Structure-Activity Relationship of Defective-Graphene-Based FeN_(4) Catalyst for Hydrogen Fuel Cells

Abstract

A. Fuel Cells Fuel cells have several benefits over conventional combustion-based technologies currently used in many power plants and passenger vehicles. Since concerns about clean energy resources have been elevated, fuel cells have received considerable attention because of the clean emissions in comparison with petroleum-based fuels which are on the brink of depletion. When hydrogen is fed as fuel, the fuel cell only generates electricity, water, and some heat. There are no carbon dioxide emissions and no air pollutants that create smog and cause health problems at the point of operation. In addition, fuel cells can operate at higher efficiencies than combustion engines, converting chemical energy in the fuel into electrical energy directly with efficiencies up to 60%. Thus, fuel cells will help solve the global problems of energy supply and clean environment. Fuel cells are expected to be used in a wide range of applications, including transportation, portable power generation and so on. A fuel cell is an electrochemical cell that converts the chemical energy of a fuel and an oxidizing agent into electricity through a pair of redox reactions. There are many types of fuel cells, and they all consist of an anode, a cathode, and a electrolyte that allows ions, often positively charged hydrogen ions (protons), to move between the two sides of the fuel cell. Among all the existing fuel cells, the proton exchange membrane fuel cell (PEMFC) has been actively developed for use in vehicles, portable electronics, and combined heat and power systems due to its simplicity, low working temperature, high power density, and quick start-up. In PEMFC, thin polymer membranes are used as electrolytes and protons act as carriers that carry charges within the membrane. PEMFCs are especially well suited as the main power sources for automobiles and buses. Fuel cell vehicles (FCVs) have been considered as one of the final solutions for automotive business and have profound advantages over battery powered electric vehicles. Indeed, the first mass produced FCVs, the Toyota Mirai, have been commercially sold in Japan since 2014 and are going to be available in North America in 2015 at a price of ∼57 000 US dollars. One of the main reasons for the high sale price of the Mirai is the high Pt loading in the fuel cell stacks. Recently, nearly all commercial hydrogen fuel cell systems use catalysts prepared from platinum supported on carbon materials (Pt/C). Both the anode and the cathode electrodes consist of highly dispersed Pt-based nanoparticles on carbon black to promote the reaction rates of the hydrogen oxidation reaction (HOR) and the oxygen reduction reaction (ORR). At the cathode, oxygen is reduced by reaction with protons and electrons to produce water (1/2O_(2) + 2H^(+) + 2e^(−) → H_(2)O), while hydrogen is oxidized (H → 2H^(+) + 2e^(−)). At the same time, electrons and protons are produced that are transferred to the cathode through an external circuit and the proton exchange membrane, respectively. However, the high cost of platinum limits the wide-scale application of Pt-based hydrogen fuel cell systems. This problem is related to the low rate of the cathode reaction where oxygen is reduced. The reaction rate of HOR on Pt is fast enough so that the Pt loading at the anode can be reduced to less than 0.05 mg cm^(-2). On the other hand, the sluggish reaction kinetics of ORR at the cathode requires considerably higher Pt loading (~0.4 mg cm^(-2)) to achieve