Nanomaterials 10 01415

nanomaterials
Article
Spectroscopic Properties of Si-nc in SiOx Films
Using HFCVD
Zaira Jocelyn Hernández Simón 1 , Jose Alberto Luna López 1, * ,
Alvaro David Hernández de la Luz 1 , Sergio Alfonso Pérez García 2 , Alfredo Benítez Lara 3 ,
Godofredo García Salgado 1 , Jesus Carrillo López 1 , Gabriel Omar Mendoza Conde 1 and
Hayde Patricia Martínez Hernández 4
1 Centro de Investigaciones en Dispositivos Semiconductores (CIDS-ICUAP), Benemérita Universidad
Autónoma de Puebla (BUAP), Col. San Manuel, Cd. Universitaria, Av. San Claudio y 14 Sur, Edificios IC5 y
IC6, Puebla 72570, Mexico; imezaira@gmail.com (Z.J.H.S.); joalvada1@hotmail.com (A.D.H.d.l.L.);
godgarcia@yahoo.com (G.G.S.); jesus.jecarril@gmail.com (J.C.L.); gaomec13@gmail.com (G.O.M.C.)
2 Centro de Investigación en Materiales Avanzados (CIMAV), Alianza Norte 202, Parque de Investigación e
innovación Tecnológica Apodaca, Nuevo León 66600, Mexico; alfonso.perez@cimav.edu.mx
3 CONACYT-Centro de Investigaciones En Óptica AC (CIO), León, Guanajuato 36000, Mexico;
alfredbl@cio.mx
4 Departamento de Ingeniería Eléctrica y Electrónica, Instituto Tecnológico de Apizaco (ITA), Fco I Madero s/n,
Barrio de San José, Apizaco, Tlaxcala 90300, Mexico; pathaymh@yahoo.com
* Correspondence: jose.luna@correo.buap.mx; Tel.: +52-222-2295500

Received: 21 February 2020; Accepted: 28 June 2020; Published: 20 July 2020
Abstract: In the present work, non-stoichiometric silicon oxide films (SiOx ) deposited using a hot
filament chemical vapor deposition technique at short time and simple parameters of depositions are
reported. This is motivated by the numerous potential applications of SiOx films in areas such as
optoelectronics. SiOx films were characterized with different spectroscopic techniques. The deposited
films have interesting characteristics such as the presence of silicon nanoclusters without applying
thermal annealing, in addition to a strong photoluminescence after applying thermal annealing in the
vicinity of 1.5 eV, which may be attributed to the presence of small, oxidized silicon grains (less than
2 nm) or silicon nanocrystals (Si-nc). An interesting correlation was found between oxygen content,
the presence of hydrogen, and the formation of defects in the material, with parameters such as the
band gap and the Urbach energies. This correlation is interesting in the development of band gap
engineering for this material for applications in photonic devices.
Keywords: SiOx films; Si-ncs; band gap engineering; spectroscopic characterizations; ellipsometric
spectroscopy; photoluminescence
1. Introduction
The emission of visible light in nanostructured oxides (applicable to optoelectronic devices) [1–3],
light absorption effects such as the so-called down conversion effect (solar cells) [4,5], and resistive
switching (non-volatile memories of multiple states) [6–9] are some of the characteristics that make
SiOx a material that needs to be studied deeply, also taking into account that silicon is the second
most abundant material on our planet. Another favorable feature that this material provides is that it
can be obtained using multiple synthesis techniques [10–14]. Among these techniques, it has been
found that the technique of chemical vapor deposition activated using a hot filament (HFCVD) offers
important characteristics due to the ease of obtaining thin films and powders with diverse electrical
and optical characteristics. The HFCVD technique consists of a chemical process that uses high purity
molecular hydrogen as a reagent, which is dissociated to work in its atomic form. Such dissociation is
Nanomaterials 2020, 10, 1415; doi:10.3390/nano10071415 www.mdpi.com/journal/nanomaterials

Nanomaterials 2020, 10, 1415 2 of 21
achieved thermally through an incandescent filament at ~2000 ◦ C; the chemical reaction of this process
is described in references [15,16].
Although there are several reports of the behavior of this material using various techniques [10–15],
in this investigation the spectroscopic and compositional behavior of SiOx films are analyzed, where the
material is obtained using the HFCVD technique with a short deposition time of 3 min. This material
is suitable for large-scale applications and may be important for future applications, such as light
absorbing devices (photosensors and solar cells), as well as for light emitting devices.
In addition, in this research we try to explain phenomena such as the characteristic light emission
of SiOx in approximately 1.5 eV that appears when the material has been annealed, which has been
attributed to phenomena such as the quantum confinement that is presented by a phase separation
and formation of nanocrystals or due to the formation of various defects in the material; however,
no agreement has been reached between these theories. An analysis of how deposit conditions affect
optical properties of great technological interest such as the band gap is also carried out.
