Carbon Deposition Process Research Articles

The carbon deposition process on NiFe catalyst in CO methanation: A combined DFT and KMC study

Carbon deposition often occurs in CO methanation on NiFe catalyst, resulting in the deactivation of the catalyst, which is an urgent problem to be solved in industrial production. In this work, the carbon deposition process on the surface of NiFe catalyst was systematically investigated by the combination of Density Functional Theory (DFT) and Kinetic Monte Carlo (KMC). The results of DFT calculations show that for C1 ∼ C9, the most stable adsorption configurations are carbon chains while for C10 and C11, the carbon rings are the most stable. Besides, in the reaction network of carbon deposition, the formation of C2 is the key reaction with high activation energy. KMC simulation shows that temperature can significantly affect the occurrence of carbon deposition and C2 will accumulate and react to form larger carbon species instead of decomposing. In addition, the growth rate of carbon chains is faster than that of carbon rings, and carbon species mainly react with C1 and C2 to form longer carbon chains. Therefore, the anti-carbon deposition performance of the catalyst can be improved by increasing the activation energy of carbon–carbon coupling and inhibiting the formation of C2.

Development of a dynamic gas lock inhibited model for EUV-induced carbon deposition.

The optical surface of extreme ultraviolet (EUV) lithography machines is highly vulnerable to contamination by hydrocarbons, resulting in the formation of carbon deposits that significantly degrade the quality and efficiency of lithography. The dynamic gas lock (DGL) has been proven as an effective approach to alleviate carbon deposition. However, the majority of existing studies on carbon deposition neglect the influence of the DGL. This paper is dedicated to investigating the phenomena of hydrocarbon adsorption, desorption, and cleavage with considering the effects of the DGL. A comprehensive mathematical model of the carbon deposition process is established, and the impact of radiation intensity, temperature, and hydrocarbon types on the depositing rate is considered. The results suggest that the primary cause of carbon deposition is the direct cracking of hydrocarbons induced by photons with a wavelength range between 12.5 and 14.5nm. Additionally, it has been observed that the carbon deposition rate decreases exponentially as clean gas flow increases when EUV radiation intensity exceeds 50 mW/mm2. Conversely, at low EUV radiation intensity, clean gas flow has little effect on the carbon deposition rate. An effective approach to mitigate carbon deposition is to elevate the temperature of the optical surface and employ light hydrocarbon materials in the EUV process.

The influence of the hydrometeorological factors on the CO2 fluxes from the oligotrophic bog surface.

The influence of the hydrometeorological factors on the CO2 fluxes from the oligotrophic bog surface.

Unraveling the effect of particle size of active metals in Ni/MgO on methane activation and carbon growth mechanism.

For dry reforming of methane, the active metal particle size of the catalyst has a significant effect on both the reaction activity and the resistance to carbon deposition. In this study, nickel particles of different sizes (Ni13, Ni25, and Ni37) supported on the MgO(100) slab are used to study the mechanism of CH4 activation and carbon growth based on DFT theoretical calculations. According to the results, the energy of adsorption for reaction intermediates changes depending on the size of the active metal. The adsorption process of CH3, CH2, CH and C on Ni25/MgO has a maximum exothermic value. Furthermore, the energy barriers of CH4 four-step dehydrogenation are lowest on Ni25/MgO during the CH4 activation process. The growth process of carbon deposition on the catalysts is also investigated in this work. The results indicate that the growth of carbon from C5 to C6 is difficult to proceed on Ni13/MgO due to size and active site limitation. Additionally, with an increase in particle size of the active metal, the absolute value of growth energy and average carbon binding energy of Cn increase on both Ni25/MgO and Ni37/MgO. It is proved that smaller particle size presents better resistance to carbon deposition. In the studied size range, Ni25/MgO is demonstrated to have greater catalytic activity and better resistance to carbon deposition.

