Transparent conducting oxide thin films of Si-
doped ZnO prepared by aerosol assisted CVD
Dominic B. Potter, Michael J. Powell, Jawwad A. Darr, Ivan P. Parkin
and Claire J. Carmalt
*
For the first time, aerosol assist ed chemical vapour deposition (AACVD) was used to deposit Si-doped ZnO
thin films on glass. Depositions were done at a temperature of 450
C. The precursor solut ion was made
by dissolving the air-stable compounds zinc acetylacetonate and tetraethyl orthosilicate in methanol with
a small addition of acetic acid to aid solubility. The do pant concentration in the precursor solution was
optimised to find the best opt oelectroni c properties. The incorpo ration of Si into the ZnO lattice was
confirmed by unit cell volumes calculated from X-ray diffraction (XRD) data and by X-ray
photoelectron spectroscopy (XPS). The films consisted of pure phase wurtzite ZnO, with preferred
orientation in the (002) plane. Scanning electron microscopy (SEM) was used to examine the surface
morphology of the films. The optical properties of the films were ana lysed using UV/vis spectr oscopy
and indicated that the average tr ansmittance in the visible part of the spectrum (400–700 nm) varied
between 72% and 80%. The electr ical properties of the filmswereobtainedfromHalleffect
measurements using the van der Pauw method. The incorporation of Si into the films resulted in
a decrease in resistivity down to a minimum value of 2.0 10
2
U cm for the film deposited from a 4
mol% Si : Zn ratio in the precursor solution. This cond uctive film was a significant improvement ove r
the non-conductive undoped ZnO film.
Introduction
Transparent conducting oxides (TCOs) are an important class of
semiconductor material that combine the properties of low
electrical resistivity (<10
3
U cm) and high optical trans-
mittance (>80%) in the visible region. These desirable charac-
teristics have led to the employment of TCO materials in several
optoelectronic applications, including solar panels, liquid
crystal displays (LCDs), and light emitting diodes (LEDs).
1,2
TCO thin lms have been prepared via magnetron sputter-
ing, pulsed laser deposition (PLD), atomic layer deposition
(ALD), spray pyrolysis, sol–gel deposition, and chemical vapour
deposition (CVD).
3–10
Atmospheric pressure chemical vapour
deposition (APCVD) is regularly used for industrial depositions.
This technique involves the vaporisation of volatile precursors
within a bubbler, before transporting them to a heated
substrate via a carrier gas.
A useful variation of APCVD is aerosol assisted chemical
vapour deposition (AACVD). With AACVD, rather than vapor-
ising volatile precursors, soluble precursors are dissolved in
a suitable solvent. An aerosol ‘mist’ is then generated from the
solvent, usually with a piezoelectric humidi er. The mist is then
carried to the heated substrate, where the solvent evaporates
away, leaving gaseous precursor compounds. A deposition
similar to APCVD can then occur – typically via nucleation of
precursors on the substrate surface, followed by surface reac-
tion, and then lm growth.
AACVD has several important advantages over APCVD.
APCVD relies on the use of volatile precursors, whilst AACVD
relies on the use of soluble precursors. Thus, if there are no
appropriate precursors available for APCVD, a CVD-type depo-
sition can still be performed via AACVD, using alternative
precursors. Additionally, by varying the solvent used to make up
the precursor solution, the morphology can be controlled,
which can in turn drastically alter the lm properties such as
conductivity.
11
Furthermore, AACVD is relatively inexpensive, as
it simplies the precursor vapour generation and delivery
process in comparison to APCVD. In APCVD, the bubbler and
the piping leading to the reaction chamber must all be heated to
prevent condensation of the vaporised precursors before they
reach the substrate. In AACVD, only the substrate needs to be
heated. AACVD can also be conducted in an open atmosphere,
and thus it does not require a complicated reactor system.
11–13
Indium tin oxide (ITO) and uorine tin oxide (FTO) are
currently the most commonly used TCO materials in
industry.
10,14,15
However, as a result of the increasing scarcity of
indium and tin, and the present high cost, alternative, more
sustainable materials for TCO applications are highly sought
aer. Doped ZnO materials have been widely investigated for
Department of Chemistry, University College London, 20 Gordon Street, London,
WC1H 0AJ, UK. E-mail: c.j.carmalt@ucl.ac.uk
Cite this: RSC Adv.,2017,7, 10806
Received 4th December 2016
Accepted 5th February 2017
DOI: 10.1039/c6ra27748a
rsc.li/rsc-advances
10806 | RSC Adv.,2017,7, 10806–10814
This journal is © The Royal Society of Chemistry 2017
RSC Advances
PAPER
Open Access Article. Published on 09 February 2017. Downloaded on 8/26/2022 10:57:15 AM.