2. Materials and Methods

The SiOx films were deposited on p type silicon substrates with a diameter of 2 in, orientation
(1 0 0), and resistivity of 1–5 ohm-cm. Four SiOx films were deposited using the HFCVD technique,
the parameters that were changed during the deposit process were the distance from source to substrate
(ssd) (5 mm and 8 mm) and the hydrogen flux levels (100 sccm and 25 sccm), whereas the parameters
that were kept constant are the deposit time (3 min), the voltage applied to the filaments (74 V),
the filament-to-source distance (fsd) (6 mm), and the system pressure (1 atm). A thermal annealing
was subsequently applied to the films obtained, which was applied at 1050 ◦ C. This temperature is
used because it has been reported that after thermal annealing at temperatures above 1000 ◦ C, SiOx
reacts to produce silicon nanocrystals or nanoclusters and structures with different oxidation states
with or without defects [17–19]. This thermal annealing was carried out in a tubular furnace with a
nitrogen flowing environment for 1 hour. In total, eight films were obtained; their labeling is described
in Table 1.
Table 1. Description of the labeling of the eight SiOx films obtained.
Hydrogen Flow Level: 25 sccm Hydrogen Flow Level: 100 sccm

Source-Substrate
Without Annealing With Annealing Without Annealing With Annealing
Distance:
5 mm A1 A1’ D1 D1’
8 mm A2 A2’ D2 D2’
Regarding the chosen parameters for the deposit, it is important to consider that both the deposit
temperature and the hydrogen gas flow level are important in the formation of thin films. The deposit
temperature is regulated by the distance between the source and substrate, while the filament-to-source
distance regulates the number of precursors.
The temperatures reached for the substrate during the deposit are shown in Figure 1, as well as
the thicknesses obtained, which were measured using profilometry. The trend in the temperatures
reached by the substrate follows what was mentioned regarding the distance from source to substrate,
while at a lower flow the substrate temperature increases. The temperature follows a linear trend for all
films; however, in the case of film, D2 is disturbed in the last seconds of deposit. This could be caused
by the buckling of some filament causing a decrease in the filament-to-source distance, increasing
temperature. In the case of the thicknesses, these increase with annealing. This abnormal behavior
could be explained as an additional formation of oxide caused by diffusion of residual O2 as well as
interstitial, as mentioned in [20], where despite performing a heat treatment in an Ar environment,
there is an increase in thickness.
Nanomaterials 2020, 10, x FOR PEER REVIEW 3 of 22
Figure 1. Substrate
Figure 1. Substrate temperature
temperature during
during SiO
SiOxx film
film deposition
deposition and
and thickness
thickness reached
reached for
for each
each film.
film.
To
Toobtain
obtainthe theoptical
opticalcharacteristics
characteristicsof ofthe
thedeposited
depositedSiO SiOxx films,
films, measurements
measurements of of the
the refractive
refractive
index
index (n) and extinction coefficient (k) were performed using spectroscopic ellipsometry on aa Horiba
(n) and extinction coefficient (k) were performed using spectroscopic ellipsometry on Horiba
UVISEL ◦ . In addition, the IR
UVISELellipsometer
ellipsometerwith withaaspectral
spectralrangerangeofof
(0.6–4.8
(0.6–4.8eV)eV)and
andincidence
incidence angle of 70
angle of 70°. In addition, the
absorption
IR absorptionspectra (FTIR)
spectra of the
(FTIR) ofSRO
the SROfilmsfilms
were were
obtained with awith
obtained Bruker Vector Vector
a Bruker 22 spectrometer in a range
22 spectrometer in
from 4000 cm −1 to 400 cm −1 with a resolution of 1 cm −1 . The photoluminescence (PL) spectra were also
a range from 4000 cm to 400 cm with a resolution of 1 cm . The photoluminescence (PL) spectra
−1 −1 −1
measured
were also with a FluroMax
measured with a3 Horiba
FluroMax Jobin Yvon Spectrofluorometer
3 Horiba with a 150 W xenon
Jobin Yvon Spectrofluorometer withexcitation lamp,
a 150 W xenon
0.3 nm resolution, range from 370 to 1000 nm, and high sensitivity emission detector.
excitation lamp, 0.3 nm resolution, range from 370 to 1000 nm, and high sensitivity emission detector. The excitation
line
The used to obtain
excitation linethe photoluminescence
used spectra was 335 nm.
to obtain the photoluminescence In addition,
spectra was 335XPS measurements
nm. In addition,wereXPS
made using an Escalab250Xi Thermo Scientific spectrometer
measurements were made using an Escalab250Xi Thermo Scientific spectrometer using a monochromatic Al XR15 X-raya
using
source. The depthAlprofile
monochromatic XR15was X-rayacquired
source.using Ar+ etching.
