Research Interests Cem

Ozdo˘gan January 17, 2017

Function al Materials: Modelling and Simulation

My research activities involve several areas of materials physics and nanoscience with mod-

elling and simulation and high performance computing techniques. These techniques, e.g.

electronic structure calculations, are important tools for designing of new and/or functional

molecules, materials, and devices and investig ating the physical and chemical properties o f

those. Below, I state my ongoing research projects, research accomplishments and give an

outlook on future research perspectives.

1 Current Research: Supercapacitors, Functionalizing 2D material s p1,2

2 Past Research: Sintering, Functionalizing 2D materials, Boron Clusters, Surface

Modiﬁcations, Carbon Nanotubes p2

3 Past Research: Data Mining & Big Data, High Performa nce Computing p4

4 Future Plans p5

1 Current Research

Design of Planar and Tubular Nanostructured Hetero Double Layer

Electrostatic Supercapacitors and Investigation of Energy Storage

Capabilities

Eﬃcient and sustainable clean energy resources has became serious need and research on

this ﬁeld takes more attention day by day. Developing technologies has created a workspace

about energy storage and conversion systems. This energy storage and conversion workspace

includes batt eries, fuel cells a nd supercapacitors. In the scope o f present research, modelling

and simulation of planar and tubular geometries based nanostructured double layer elec-

trostatic supercapacitors will be examined. The quantum mechanical calculations will be

done by Density Functional Theory (DFT) based non-equilibrium Green f unction method

(NEGF) (utilized package: ATK/VNL). A potential diﬀerence applied to the electrodes

causes a current by contact, from contact region t o target region. Quantum transport quan-

tities such as transport spectrum, conductivity, current-voltage characteristics (I-V graph),

and charge distribution will be calculated and by these quantities ﬁnally the capacitance of

the system is attained.

(a) (b)

Figure 1: (a) Graphene + h-BN Supercapacitor model: Au(111) electrode/G/h-BN/G/Au(111)

electrode. (b) Stacking order.

The supercapacitor models that will be used in this study are deﬁned in two material cat-

egories; 2D planar electrode materials and dielectric materials. The f ormer one includes

semi-metal graphene (G) that has thin layers, hexagonal boron nitride (h-BN) with broad

band-gap that is known as white graphene, and silicine that has the same electronic structure

with gr aphene and the latter category includes h-BN, graphene, silicine layers, carb on and

boron nitride nanot ubes. The capacitance and energy storage capacity of the various systems

http://boron.physics.metu.edu.tr/ozdogan 1/5

Research Interests Cem

Ozdo˘gan January 17, 2017

(e.g., Figure 1) will be studied by changing the width, length, the number and stacking of

layers, and chirality parameters. Systems with the best theoretical quantum capa cita nce will

be given to the literature as possible functional nanodevices. This project is supported

by T

ITAK (MFAG 115F137 Supervisor 2015).

Structural, electronic and magnetic properties of h-BN/G hybrid

and hetero structures, and design of nanosystems functionalized

with defects

The discovery of graphene in 2004 has led to investigation of new 2D materials with honey-

comb lattice structure. The most attractive materials similar to graphene are the transition

metal dichalcogenides, h-BN, and topological insulator s. The perfect lat tice match between

the h- BN (band gap E

= 5 − 7 eV ) and graphene (semi-metal) make it possible to fabricate

the h-BN-graphene hybrid and heterostructure. The ultimate purpose in making BCN hybrid

nanostructures is to ﬁnd new compounds whose electronic properties can be switched between

conducting and insulating phases. The h-BN/ graphene heterolayer structures are considered

as promising device architectures in which the h-BN acts as an insulator substrate and/o r

thin dielectric separative layer in graphene-based ﬁeld eﬀect transistors (FETs). Project

consists of two part s. In the ﬁrst part, in-plane BCN hybrids (e.g., Figure 2) and in-pla ne

G/h-BN heterostructure will be studied by employing the ﬁrst-principles calculations within

the DFT f r amework (utilized package: VASP).

Figure 2: Graphene hexagon embed-

ded into h-BN sheet.

Our ﬁrst goal is to reduce or completely close the band

gap or be able to tune the E

value in order to use

them as conductor or n- or p- type semiconductors in

electronic applications. If we can successfully tune the

, the outcome will be formatio n of semiconducting

regions in conductor graphene. This indeed means to

have nanostructures that can be used as circuit ele-

ments. Our second goal is to search for the possibility

of using the structures with gained magnetic proper-

ties in spintronics applications. Magnetic moments can be observed in BCN hybrids and

systems having defects (such as va cancies C

, B

and N

), where the symmetry is broken.