This article is licensed under a
Creative Commons Attribution 3.0 Unported Licence.
View Article Online
View Journal
| View Issue
TCO applications, due to their wide direct band gap and low
resistivity.
16–19
Additionally, there are many relatively inexpen-
sive Zn precursors available, due to the higher natural abun-
dance of Zn compared to both In and Sn.
20
Therefore, ZnO-
based TCO materials can usually be deposited inexpensively.
The main dopants for ZnO are the group 13 elements – Al,
Ga, and In.
21
These dopants have been used by many groups to
consistently prepare highly conductive, high-quality n-type ZnO
lms.
17
Minami et al. established Si as a dopant for ZnO, as it
was suggested that Si doping would have a less detrimental
effect on the amorphous silica layer found in solar cells.
22
Furthermore, Si is inexpensive in comparison to both Ga and In.
Computational studies have shown that Si will substitute for Zn
in the ZnO lattice due to the low defect formation energy.
23
An
advantage of using Si as a dopant is that it can act as a multi-
electron donor. This is benecial because each dopant ion that
is incorporated into a crystal acts as a scattering centre, so
multielectron donors can provide higher charge carrier
concentrations, whilst keeping the scattering centres to
a minimum, thus leading to high conductivity.
24
Si-doped ZnO
(SZO) thin lms have been deposited previously by various
techniques, including spray pyrolysis,
25
pulsed laser deposition
(PLD),
26
direct current (DC) magnetron sputtering,
27
and atomic
layer deposition (ALD).
7
In this work, SZO thin lms were deposited on glass
substrates via AACVD for the rst time. The electrical properties
of the lms were greatly enhanced in comparison to undoped
ZnO deposited in the same conditions.
Experimental
Film synthesis
AACVD depositions were carried out as detailed in previous
work.
28
All chemicals were used as bought: zinc acetylacetonate
(Zn(acac)
2
) (Sigma Aldrich, Dorset, UK), tetraethyl orthosilicate
(TEOS) (98%, Sigma Aldrich, Dorset, UK), acetic acid (99%,
Fisher, Leicestershire, UK), methanol (99.9%, Fisher, Leices-
tershire, UK) and nitrogen gas (99.99%, BOC, Surrey, UK).
A typical precursor solution was made by dissolving
Zn(acac)
2
(0.50 g, 1.90 mmol) in methanol (20 mL), and then
adding a dopant quantity of TEOS. Acetic acid (1 mL) was
added to improve the solubility of the Zn(acac)
2
. The solution
was stirred for at least 10 minutes in a bubbler. The substrate
was a 3.2 mm thick oat glass plate (Pilkington Technology
Management Limited, Lancashire, UK), precoated with a 50 nm
thick SiO
2
barrier layer to prevent leeching of ions between the
substrate and the lm. The glass was cut to an area of 15 cm
4 cm, and was then washed using soapy water, acetone and
isopropanol. The substrate was then laid horizontally on
a carbon heating block, and heated in a quartz tube to 450
C,
with a top plate suspended approximately 8 mm above, parallel
to the substrate, to ensure laminar ow of the aerosol. An
aerosol mist of the precursor solution was generated using
a ‘Liquifog’ piezo ultrasonic atomizer from Johnson Matthey,
which uses an operating frequency of 1.6 MHz to produce
a mode droplet size of 3 mm. Nitrogen gas was used as a carrier
gas to transport the aerosol to the heated substrate, at a rate of
1 L min
1
. The reactor exhaust was vented into a fume
cupboard. When the precursor solution and associated aerosol
mist had been completely emptied from the bubbler, the coated
substrate was cooled under a continuous ow of N
2
gas, until
the temperature was below 100
C before it was removed from
the reactor.
Film characterisation
X-ray diffraction (XRD) patterns were recorded using a Bruker
D8 Discover X-ray diffractometer using monochromatic Cu K
a1
and K
a2
radiation of wavelengths 1.54056 and 1.54439
˚
A
respectively, emitted in an intensity ratio of 2 : 1 with a voltage
of 40 kV and a current of 40 mA. The incident beam angle was in
a grazing setup at 1
and data was collected between 10
and 66
2q with a step size of 0.05
at 2 s per step. Lattice parameters
were calculated from the XRD data using GSAS and EXPGUI
soware.