The depth profile The
wasbeam energy
acquired wasAr
using 4 keV, and the
+ etching. Thesputter
beam
current
energy was
was 2.5 The the
µA. and
4 keV, sizesputter
of the raster
currentwas approximately
was 2.5 μA. The size15 times
of thelarger
rasterthan
wasthe XPS-measured
approximately 15
area.
timesFinally, micrographs
larger than were obtained
the XPS-measured area. using
Finally,High Resolutionwere
micrographs Transmision
obtained Electronic
using HighMicroscopy
Resolution
(HRTEM)
Transmision in Nanotech
ElectronicJEOL JEM-2200FS
Microscopy + Cs equipment
(HRTEM) in Nanotech with
JEOLa spherical
JEM-2200FSaberration corrector on
+ Cs equipment withthea
condenser lens and operated at an acceleration voltage of 200 kV.
spherical aberration corrector on the condenser lens and operated at an acceleration voltage of 200
kV.
3. Results
The spectroscopic ellipsometry technique allows the optical parameters of interest to be known
3. Results
by obtaining the change in the amplitude of the light after reflection (Ψ) and the phase change of
The spectroscopic ellipsometry technique allows the optical parameters of interest to be known
this (∆), through the mathematical inversion of the data (Ψ, ∆), which are converted directly into
by obtaining the change in the amplitude of the light after reflection (Ψ) and the phase change of this
the optical constants refractive index (n) and extinction coefficient (k) [21]. With the values n and k
(Δ), through the mathematical inversion of the data (Ψ, Δ), which are converted directly into the
obtained by spectroscopic ellipsometry the transmittance and reflectance of all deposited SiOx films
optical constants refractive index (n) and extinction coefficient (k) [21]. With the values n and k
were calculated by means of the reflection and transmission coefficients for the s and p polarizations
obtained by spectroscopic ellipsometry the transmittance and reflectance of all deposited SiOx films
which were calculated using Snell’s law [22].
were calculated by means of the reflection and transmission coefficients for the s and p polarizations
Figure 2 shows the reflectance and transmittance obtained specifically from A1 and A1’ SiOx films.
which were calculated using Snell's law [22].
The transmittance and reflectance of the films A1 and A1’ are shown as a sample of the results obtained,
Figure 2 shows the reflectance and transmittance obtained specifically from A1 and A1´ SiOx
however, Figure 3 shows the complete results of the absorption coefficients obtained. In Figure 2,
films. The transmittance and reflectance of the films A1 and A1´ are shown as a sample of the results
transmittance less than 85% is observed and likewise reflectance less than 35%, in addition to the typical
obtained, however, Figure 3 shows the complete results of the absorption coefficients obtained. In
oscillations caused by the interference of the electromagnetic modes confined inside the films [23].
Figure 2, transmittance less than 85% is observed and likewise reflectance less than 35%, in addition
On the other hand, to analyze deeply these results, the absorption coefficient α is calculated using the
to the typical oscillations caused by the interference of the electromagnetic modes confined inside the
equations [23]
films [23]. On the other hand, to analyze deeply these results, the absorption coefficient α is calculated
T (1 − R)e−αt (1)
using the equations [23]
ln(T ) ln(1 − R) − αt (2)
≅ (1 − ) (1)
( )≅ (1 − ) − (2)
tendency prevails in all films, in addition to a slight increase in α for films deposited with a hydrogen
flow level of 100 sccm. Such behavior results due to thermal effects provoke an atomic structural
rearrangement which enriches the absorption mechanisms inside the material. Although this
phenomenon has been previously reported in this material, it is still not well understood what type
of arrangement
Nanomaterials occurs
2020, 10, 1415 to increase the absorption coefficient when applying thermal annealing. 4Itofhas 21
been suggested that it may be due to the formation of ultra-small amorphous silicon grains that
induce disorder in the network or due to the formation of nanocrystals since the optical properties of
bulk and nanoscale materials differ [24], orαit=has (1 − R)
1 also
ln been, suggested that it may be due to an increase (3)
in the Si dangling bonds given the hydrogen desorption t T [23] that is observed later. In the case of the
which
presentinvolve the values
investigation, theofthree
reflectance and transmittance
previously experimentally
mentioned phenomena occur,obtained.
for whichSince
all ofthe specified
them could
transmittance and reflectance data have multiple interferences, obtaining the optical
contribute to the increase of the absorption coefficient. Theoretical studies have been carried out in absorption
coefficient is a process
other research works that
[25]may include
where a certain degree
the reactions of error.
that occur whenForapplying
this reason, it is clarified
thermal that the
treatment are
values
modeled.obtained are parameters that describe a particular sample and not the physical property.
