-inspired electrocatalysts derive their inspiration more broadly from the protein scaffold surrounding the active metal center of metalloenzymes, which is responsible for their high activity and selectivity.Their structural characteristic is the presence of primary and secondary (or outer) coordination spheres, which contribute significantly to the function of metalloenzymes.The primary coordination sphere is dominated by covalent interactions between ligands and metal ions; the number of ligands and the oxidation state of the metal affect spin-state ordering and reactivity. [27]The secondary coordination sphere represents the residues that do not directly bind to the active metal center but interact with the primary ligands via long-range interactions, modulating the catalytic properties of the metalloenzymes.

Dehydrogenase-Inspired Electrocatalysts
20a] Thus, to circumvent these issues, the engineering of electrocatalysts inspired by the structure of these enzymes is required.
MoÀ Cu and NiÀ Fe are commonly used as the active metal centers of these bio-inspired electrocatalysts for the reduction of CO 2 to C 1 products (formate and methane, respectively). [29]29c] Thus, the design of such electrocatalysts is extremely challenging, as there are several intertwined factors affecting their activity and selectivity.Their activity is also influenced by the presence of O atoms, and the position of a positively charged group or amide pendants in the outer coordination sphere, while their selectivity is influenced by the size of functional groups in the secondary coordination sphere (Figure 4a and b).
28a] In D100R mutant, aspartic acid is replaced by arginines, which are placed near the active metal center; this positively charged group attracts CO 2 , increasing its concentration near the metal center and, hence, enhancing the reaction rate.Even though the position of a positively charged group affects the activity of these artificial enzymes, there are several additional factors that must be considered as well: the accurate positioning of the charged group may be inhibited due to the presence of hydrogen bonds or salt

Minireviews
bridges, which prohibit the incorporation of residues near the active metal center and, thus, do not increase the reaction rate.Water distribution around the active metal center may influence the catalytic activity too, according to molecular dynamic simulations.28a] To examine the effect of the position of amide pendants in the secondary coordination sphere on CO 2 RR to CO (Figure 4a), iron tetraphenylporphyrin (Fe-TPP) derivatives bearing amide pendants at various positions (ortho-1, À 2, and para-1, À 2) at the periphery of the metal core are investigated. [30]Fe-TPP derivatives with proximal and distal amide pendants on the ortho-position of the phenyl ring demonstrate significantly higher TOF CO values compared to their counterparts with amide pendants on the para-position (TOF CO � 5.5 • 10 6 s À 1 and � 6.8 • 10 3 s À 1 for Fe-ortho-2-amide and Fe-para-2-amide, respectively using 0.1 M tetrabutylammonium hexafluorophosphate (TBAPF 6 ) in dimethylformamide (DMF) as electrolyte solution). [30]Fe-ortho-2amide has the smallest HÀ O distance ( � 1.45 Å) or largest through-space interactions, indicating that properly posi-tioned amide pendants increase the affinity for CO 2 via hydrogen bonding interactions. [30]part from amide pendants, amine pendants (NHÀ R) or phenolic groups can be introduced in the ortho positions of the phenyl ring resulting in � 98 % and � 90 % FE CO , respectively (0.1 M tetrabutylammonium tetrafluoroborate (n-Bu 4 NPF 4 ) in DMF as electrolyte solution). [31]31b] The presence of O atoms in the secondary coordination sphere (Figure 4a) also plays a role in the activity of the catalyst towards CO 2 RR. [32]Oxymethyl-ether (À OMe) and hydroxyl (À OH) pendant groups are incorporated into the outer coordination sphere of iron-bipyridine compounds and their activity towards CO 2 RR to formate is examined.32a] In terms of the selectivity of the bio-inspired electrocatalyst (Figure 4b), the size of the functional groups [33] and their combined tuning with the active metal center [34] are crucial parameters.
33b] The incorporation of bipyridine-modified ligands, in which two benzylic amines are positioned in the secondary coordination sphere of the active metal center can alter the selectivity of the electrocatalyst towards specific products of CO 2 reduction.These benzylic amines serve as proton transfer relays and form ([Metal]-H) units, which, in turn, create "formato" compounds ([Metal]-O 2 CH) upon interaction with CO 2 , resulting in the formation of formic acid (HCOOH).If the active metal center is rhenium (Re) or ruthenium (Ru), the binding of CO 2 onto the metal is more favorable than the formation of ([Re/Ru]-H) and it exhibits

