Double diffusive convection (DDC) refers to a ubiquitous phenomenon in the ocean where the convection flow is driven by both salinity and temperature gradients in the vertical direction. DDC plays an important role in ocean mixing. DDC can induce the so-called thermohaline staircase, which consists of a stack of alternating convection and finger layers. In this research we conducted large-scale simulation and successfully realize the thermohaline staircase with one convection layer between two finger layers. This simulation provides valuable information about the dynamics and vertical fluxes of the thermohaline staircase.
The simulation is run on Cartesius. The visualization is done on Elvis cluster. This particular simulation needs to be run for very long time, in order to prove the three-layer state is stable up to the turbulent diffusive time scale. Cartesius system plays an essential role for this study.
The visualization provides illustrative information about the thermohaline staircase. From this figure, one can clearly observe distinct flow structures in different layers, i.e. salt fingers in the finger layer and large-scale convection role in the convection layer. Especially, the figure shows interesting phenomena at the interfaces between layers. As the fingers grow, they cannot sustain the self-organized state and gather into large clusters, which drive the large-scale roles in the convection layer.
|When the top water is warmer and saltier than the bottom water, vertically elongated salt fingers can occur even when the density is smaller at top. For certain conditions, salt fingers cannot occupy the whole height of the domain and multi-layer state appears. This state resembles the thermohaline staircase, which is very common in the ocean. Here we show a numerical realization of such thermohaline staircase between two parallel plates. The top plate has higher temperature (with Prandtl number 7) and salinity (with Prandtl number 21). The temperature and salinity differences between two plates are measured by Rayleigh numbers, which are 4x1010 and 1x1011, respectively. For this set of control parameters starting from uniform scalar distributions, we obtain a three-layer state, as shown by the three-dimensional volume rendering of the salinity field. The color and opacity are both set by salinity. Clearly, adjacent to the top and bottom plates two finger layers exist, and in between there is a convection layer. In the finger layers we observe salt fingers, which have very small horizontal size and large height. In the convection layer large-scale convection roles emerges and the fluid is well mixed.|
Heart diseases are one of the leading causes of mortality all around the world. Numerical simulations and visualizations of the cardiovascular system can help surgeons in understanding how different surgical solutions affects blood circulation and be a guide for the most appropriate procedure for a specific patient or type of patients. Using our novel fluid-structure interaction algorithm, we perform simulations of blood flow inside the left ventricle of the human heart with prosthetic mechanical mitral valves. We study the flow behavior for different valve types and configurations to understand its effect on the overall flow dynamics inside a human heart.
The left ventricle of the human heart is a crucial part of our body as it is responsible for pumping oxygenated blood to all the tissues. Only through these visualizations we have been able to identify the splitting of blood flow into different streams inside the ventricle by the mechanical valves. This is crucial since it helps surgeons in making a choice between the various kinds of valves available for patients with weak hearts. By changing the type of valves, its material and visualizing the flow structure in different simulations, we have been able to provide information to medical surgeons on the blood flow dynamics inside the left ventricle which cannot be obtained using any other technique.
Numerical simulations were performed on a local cluster while the visualization was performed on Elvis visualization cluster using 2 cores along with GPU acceleration.
|A cut-section of the left ventricle of human heart is shown along with prosthetic mechanical valves; the valves are currently in closed position. Blood flows into the ventricle from the left atrium. The mechanical valves rotate about fixed hinges depending on the forces exerted by the inflowing blood. The left ventricle can expand or contract based on the hydrodynamic forces exerted by the blood flow into the ventricle. The imposed flow rate of blood into the ventricle is similar to what is observed in the human heart. At the start of the simulation the ventricle is filled with blood, which is at rest.|
|A snapshot of the iso-surface of the velocity magnitude of blood flowing into the left ventricle. This is during the initial phase of the expanding (diastolic) stage of the heart cycle. The hydrodynamic forces exerted by the flow of the blood as it enters the ventricle rotates the prosthetic mechanical valves. The stream of blood is split into different jets by the mechanical valves as can be seen from the flow pattern in the figure. Also notice that the left ventricle has expanded in comparison to its shape in image_1.png due to the rushing inflow of blood.|
|This figure shows blood flow into the ventricle in an advanced phase of the expanding (diastolic) stage of the heart cycle. The left ventricle has further expanded due to the influx of blood. Two major streams can now be observed, while the smaller streams have dissipated. Splitting of blood flow into streams is not desired as a lot of energy is dissipated and thus a lot of blood is left behind in the ventricle during the expansion and contraction phase of the heart cycles.|
eWaterCycle is a project that provides detailed hydrological information for water management challenges around the globe. We calculate how much water is available in each part of the world at present and in the near future. Using the most advanced scientific models and high performance computing, we are able to make statements about the water availability at every square kilometer of the land surface. Unique to this project is that in situ and satellite data is being used to update and improve the forecast, not unlike the setting of a watch once every day. It is the first operational global hydrological model that runs forecasts up to ten days ahead.