Figure 2. Reflectance and transmittance of the A1, A1’

Figure 2. A1´ SiO
SiOxx films without
without (A1)
(A1) and
and with
with thermal
thermal
Nanomaterials
annealing
annealing2020, 10, x FOR PEER REVIEW
(A1’).
(A1´). 5 of 22
Figure 3. The absorption coefficients of the SiOx films without

without (A1, A2, D1, D2) and with thermal
annealing (A1’,
(A1´, A2’,
A2´, D1’,
D1´, D2’)
D2´) considering
considering two
two source-to-substrate
source-to-substrate distances (ssd): 5mm (A1, A1´,
A1’, D1,
D1’)
D1´) and
and 8mm
8mm (A2,
(A2, A2’,
A2´, D2,
D2, D2’).
D2´).
The
Using α results calculated
the values of the for all deposited
absorption films are
coefficient (α),shown in Figure
we estimate the3,band
in which
gap the inset(BG)
energy shows the
using
absorption coefficients of the films deposited at a higher sdd. From this figure
the Tauc relation [26] where the BG was obtained using linear regression considering the equation it is observed that the
absorption coefficient displays values which are similar to those reported in SiOx films obtained using
(
other methods of analysis, such as the transmittance ) ⁄ = obtained − using UV-visible spectroscopy [16]. (4) It is
worth noting that thermal annealed
where α is the absorption coefficient, films increase their absorption coefficient
is the energy of the impinging photon, and such a tendency
is the BG, and A
prevails in all films,
is an arbitrary in addition
constant. We also to choose
a slight the
increase in α for
exponent ½, films
whichdeposited
corresponds withto
a hydrogen
an indirectflow level
allowed
of 100 sccm. Such behavior results due to thermal effects provoke
transition according to what other authors have used [27,28] for this material. an atomic structural rearrangement
whichAsenriches
is widely theknown,
absorption mechanisms
an amorphous insideexhibits
material the material. Although
a complex this of
structure phenomenon
the density has been
of states,
previously
such is the reported
case for thein this
SiOxmaterial,
films. For it this
is still not well
reason, theunderstood what type of
analysis of absorption arrangement
mechanisms in occurs
this typeto
increase the absorption coefficient when applying thermal annealing. It has been
of material is so complex; however, we can realize an approximation of the absorption coefficient suggested that it may
through the presence of electronic states inside the band gap energy in the proximity of the valence
and conduction band edges, as is proposed by the Urbach energies, which is now applied for the
deposited SiOx films using Urbach’s empirical rule, given by the equation [29]
( )
= . (5)
be due to the formation of ultra-small amorphous silicon grains that induce disorder in the network
or due to the formation of nanocrystals since the optical properties of bulk and nanoscale materials
differ [24], or it has also been suggested that it may be due to an increase in the Si dangling bonds
given the hydrogen desorption [23] that is observed later. In the case of the present investigation, the
three previously mentioned phenomena occur, for which all of them could contribute to the increase
of the absorption coefficient. Theoretical studies have been carried out in other research works [25]
where the reactions that occur when applying thermal treatment are modeled.
Using the values of the absorption coefficient (α), we estimate the band gap energy (BG) using the
Tauc relation [26] where the BG was obtained using linear regression considering the equation
(αhν)1/2 = A hν − E g

(4)
where α is the absorption coefficient, hν is the energy of the impinging photon, E g is the BG, and A
is an arbitrary constant. We also choose the exponent 12 , which corresponds to an indirect allowed
transition according to what other authors have used [27,28] for this material.
As is widely known, an amorphous material exhibits a complex structure of the density of states,
such is the case for the SiOx films. For this reason, the analysis of absorption mechanisms in this type
of material is so complex; however, we can realize an approximation of the absorption coefficient
through the presence of electronic states inside the band gap energy in the proximity of the valence and
conduction band edges, as is proposed by the Urbach energies, which is now applied for the deposited
SiOx films using Urbach’s empirical rule, given by the equation [29]
(hv)
α = a0 exp . (5)
Eu
where α is the absorption coefficient, a0 is a constant, and Eu is Urbach’s energy, from (5) we obtain the
following equation:
(hv)
ln α = ln a0 + (6)
UE
The Urbach energy (UE) can be obtained from the inverse slope of the straight line of the path ln α
versus hv energy of the incident photon.