The second part of our project will concentrate on the in-plane h-BN/G/h-BN and G/h-

BN/G sandwiched heterostructure. Our ultimat e task is to obtain nanodevices which can be

used as circuit elements by tuning the electronic properties and band ga p of the heterostruc-

ture via the vacancies. Our ult imate g oal is to design in-pla ne hybrid/heterostructure to

be employed in technology project. This project is support ed by T

ITAK (MFAG

114F426 Researcher 2015).

2 Past Research

• Sintering Mechanism and Determination of Melting Points of Silver Nanopar-

ticles Sintering of Ag nanopa r ticles (NPs) is increasingly being used as a driving mechanism

for joining in the microelectronics industry. We performed molecular dynamics (MD) simu-

lations based on the embedded atom method (EAM) to study pressureless sintering kinetics

of two Ag NPs (range of 4 to 20 nm), and sintering of three and four Ag NPs. We found that

the sintering process passed throug h three main stages. Sintering rates obtained by our sim-

ulation were higher than those obtained by theoretical models generally used for predicting

http://boron.physics.metu.edu.tr/ozdogan 2/5

Research Interests Cem

Ozdo˘gan January 17, 2017

sintering rates of microparticles. Furthermore, diﬀerent sizes of single-crystal Ag NPs to ﬁnd

complete melting and surface premelting points are studied by using a MD simulation ba sed

on the EAM. Our model can predict both complete melting and surface premelting points

for a wider size range of NPs. Melting kinetics showed three diﬀerent trends. (utilized

package: developed by us)

• Functionalized graphene by embedded boron clusters The doping of graphene by

embedded boron clusters was studied with DFT as a po ssible way of modifying the intrinsic

properties of graphene (utilized package: VASP). We ﬁnd that B

clusters embedded

into graphene and graphene nanoribbons are structurally stable a nd locally metalized t he

system. A linear chain of boron clusters forms a metallic “wire” inside the graphene matrix.

Thus, the doping of graphene with the cluster implantation technique mig ht be a viable tech-

nique to locally metalized graphene without destroying its attractive bulk properties. The

stability of edge states and edge magnetism in zigzag edge graphene nanoribbons (ZGNRs)

is also discussed. We point out that magnetic edge states might not exist in real systems.

Even if systems with magnetic edge states could be made, the intrinsic mag netism would

not be stable at room temperature. Charge doping and the presence of edge defects further

destabilize the intrinsic magnetism of such systems. This project is supported by FP6

(RII3-CT-2003-506079 HPC-EUROPA Supervisor) .

• Boron Clusters Boron is a promising material that has ability to build strong and

highly directional bonds with boron itself. As a result, boro n atoms form diverse struc-

tural motifs, ultimately can yield distinct nanostructures, i.e., it can take almost a ny shape.

The electronic and geometric structures, total and binding energies, ﬁrst and second energy

diﬀerences, harmo nic frequencies, p oint symmetries, dipole moments, chemical bondings

and highest occupied molecular orbital-lowest unoccupied molecular orbital(HOMO-LUMO)

gaps, fragmentation channels, io nization energies, and the Coulomb explosion of small neu-

tral, charged and B

(n = 2 − 20) clusters, solid α − B

and γ − B

and α, γ and hydrogen

bonded triangular buckled sheets have been investigated using ab initio quantum chemical

and DFT methods (B3LYP with 6-311++G(d,p) basis set) (utilized packages: Gaussian

& VASP). Within this size range, the planar and quasi-planar (convex) structures have the

lowest energies. The structural transition from 2D to 3D is found at the size of 20. Highly

charged unstable clusters dissociate spontaneously into several neutral or charged fragments,

and large a mo unts of energy are produced, depending on the charg e of the parent cluster.

