...

Modeling the Growth of ZnO Nanoclusters in Water Abdullah Alsunaidi

by user

on
Category: Documents
7

views

Report

Comments

Transcript

Modeling the Growth of ZnO Nanoclusters in Water Abdullah Alsunaidi
2010 International Conference on Nanotechnology and Biosensors
IPCBEE vol.2 (2011) © (2011) IACSIT Press, Singapore
Modeling the Growth of ZnO Nanoclusters in Water
Abdullah Alsunaidi
Physics Department, King Fahd University (KFUPM)
Dhahran 31261, Saudi Arabia
[email protected]
In this work, we carry out classical molecular dynamics
simulations to investigate the process of evolution of small
ZnO nanoparticles (nano-clusters) in water. We use the same
empirical potentials used in a previous study [9] to produce
these clusters in vacuum. Since these potentials do not take
into account the possibility of dissociation of the H2O
molecules, we also carry out density-functional calculations
(DFT) in order to comment on the stability of the
intermediate and final structures of the clusters.
Abstract— We carried out molecular dynamics simulations to
investigate the process of formation of (ZnO)n nanoclusters
with size up to n=16. We started with Zn-O units randomly
mixed with 400 water molecules and interacting via empirical
potentials. Each Zn atom immediately binds to up to three
water molecules forming ZnO(H2O)3 complexes. The cluster
then grows by breaking these Zn-water bonds and forming
rings with structures like Zn2O2(H2O). Finally, these structures
aggregate and re-structure to form ring-like or bubble-like
ZnO nanocluster. The obtained nanoclusters are composed of
squares and hexagons and are similar to the clusters produced
in vacuum and reported using evolutionary algorithm. In
order to test the accuracy of the classical potentials, we carried
out density-functional calculations at the GGA level to
investigate the interaction of the intermediate and the final
ZnO clusters with H2O molecules. It was found that H2O
favorably adsorbs and sometimes dissociates when it
approaches the ZnO clusters.
II.
We study the processes of formation and growth of ZnO
clusters in water, using molecular dynamics simulations (MD)
in the NVT ensemble, employing the DL_POLY code.
Initially, Zn-O units are distributed randomly in a cubic box
of 24 × 24 × 24 Å3 size with 400 water molecules. The
influence of the temperature is investigated for few cases by
performing simulations at two temperatures (300 and 400 K).
Our chosen interatomic potentials consist of a Coulomb part
qi q j
U ijCoul = k
(1)
rij
Keywords-Zinc-Oxide, nanoparticles, molecular dynamics
simulation, Density-functional theory
I.
COMPUTATIONAL METHODOLOGY
INTRODUCTION
where qi is the point charge representing the ion, k a
dimensional constant, and rij the distance between two point
charges, i and j, and of a short-range part, the form of which
is dependent on the value of rij and is composed of three
different potential functions: (i) a Buckingham potential
⎛ rij ⎞ ⎛ Cij ⎞
(2)
U ijBuck = Aij exp ⎜ − ⎟ − ⎜ 6 ⎟
⎜ ρ ⎟ ⎜r ⎟
ij ⎠ ⎝ ij ⎠
⎝
In recent years, there has been growing interest in
studying the properties of metal oxide nanostructures.
Considerable attention has been given to zinc oxide as a
promising multifunctional material with wide-ranging
technological applications including electrical and lightemitting devices, gas sensors, and catalysis [1,2]. In the last
few years, several experimental studies were reported on the
interaction of H2O with polar ZnO(0001) and non-polar
ZnO(1010) surfaces. Dissociation of H2O on the OZnO(000-1) surface was observed by Kunat et al. [3]. The
dissociation of the water molecules is proposed to take place
at the O-vacancies of the O-ZnO surface. Measurements
made by Meyer et al.[4] showed that the most stable
structure of water on the ZnO(1010) surface turned out to be
a configuration in which every second water molecule is
dissociated. Similar results were obtained by Dulub et al. [5]
who studied the structure of water monolayers adsorbed on
ZnO(1010) at room temperature using a scanning tunneling
microscopy (STM) and density-functional theory (DFT).
Most of the water is in a lowest-energy configuration where
every second molecule is dissociated. DFT calculations
reveal that water molecules repeatedly associate and
dissociate in this sustained dynamical process. Other DFT
calculations carried out by many groups [6,7,8] were in
agreement with experimental findings.
(ii) a Lennard-Jones potential
Bij Dij
U ijLJ = 12 − 6
rij
rij
(3)
and (iii) a polynomial potential
U ijPoly = aij + bij rij + cij rij2 + d ij rij3 + eij rij4 + f ij5
(4)
The species dependent potential parameters, A, B, C, D, ρ,
a, b, c, d, e, and f, are given in Ref.[10]. For the interaction
of Water with ZnO units, we used the consistent valence
force field water model [11] in which the O and H atoms are