Minireviews
34a] On the contrary, if Mn serves as the active metal center, then the formation of ([Mn]-H) compound is more favorable than the binding of CO 2 onto the metal and it results in high selectivity towards the production of HCOOH (FE HCOOH � 80 %, TOF > 4000 s À 1 , 0.2 M Bu 4 NBF 4 in MeCN electrolyte solution). [34]

Dehydrogenase-Inspired Electrocatalysts with Pseudo-Secondary Coordination Sphere
Apart from the utilization of different ligands for the formation of a secondary coordination sphere, another successful strategy is the employment of cationic buffers (such as imidazole, bicarbonate, phosphate, and triethanolamine) replacing the long-chain ligands in the outer coordination sphere of DHases.In this catalyst design, [Ni(cyclam)] 2 + (cyclam = 1,4,8,11-tetraazacyclotetradecane) is the active metal center which catalyzes the reduction of CO 2 to CO or formate via a proton-coupled, ECEC (electron transfer-chemical step-electron transfer-chemical step) pathway; its activity and selectivity are increased via pyridine-or imidazole-binding. [35]Imidazole preferentially binds to the Ni II of [Ni(cyclam)] 2 + , stabilizes the divalent oxidation state, and decreases the required reduction potential.Once [Ni(cyclam)] 2 + is reduced, imidazole buffer plays the role of histidine ligands in DHases, transferring protons to and from the bound substrate, and hence, acting as a pseudo-secondary coordination sphere. [35]he highest turnover frequencies (TOF CO ) of � 50 s À 1 (100 μM [Ni(cyclam)] 2 + with 100 mM potassium chloride (KCl) as electrolyte solution) are observed for imidazole (red diamond, Figure 5) and imidazole-derived buffers (purple diamond, piperazine, Figure 5).However, imidazole buffer exhibits the highest negative electrocatalytic overpotential for CO 2 reduction ( � 0.77 V vs. NHE), compared to bicarbonate buffer (blue diamond) demonstrating the least negative electrocatalytic potential of � 0.69 V (vs.NHE), similar to the value for the reduction of CO 2 to CO at pH = 7. [35b]

Application of NICE Methodology
The above examples do not explicitly use a systematic design framework, as the one provided by NICE, but instead represent a set of isolated inspirations from natural systems.
Our NICE methodology can address this issue, as it deploys a thematic approach that comprises four stages to bridge nature with technology in applications, namely: nature-inspired concept, design, and prototyping for experimental realization (Figure 2).
First, the nature-inspired concept, is identified as a mechanism found commonly in nature, which is the crucial ingredient to solve an issue in nature (e.g., the properties of the hierarchical transport network leading to scalability, within T1) that is also pertinent in the envisioned techno-logical application, despite the often different contexts of nature and technology (e.g., a lung and a fuel cell). [19]econd, the nature-inspired design stage translates this concept into a design that is specifically formulated for the intended application; hence, an (abstract) mechanism is adopted but its (concrete) realization needs to be adapted for technological use (e.g., the self-similar branching of the lung and the dimensioning of its channels for scalable, minimum entropy production, need to be adapted in the context of the different environment and production requirements of a hydrogen fuel cell flow plate, even though both require scalable, maximally efficient air distribution).
Finally, experimental realization is achieved by manufacturing prototypes based on the nature-inspired, often computationally assisted design through experimental testing and characterization to investigate performance under technologically relevant conditions.Prototyping is an iterative process that embraces new synthesis and (e.g., digital and additive) manufacturing techniques.
The design of bio-inspired CO 2 RR electrocatalysts reported in the literature relates to the T2 theme of NICE methodology (Section 1.2.), namely force balancing, achieved through nano-confinement.The employment of a secondary sphere to modulate active metal center-ligand interactions or forces is indeed a successful strategy (Section 3.1.)for the enhancement of their catalytic activity and selectivity (Figure 6).