The base versus tip growth mode of carbon nanotubes by catalytic hydrocarbon cracking: Review, challenges and opportunities

Hydrogen gas production using catalytic hydrocarbon cracking on metal nanoparticles has become a vital bridge technology, as this process, unlike the traditional methane-based reforming process, does not co-produce carbon gasses (e.g., CO, CO2). The major benefit of direct catalytic methane cracking using a nanostructured catalyst is the formation of solid carbon in high-value products, such as carbon nanotubes (CNTs) or filaments. This solid carbon can then be removed physically and valorized. However, it is challenging to design a catalyst capable of sustaining its activity after solid carbon has started to deposit and grow, meanwhile preventing the formation of coke. For CNTs, in particular, the base growth mode of CNTs is the desired pathway, as the catalyst can then be regenerated and re-used. If the CNTs are formed under the tip-growth mode, the catalyst particle will be lifted off the support of the regeneration and reusability is lost. Therefore, the study of CNTs growth modes is a vital topic, both experimentally and theoretically, of designing appropriate catalysts. To date, enormous efforts have been made to investigate conditions where the CNTs base growth mode can be maintained during the carbon deposition process. Several possible correlations and mechanisms regarding the base growth mode have been explored and established. Theoretical calculations and numerical simulations across length scales were conducted to investigate the nucleation mechanism and growth of CNTs. Density functional theory (DFT), classical and quantum-based molecular dynamics (MD) and Monte Carlo (MC) simulations were also carried out to study the initial CNT cap formation and its encapsulation during the catalytic hydrocarbon cracking process. A thermodynamics-based nucleation formula of CNTs for both the base and tip growth modes was also established. However, there is still no consensus on what determines the CNT growth modes and what the roles are of the various influencing factors such as the nanoparticle size, the oxidation state of the catalyst, the mechanical properties of CNT, and the catalyst-support interaction. This paper reviews the current status of CNT's growth mechanism development. The benefits and limitations of theory and modeling approaches concerning CNT growth modes are discussed.

Oxidized graphene fabrication from oil palm shell at different exposure time

Biochar made from Oil Palm Shell (B-OPS) can release bio-gaseous compounds such as methane and carbon dioxide during thermal pressure. A percentage of these bio-gaseous compounds can be used as precursors in the carbon deposition process during Chemical Vapor Deposition (CVD) to synthesize oxidized graphene. By manipulating the exposure time during the CVD process, one can observe varying nucleation, growth, and exfoliation differences. The highest exposure time results in exfoliation, observable crystallinity, high carbon content, low defects, and a reduction in oxygen within the intercalated layers. Furthermore, oxidized graphene synthesized at different exposure times can be characterized using various methods such as Atomic Force Microscope (AFM), Scanning Electron Microscope – Field Emission Gun (SEM- FEG), Energy Dispersive X-ray (EDX), micro-Raman, X-Ray Diffraction (XRD), and X-ray Photoelectron Spectroscopy (XPS). Observing the carbon characteristics at 30-minute intervals allows one to speculate on the relationship between carbon deposition and exposure time. As the exposure time increases, the initial amorphous carbon layer deposition changes to a higher sp2 hybridized layer deposition. G-65 has shown optimized properties, such as a high hybridization trend, a high surface area porous-wavelike structure, an observable crystallinity, and the highest sp2 compared to sp3 deconvoluted peaks.

Identification of internal polarization dynamics for solid oxide fuel cells investigated by electrochemical impedance spectroscopy and distribution of relaxation times

Reliability and durability are two major problems restricting for Solid oxide fuel cells (SOFCs) industrial applications. To ensure the long-term stability, understood the internal operation mechanism of SOFC was necessary. The polarization curve combined with the distribution of relaxation times (DRT) method to analyze the electrochemical impedance spectroscopy (EIS) was used to identify the SOFCs modulation and degradation effect through the excitation frequency. Dynamic evolution of carbon deposition process and specific time constants were determined employing EIS and DRT. Special attention was paid to the dependency between the complex reaction processes that occur during the different operation of SOFC. The corresponding frequencies of concentration and activation polarization resistances in EIS were investigated under various gas compositions. The high concentration polarization resistance caused by the addition of nitrogen and hydrogen starvation was the key limiting factor for high cell power density. The total polarization resistance increased from 1.50 Ω/cm2 to 5.26 Ω/cm2 when the H2:N2 ratio decreased from 1:0 to 1:5. The anode did not deteriorate due to the poisoning of Ni particles in a short-term tested under 0.5 A/cm2 current density when the fuel was methane.