29,30
X-ray photoelectron spectroscopy (XPS) was done
using a Thermo Scientic K-alpha spectrometer with mono-
chromated Ka radiation, a dual beam charge compensation
system and constant pass energy of 50 eV, with a spot size of 400
mm. Data was tted using CasaXPS soware. Scanning electron
microscope (SEM) images were obtained using a JEOL JSM-
6301F SEM at an acceleration voltage of 5 kV. UV/vis spec-
trometry was done using a Perkin Elmer Lambda 950 UV/Vis/
NIR Spectrophotometer in both transmission and in diffuse
reectance mode. Room temperature Hall effect measurements
were carried out on an Ecopia HMS-3000, which utilises the van
der Pauw method. Measurements were taken using a 0.58 T
permanent magnet and a current of 1 mA.
Results and discussion
Film synthesis
SZO thin lms were successfully deposited on glass substrates
via AACVD. Zn(acac)
2
and TEOS were used as the Zn precursor
and the Si precursor respectively. Zn(acac)
2
can be purchased at
a lower cost than several other commonly used Zn precursor
compounds, including diethyl zinc,
31–33
zinc acetate,
34–36
and
Zn(thd)
2
(thd ¼ 2,2,6,6-tetramethyl-3,5-heptadionate).
21
Furthermore, diethyl zinc, which is perhaps the most
commonly used Zn precursor, is highly pyrophoric, which
makes its use hazardous and non-trivial. As Zn(acac)
2
is an air
stable solid compound, it is very safe and easy to handle, which
makes it attractive for use in industry.
Rashidi et al. deposited SZO lms via the related spray
pyrolysis technique, however the solvent used was a mixture of
water and isopropanol.
25
In this work, methanol was used,
which is a more reducing solvent and hence should promote
oxygen vacancies.
The methanol solution was carried to the substrates using N
2
carrier gas. Depositions were performed at 450
C and took ca.
40 minutes. The ease of synthesis is an important factor when
considering the merits of the lms, as they were deposited in an
open atmosphere, from an inexpensive, air-stable solution. The
resultant lms were highly stable, were adherent to the glass
substrates, and appeared optically transparent ($72%).
This journal is © The Royal Society of Chemistry 2017 RSC Adv.,2017,7, 10806–10814 | 10807
Paper RSC Advances
Open Access Article. Published on 09 February 2017. Downloaded on 8/26/2022 10:57:15 AM.
This article is licensed under a
Creative Commons Attribution 3.0 Unported Licence.
View Article Online
As the 6 mol% SZO lm displayed the highest transparency
value of 80% as well as good electrical properties (Table 1), this
deposition was repeated at 500
C and 550
C to investigate
whether a higher deposition temperature would improve the
lm properties. It was observed that with increasing deposition
temperature, the lms appeared visibly darker, which was most
likely due to an increase in the amount of carbon being incor-
porated into the lms. Additionally, the electrical properties
diminished signicantly; the lms deposited at higher
temperature were too resistive to give any Hall values using the
van der Pauw technique. Due to the considerable reduction in
optoelectronic properties, these lms were not selected for
further analysis, and thus will not be included in the remainder
of this section.
Crystal structure
The crystal structure of the lms was determined using X-ray
diffraction (XRD), as shown in Fig. 1a. All of the as-prepared
lms consisted of pure-phase wurtzite ZnO. Due to the nature
of thin lms, strain is oen experienced during growth, which
leads to preferred orientation of certain crystal planes. For these
lms, the preferred orientation was in the (002) direction,
perpendicular to the surface of the substrate. This c-axis
orientation has been observed previously in SZO thin lms
deposited by various techniques.
7,25–27
In order to extract the unit cell volumes of the lms, LeBail
renement was performed on the diffraction patterns using
GSAS and EXPGUI.
29,30
It was observed that the unit cell volumes
decreased linearly as the Si concentration was increased in the
precursor solution (Fig. 1b). This can be attributed to the
smaller ionic radius of Si
4+
(0.4
˚
A) in comparison to Zn
2+
(0.74
˚
A).
7,37
It has been suggested through computational studies that
Si
s(Zn)
has a lower formation energy than Si
s(O)
,Si
i(tet)
,or
Si
i(oct)
.
23,24
With increasing Si concentration, there will be an
increase in the substitution of Zn
2+
for the smaller Si
4+
. This will
result in a reduction in the size of the unit cell. The observation
of a decrease in unit cell volume with increasing Si concentra-
tion implies that the Si had been successfully incorporated into
the ZnO lattice. The linear reduction in unit cell volume
suggests that the amount of Si incorporated into the ZnO lattice
was strongly dependant on the initial amount of Si used in the
precursor solution.