a desirable fuel cell performance. Pt is a scarce and expensive metal and Pt-based catalysts contribute by about 25% to the overall fuel cell costs. Therefore, reducing its loading or even completely replacing it with an abundant and cheap metal would be a key in the design of an eco-friendly and energy-efficient fuel cell. B. Non-Precious FeNC Catalyst There are two approaches to drop the cost of fuel cells: development of (i) less expensive Pt-based catalysts which includes advanced Pt alloys, core-shell structures, transition metal oxides and chalcogenides, and carbon-based non-noble metal composite catalysts, and (ii) alternative Pt-free electrocatalysts that exhibit higher activity and incur lower production cost than Pt-based catalysts. In the case of the latter, non-precious transition metals (e.g. Fe or Co) and nitrogen (N) co-doped carbon (denoted as M-N-C) have shown great promise as Pt-free electrocatalysts for PEMFCs. In particular, the FeN_(4) moieties embedded in carbon networks, denoted as FeNC catalyst, represented the Pt-free catalyst, which has exhibited catalytic activity approaching to that of Pt/C catalysts in acid. Lefevre et al. reported that the current density of iron-based catalyst for oxygen reduction reaction (ORR) was equal to that of Pt catalyst at a cell voltage of ≥0.9 V. Moreover, a non-precious metal catalyst derived from polyaniline, iron, and cobalt was found to catalyze the ORR at onset potentials within ~60 mV of that delivered by a state-of-the-art Pt/C catalyst. The FeNC catalysts have been extensively studied as a cathode material to replace Pt/C catalyst. The structure of the FeN_(4) moiety is similar to that of heme, an iron-porphyrin complex, which is contained in many proteins as active sites closely related to O_(2), such as hemoglobin and catalase. In the respiration of aerobic organisms, O_(2) is reduced to H_(2)O by catalysis of a heme-containing enzyme, generating the energy-rich molecule (adenosine triphosphate, ATP). The active site of the enzyme consists of one of heme and Cu(I) coordinated by three histidines through their nitrogen atoms and situated ~0.5 nm apart from the heme. It has been known for a long time that cathodic O_(2) reduction is catalyzed by heme adsorbed on an electrode surface or carbon particles used as catalyst supports in contact with electrodes in acidic, neutral, or alkaline electrolytes. The O_(2) molecule receives protons from the electrolyte and electrons from the electrodes through the heme. This phenomenon implies the ability of heme as an active site to reduce O_(2) in the absence of an environment of protein structures. The FeNC catalyst can be prepared through the direct pyrolysis of a mixture of elemental precursors containing carbon, nitrogen, and iron at high temperature. The synthesized FeN_(4) active sites are responsible for the high catalytic activity toward the ORR due to the characteristic interaction between iron and oxygen. Both X-ray absorption near edge structure near edge structure and aberration-corrected scanning transmission electron microscopy characterization studies revealed that the FeN_(4) sites in carbon layers are most possibly responsible for the observed ORR activities of the FeNC catalysts. The ORR activity of the FeNC catalyst has been known to be affected by changes in the surrounding carbon network as mentioned above. In addition, a wide range of ORR activities have been reported for the FeNC catalysts prepared under different synthesis conditions, including precursor types and concentrations, and varying thermal treatment conditions. This