modeled by point charges connected through two and threebody forces.
After that we also examine the stability of the
intermediate and final clusters by carrying out calculations
based on the density-functional theory (DFT). This step will
109
help show how water molecules adsorb (or dissociate) to the
cluster at and how the cluster structure could change. We
have done these simulations using the Dmol3 code
implemented in the Materials Studio package. The
generalized gradient approximation (GGA) was adopted to
describe the exchange correlation interaction with the
parametrization by Perdew, Burke, and Enzerhof (PBE).
Density functional semi-core pseudopotentials (DSPP) fitted
to all-electron relativistic DFT results, and double numerical
basis set including d-polarization functions (DND) were
employed. Spin-unpolarized self-consistent field (SCF)
calculations were carried out with a convergence criterion of
10−6 au. for total energies. Geometry optimization was
performed using the Broyden-Fletcher- Goldfarb-Shanno
(BFGS) algorithm with a convergence criterion of 2 × 10−3
au. for the maximum force and 5 × 10−3 Å for the maximum
displacement.
III.
(a)
(b)
RESULTS AND DISCUSSION
We monitored the growth of clusters of sizes n=2-16 at
the two temperatures T=300 and T=400 K. The initial stages
of growth are similar for all the clusters so we take the
growth of n=8 as an example. Figure 1 shows different
stages of growth for n=8 at T=300 K. After 0.06 ns, we
noticed that the ZnO units are hydrated with one, two or
three H2O molecules forming ZnO(H2O), ZnO(H2O)2 and
ZnO(H2O)3 complexes. The stability of these complexes is
found by calculating their binding energies (BE). The
ZnO(H2O)2 has the largest binding energy of 0.97 eV,
followed by ZnO(H2O)3 with BE=0.875 then the last is
ZnO(H2O) with BE=0.636 eV. Of course these values may
change if the bond lengths and angles change. We observed
that after 0.42 ns, square rings of ZnO started forming. These
have high binding energy (about 8 eV) compared to the
previous structures. Rings then start coalescence forming
open structures. After 1.17 ns, the open structures start
forming bubble-like structures.
Because of the higher kinetic energy of the molecules,
the growth process at T=400 K is faster. Similar to the
previous case, the complexes ZnO(H2O), ZnO(H2O)2 and
ZnO(H2O)3 form after 0.06 ns. However, after 0.27 ns,
hexagonal (ZnO)3 and octagonal (ZnO)4 rings start forming
until we get two (ZnO)4 rings at time 0.42 ns. This shows
that the (ZnO)2 square rings are not stable at high
temperatures. In fact, it is known (to be published) that
clusters composed of squares are less stable than those
composed of hexagons. In order to get the most stable
structures, bonds keep breaking because of the high
temperature and we finally get two clusters after 0.63 ns,
which makes the system ready to take the final shape after
these two clusters merge.
(c)
(d)
Figure 1. snapshots for the growth process for the n=8 clusters at T=300
K taken at (a) 0.06 ns, (b) 0.42 ns, (c) 0.51 ns, (d) 1.17 ns. (Oxygen:Red
and Blue, Zinc:Grey, Hydrogen:White)
110
We show in Figure 2, a list of the final clusters obtained
in this study (at T=300 K) along with the most stable
structures as calculated by DFT. For the small rings (n=2,3,4)
the MD method gives the global energy structures predicted
by DFT. Remember that when searching for all the possible
structures, the evolutionary algorithm (EA) [9] gives many
other structures. For example, it gives a cube and other
irregular structures for the n=4. But these structures are less
stable than the ring. In the present study, it would be time
consuming to start with different initial configurations to
search for all possible structures. However, for n=5, we have
started from two different initial configurations and MD gave
us two minimum energy structures: a ring and a chair-like
structure similar to those found by evolutionary algorithm [9].
The most stable of these, as found by DFT, is the ring
structure. For the n=6 (not shown), we got a drum-like
structure composed of two n=3 rings. This is also found by
EV, but it is not the most stable structure. For n=7, the MD
did not give us a ring structure but rather a chair-like
structure. This is also similar to what was found by the
evolutionary algorithm. This in fact occurs because a drumlike (n=6) structure was formed first before a ZnO unit
attaches to form this structure. For n=8, the MD method gave
the predicted minimum energy structure. Notice that for n=8,
the EA gives many different structures in a small energy
range. For larger structures, like n=16, the system is stuck in
one of the minimum energy structure and could not find the
expected spheroid structure. The structure shown in the
figure is also predicted by the EA but is less stable. This puts
a limitation on the cluster sizes that the MD can predict.