Minireviews
Are there any additional elements in the design space that could further improve the properties of bioinspired CO 2 RR electrocatalysts?NICE could be employed as an innovation accelerator for the design and development of such materials by following each stage of its methodology, and recognising universal aspects captured by the NICE Themes.This allows us to avoid fragmentation in case-bycase catalyst developments.
18b] This structural geometry provides control over the chemical environment where the reaction occurs and the transport of products to new active sites for cascade (substrate channeling).A recent example is an ORR electrocatalyst (PtNi) with similar structure to metalloenzymes: [36] isolated substrate channels are formed between the surface and the center of the nanoparticles, while their exterior surface is passivated by a surfactant to ensure that the electrochemical reactions take place in nanoconfined substrate channels.36a] The diameter of their substrate channels greatly impacts the activity of these bioinspired PtNi nanoparticles.36b] Thus, an electrocatalyst consisting of an optimal combination of nanochannels (between 1-4 nm diameter) to leverage nano-confinement, a hierarchical transport network to minimize transport limitations, [37] and a coordination sphere to improve the properties of its active metal center could be highly active and selective towards CO 2 RR.Computationally assisted models (nature-inspired design), a step that is often neglected in the reported literature, should be developed first to aid in the design of bio-inspired electrocatalysts.Molecular modelling could help fundamental understanding to achieve customized selectivity through cascade reactions, lowering the free energy barrier of CO 2 RR and, hence, its overpotential. [1,38]Synthesis procedures should then be carefully chosen to create these CO 2 RR electrocatalysts and evaluate their activity, stability, and selectivity in a CO 2 RR device (prototyping for applications).

Conclusions and Outlook
In summary, nature is an ideal source of inspiration for the design of artificial CO 2 RR electrocatalysts, as there are plenty of biological examples that efficiently catalyze the same reactions and have robust structures, deployed at scale.The examples presented in Sections 2 and 3 demonstrate that bioinspired design prevails over narrow biomimetics (or bio-imitation), since it considers the structure and function of the biological example and the different context between nature and technological applications.However, thus far, most bioinspired examples represent a set of isolated inspirations from natural systems leading to non-optimal Figure 6.Systematic methodology for the design of nature-inspired CO 2 RR electrocatalysts.The unit of TOF max values is s À 1 .The figures under "Prototype" and "Application" are reproduced from ref. [30] with permission from the Royal Society of Chemistry.
3a,38-39] The adoption of NICE could lead to a more systematized bioinspired design strategy and accelerate the development of highly efficient bioinspired CO 2 RR electrocatalysts.The NICE methodology could become even more effective when paired with further advances in the fundamental understanding of the CO 2 RR mechanism.3a] Finally, synthesis protocols viable for mass production must be developed to enhance the low yield of electrocatalysts prepared by conventional approaches.
Apart from electrocatalysts, the commercialization of this technology is also contingent upon the significant improvement of the design of CO 2 RR devices to enhance their energy efficiency (i.e., high activity and selectivity at low overpotential).Mass transport limitations within the device must be resolved, while product separation and tolerance towards gas inlet purity must be improved. [39]ature provides examples that can be leveraged to circumvent these issues, through its intrinsically scaling hierarchical transport networks.

Figure 1 .
Figure 1.Overview of possible CO 2 reduction pathways for C 1 and C 2 + products.