Figure 4 shows an example of the procedure used to calculate the BG, while the inset in this figure
shows the value of the corresponding UE for D1’ film. To obtain the BG, the linear region belonging
to the high absorption edge was taken into account in such a way that in Figure 4 it corresponds to
energies from 3 eV to 3.5 eV, while to calculate the UE we utilize the energy region that is below the
high absorption edge which corresponds to the energy range from 2.2 eV to 3 eV. This procedure
is applied to each of the deposited films, and the results are displayed in Table 2. From the results
exhibited in Table 2, it is remarkable that when we have a higher ssd in the deposit or a greater flow of
hydrogen the value of BG energy increases and the UE decreases, a trend that is accomplished for all
the films except for the D2 film, this could be due to the increase in the temperature of the substrate in
the last seconds of deposit of the film D2, whereby there was a change in the deposit conditions, so the
trend is not followed.
As can be seen, in general the thermal annealing increases the BG value and decreases the UE,
which is an indication that a structural rearrangement occurs in the atomic lattice. Due to thermal
effects such a phenomenon favors the crystallinity of the material. For that reason the UE is reduced,
indicating that the amorphous phase is also reduced. This is corroborated using the FTIR technique.
Figure 5a shows the FTIR spectra offset in the absorbance axis for all SiOx films in a range from
400 cm−1 to 1400 cm−1 . The solid lines correspond to the FTIR spectra of the films without thermal
annealing and the dotted lines to the FTIR spectra with thermal annealing. By virtue of the intensity of
the localized peak in the range from 950 cm−1 to 1350 cm−1 , it is difficult to appreciate clearly both the
shape and position of the peaks having less intensity. For this reason, in Figure 5b) an amplification of
2.1 eV 2.21 eV
Without annealing
UE: 0.53 eV UE: 0.47 eV
2.35 eV 2.43 eV
With annealing
UE: 0.40 eV UE: 0.39 eV
Hydrogen flow: 100 sccm
D1 D2
the intensity of these peaks is shown by placing the 2.36 eV on a scale
spectra 2.13 eV 400 cm−1 to 950 cm−1 and
from
−1 Without
−1 annealing
from 2150 cm to 2400 cm . All the spectra shown haveeV
UE: 0.44 been normalized
UE: 0.50 eV with the purpose of not
considering the variation of the thickness of the films
2.39and
eV only taking
2.49 into
eV account the variations or
With annealing
shifts in wave number in the spectra. UE: 0.42 eV UE: 0.38 eV
Figure 4. Energy band gap calculated for D1’

D1´ film with
with its
its corresponding
corresponding Urbach
Urbach energy.
energy.
Table 2. Energy band gaps and Urbach energies.

As can be seen, in general the thermal annealing increases the BG value and decreases the UE,
which is an indication that a structuralHydrogen
rearrangement
flow: 25 occurs
sccm in the atomic lattice. Due to thermal
effects such a phenomenon favors the crystallinity ofA1 the material. ForA2
that reason the UE is reduced,
indicating that the amorphous phase is also reduced. This is corroborated using the FTIR technique.
2.1 eV 2.21 eV
Figure 5a shows the FTIR Without
spectra offset in the absorbance axis for all SiOx films in a range from 400
annealing
cm−1 to 1400 cm−1. The solid lines correspond UE: to 0.53
the eV UE: 0.47 of
FTIR spectra eV the films without thermal
annealing and the dotted lines to the FTIR spectra2.35 with eVthermal annealing.
2.43 eV By virtue of the intensity
With annealing
of the localized peak in the range from 950 cm UE: −1 to 1350 cm , itUE:
0.40 eV −1 is difficult
0.39 eV to appreciate clearly both
the shape and position of the peaks having less intensity.
Hydrogen flow: 100 sccm For this reason, in Figure 5b) an
amplification of the intensity of these peaks is shown by placing the spectra on a scale from 400 cm−¹
D1 D2
to 950 cm−¹ and from 2150 cm−¹ to 2400 cm−¹. All the spectra shown have been normalized with the
2.36 eV 2.13 eV
purpose of not consideringWithout
the variation of the thickness
annealing of the films and only taking into account the
variations or shifts in wave number in the spectra. UE: 0.44 eV UE: 0.50 eV
In the obtained FTIR spectra, the characteristic2.39absorption
eV peaks
2.49 eVof SiO2 were identified. These
With annealing
peaks correspond to the vibration modes Si–O–Si rocking
UE: 0.42 eV (R) at 458
UE: 0.38cmeV
−1, Si–O–Si bending (B) at 812
cm and Si–O–Si stretching (S) at 1082 cm [15]. There are shifts in the mentioned peaks indicating
−1, −1
a stoichiometry different from that of SiO₂, which confirms the fact that the films are made up of SiOx.