Investigating the influences of metal-support interaction in Ni[sbnd]Fe catalysts on the quality of carbon nanomaterials from waste polypropylene

Pyrolysis-catalysis is a promising method to realize the recycling of waste plastic. Thereinto, metal-support interaction (MSI) has a crucial impact on the properties and performances of catalysts. Different active metal components (Ni or/and Fe) and pre-reduction process were introduced to adjust the metal-support interaction in this work. The results showed that moderate MSI was obtained from NiFe while Ni presented the weaker MSI and a stronger one was revealed for Fe. After pre-reduction, the promotion of MSI was observed for all catalysts, leading to the increase of active metal particle size. For the carbon deposition process, a higher CNMs yield (44.4 wt%) was obtained from NiFe, followed by 30.0 wt% of Fe and 18.2 wt% of Ni. The moderate MSI and synergistic effects of two metal species contributed to the CNMs production. However, different results were observed after fresh NiFe catalysts were reduced. Less carbon was formed on reduced monometallic Fe or bimetallic NiFe catalyst. This might be ascribed to the stronger MSI promoted by the pre-reduction of Fe-containing catalysts, inhibiting the catalyst activity. In addition, the strengthened MSI made the CNMs growth mechanism shift from tip-model to base-model and the graphitization degree of CNMs was also decreased. Our study may provide specific guidance for the production of high-valued CNMs recycling from waste plastics.

Catalytic Upgrading of Biomass-Gasification Mixtures Using Ni-Fe/MgAl2O4 as a Bifunctional Catalyst.

Biomass gasification streams typically contain a mixture of CO, H2, CH4, and CO2 as the majority components and frequently require conditioning for downstream processes. Herein, we investigate the catalytic upgrading of surrogate biomass gasifiers through the generation of syngas. Seeking a bifunctional system capable of converting CO2 and CH4 to CO, a reverse water gas shift (RWGS) catalyst based on Fe/MgAl2O4 was decorated with an increasing content of Ni metal and evaluated for producing syngas using different feedstock compositions. This approach proved efficient for gas upgrading, and the incorporation of adequate Ni content increased the CO content by promoting the RWGS and dry reforming of methane (DRM) reactions. The larger CO productivity attained at high temperatures was intimately associated with the generation of FeNi3 alloys. Among the catalysts’ series, Ni-rich catalysts favored the CO productivity in the presence of CH4, but important carbon deposition processes were noticed. On the contrary, 2Ni-Fe/MgAl2O4 resulted in a competitive and cost-effective system delivering large amounts of CO with almost no coke deposits. Overall, the incorporation of a suitable realistic application for valorization of variable composition of biomass-gasification derived mixtures obtaining a syngas-rich stream thus opens new routes for biosyngas production and upgrading.

Unveiling the existence and role of a liquid phase in a high temperature (1400 °C) pyrolytic carbon deposition process

Based on experiments aiming at depositing pyrolytic carbon onto carbon nanotubes by means of a chemical vapor deposition (CVD) process at 1400 °C, the work we report here demonstrates that the deposition involves the transient formation of a liquid phase (here organic), and that wetting physics is still able to apply in spite of the high temperature and the nanoscale, i.e., far beyond the condition range usually investigated. This was unexpected, because the high temperature makes all the physical and chemical processes involved transitory, including the fact that the liquid turns itself into solid carbon because of the ongoing carbonization process. The observations provide an estimate of the time scales of the involved processes, as they have to be short enough to complete in spite of the high temperature conditions. From a practical point of view, as the resulting material is a solid, all-graphenic carbon, the work demonstrates that using wetting physics in high temperature CVD can be a way to dramatically modify the surface energetics and the nano/microscale morphology of carbon nanofilaments. The statement is assumed to also apply to other chemical systems.

Analysis of Soot Deposition Mechanisms on Nickel-Based Anodes of SOFCs in Single-Cell and Stack Environment

Solid oxide fuel cells (SOFCs) can be fueled with various gases, including carbon-containing compounds. High operating temperatures, exceeding 600 °C, and the presence of a porous, nickel-based SOFC anode, might lead to the formation of solid carbon particles from fuels such as carbon monoxide and other gases with hydrocarbon-based compounds. Carbon deposition on fuel electrode surfaces can cause irreversible damage to the cell, eventually destroying the electrode. Soot formation mechanisms are strictly related to electrochemical, kinetic, and thermodynamic conditions. In the current study, the effects of carbon deposition on the lifetime and performance of SOFCs were analyzed in-operando, both in single-cell and stack conditions. It was observed that anodic gas velocity has an impact on soot formation and deposition, thus it was also studied in depth. Single-anode-supported solid oxide fuel cells were fueled with gases delivered in such a way that the initial velocities in the anodic compartment ranged from 0.1 to 0.7 m/s. Both cell operation and post-mortem observations proved that the carbon deposition process accelerates at higher anodic gas velocity. Furthermore, single-cell results were verified in an SOFC stack operated in carbon-deposition regime by dry-coupling with a downdraft 150 kWth biomass gasifier.