The Cartesius Dutch National Supercomputer is used to create all data visualized. The Cartesius is used to pre-process the global weather data used, run our forecast model in parallel on a number of so-called fat nodes, and process the results into the final output. The results are also stored on the SURFsara Data Archive for later long-term verification of the forecast.
The visualization contributes to the eWaterCycle research by showing the global hydrology community that modeling on a global scale is possible and useful for those areas that might not otherwise be serviced by hydrological science. The web interface also allows the researchers to explore and verify the results of the simulations directly, without having to copy the massive amounts of data from the High performance Computing infrastructure.
|The web interface to visualize the global hydrological data is continuously fed data directly from the Cartesius Dutch National Supercomputer, where the forecast model is run once per day. The data is transformed into image tiles based on the preferences of the user, and sent over the web using the widely used WMS format.|
|Detail image showing additional possibilities for colour schemes and base layer. The (interactive) graph on the left shows the amount of water (discharge) over the coming week at the selected map location, including uncertainty bands based on running an ensemble of model realisation in parallel. The user can interact with the graph to select the time used for the image layer of the map.|
|As a truly global forecast, the model even shows discharge in remote areas where hydrological modeling is not commonly available. This detail image shows the flow of the river Congo.|
People have often sought solitude in the starry night sky, and it is an appropriate place for that. Looking at the dark night sky our sun seems very lonely. Neighboring stars are so far away that they look like mere specks of light, and more distant stars blur together into a feeble glow. Our fastest space probes will take tens of thousands of years to cross the distance to the nearest star. Space isolates us like an ocean around a tiny island. But this is not how the Sun was born. In our study to the formation and evolution of the Solar System, we strive at understanding the circumstances and environment in which this all happened. We now think that the Sun was born in a star cluster of a few thousand stars, that formed from a large molecular structure. When the first stars appeared, the cluster must have been composed of a mixture of young budding stars and gas. We developed a large multi-scale and multi-physics software environment, called the Astronomical Multi- purpose Software Environment (or AMUSE for short), with which we study the origin of the Solar System. Our calculations include a wide variety of physical ingredients, starting with a giant cloud composed of molecular gas. The cloud collapses under its own gravity leading to high-density enhancements. These 'clumps' grow in mass until the first stars are born. The star formation process continues for a few million years, after which the first planetary systems form in gaseous and debris disks around the young stars. During this process, the stars interact gravitationally, and they start developing a dense stellar wind that eventually blows away the residual gas. At the same time, the Milky Way Galaxy pulls at the stars, which eventually all go their own way. leaving the current Solar System is lonesome. This story on the formation of the Solar System has been presented for the first time as a coherent story in Portegies Zwart popular article "the long lost siblings of the sun" in November 2009 in Scientific American and in a followup article in the same journal in 2014. In later work he and Inti Pelupessy continued to work on this coherent understanding on the formation of the Solar System and the ability of Earth. This work has resulted in a dozen scientific publications in a variety of journal, including IEEE Computer, the Astrophysical Journal and the Monthly Notices of the Royal Astronomical Society.
We present three images of the formation and evolution of our large scale simulations on the formation of the Solar System in its parental molecular cloud. Our simulations started with hydrodynamics of the gravitational collapse of the molecular cloud. As soon as the first stars form, separate gravitational dynamics and stellar evolution codes are automatically switched on, including the hydrodynamics code to resolve the winds of the young stellar population and a disk-evolution code to follow the formation of planets in the protoplanetary disks. Eventually, all the gas is ejected, mainly by the stellar winds and the protoplanetary disks turn into planetary systems. The colors of the images are carefully matched to a model for human color perception, and therefore colors in the images represent how an observer would perceive the stellar light. This gives the images an artistic appeal, as well scientific accuracy. The images show the molecular cloud at an age of 0.7 Myr, 1.6 Myr and 2.5 Myr.