We argue t hat this mechanism makes boron clusters a clean, safe, and cheap energetic mate-

rial. The structure of the B

100

fullerene exhibits unusual stability among all noninteracting

free-standing clusters. T his project is supported by T

ITAK ( T BAG 105T084

Researcher)

• Surface Modiﬁcations Ion beams are used in a wide range of applications, e.g., cleaning

crystal surfaces, depth proﬁling the atomic composition a nd chemical analysis of the surface

etc. An Ar

− Ni(100) collision system at 1 keV impact energy was investigated by using

realistic isoenergetic MD simula t ions to understand the mechanism behind the onset of sput-

tering (utilized package: developed by us). To this end, a sequential MD simulation

program is converted into a linear scaling pa r allel code. Several properties such as penetra-

tion depths, angular and energy distributions of the reﬂected Ar and sputtered Ni atoms,

dissociation time, embedded, scattering, sputtering patterns and geometries of the sputtered

clusters are reported. The calculated sputtering yield is found to be in good agreement with

the available experimental results.

• Carbon Nanotubes (PhD Work) Single walled carbon nanotubes (SWNT) are ob-

http://boron.physics.metu.edu.tr/ozdogan 3/5

Research Interests Cem

Ozdo˘gan January 17, 2017

served to have high ﬂexibility, strength, and stiﬀness, very similar to those of individual

graphene sheets. The O(N) and parallelization techniques have been successfully applied in

tight-binding MD (TBMD) simulations of metallic and semiconducting behavior o f armchair

and zigzag SWNTs (utilized package: developed by us). We converted a sequential

O(N

) TBMD simulation program into an O(N) parallel code to simulate SWNTs under

tensile loading. Young’s modulus, tensile strength, and the Poisson ratio are calculated. The

stress-strain curve is obtained. The elastic limit is observed at a strain rate of 0.09 while the

breaking point is at 0.23. The frequency of vibration for the pristine 10× 10 carbon nanotube

in the radial direction is 4.71 × 10

GHz and found that it is sensitive to the stra in rate.

This project is supported by T

ITAK (TBAG 1877 (199T106) Researcher)

and METU ( AFP-2000-07-02-11 Researcher).

3 Past Research

• Data Mining & Big Data There has been an explosive g rowth of very large or huge

datasets in scientiﬁc and commercial domains. Cluster analysis has become crucial task

for mining of the datasets. A novel linear scaling para llel clustering algorithm based on

wavelet transforms is implemented (utilized package: developed by us). Our results

demonstrate that developed parallel WaveCluster algorithm exposes high speedup and scales

linearly with the increasing number of pr ocessors. We further implemented our WaveCluster

algorithm to graphical processing units (GPUs) level parallelization and investigated the

parallel performance for very large spatial da t asets. Divide and conquer approa ch is followed

on t he implementation of wavelet transform and multi-pass sliding window approach on the

implementation of connected component labeling.

Data reduction is perhaps the most critical component in retrieving information from big

data in many data-mining processes. We examined the mot ivations behind why t he reduction

techniques are important in the analysis of big datasets by presenting several basic reduction

techniques in detail, stressing the advantages and disadvantages of each. A special emphasis

were given t o parallel wavelet-based multi-resolution data reduction techniques on distributed

memory systems using Message Passing Interface (MPI) and shared memory architectures

on GPUs along with a demonstration of the improvement o f performance and scalability for

one case study.

• High Performance Computing High performance computing methods and tools are

becoming as indispensable for many research ﬁelds including materials science, nano science

and technology. I have made use of the concepts/tools/libraries such as MPI, Parallel Virtual

Machine (PVM), ScaLAPACK, LAPACK, BLAS during writing or parallelization of MD

codes for distributed memory systems.

http://boron.physics.metu.edu.tr/ozdogan 4/5

Research Interests Cem

Ozdo˘gan January 17, 2017

4 Future Research Per spectives

• Borophene It is very recently reported that crystalline 2D boron sheets are synt hesized

and show metallic characteristics ( Science 350, 267, 1513-1516 (December 2015)).

• Azobenzene Azobenzene has the property of the photoisomerization of trans and cis iso-

mers by lig ht. Transport theory calculations mig ht reveal the possible usage as a molecular

switch.

• Blueprint: Doping 2D Materials Modelling and simulation of defect creation mecha-

nism (mimicking electron beam irradiatio n) and following deposition of foreign atoms into

defects by using ab initio molecular dynamics simulations. Then, performing transport cal-

culations to investigate that possible functional nanodevice as a ultimate goal.

• TBMD I am planning extend our O(N) parallel TBMD code. Target systems are silicon

carbide for simulation of the production epitaxial graphene and multiwall CNTs for large

scale simulations.

http://boron.physics.metu.edu.tr/ozdogan 5/5