In the obtained FTIR spectra, the characteristic absorption peaks of SiO2 were identified.
The absorption peaks attributed to the Si–H bond were also found in the H–Si≡O 3 configuration,
These peaks correspond−1to the vibration modes Si–O–Si rocking (R) at 458 cm−1 , Si–O–Si bending (B)
which are B at 880 cm and S at 2250 cm ; in addition, with a lower intensity absorption peaks
−1
at 812 cm−1, and Si–O–Si stretching (S) at 1082 cm−1 [15]. There are shifts in the mentioned peaks
attributed to the Si-H bond in the H–Si≡Si3, H–Si≡O2, H–Si≡O1 configuration were found, which are S
indicating a stoichiometry different from that of SiO2 , which confirms the fact that the films are
made up of SiOx . The absorption peaks attributed to the Si–H bond were also found in the H–Si≡O3
configuration, which are B at 880 cm−1 and S at 2250 cm−1 ; in addition, with a lower intensity absorption
peaks attributed to the Si-H bond in the H–Si≡Si3 , H–Si≡O2 , H–Si≡O1 configuration were found, which
are S at 2100 cm−1 , 2156 cm−1 , and 2119 cm−1 , respectively. These configurations are reported in the
literature [29].
at 2100 cm−1, 2156

Nanomaterials cm10,
2020,
−1, and 2119 cm−1, respectively. These configurations are reported in the literature
1415 7 of 21
[29].
a) b)
Figure
Figure 5. (a) 5.
The(a)Fourier
The Fourier Transform
Transform Infrared
Infrared (FTIR)
(FTIR) spectra
spectra obtained
obtained forSiO
for the thex films
SiOx films in a range
in a range from from
400
−1 cm −1 to cm
1400
−1. cm −1 . (b) spectra
FTIR spectra obtained forx films
SiOx films in a range −1 950−1cm−1
400 cm to 1400 (b) FTIR obtained for SiO in a range fromfrom 400−1cm
400 cm to cm
to 950
and from −1 −1
and from 2150 2150
cm−1 cm
to 2400 tocm2400
−1. cm .
TableTable 3 shows
3 shows the positions
the positions of the
of the absorption
absorption peaks
peaks present
present in in
thethe deposited
deposited SiO
SiO x films
x films asas well as
well
the vibrational modes and the type of molecule to which they are attributed.
as the vibrational modes and the type of molecule to which they are attributed.
TableTable 3. Peak
3. Peak positions
positions present
present in the
in the spectra
spectra of Figure
of Figure 5, 5,
asas wellasasthe
well thevibrational
vibrationalmodes
modes and
and the
the type
of molecule to which they are attributed.
type of molecule to which they are attributed.
A1A1 A2A2 D1D1 D2

D2 A1’
A1´ A2’
A2´ D1’
D1´ D2’
D2´
Rocking
Rocking Si–O–Si 451
451cm
cm−1 453 cmcm 451 cm 455
455 cm 461 cm−1-1 461 cm−1-1 463 cm 463 cm
-1 -1 −1 -1 -1 −1-1 −1-1
Si–O–Si 453 451 cm−1 cm−1 461 cm 461 cm 463 cm 463 cm
Bending Si–O–Si 808 cm-1 802 cm-1 802 cm-1 800 cm-1 812 cm-1 812 cm-1 812 cm-1 812 cm-1
Bending 1074 cm-1−1 1068 cm-1 −1 1070 cm-1−1
Stretching Si–O–Si 1060 cm−1 1089cm cm 1089cmcm 1091 cm 1091
Si–O–Si 808 cm 802 cm 802 cm 800 cm -1
812 −1-1 812 −1-1 812 cm −1-1 812 cmcm
−1-1
Bending H–Si≡O3 883 cm-1 881 cm-1 881 cm-1 879 cm-1 - - -
Stretching -1 −1 2258 -1 −1 2256 -1 −1
Stretching H–Si≡O
Si–O–Si3 2258
1074cmcm cmcm
1068 cm
1070 cm 1060 cm-1−1
2256 - −1
1089 cm 1089 cm−1 - −1
1091 cm - −1
1091 cm
Bending
H–Si≡O 883 cm−1 881 cm−1 881 cm−1 879 cm−1 - - -
We can observe
3 that, compared to the peak corresponding to the vibrational mode R of SiO2, the
peak of Stretching
the SiOx films without thermal annealing
H–Si≡O3 2258 cm −1 2258 cm−1 2256 cm−1 shows 2256 cma−1slight shift- towards lower wave - numbers, -
this indicates a slight increase in silicon composition in some Si–O bonds, whereas with thermal
annealing we have a condition closer to that of the SiO2 stoichiometry.
