ZnOgrowth

CVD for ZnO synthesis in polystyrene opals

Here is a controlled process to fabricate large-area high quality ZnO/Polystyrene (PS)
composites and ZnO inverted opals with fine control over filling volume (nanometric
in thickness) by a modified metalorganic chemical vapour deposition (MOCVD)
method. Furthermore, once the ZnO structures are obtained, a sulfidation process
is shown capable to produce ZnS inverted opals maintaining the crystal quality
of the original opal structures. ZnO has been grown in thin PS opals films on
silicon substrates (500 nm sphere diameter and total thickness from 10 to 40
monolayers, 5 to 20 microns) by using a diluted solution of DMZn in hexane as
zinc precursor. Double Distilled Water (DDW) has been used as oxygen precursor.
The diluted metalorganic compound makes its manipulation easier compared with
the pure, highly reactive one. It also provides the key to finely control the
ZnO growth inside the opal. Both sources were kept separate in glass bubblers at
room temperature. The precursors are alternately fed into the reactor by using
Nitrogen as carrier gas at a fixed flow rate. Similar setups were used to grow
SiO2 in opals at room temperature. In our case, the reactor was kept
around 90º C which is a compromise temperature to obtain smooth deposits with
fine grains, avoiding the softening of PS. Room temperature conditions yield
rough and inhomogeneous deposits. The ZnO film thickness deposited through an
alternating exposure to both sources in succession (a cycle) depends on the
carrier gas flow rate, temperature and exposure time of each source. Thus, a
bare opal can be filled in a single or even in several cycles. This fact is of
paramount importance in order to tune the photonic properties to the desired
specifications for optimal performance. Most of the times, theory predicts
peculiar or functional behaviour for a given filling fraction. It is worth
mentioning that due to the conformal growth and geometrical constrains the
maximum value achievable for the filling fraction is close to 86%. In addition,
conformal growth permits the fabrication of inverted structures even for
extremely low ZnO filling fractions. Inverted opals were obtained upon
calcination of PS-ZnO composites at 450º during 3 hours in air. This process
completely removes the PS matrix but also recrystallizes the ZnO, as will be
discussed later. Scanning electron microscopy (SEM) inspections and optical
spectroscopy have been performed in order to monitor the infiltration process
and X-ray diffraction to confirm the crystal quality and composition of the
deposits.


Figure
1.
SEM images of cleaved edges in ZnO infiltrated (a) and inverted (b and c) opal
structures from 500 nm polystyrene spheres.

Figure 1 shows SEM images of ZnO based opals grown on silicon substrates. Figure
1a shows the appearance of a cleaved edge of a ZnO-PS composite with an
infiltration degree of 60% of the pore volume. In this image, and due to the
high contrast between conductive ZnO and the insulating PS spheres, the latter
become hardly
visible seeming an inverted-like structure where the contact points between
spheres can be observed as dark spots. Figures 1b and 1c show actually inverted
ZnO opals from cleaved edges at different magnifications. The high quality of
the smooth and homogeneous coating can be observed. Surface inspections reveal
homogeneous growth over millimetre areas. The inverted replica order is only
disturbed by the usual cracks (hundreds of microns apart) present in this type
structures not by the ZnO growth. The empty interior of the shells in the
inverted replicas are only observed in cracks and the cleft edges. From those we
can give a rough estimate for the ZnO shells thickness of about 20-30 nm, in
good agreement with the expected value. The elemental composition of the ZnO
inverted structures was studied by Energy Dispersive X-Ray analysis (EDX) (not
shown). The absence of a carbon peak reveals the quantitative removal of PS.


Figure
2.
From top to bottom: Specular reflectance spectra for bare PS opal and five CVD
cycles for ZnO infiltration.

By a
sequence of several cycles it is possible to infill the opal with fine control
over the filling fraction. The refractive index of opal free ZnO films was
measured by ellipsometric techniques from 0.4 to 1.6 mm,
varying from 2.2 to 1.9. The infilling process has been monitored by the optical
reflectance at near normal incidence with respect to the (111) fcc planes taken
with the help of an optical microscope from a circular area of about 375 microns
of diameter. The red-shift in the first order Bragg peak as the infiltration
proceeds accounts for the increase of the average refractive index of the
structure due to ZnO infilling. Figure 2 shows the reflectance spectra for 5 CVD
cycles. The bare PS opal spectrum (top) shows a peak with 55% reflectivity at
1160 nm corresponding to the photonic pseudogap at L point due to (111) planes.
As we increase the number of cycles this peak moves to lower energies (higher
wavelengths) till a value of around 1330 nm. This latter value is hard to
ascertain with precision because at values near 80% of pore volume infiltration
the average refractive index of the pore (ZnO/Air) reaches that of the
polystyrene (n = 1.57) producing an index matching effect which hides the Bragg
peak. However, high energy features (around 0.6 microns) survive even at near
index matching conditions (small arrows in figure 2) indicating the persistence
of the photonic effects and the high quality of the structures.


Figure 3.
Evolution of the ZnO shell thickness around the PS
spheres and pore infilling (in
percentage of the available volume) as a function of the number of cycles.