Anti-carbon deposition mechanism of H2O on nickel-based anode of SOFC: A ReaxFF molecular modelling

Carbon deposition occurs when Dimethyl ether (DME) fuel is used for SOFC, leading to battery degradation. In order to study the effect of water addition on carbon deposition, this work used reactive force field molecular dynamics (Reaxff MD) to simulate the process of carbon deposition with or without water addition, and analyze its anti-carbon deposition mechanism on nickel-based anode.It is found that the number of carbon atoms on nickel can be effectively reduced by mixing water with fuel. As the H2O/DME ratio increases, there are fewer carbon atoms on the nickel anode. And there are two main ways for water molecules to resist carbon deposition. First is that the OH group generated by decomposition of water molecules at high temperature reacts with CH component to form aldehyde group, which reduces the formation of carbon deposition precursor. The other is that the increase of water molecules introduces more oxygen atoms into the system, and the carbon atoms formed by DME molecules combine with oxygen atoms to form CO, thus reducing carbon deposition. This study is helpful to promote the industrialization of DME as SOFC fuel.

Mechanical response of nickel-based anodes in solid oxide fuel cells during carbon deposition using reaction molecular dynamics

Carbon deposition of solid oxide fuel cell (SOFC) anode is a key factor to understand the risk of structural cracking due to mechanical stress. In this study, Grand Canonical Monte Carlo (GCMC) was used under ReaxFF to construct the corresponding carbon deposition model and migration law of carbon atoms. The law of heat generation, diffusion coefficient of volume expansion and mechanical behavior of the nickel carbide composition in the simulation are close to the experimental and theoretically results, which validates the model built by ReaxFF. During analysis, the change in structural characteristics and mechanical properties of the anode caused by the carbon deposition were observed. The variation in volume of carbon deposition during diffusion process was detected, and the structure of the carbon deposition process was studied by combining ohmic loss and radial distribution function (RDF). Influence of carbon deposition on the mechanical properties of the uniaxial tensile process using different models was compared and the behavior of carbon content, expansion/compression rate, poison ratio and nickel carbide formation was obtained. The relationship between Young's modulus and carbon deposition of mechanical properties such as yield stress is clarified.

Theoretical Insights into the Structure and Activity of Cobalt Modulated by Surface and Subsurface Carbon in Operando Conditions

In order to understand the effect of in-situ carbon deposition and permeation process on cobalt-catalyzed heterogeneous reactions, the thermodynamics and kinetics of the carburization process of te...

Synthesis of nitrogen doped segmented carbon nanofibers via metal dusting of Ni-Pd alloy

A single-phase micro-disperse Ni-Pd (5 wt.%) alloy prepared by coprecipitation technique was studied as a catalyst precursor for the synthesis of the nanostructured carbon product. Catalytic synthesis of N-doped carbon nanofibers was implemented by the joint decomposition of C2H4Cl2 and CH3CN vapors in H2 excess at 600 °C. The in situ kinetic experiments performed in a flow gravimetric setup with McBain balances showed that the carbon deposition process is characterized by an induction period (18−20 min). The intensive metal dusting of the Ni-Pd alloyed precursor in reaction conditions was found to lead to its complete disintegration. The obtained submicron-sized Ni-Pd particles were shown by XRD to have the same composition as the starting Ni-Pd (5%) precursor. The growth of carbon filaments accompanied the rapid disintegration of Ni-Pd alloy. The addition of acetonitrile vapors into the reaction mixture was found to enhance carbon yield from 24 (0%) to 40.5 (5 vol.%) and 50.9 (8 vol.%) grams per 1 g of catalyst for 2 h of reaction. Joint catalytic processing of C2H4Cl2 and CH3CN was shown to be sufficient for the production of carbon nanofibers doped with nitrogen. According to XPS and EDX data, the average amount of nitrogen within the obtained N-CNF was about 1.8 wt.%. It was revealed by microscopic studies (SEM, TEM) that the produced N-doped carbon filaments are characterized with a rather regular segmental structure similar to that observed for a reference sample of N-free CNF. The synthesized N-CNF samples are characterized by ID/IG ∼ 1.3 (Raman spectroscopy) and high textural parameters (SBET ∼ 390−450 m2/g, Vpore ∼ 0.6 cm3/g).