Elemental analysis
The elemental concentrations were obtained using X-ray
photoelectron spectroscopy (XPS). The concentrations were
obtained both at the lm surfaces, and within the bulk of the
lms aer etching. XPS analysis conrmed the presence of Zn
in each sample, with a binding energy of 1022.5 eV for the Zn
2p
3/2
peak, which closely matches literature values of ZnO
(0.4 eV).
38,39
Incorporation of Si into the ZnO structure was also
conrmed, with the Si 2p
3/2
peaks generally being centred
around 102.2 eV. This is within 1.0–1.5 eV of literature values of
pure SiO
2
.
40–42
The larger discrepancy of the Si 2p binding
energies in comparison to the values measured for the Zn 2p
could be due to the delocalisation of the Si
4+
electrons into the
ZnO structure, resulting in a lower binding energy.
The concentration of Si in the lm, both at the surface and in
the bulk, increased as the amount of Si used for the precursor
solution was increased. Again, this suggests that the amount of
Si incorporated into the ZnO lattice was strongly dependant on
the initial amount of Si used in the precursor solution (Fig. 2).
Comparison of the Si concentration obtained at the surface
and the concentration aer etching indicate that there has been
asignicant segregation of the dopant towards the uppermost few
nanometres of the lm (Fig. 3). This is likely due to the competing
reaction of the formation of a thin surface layer of SiO
2
,which
would be amorphous as it was not observed by XRD. This non-
conductive surface oxide layer could also explain why the resis-
tivity values of the lmsaren'taslowasSZOlms prepared by
other methods,
16
however the ease of preparation makes AACVD
a scalable technique. This was analogous to Al-doped ZnO thin
lms prepared previously via AACVD, which also used methanol
solutions containing Zn(acac)
2
as the Zn precursor.
28
Surface morphology
The surface morphologies of the lms were analysed using
scanning electron microscopy (SEM). The morphologies were
fairly consistent, with the lms displaying a grain structure
consisting of well-dened, layered, plate-like structures. These
grains appeared to be hexagonal in shape and were approxi-
mately 1–2 mm in diameter (Fig. 4).
From the SEM images of the 6 mol% Si-doped ZnO lm,
new hex agonal layers can be se en growing from the centres of
the surfaces of the hexagonal grains, indicating the l ayer-by-
Table 1 Optoelectronic properties of the SZO films deposited via AACVD. % Si, silicon molar concentration in the precursor solution in
comparison to Zn; V, unit cell volume, with the number in parentheses representing the standard deviation; T
l400700
, average transmittance
over 400700 nm; E
g
, band gap energy; r, resistivity; R
Sh
, sheet resistance; n, bulk carrier concentration; m, carrier mobility
%Si V/
˚
A
3
T
l400–700
/% E
g
/eV r/10
2
U cm R
Sh
/U,
1
n/10
19
cm
3
m/cm
2
V
1
s
1
0.0 47.878 (7) 74 3.16 N/A N/A N/A N/A
0.2 47.643 (4) 75 3.19 24.0 2400 0.44 5.9
0.5 47.638 (3) 72 3.18 11.7 1170 1.14 4.7
2.0 47.617 (5) 73 3.19 2.1 212 2.49 11.9
4.0 47.610 (1) 75 3.19 2.0 201 2.64 16.5
6.0 47.595 (6) 80 3.20 2.5 254 1.63 15.1
8.0 47.59 (2) 76 3.18 8.1 809 1.02 7.6
10.0 47.56 (2) 77 3.09 N/A N/A N/A N/A
10808
| RSC Adv.,2017,7, 10806–10814
This journal is © The Royal Society of Chemistry 2017
RSC Advances Paper
Open Access Article. Published on 09 February 2017. Downloaded on 8/26/2022 10:57:15 AM.
This article is licensed under a
Creative Commons Attribution 3.0 Unported Licence.
View Article Online
layer growth mechanism of the grains (Fig. 5a). This layered
hexagonal grain st ructur e is similar t o SZO lms deposited by
spray pyrolysis at 450
CbyRashidiet al., however the grains
therein were approxi matel y 10 times smaller, at ca. 200 nm in
diameter.
25
Thismaybeduetotheshorterresidencetimethat
the precursor solution experiences in a spray pyrolysis depo-
sition, as the solution is sprayed directly at the heated
substrate. In AACV D, th e aero sol m ist is c arr ied mo re g ently
over the heate d sub strate, w hic h can allow for more time for
molecular mixing and for grain growth. This is signicant, as
a larger grain size is oen desired for TCOs, due to the
reduction in grain boundary scattering, and hence the
increase in carrier mobility.