On the Weother
can observe
hand, thethat,
peakcompared
located atto 812 thecmpeak corresponding
−1 (Si–O–Si B) appearstointheall vibrational
films; however, mode in R of SiO2 ,
films
the peak
without thermalof the SiOx films
annealing, without
this peak isthermal
attached annealing shows
to the peak a slightatshift
located 880 towards
cm−1, whichlower wave numbers,
corresponds
this indicates a slight increase in silicon composition in
to the H–Si≡O3 B molecule. This peak is closely related to the peak located at 2256 cm some Si–O bonds, whereas with thermal
−1 that
annealing we have a condition closer to that
corresponds to the vibrational mode S of the H–Si≡O3 molecule. of the SiO 2 stoichiometry.
Having both vibrational modes in
On the other hand, the peak located at 812 cm −1 (Si–O–Si B) appears in all films; however, in films
mind we corroborate the existence of the H–Si≡O3 configuration in the films without thermal
withoutwhichthermal annealing,tothis −1 , which corresponds to
annealing, is attributed thepeak is attached
hydrogen to the peakinlocated
incorporation at 880 cmprocess.
the deposition In addition,
the H–Si≡O −1 that corresponds
a stronger 3 B molecule.
intensity This peak ismodes
of these vibrational closelyisrelated
observed to the
forpeak located
the films at 2256 cm
deposited at a larger ssd. It
to the vibrational mode S of the H–Si≡O 3 molecule. Having
is also observed that when increasing the H–Si≡O3 B peak the Si-O-Si R peak decreases, which both vibrational modes in mind is we
corroborate the existence of the
highly linked to the increase of silicon contentH–Si≡O 3 configuration
in the material [30]. After the film is annealed, both is
in the films without thermal annealing, which
peaks attributed
disappear todue
the to
hydrogen incorporation
the desorption of hydrogenin theat deposition process. In
high temperatures addition,
and a stronger
the intensity of theintensity
Si–
of these vibrational modes is observed for the films
O–Si B mode increases, which is, as already mentioned, characteristic of SiO2. deposited at a larger ssd. It is also observed
that
The when increasing the H–Si≡O
peak corresponding 3 B peak the
to the vibrational mode Si-O-Si R peak information
S provides decreases, which
aboutisthe
highly linked to the
composition
of theincrease
deposited of silicon contentthis
films. When in the
peak material
displaces [30]. After the
towards filmwave
higher is annealed,
numberboth peaks
values disappear
it indicates due to
that
the desorption of hydrogen at high temperatures and the intensity of the Si–O–Si B mode increases,
which is, as2020,
Nanomaterials already mentioned,
10, x FOR PEER REVIEW characteristic of SiO2 . 8 of 22
The peak corresponding to the vibrational mode S provides information about the composition of
the
the density
deposited of films.
the Si–O–Si
Whenbonds this peakis increased,
displaces and suchhigher
towards films arewavemore stoichiometric
number [31]. In films
values it indicates that
without
the densitythermal
of the annealing,
Si–O–Si bondsthe peak assigned toand
is increased, the such
vibrational
films aremode of stretching
more stoichiometric is shifted
[31]. towards
In films
smaller
without wave
thermal numbers,
annealing, suggesting
the peak aassigned
higher to proportion of silicon
the vibrational modeinofthe amorphous
stretching phasetowards
is shifted and a
greater
smaller amount of unbound
wave numbers, oxidation
suggesting states.
a higher However,
proportion ofwhen
siliconapplying thermal annealing,
in the amorphous phase and the peak
a greater
goes
amounttowards larger wave
of unbound numbers,
oxidation states.which indicates
However, whenan increase
applyinginthermal
the Si–O–Si bonds; the
annealing, therefore,
peak goes the
oxygen
towardsbonds are also
larger wave increased
numbers, which and it indicates
indicates a change
an increase in theinSi–O–Si
the silicon
bonds; excess, this makes
therefore, the oxygen the
stoichiometry
bonds are alsoof the filmsand
increased closer to that of
it indicates the SiOin
a change 2 films.
the silicon excess, this makes the stoichiometry of
The XPS
the films closertechnique
to that of allows
the SiOinformation
2 films. to be obtained on the binding energy of the elements
present
TheinXPSa material.
techniqueFrom the XPS
allows measurements,
information the oxygen
to be obtained andbinding
on the silicon composition
energy of the ofelements
the SiOx
films
presentandinthe oxidationFrom
a material. statestheSin+ present
XPS were obtained.