However, we believe that at higher temperatures and lower
concentrations of ZnO, the MD can do better in predicting
the stable ZnO structures for larger clusters.
There is an important issue that should be addressed here,
when comparing the DFT calculations with the MD results.
Usually to test the interaction of water with ZnO, the DFT
calculations are carried out with only one H2O molecules
adsorbed to the ZnO. In this case, the effect of hydrogen
bonding between the H2O molecules is not considered. Here
we give some examples to check the stability of the clusters.
In Figure 3, we show some of our DFT calculations for
intermediate clusters. In Figure 3 (a), we optimized the
structure of the ZnO(H2O)3 complex surrounded by one layer
of H2O molecules. We found that the optimum structure
would result in the transfer of an H atom from one of the
H2O molecules to the bare O atom, while the other two
molecules are only adsorbed to the Zn atom.. This
dissociation can not be observed using our interatomic
potentials. When we carried out this calculation for the
cluster not surrounded by the water layer, we noticed that the
H2O molecules break the bonds with the ZnO unit, leaving
the ZnO not hydrated. We noticed the dissociation of the
H2O molecules also in the case of the n=2 rings found in
Figure 1 (see Figure 3 (b)). For the n=3, ring, we got the
structure in Figure 3 (c) irrespective whether the complex is
surrounded by a water layer or not. Finally for the n=8 final
cluster, we noticed that when we have only few H2O
molecules surrounding the cluster, the cluster tends to break
apart and one or two H2O molecules dissociate depending on
n=2
n=3
n=4
n=5
n=7
n=8
n=16
Figure 2. Minimum energy structures for some of the ZnO nano-clusters
as found by DFT (left) and those found by MD simulations (right).
(Oxygen:Red, Zinc:Grey, Hydrogen:White)
111
their positions. This cluster is however stable (as shown in
Figure 3 (d)) when geometry optimization is carried out for
the cluster surrounded by a layer of H2O molecules. This
study shows that DFT calculations should be done with the
clusters surrounded by layers of H2O in order to compare
with results of MD simulations. This is currently not feasible
considering the time and memory needed for such
calculations.
IV.
CONCLUSION
We have studied the growth process for small nanoclusters of ZnO using molecular dynamics simulation
method with empirical potentials. At T=300 K, the
intermediate clusters are more open, forming small rings that
merge to form the final cluster. At the higher temperature
T=400 K, the intermediate structures are hexagons and
octagons and they close only at late stages forming bubblelike structures. The final structures are similar to those found
by evolutionary algorithm [9] using the same potentials but
some of them are different from those predicted by DFT
calculations. We have also investigated the effect of
hydrogen bonding between H2O molecules on the stability of
the intermediate and final clusters. We found that carrying
out DFT calculations for the clusters surrounded by a layer
of water molecules gives different results from calculations
that consider only a single H2O molecule attached to the
cluster. We have also observed dissociation of the H2O
molecules at the intermediate stages. However, this
dissociation does not change the final structure of the clusters.
(a)
ACKNOWLEDGMENT
The author would like to thank King Abdul-Aziz City for
Science and Technology (KACST) for funding this work
through project number APR-27-110
(b)
REFERENCES
[1]
Ozgur, U.; Alivov, Ya. I.; Liu, C.; Teke, A.; Reshchikov, M.
A.;Dogan S.; Avrutin, V.; Cho, S.-J.; Markoc¸, H. J. Appl. Phys. 2005,
98,041301.
[2] Pearton, S. J.; Norton, D. P.; Ip, K.; Heo, Y. W.; Steiner, T. Prog.
Mater. Sci. 2005, 50, 293.
[3] M. Kunat, St. Gil Girol, U. Burghaus, and Ch. Wo1ll, J. Phys. Chem.
B 2003, 107, 14350-14356
[4] Bernd Meyer, Hassan Rabaa and Dominik Marx, Phys. Chem. Chem.
Phys., 2006, 8, 1513–1520
[5] Olga Dulub,1 Bernd Meyer,2 and Ulrike Diebold, Phys. Rev. Lett., 95,
2005, 136101
[6] Y. Yan, and M. Al-Jassim, Phys. Rev. B, 72, 2005, 235406
[7] Bernd Meyer,Dominik Marx, Olga Dulub, Ulrike Diebold, Martin
Kunat, Deler Langenberg, and Christof Woll, Angew. Chem. Int. Ed.
2004, 43, 6641.
[8] Arrigo Calzolari and Alessandra Catellani, J. Phys. Chem. C 2896
2009, 113, 2896–2902.
[9] A. Al-Sunaidi, A. Sokol, C. Richard A. Catlow, and S. M Woodley, J.
Phys. Chem. C 18860 2008, 112, 18860–18875.
[10] L. Whitmore, A. Sokol, C. R. Catlow, Surf. Sci. 2002, 498,135.
[11] K. Lau, H. Alper, T. Thacher and T. Stouch, J. Phys. Chem., 1004, 98,
8785.
(c)
(d)
Figure 3. Geometry optimization for some of the ZnO structures found by
MD simulations (a) ZnO(H2O)3, (b) (ZnO)2(H2O)3, (c) (ZnO)3(H2O)3, (d)
(ZnO)8. (Oxygen:Red, Zinc:Grey, Hydrogen:White)
112
Fly UP