range of measured ORR activities indicated that there are significant changes in the surrounding chemical environment around the Fe active site depending on the synthesis conditions. Among the many possible variations in the surrounding chemical environment of Fe active sites, a disordered carbon network, such as a Stone–Wales defect, interstitial, vacancy, or dislocation, can be introduced during the synthesis process. In fact, special processes are required to produce defect-free graphene materials, and most mass-produced graphene structures include native defects. Likewise, the synthesis of FeN_(4) containing graphene is accompanied by the formation of native defects in the carbon network. To improve the performance of the FeNC catalysts, atomic-level engineering based on a fundamental understanding of the reaction mechanism is required. A detailed understanding of the correlation between the chemical environment and ORR activity of FeNC catalysts can provide insights into the design of the FeNC catalysts with high ORR activity. Many experimental and theoretical studies have attempted to determine the structure–activity relationship of the FeNC catalysts. In particular, theoretical density functional theory (DFT) calculations are typically employed to investigate the catalytic activity of an FeN_(4) moiety embedded in a basal plane, or edges of perfect graphene, to provide fundamental insights into the ORR mechanism by identifying possible reaction pathways and energetics. For Pt-based catalysts, first-principles DFT calculations have been successfully applied to predict the energetics, evaluate the activity, and elaborate on the catalytic mechanisms of ORR on their surfaces. These calculations could complement experimental analysis which had have difficulty in specifying the structure of the active sites and figuring out the clear correlation between the heat treatment conditions and the properties of active sites due to the variations of macroscopic structure or conductivity. Given that graphene is susceptible to defects, detailed assessment of changes in the ORR activity of FeNC catalysts in response to the systematic variations of carbon networks is needed. In this study, we performed a series of first-principles calculations and demonstrated that structural changes in the carbon network around the FeN_(4) moiety resulted in significant perturbation in the electronic structure of the Fe active site and the corresponding ORR activity. To this end, structural variations in the carbon network around the FeN_(4) moiety were introduced based on the systematic vacancy defects of graphene materials. Vacancies are believed to be the predominant defects in graphene and can be formed during growth or upon irradiation. Among a variety of the vacancy defects, we chose divacancy (V_(2)) defects to be introduced because they are not only the simple vacancies that can be present in graphene, but also thermodynamically more stable than single vacancies. A single vacancy has dangling bond that always remains due to geometrical reasons when the odd number of carbons are missing from the perfect graphene. These V_(2) defects include V_(2)(5-8-5), V_(2)(555-777), and V_(2)(5555-7777) previously reported by Lee et al. For ideal- and defective-FeNC models, we evaluated the binding free energies of key reaction intermediate species (O_(2), OOH, O, and OH) for ORR and converted them to ORR activity to reveal the intrinsic structure–activity relationship. The *OH binding free energy proved to be significant in determining ORR activity, resulting in a clear volcano plot. It