Production of carbon coated Al-foams and evaluation of their mechanical response

The feasibility of a cost-efficient method to produce carbon coated Al-foams in a single-step heating process is investigated. In this context, a dissolution-sintering process for manufacturing metal foams is developed using raw cane sugar as both the space holder material and the carbon generating medium. The structural characteristics of the produced carbon coated Al-foams were evaluated by XRD, Raman and SEM analyses. Static compression tests were conducted to determine the mechanical response of the produced carbon coated Al-foams. The produced carbon coated Al-foams were fully covered with an external carbon layer having a relatively uniform thickness, despite the inherent complex 3D surface of the foam. The carbon deposition process led to a significant increase of the compression strength of the Al-foam. A strain softening behavior was observed for the carbon coated Al-foams attributed to micro-cracking of aluminum carbides. The produced carbon coated Al-foams can potentially be used as a catalyst support structure.

Dependence of the Pyrocarbon Structure on the Parameters of the Process of Pyrolysis of Hydrocarbon Gases in an Electrothermal Fluidized Bed

One of the variants of the nuclear reactors with increased passive protection being designed at the present time is the reactor with fuel microelements (coated particles) and ball fuel elements. Among the principal components of this type of nuclear fuel is pyrocarbon. Owing to its unique properties, pyrocarbon can also be used in various scientific and industrial areas. Over several years, at the Institute of Gas of the National Academy of Sciences of Ukraine, investigations were carried out on obtaining pyrocarbon coatings by means of pyrolysis of hydrocarbon gases in reactors with an electrothermal fluidized bed. Based on theoretical and experimental data, the present article considers the dependence of the structure of a pyrocarbon coating on the pyrolysis process parameters. Comparing experimental data with theoretical results depending on the temperature and concentration of gaseous hydrocarbon, the structure of the pyrocarbon deposited in an electrothermal fluidized bed can be divided into three forms: laminar, isotropic, and granular (dropping). Among the factors that influence the process of carbon deposition are the content of admixtures and point overheating (cathode spots). Admixtures (mainly iron compounds) can behave as a catalyst accelerating the process of formation of nanotubes and nanofibers. Controlling the temperature and composition of the original material, it is possible to obtain certain allotropic modifications of carbon with assigned structure.

Оценка биологической продуктивности лесной среды в условиях урбанизации (на примере Воронежской нагорной дубравы)

Fast-paced rates of urbanization reduce the area of forestlands. As a result, the volumes of photosynthesis, metabolic processes, oxygen production, carbon deposition and other processes are decreasing. All of this gives pause for thought about the existence of forests on the planet. The Voronezh upland oak forest with a total area of more than 7 ths ha is the central organizing element of the entire urban ecosystem of Voronezh. The forest stand is represented by coniferous crops along the left bank of the terraces above the flood-plain, as well as broad-leaved forest plantations extended over the riverside slopes of the watershed. The oak forest is a state nature reserve of regional significance and has the status of a specially protected natural area. One of the main functions of the nature reserve is lowering the level of anthropogenic impact on unique landscape systems and maintaining the ecological balance in the region. Through the example of the oak forest the basic forest types are characterized and their total biomass is calculated, the above-ground organic biomass is compared with the mass of soil humus and total biomass with the clay biomass. It was found that the total biomass reserves correlate with the mass of physical clay in the root layer. So, pine crops growing in habitats containing up to 900 t/ha of physical clay produce 121.57 t/ha of biomass; its double increase boosts the productivity of the forest to 288.92 t/ha. Herewith, in the first case, the yield class of pine crops ranges from the second to the third; in the second case it reaches the first class. The oak forest is losing its forest-growing potential, its sanitary and hygienic properties are deteriorating. As well as, there is a general decline in the economic value of the specially protected natural area in the context of ecosystem services. For citation: Odnoralov G.А., Tikhonova е.N., Golyadkina I.V., Malinina т.А. Assessment of Urban Forest Biological Productivity (Case Study of the Voronezh Upland Oak Forest). Lesnoy Zhurnal [Russian Forestry Journal], 2020, no. 2, pp. 60–72. DOI: 10.37482/0536-1036-2020-2-60-72