43
The lm thicknesses were determined using side-on SEM,
and consistently shown to be 1 mm (Fig. 5b), thus indicating
a growth rate of approximately 1.5 mm per hour. The consistent
morphologies and lm thicknesses indicate that the inclusion
of Si in the precursor solution did not affect the solubility of Zn
ions in the solution, nor did it hinder the delivery of the aerosol
to the substrate.
The highly textured surface morphologies could be advan-
tageous for applications such as solar cells, in which a rough
surface morphology is desired in order to promote the scat-
tering light and minimise losses through reection.
10,44,45
Optical properties
The optical properties of the lms were analysed using UV/vis
spectroscopy. The average transmittance across the visible
Fig. 1 (a) XRD patterns of simulated bulk wurtzite ZnO (ICSD #82028), as well as undoped, and Si-doped ZnO films deposited via AACVD. The
apparent rise of an amorphous feature in the 10% SZO film is simply because the peak intensities for that diffraction pattern were lower, resulting
in a slight “stretching” of the image, relative to the other diffraction patterns. (b) The trend shown in the unit cell volumes of the SZO films upon
increasing Si concentration in the precursor solution.
Fig. 2 Si : Zn ratios at the surfaces and within the bulks of the films, as
determined by XPS. Despite the gradual increase at both the surface
and the within the bulk of the films, the amount of Si increases more
rapidly at the surface, indicating a competing, secondary Si phase
formation.
This journal is © The Royal Society of Chemistry 2017 RSC Adv.,2017,7,10806–10814 | 10809
Paper RSC Advances
Open Access Article. Published on 09 February 2017. Downloaded on 8/26/2022 10:57:15 AM.
This article is licensed under a
Creative Commons Attribution 3.0 Unported Licence.
View Article Online
part of the spectrum (400–700 nm) uctuated between 72–
80%, although it generally increased with Si concentration
(Table 1). The lm with the highest transmittance was
the 6 mol% SZO lm, which d isp layed an average tra ns-
mittance of 80% across the visible part of the spectrum . This
is signicant, as it achieved the industrial requirement
of 80% transmittanc e across the visible part of the spec-
trum.
10,31
The undoped lm achieved an average trans-
mittance of 74% across the visible part of the spectrum . This
is notably higher than the undoped ZnO lm deposited
by spray pyrolysis by Rashidi et al., which displays a r elatively
low transmittance across the visib le range of approximately
50–60%, possibly indicating high amounts of carbon
contamination.
25
The UV/vis spectra for the lms with different Si concentra-
tions were fairly consistent across the range of wavelengths that
were scanned. None of the spectra showed a signicant
decrease of transmission at longer wavelengths, nor a signi-
cant increase of reectance at longer wavelength; rather, the
transmission spectra and the reectance spectra only uctuated
within an approximate range of 10% and 4% respectively
across the entire infrared region that was measured (Fig. 6).
This indicates that these SZO lms would not be appropriate for
low-emissivity coatings, which require a high reectance in the
IR range.
10,32,46
Tauc plots were used to extract the band gaps from the
transmission-reection spectra (Fig. 7). The band gap increased
from 3.16 eV for the undoped ZnO lm to between 3.18 and
3.20 eV for the SZO lms (up to 8% Si). The observed band gap
widening is due to the Burstein–Moss effect, whereby electrons
provided by the Si occupy the conduction band, thus raising the
Fermi level, E
F
.
8,47–49
Hence, Si doping was a good route to
improve the electrical properties of ZnO thin lms, whilst
maintaining high visible light transmission. The exception to
this was the 10 mol% SZO lm, in which the band gap dropped
to 3.09 eV. This could be due to the band gap narrowing effect,
which is a result of many-body interactions involving charge
carriers and impurities.
26,49,50
The exponential decay of the Tauc plot at lower energy values
is known as the Urbach tail. It is commonly associated with
Fig. 3 XPS spectra for the 10 mol% SZO film, showing the (a) Si 2p signal measured at the surface, (b) Si 2p signal measured after etching, (c) Zn 2p
signal measured at the surface, and (d) Zn 2p signal after etching.
10810
| RSC Adv.,2017,7, 10806–10814
This journal is © The Royal Society of Chemistry 2017
RSC Advances Paper
Open Access Article. Published on 09 February 2017. Downloaded on 8/26/2022 10:57:15 AM.
This article is licensed under a
Creative Commons Attribution 3.0 Unported Licence.
View Article Online