was found that the defective carbon network makes a negative contribution to the ORR activity, and that the ideal-FeNC model showed the best activity. Interestingly, the robust linear scaling relationships between the binding free energies of the reaction intermediates restricted the ORR activity window. Thus, a more powerful chemical perturbation to the Fe active site is necessary to break or shift the existing correlation, thereby enhancing the intrinsic activity of the FeNC catalyst system. This fundamental understanding provides an important insight into how to control the atomic structure of Fe active sites to optimize the performance of the FeNC-based fuel cell catalysts.;고분자 전해질 연료전지는 효율이 높고 유해물질 및 온실가스 유발 물질 배출이 거의 없으며, 상용화에 가장 근접해 있는 기술이다. 고분자 전해질 연료전지의 효과적인 상용화를 위해서는 연료전지의 양극(cathode)에서 일어나는 산소 환원 반응(oxygen reduction reaction, ORR)에 대한 낮은 촉매의 활성 및 내구성 문제가 해결되어야 한다. 현재 ORR 촉매로 가장 많이 사용되는 백금 기반 촉매는 고비용 가격 문제가 있으므로 이를 해결하기 위해 비귀금속 촉매의 활성을 증진시켜 백금을 대체하도록 하는 연구가 활발히 이뤄지고 있다. 여러 비귀금속 촉매 후보들 중, FeNC 촉매는 백금 기반 촉매를 대체할 만한 물질로서 각광받아 왔다. FeNC 촉매란 탄소 네트워크에 FeN_(4) 활성점이 박혀있는 구조의 단원자 촉매를 말하는데, 헤모글로빈의 포르피린 고리와 흡사한 구조를 가지며 헤모글로빈과 마찬가지로 산소가 활성점에 가역적으로 흡착하기 때문에 산소 환원 반응에 대한 활성이 좋아 이에 대한 많은 연구가 진행되었다. FeNC 촉매는 일반적으로 Fe, N, C가 포함된 전구체를 열처리하여 합성할 수 있다. FeNC 촉매를 다룬 기존 이론 연구들은 결함이 없는 (perfect) 그래핀에 FeN_(4) 활성점이 embedded된 이상적인 모델을 기반으로 이루어져 왔다. 그러나, 합성 과정에서 Fe 활성점 주변의 탄소 네트워크에 결함이 생길 확률이 높으며 이러한 현상들이 실험적으로 발견되어 왔다. 따라서 본 연구에서는 원자/분자 스케일의 밀도범함수(density functional theory, DFT) 기반 계산화학 시뮬레이션을 통해 탄소 네트워크의 다양한 결함이 FeN_(4) 활성점의 전자구조 및 ORR 활성에 미치는 영향을 분석함으로써 열처리 조건과 ORR 활성 간의 근본적인 관계를 밝히고자 하였다. 본 연구진은 그래핀에서 흔하게 발생할 수 있는 결함 중 가장 간단한 divacancy defect를 가진 3개의 그래핀 베이스 모델(V_(2) 5-8-5, 555-777, 5555-7777)을 다루었다. 결함 주변에 FeN_(4) 활성점을 도입하여 총 12개의 defective model을 구성하였다. Ideal model을 포함한 총 13개의 모델에서 중간반응물의 흡착에너지를 계산한 결과, 흡착 에너지 사이에 선형관계가 존재하는 것을 밝혔으며, 이를 통해 FeN_(4) 활성점의 전자구조가 주변의 결함에 의해 systematically 변화한다는 것을 알 수 있었다. 중간반응물 중에서도 흡착물 OH는 FeN_(4)의 전자구조에 매우 responsive하기 때문에, ORR 이론전위값을 OH의 흡착 에너지에 대해 플롯하였다. 그 결과 volcano plot 형태를 보이며, ideal model에서 가장 좋은 활성이 도출되는 것을 확인하였다. 이는 활성점 주변의 결함이 ORR 활성에 부정적인 영향을 미치는 것을 의미하며, 따라서 FeNC 촉매가 최대의 활성을 내기 위해서는 합성 과정에서 완전히 탄소화(graphitization)된 탄소 네트워크가 생성되도록 적절한 열처리가 필요함을 시사한다. 실제로, FeNC 촉매의 열처리 조건에 따라 ORR 활성이 조절되는 실험 연구 결과들이 보고되어 왔다. 일반적으로 높은 온도(800 C)에서 충분한 탄소화가 이루어질 때 활성이 최대화된다고 보고하고 있고, 열처리 조건에 따라 활성점 주변의 local carbon/nitrogen network의 변화를 평균적인 XPS 분석을 통해 예측하고 있다. 실험 분석을 통해 활성점 구조를 특정하기 어렵고, 실험적 활성에는 활성점의 intrinsic activity뿐만 아니라 열처리에 따른 거시적인 입자 구조 및 전도성의 변화 등 다양한 요인이 작용하기 때문에 열처리 조건과 활성점 구조간의 상관관계를 명확하게 규명하기 어렵다. 본 이론 기반 연구를 통해 ideal FeN_(4)의 intrinsic activity가 가장 우수하다는 결론을 낼 수 있었고, 이는 실험적인 최적 열처리 조건에서 ideal FeN_(4)가 높은 확률로 형성되는 것과 크게 연관된 결과임을 추측할 수 있었다. 또한, volcano plot을 O_(2)와 OH의 흡착 에너지에 대한 2D 활성 맵으로 확장시킨 결과, 선형관계가 존재할 때 얻을 수 있는 최대 전위값이 0.68 V이고 ideal model의 값이 이에 이미 근접한 것을 확인하였다. 이를 통해, 탄소 네트워크 구조 변화를 통해 FeN_(4) 활성점의 전자구조를 튜닝함으로써 산소 환원 반응의 활성을 높이는 전략에는 활성 향상의 마진에 한계가 있음을 알 수 있었다. 이는 중간반응물들의 흡착 에너지 간의 견고한 linear scaling relationship이 존재하기 때문이다. 네 개의 전자 전달 반응으로 진행되는 ORR의 열역학 전위(thermodynamic potential)은 각 단위 전자 전달 단계의 반응에너지들 중 가장 less favorable한 단계에 의해 결정되고, 각 반응에너지는 연관된 반응중간체의 흡착에너지로 결정된다. 일반적으로 반응중간체의 흡착에너지 사이에 linear scaling relationship이 존재하면서 특정 반응중간체를 선택적으로 안정하게 또는 불안정하게 만들기 어렵기 때문에 ORR 열역학 전위의 향상에 한계가 발생하게 된다. 예를 들어, *OO → *OOH 단계가 과전위 발생 단계인 경우, 이 단계의 반응에너지를 낮추기 위해서는 *OO에 비해 *OOH를 상대적으로 안정화시켜야 한다. 그러나 두 중간반응체 사이에 linear scaling relationship이 존재한다면 활성점의 전자 구조 변화에 따라 *OOH가 안정화될 때 *OO도 같이 안정해지므로 과전위를 낮추는 데에 한계가 발생하게 된다. 따라서 FeNC 촉매의 활성을 ideal model 이상으로 향상시키기 위해서는 이러한 linear scaling relationship을 깰 수 있는 특별한 구조의 활성점 설계에 대한 연구가 향후 이루어져야 할 것이다.