In the images, we strive at mimicking the observations to our best knowledge. The gas distribution, stellar positions, and intrinsic stellar parameters are calculated self-consistently using the AMUSE framework. The visualization of the stars and the interaction of the star light with the gas is realized using approximate radiative transfer calculations. The colors of the stars match their intrinsic colors calculated using a stellar evolution code. The interstellar gas and dust absorbs and reflects the starlight. These calculations are realized using hydrodynamics and radiative transfer. We took the proper gas-to-dust ratio of the interstellar gas properly into account in calculating the reflectiveness and absorption features. The reddening of the stars due to the obscuring dust is, therefore, real, and not an artifact of the calculation. The characteristics of gas and dust scattering depend on the composition of the gas, and on the intensity and wavelength of the starlight. Taking these effects into account properly is challenging, and an accurate solution requires full 6-dimensional radiative transfer calculation. This is an extraordinarily expensive operation, which requires hours of computer time. The tool we developed to mimick the observations taken somewhat simpler but sufficiently accurate approach, but enables us to visualize the simulation output directly with an interactive viewer. We mimicked observing the simulations with the Hubble Space Telescope. The telescope has a wide range of characteristics which we took into account in producing these images. These characteristics include the 4chromatic aberration, deformations of the lense and the artifacts produced in the mirror mount in the spacecraft. We accounted for each of this effect, and they are visible in any HST image, including our mimicked images. The chromatic aberration is best seen in the rainbows in the brightest stars. The effect of spherical aberration is visible in the large the visual impression of the brightest stars (the so-called point-spread function), and the mirror mount results in the 'spikes' seen in the stellar images. Our images, therefore, reproduce to good approximation images captured by HST. At the moment our visualization package is still in development, but a paper is in preparation and as soon as this is published the entire package will become online available for free via github.com.
The images shows here are the result of large and complex calculations. Many of these calculations are performed in a wide variety of computers, including local workstations, the Little Green Machine, SURFsara's LISA, and the Huygens and Cartesius supercomputers at SURFsara.
|Proto star cluster at an age of 0.7 Myr after the start of the simulation. The first stars just formed some 10000 years go and they are deeply embedded in the gas cloud from which they formed. These first stars are insufficiently bright and massive to penetrate the encapsulated gas.|
|At an age of 1.6 Myr after the start of the simulation, almost all the stars in the cluster have formed. Each star is surrounded by a protoplanetary disk, but these are still too young to actually form planets. The sizes of the disks are rather strongly affected by close encounters with other stars with disks. By this time the cluster is composed of about 3300 stars, but on the image only the brightest 20 -or so- are able to penetrate the still surrounding dense molecular gas.|
|At an age of 2.5 Myr star formation has almost completely stopped, and planet formation is well underway. About 100 stars are orbited by Jupiter-like gas planets. There are more than 600 planets in the image, but none of these is easily seen because they are simply too small and too dark to be detectable in this rather crude image. There is still some residual gas, but the brightest somewhat older (> 2 Myr) stars (concentrated to the left) easily shine through the gas. The less massive young (< 1 Myr) stars (to the right) are still partially embedded, which is visible from their reddened colors. In fig. 2 this bimodality is not yet visible, but this is also a projection effect. We slightly moved the camera for this picture to show the internal cluster structure.|
We are developing Hemocell, a numerical blood flow simulation framework to explore the transport properties of whole blood on a cellular level. Hemocell is quite flexible and with the accurate transport characteristics we plan to develop a thrombus formation model. To this end, we have started to extend the framework with biochemical processes such as bond formations between platelets.
The initial position and rotation of the cells were computed on Lisa (gaborz) using simple mechanical model where the cells are positioned using a uniform random spatial distribution. From this position they start to repel each other based on the volume of overlap until they reach a force-free distribution. The cellular flow was calculated on Cartesius (zega) on 432 cores. It consists of a lattice Boltzmann CFD coupled with the deformable particles through immersed boundary method.
The dynamical behaviour of the cells as an aggregate is difficult to study by only looking at the statistics (aka. "the numbers"). Different regions behave differently, e.g. a cell close to the wall experiences a substantially different shear environment than one travelling in the middle of the channel. Applying 3D visualization is an effective way to spot interesting events such as spontaneous cluster formations or inter-particle collisions. This process can be further enhanced by producing videos that help to understand the temporal effects as well.
|Fully resolved blood flow with deformable particles in a small vessel.|
Rotating convection is an important phenomenon in the atmosphere around the earth and in many industrial processes. To study rotating convection a Rayleigh-Bénard convection (RBC) setup is used, where a fluid in a container is heated from below and cooled from above. Rotating this container around its vertical axis induces transitions in flow structures and in the heat transfer between the hot and cold plate. We analyze these effects of rotation by tracking particles in rotating RBC, using direct numerical simulations. Lagrangian statistics of passive tracer particles are used to characterize flow structures. The case of inertial particles, which are two-way coupled to the flow, is studied in order to learn how to actively tune flow structures and heat transfer and possibly trigger a transition to enhanced heat transport, which can be beneficial for industrial applications.