Synthesis and (some) applications of carbon-nanotube-supported pyrolytic carbon nanocones

Graphene-based cones with nanosized apex can be obtained by means of a high temperature pyrolytic carbon depositionprocess using methane and hydrogen as gaseous feedstock and single carbon nanotubes as deposition substrates. Aside thecones, micrometer-sized carbon beads or fibre segments are deposited meanwhile which are a key morphological componentfor allowing handling and mounting the carbon cones and then using them for various applications. Based on both theliterature dealing with pyrolytic carbon deposition processes and experimental observations, a peculiar depositionmechanism is proposed, involving the transient formation of pitch-like liquid phase droplets which deposit onto theindividual carbon nanotubes. In this picture, it is believed that a key parameter is the diameter ratio for the droplets and thenanotubes, respectively. The cone concentric texture and perfect nanotexture are shown by high resolution transmissionelectron microscopy, which allows interesting mechanical and conducting properties to be predicted. Correspondingly,applications of the carbon nanocones as electron emitters for cold-field electron sources on the one hand, and as probes forvarious modes of near-field microscopy on the other hand, have been tested.

Single Particle Characterization of Dynamic Morphological and Phase Changes in Selenium-Doped Germanium Anode during (de)Lithiation Processes

Germanium (Ge) is a promising candidate for high-capacity Li-ion battery anode due to its high theoretical capacity of 1384 mAh/g (for the charged Li15Ge4 phase), low operation voltage, fast bulk Li diffusion, and high electrical conductivity. However, the major challenge in the development of Ge anode is the large volume change involved in the reaction scheme, which could result in rapid loss of specific capacity. We recently reported that selenium (Se) -doped Ge (GeSe) forms an inactive phase that buffers the volumetric expansion of Ge which provides better cycling life and performance (1). However, the dynamic morphological and chemical transformations of GeSe that occur during battery cycling processes is still elusive. The traditional in-situ coin cell has rotating limitations which makes three-dimensional experiment difficult except via limited angle tomography. It also suffers from X-ray damage problems to the binder. To understand the dynamic morphological and chemical changes of GeSe, we developed a single particle battery cell for operando TXM and XANES. A micron-sized GeSe particle was attached to a pre-coated tungsten probe by ion beam carbon deposition on a Zeiss Nvision 40 FIB-SEM at the Center for Nanoscale Materials, Argonne National Laboratory. The single particle battery cell was assembled in an argon-filled glovebox. A small piece of Li metal was attached to a copper wire as the counter electrode. Both the tungsten and lithium electrodes were carefully inserted into a funnel-shaped open-ended silica capillary and stabilized via epoxy. The quartz housing was fully filled with 1 M LiPF6 EC/DEC electrolyte through a micro-syringe and sealed with torr seal epoxy. This binder free operando single particle cell eliminates discrepancies caused by particles overlapping in the in-situ coin cell during TXM and the impact of X-ray damage to the binder (2). It provides chemical and morphological information in three-dimensional space via computed nanotomography coupled with XANES. Figure 1 shows the performance of operando GeSe cell and the 2D morphological change of a GeSe particle attached to the tungsten probe at 11.113 keV. Regardless of the defect caused by the carbon deposition process, the results indicate that GeSe undergoes a different dynamic chemical transformation compared to pure Ge. The selenium-doped germanium has the ability to reinstate its original shape after cycling. This study can help to understand the mechanical stability and degradation mechanism of high capacity lithium alloy anode with additional stress-accommodating phase. K. C. Klavetter, J. Pedro de Souza, A. Heller, C. B. Mullins, High tap density microparticles of selenium-doped germanium as a high efficiency, stable cycling lithium-ion battery anode material. Journal of Materials Chemistry A 3, 5829-5834 (2015).L. Y. Lim, N. Liu, Y. Cui, M. F. Toney, Understanding Phase Transformation in Crystalline Ge Anodes for Li-Ion Batteries. Chemistry of Materials 26, 3739-3746 (2014). Figure 1