By comparing
these spectra with photonic bands calculations we can estimate the ZnO grown in
each CVD cycle. Figure 3 shows the filling volume and the associated layer
thickness as a function of the number of cycles. As can be observed, ZnO layer
thickness increases with the number of cycles, from 5 nm in the first cycle (18%
of the pore volume) to about 40 nm for the last one (around 85% of the pore
volume). Linear fit of the experimental points indicates a growth rate of 8 nm
per cycle in good agreement with ellipsometric measurements for plain thin ZnO
films. These figures give an idea of how fine this growth process is in terms of
ZnO shell thickness control.


Figure
4.
Specular reflectance spectra of a PS bare opal (b), 80% ZnO/PS opal composite (c)
and ZnO/air inverted structure (d) together with the corresponding photonic
bands calculations (a) for bare (solid), composite (dash) and inverted
(dash-dot) structures.

Optical
properties of ZnO inverted structures have also been measured. Figure 4 shows
the reflectance spectra of bare PS opal (panel b), ZnO/PS composite with 80% of
the pore volume infilling (panel c) and the same structure after PS calcination
(panel d) respectively together with photonic bands calculations (panel a). The
parameters used for the calculations were the following: for the bare opal it
was used PS spheres with a refractive index of nPS = 1.59 arranged in
fcc lattice (solid line in figure 4a); for the ZnO infiltrated structure (dash
line in figure 4a) a layer of 30 nm around the PS spheres with a refractive
index of nZnO = 1.93 (taken from ellipsometric data) was added;
finally, for the inverted ZnO replica (dash-dot line in figure 4a) the spheres
refractive index was substituted for 1 (air). As can be clearly seen from the
spectra, in figure 4c the Bragg peak disappears due to index matching (notice
the pseudo gap disappearance in dash line of panel a) and our system behaves as
a homogeneous medium with a refractive index of around nave = 1.65
and 5 microns thick. After removal of PS by calcination the Bragg peak shifts to
higher energies (blue-shifted) due to the lowering of the average refractive
index (from nave = 1.65 to nave = 1.2) and gets broader
(almost double) because the increase of the refractive index contrast (nZnO/nair
= 1.9). In addition to this procedure to fabricate ZnO inverted opal structures,
it is possible to make a step further. Sulfidation of ZnO is a method

[1]

which has been used to fabricate ZnS nanotubes recently.

[2]

Here, ZnO inverted opals can be sulfidated to produce ZnS inverted replicas.


Figure
5.
X-ray diffractograms from ZnO/PS composite (a), and ZnO inverted replica after
calcination at 450ºC for 3 hours (b) and ZnS inverse opal after sulfidation.
Silicon reflection arises from the substrate.

X-Ray
diffraction was performed on ZnO/PS composites, ZnO inverted replica and ZnS
after sulfidation, as can be seen in figure 5. Panel 5a corresponds to an
as-grown ZnO/PS composite, panel 5b shows the X-Ray diffractogram from the
inverted ZnO opal after calcination and panel 5c corresponds to ZnS after
sulfidation of the previous. As can be observed, upon crystallization, sharp
peaks are obtained in agreement with an increase in the grain size. All samples
present a polycrystalline nature the grain size increasing from 4 to 40 nm upon
calcination due to crystallization and 53 nm after sulfidation due to 30%
difference in ZnS density which provokes the increase of the semiconductor
volume from that of ZnO.


Figure
6.
Specular reflectance for bare opal (a), ZnO/PS Composite (b), ZnO inverted opal
(c) and ZnS after sulfidation.

This
process can again be monitored by optical spectroscopy. In figure 6 four
specular reflectance spectra are shown that correspond to bare PS opal (6a);
composite ZnO/PS opal with 60% of the pore volume infiltrated (6b); ZnO inverted
opal after calcination (6c); and ZnS inverted structure after sulfidation of the
previous (6d). As expected, the Bragg peak centred ca. 0.6 (in reduced unites
a/l,
where a is the lattice constant) for the bare PS opal, shifts to lower
energy and narrows after ZnO infiltration. After Calcination, this peak moves to
higher energies due to lower average refractive index. After sulfidation, the
peak moves again to lower energies for two reasons: Firstly, due to a 30%
specific volume increase from ZnO to ZnS the lattice parameter grows a 10% which
contributes to this red-shift; secondly, ZnS refractive index, nZnS =
2.3, is higher than that of ZnO. This latter fact also contributes to widening
the peak due to a higher dielectric contrast. This demonstrates not only that a
complete control over the photonic band gap properties is provided by the
infiltration procedure but also that the structures, even when shells as thin as
30 nm are synthesized, are robust enough to withstand inversion. This robustness
makes them amenable to further chemical processing for regrowth or sulfidation
as shown.

[1]

S. Lew, A. F. Sarofim, and M. Flytzanistephanopoulos, Industrial &
Engineering Chemistry Research 1992, 31, 1890.

[2]

L. Dloczik, R. Engelhardt, K. Ernst, S. Fiechter, I. Sieber, and R.
Konenkamp, Applied Physics Letters 2001 78, 3687.