The Rayleigh-Bénard simulations are run on Cartesius, using a finite-difference code originally developed by R. Verzicco and P. Orlando (R. Verzicco and P. Orlando, J. Comput. Phys., 1996). This code was extended by a particle routine, which allows us to numerically integrate different types of particles in the flow, from simple tracers to two-way coupled inertial particles. The resources of Cartesius were needed to fulfill the required grid size and the necessary amount of time steps for statistical convergence. Each n time steps, an HDF5 file including the position, velocity and acceleration of the particle was created. A python script, which was also run on Cartesius, was used to pick 100 particles of these HDF5 files and copy its position and velocity to a text file. These text files were then used in a gnuplot script, to create the final plots.
From the visualizations of tracer trajectories in rotating and non-rotating Rayleigh-Bénard convection (RBC) we can qualitatively observe the completely different flow structures in the two cases. In non-rotating RBC a large-scale circulation (LSC) of up-going hot-fluid and down-going cold fluid is present. At a critical rotation rate this LSC breaks down and vertically aligned plumes start to dominate the flow. These typical flow structures are indeed followed by the passive tracer trajectories shown in the figures. These qualitative observations inspired us to quantify the geometry of the tracer trajectories by computing its curvature and torsion. Curvature and torsion measurements are discussed in a manuscript, which is recently submitted to Physical Review Fluids.
|Tracer trajectories, which perfectly follow the fluid motion, in non-rotating cylindrical Rayleigh-Bénard convection (RBC). The red and blue circles represent the hot bottom and cold upper plate, respectively. The colors of the particles represent vertical velocity, where blue is negative velocity and red is positive velocity. The particles perfectly follow a large scale circulation of up-going hot fluid on one side of the cell and down-going cold fluid on the other side of the cell, exactly as expected for non-rotating RBC.|
|Tracer trajectories, which perfectly follow the fluid motion, in rotating cylindrical Rayleigh-Bénard convection (RBC). The red and blue circles represent the hot bottom and cold upper plate, respectively. The colors of the particles represent vertical velocity, where blue is negative velocity and red is positive velocity. The Large Scale Circulation, observed in non-rotating RBC, has disappeared and now vertically aligned vortical plumes dominate the flow. Indeed the particles show a more localized upwards motion now.|
Twenty thousand years ago, most of Canada and the northern half of the US was covered by a vast ice-sheet that was at places over four kilometres thick. When global temperatures started rising and the ice retreated, the melt water created a vast lake over the area of present-day Hudson Bay. When the ice dam over the Hudson Strait that held the lake in place broke, all that water rushed into the northern Atlantic Ocean, raising global sea level by over two metres. Our research looks at how and why this happened. We use a computer model that simulates the ice sheet and the lake in a physical way - temperature, mechanical forces, etc. This model is a large computer program which we run on the LISA cluster.
We created a 3D print of the model result at about 10,000 years before present, showing the ice sheets over Canada and Greenland and the huge lake. The vertical scale is exaggerated but to scale - the ice is about four kilometres thick, higher than the Rocky Mountains. The ice also lowered global sea level, simply because a lot of water was not in the sea but was ice instead. This effect is visible in the print: look closely and you'll see that the Bering Sea is actually the Bering Land. Alaska and Siberia are still connected, allowing the first native Americans to enter the continent. The "blockiness" of the print is intentional: each block represents a horizontal area of 40 by 40 kilometres, which is the resolution of our model.
Our ice model is run on the LISA cluster under the project name "Reconstruction of Temperature and Sea Level" (project number MP-206). We use this system because the model is quite large (~ 30 Gb of RAM) and takes a long time to run (about 60 hours for one simulation). The model generates large data fields containing information about topography, ice thickness and sea level. We created a small program that converts these raw data fields to a standard 3D object format. The actual print was created by Shapeways.
The model produces three-dimensional data fields that can be hard to visualize on a two-dimensional computer screen. Many of the questions we want to answer (where was the ice, where was the lake, which way did they move, how much water was in the lake, where did it drain, etc.) are qualitative rather than quantitative, meaning that it can be difficult to give solid answers. Proper visualization of the results helps us understand what's going on. Also, a 3D print is extremely useful at conferences for showing other people what we do.
