|
Prasenjit Sen
Research Summary:
Over the past year (and in HRI since February 2004)
the primary aim of my research has been to
understand properties of materials from first-principles, i.e.,
by solving the
Schrödinger equation for the interacting ion-electron system for
specific materials. For this I use both effective one-body
methods like Hartree-Fock (HF) approximation and density functional
theory (DFT), and explicit many-body methods of variational and
diffusion quantum Monte Carlo (QMC). I have used these methods,
sometimes combination of them, to study surfaces, clusters and periodic
solids.
Si(211) surface1
Si is the most widely used material in microelectronic industry and it
is often used as a substrate to grow devices on. For this purpose, Si
surfaces are extremely important for material scientists, and
microscopic understanding of their properties are very crucial. Si(100)
surface has been studied extensively and its low temperature properties
are more or less well understood. However, higher index surfaces of Si
have not been studied so well, though they are ideal to grow polar
semiconductors (III-V or II-VI) on them. This prompted us to study the
Si(211) surface using plane-wave pseudopotential DFT. We studied the
possible reconstructions of this surface, which compare well with
experimental suggestions. We also studied adsorption of Te and As on
this surface. Adsorption properties of Te on this surface is important
in the context of growth of II-VI semiconductors like CdTe, HgTe or
HgCdTe, while As acts as a good surfactant for growth of other materials.
We found the most favorable points for adsorption of an isolated Te or As
atom on the ideal (bulk-terminated) and (2 × 1) reconstructed
Si(211) surfaces. More importantly, we found that there are more than one
points on the surface with comparable binding energy values for
adsorption of Te and As. This can have
important consequences for growth on this surface. We also found that
on the (2 × 1) reconstructed surface, at 0.5 monolayer coverage, As
adatoms dimerize on this surface, while Te adatoms do not. In this
respect their behaviour is exactly similar to that on the Si(100)
surface.
Metal encapsulated Si clusters1
Recently there is a lot of research interest in clusters. The
motivation stems primarily from the possibility of being able to build
cluster-assembled (solid) materials whose properties can be tuned by
tuning the clusters. Successes with carbon fullerenes has given a big
boost to such efforts. However, the motivation also comes from the
experimental finding of luminosity in very stable
Si clusters of certain sizes in the visible range of the spectrum.
This holds promise for their use as new dyes, much longer lived than
the organic ones currently used, in various medical applications.
Experiments have found that in the small size range (10-16 Si atoms), Si
clusters encapsulating a transition metal (TM) atom is more stable than a bare
Si cluster. This motivated us to study electronic and geometric
properties of TM encapsulated Sin clusters--mostly Si12, but
also Si10 and Si11 in certain cases. We explained the properties
of clusters involving a whole range of TM atoms from the 3d, 4d and
5d series. Apart from explaining the observed stability of certain
clusters, we predicted other stable structures not synthesized so far.
While most of the calculations were done using HF and/or DFT methods,
for TiSi12 clusters we used QMC methods to confirm that the ground
state of this cluster is really a spin-triplet (while that of all the
other clusters are spin-singlets). We explained this in terms of
different degree of overlap between the Si sp and TM d orbitals in
different clusters.
Our further DFT calculations indicated that some of these clusters may
form stable extended solids. However, we could not make very
definitive statements on that.
- These works were completed before I joined HRI.
Magnetism in CaB62
CaB6 is very much a focus of current research interest because of the
observation of rather unusual ferromagnetism (FM) in the system. Initial
observations indicated that the La-doped compound,
Ca1-xLaxB6, show FM with a small moment (< 0.07&muB per
La atom), but a high Curie temperature of &sim 600K. A number of
theories were proposed to explain such an unusual FM state. One of them
(the excitonic insulator model)
is based on the assumption that the pure system is a semi-metal with a
small band overlap at the X-point of the primitive Brillouin zone
(BZ). This was suggested by local density approximation (LDA) band
structure calculations. Given that LDA
always underestimates band gaps, this is questionable. Experiments
also produced conflicting results regarding whether the
system is (semi-)metallic
or semiconducting. Later it was also suggested that the origin of FM was
probably extrinsic--due to some magnetic impurities such as Fe or Ni.
To clarify this confusing scenario regarding the electronic structure
of pure CaB6 and the origin of FM in the system, I did a detailed
calculation of the system in collaboration with the group of Prof. Lubos
Mitas at the North Carolina State University, USA, using HF,
DFT and QMC methods. Using explicitly correlated many-body wavefunctions
in QMC, we have convincingly showed that the pure compound is a
semiconductor. We have also found that La impurity does not give rise to
FM in the system while Fe impurity does. On the other hand, presence of
a B vacancy destroys the FM due to any Fe impurity. With these results,
our calculations explain almost all the experimentally observed phenomena
in the CaB6 system. We are in the process of writing up our results.
-
The computations were done on the cluster in Prof. Mitas's group at the
North Carolina State University.
Publications:
- Electronic structure of Te- and As-covered Si(211), Prasenjit Sen, Inder P. Batra, S. Sivananthan, C. H. Grein, Nibir Dhar, and S. Ciraci,
Phys. Rev. B 68, 045314 (2003)3.
- Electronic structure and ground states of transition metals encapsulated in a Si12 hexagonal prism cage, Prasenjit Sen and Lubos Mitas,
Phys. Rev. B 68, 155404 (2003)3.
These papers were published before I joined HRI in Feb. 2004.
|