BioMed Xplorer utilizes and represents the medical domain data in a graph, consisting of all concepts and relationships among them, using the standardized ontology of the unified medical language system. BioMed Xplorer's primary knowledge base captures the medical peer-reviewed literature from several large data sources, and exploits "Linked Data" technologies to deal with large amounts of heterogeneous and dynamic information. It is designed to both scale up and scale out, and at present uses the SurfSara's high performance computing cloud infrastructure to gather, process, and analyze all the information.
The HPC cloud was used to generate the RDF triples and to host the service.
The visualization enables medical researchers to interact with the data in an intuitive manner without the need for compute science knowledge and with a very small learning curve.
|Graph-based visualization of BioMed Xplorer. Concepts are classified into one of 15 semantic type categories. A statement summery is also visable|
|Statement details for the " Glucose associated with Diabetes Mellitus, Non-Insulin-Dependent" relation presenting and wide range of related information|
|An expanded Graph with focus on the concept "Glucose" showing the definition of the term.|
Hyforformylation is well-known process, but today still many regio- and ennatioselectivity issues not solved. Traditional strategies for selective control rely on fine tuning ligands properties, such as electronic, steric & bite angle. In our group, we have developed novel approach to steer the regioselectivity via substrate preorganization. With the smart design of OrthoDIMphos, we got extremely high linear selectivity for 3-butenoic acid hydroformylation, which is the best results reported up to now. To get depth understanding of this novel system, we carried out mechanism studies. We obtained dimer (X-ray structure) starting from OrthoDIMphos and [Rh(nbd)2BF4] with CO babbling. However, does dimer still formed under hydroformylation conditions? With the aid of DFT calculations, we demonstrated the main active species is dimer with around 30 kcal/mol more stable than the monomer because of the internal hydrogen bonding, which is corroborated by High Pressure spectroscopy studies.
As the trigonal bipyramidal rhodium species could be either ee or ea coordination model, the monomer showed with ee coordination model is more stable the ea coordinated monomer. The hydride is at apical position poiting towards the pocket formed by two pyrrole-NH and two amide-NH hydrogen donor, which can be used to preorganized anion functionalized substrates via hydrogen bonding interaction to control the selectivity. Taking consideration of all the possible coordination model of trigonal bipyramidal rhodium species, at least three possible dimer species, i.e. ee-ee, ea-ea and ee-ea, could form under catalytic conditions. With the aid of DFT calculations, the ee-ee coordination dimer is the most stable species among all the dimer species and even around 30 kcal/mol stable than the monomer species showed in the capture. In contrast to monomer, the dimer is extremely stable because of the internal hydrogen bonding and as such the hydrogen bonding cannot bind efficiently with anions to preorganize substrates as the monomer does, which is relevant to the sensitivity of selectivity to substrates/catalyst ratio as the substrates can split the dimer to monomer leading to high linear selective hydroformylation.
All the geometry optimization were performed on LISA cluster. VMD module was used to produce all the images we produced by processing data generated from these calculations.
The current research work derives largely from the equilibrium between the monomer and dimer units, and the pre-organization of substrate within these structures drives this equilibrium which enables a regioselective control over the ratio of liner/branched ratio of product formed. The current visualization gives molecular level structural insights in the system whereon the geometric and electronic favorability of substrate pre-organization can be understood in detail.
The composition of the interstellar medium (ISM), the space between galaxies and stars, is crucial for understanding star and planetary formation. Telescopic studies reveal the precence of polyaromatic hydrocarbon molecules (PAHs) and it is believed, that these molecules cluster with each other and form small dust grains, slowly building planets and stars. Our study is aimed at characterizing the clustering of PAHs with infrared spectroscopy. Infrared radiation induces vibrational motions in molecules, which, together with quantum chemical computations, help us shed light on the structure of these PAHs.
Quantum chemical calculations (DFT) were performed with Gaussian09 under account name alemmens. These were visualized with Chemcraft.
The visualizations allow us to compare and look at vibrational modes from different calculations. In detail: we perform anharmonic calculations and harmonic calculations. The anharmonic calculations are an order of magnitude more difficult, so these are calculated on a very low level of theory. Subsequently we can apply the anharmonicity shift to the harmonic calculations, which can be performed on a higher level of theory.
|In the FELIX laboratory, we use intense infrared light to look at molecules that serve as a model for molecules in interstellar space. Irradiation of these molecules with infrared light induces vibrations of molecules. Together with quantum chemical calculations performed with SurfSara, the vibrations help to determine the structure of molecules. Also, the calculations provide us a way to visualize the motion of molecules when they are irradiated with infrared light. In this image, the structure of a cluster of 3 acenaphthene molecules is shown. The background consists of space image contained in one of our vacuum chambers, where we mimic the space environment.|
Use of simple organic molecules such as primary alcohols, carboxylic acids, amines as sources of hydrogen could hold the key towards developing a hydrogen economy. The current work focusses on dehydrogenation of alcohols for generation of electricity and production valuable chemicals such as aldehydes as side products. Dehydrogenation of alcohols can be performed in heterogeneous manner by using metal (or metal oxides/alloy) based solid surfaces as catalysts, or by using molecular catalysts (homogeneous catalysis). Heterogeneous catalysts often end up producing Carbon monoxide as a by-product which poisons the catalyst. A way to circumvent this problem is to make use of organometallic fuel cell (OMFCs)(Ref 1) systems. In OMFCs the inorganic catalysts are anchored on a conductive carbon polymer support. The electrons released from oxidation of alcohols are stored in the catalysts in form of intermediates generated in the catalytic cycle. These electrons are used to drive the fuel cells. While many inorganic catalysts are known to perform dehydrogenation of alcohols, only the Rh(trop)2N (Ref 2) catalyst is known to work in the set-up of an OMFC. Therefore, it is important to understand the underlying mechanism of the process of electron transfer. We have performed Density Functional Theory based electronic structure calculations on the intermediates in the catalytic cycles of different catalysts. Here we show how the geometric and electronic structures couple together in case of Rh(trop)2N that leads to electron flow in OMFCs. The mechanistic understanding gained from this work will be used to guide catalyst design for industrial scale development of OMFCs.
References: 1. S. P. Annen, V. Bambagioni, M. Bevilacqua, J. Filippi, A. Marchionni, W. Oberhauser, H. Schönberg, F. Vizza, C. Bianchini, H. Grützmacher, Angew. Chem. Int. Ed. Engl. 2010, 49, 7229-7233, 2. H. Grützmacher et al., Angew. Chem. Int. Ed. 2008, 47, 3245 -3249
The electronic structure visualized in terms of HOMO/LUMO distribution with and without the Phospohrous anchor clearly provide guidelines for design of catalysts which will work in an OMFC. The optimized geometries of the catalyst and the intermediates give valuable insights in to catalysts design specially comparisons are made with other known catalysts which perform the same function of alcohol dehydrogenation but are inactive in OMFCs. HOMO: Highest occupied molecular orbital (red and blue lobes), LUMO: Lowest unoccupied molecular orbital (green and yellow lobes)
LISA cluster was used to perform all the geometry optimizations and electronic structure calculations. The images were produced by the processing data generated from these calculations. For producing the images of the molecule the VMD module was used. gOpenmol software package was locally used to generate the HOMO and LUMO pictures by processing the data generated on LISA account
|Molecular level mechanistic understanding of current flow in Organometallic Fuel Cell Systems Simple molecules such as alcohols, carboxylic acid etc. are attractive sources of hydrogen and can therefore play important role in development of a Hydrogen Economy. Especially interesting is use of such molecules in inexpensive carbon polymer based organometallic fuel cells (OMFCs). OMFCs contain discrete inorganic catalysts anchored on conductive carbon support. While several inorganic molecules are known for dehydrogenation of alcohols, only Rh(trop)2N catalyst has shown activity for alcohol dehydrogenation in an OMFC. Understanding of the underlying mechanism of electron transfer in OMFCs is therefore important. (A)Schematic representation of an OMFC. A catalyst anchored to conductive carbon support via phosphorous based ligand stores 1 molecule of hydrogen and gets reduced by 2e- upon oxidation of an alcohol. The stored electrons are used to generate electricity in the fuel cell. (B) Rh(trop)2N molecule (C) Rh-H-(trop)2N-H catalytic intermediate with 1 equivalent of Hydrogen stored (D) Phosphorous based anchor added to the metal center. (E) HOMO and LUMO of B (F) addition of the phosphorous anchor delocalizes the LUMO and facilitates electron transfer from B. D-F only relevant H shown in picture.|
SURFsara | Science Park 140 | 1098 XG Amsterdam | T +31 (0)20 800 1300 | F +31 (0)20 800 1401 | firstname.lastname@example.org