measurements, the oxygen and silicon composition of the SiOx
filmsFigure
and the 6 shows
oxidation states Sin+ present
the composition profilewere
in percentage
obtained.of silicon and oxygen of the films according
to theFigure
depth of penetration,
6 shows the compositionwhereprofile
dotted line curvesofare
in percentage shown
silicon and for annealed
oxygen films,according
of the films while the to
continuous
the depth oflines represent
penetration, filmsdotted
where without linethermal
curves annealing.
are shown for With this data
annealed it was
films, possible
while to obtain
the continuous
lines
an represent
average valuefilms
of xwithout thermal annealing.
in the stoichiometric Withfor
SiOx ratio thiseach
dataofitthe
wasfilms
possible
usingtotheobtain an average
formula =
O
value of
ratio [32]. x in the stoichiometric SiO x ratio for each of the films using the formula Si = x ratio [32].
Figure
Figure 6. Compositional
Compositional profile
profile of
of silicon
silicon and
and oxygen
oxygen obtained
obtained using
using XPS
XPS according
according to
to the
the depth
depth of
of
penetration
penetration in the SiOxx film
film and
and xx values
values in
in the
the SiO
SiOxxstoichiometric
stoichiometricratio.
ratio.
In Figure
In Figure 6,
6, aa remarkable
remarkable change
change is
is observed
observed in in the
the amount
amount of of silicon
silicon and
and oxygen
oxygen present
present in
in the
the
proximity of
proximity of the
the film surface with respect to that present inside the volume. Such Such change
change isis more
more
remarkable for
remarkable forfilms
filmswith
withthermal
thermal annealing.
annealing. Particularly
Particularly for A1
for the the and
A1 D2
andfilms,
D2 films, the change
the change is more is
more abrupt. Such films were deposited at a lower ssd; in addition, for these samples
abrupt. Such films were deposited at a lower ssd; in addition, for these samples the annealing effectthe annealing
effect enhanced
enhanced the stoichiometry
the stoichiometry in thein the region
region nearsurface
near the the surface
due todue
bothto the
bothatomic
the atomic rearrangement
rearrangement and
the increment of oxygen atoms. The x values show that in general thermal annealing films present an
increase in x, which indicates an increase in the amount of oxygen and a tendency towards SiO2
stoichiometry.
To obtain the oxidation states of silicon Sin+ (with n = 1, 2, 3, 4) corresponding to the chemical
structures Si–OSi3, Si–O2Si2, Si–O3Si, and Si-O4, respectively [33], as well as the chemical state of silicon
and the increment of oxygen atoms. The x values show that in general thermal annealing films
present an increase in x, which indicates an increase in the amount of oxygen and a tendency towards
SiO2 stoichiometry.
To obtain the oxidation states of silicon Sin+ (with n = 1, 2, 3, 4) corresponding to the chemical
structures Si–OSi3 , xSi–O
Nanomaterials 2020, 10,
Si2 , REVIEW
FOR 2PEER
Si–O3 Si, and Si-O4 , respectively [33], as well as the chemical state of
9 of 22
silicon in bulk Si0 (Si2p3/2 and Si2p1/2), the XPS spectra of the Si2p peak were deconvolved to find the
corresponding bindingand
in bulk Si0 (Si2p3/2 energies attributed
Si2p1/2), to spectra
the XPS each of the previous
of the states.were
Si2p peak These energies aretoshown
deconvolved in
find the
Table 4 [33].
corresponding binding energies attributed to each of the previous states. These energies are shown
in Table 4 [33].
Table 4. Binding energies of the Si core levels and the Si oxidation states.
Table 4. Binding
Si2penergies of the Si core
Si2p Si1+levels and the Si oxidation
Si2+ Si3+ states.Si4+
3/2 1/2
Binding energies Si2p

99.5 Si2p
eV 3/2 100 eV 1/2 101Si
eV1+ SieV
101.5 2+ SieV
102.5 3+ SieV
103.5 4+
Binding energies 99.5 eV 100 eV 101 eV 101.5 eV 102.5 eV 103.5 eV

As an example, in Figure 7 the deconvolution of the spectrum corresponding to the D2 film is
shownAs
(a)anonexample, in Figure
the surface 7 the(0deconvolution
of the film of the
to 10 nm) and (b) spectrum
in the volumecorresponding
of the film (10 to
nmthe
to D2
300 film
nm).is
shown a) on the surface of the film (0 to 10 nm) and b) in the volume of the film (10 nm to 300 nm).
(a) (b)
Figure7.7.Deconvolution
Figure Deconvolutionof ofthe
theD2D2film
filmspectra
spectraobtained
obtainedusing
usingXPS
XPS(a)
(a)for
forthe
theregion
regionfrom
from00toto10
10nm
nm
and
and(b)
(b)for
forthe
theregion
regionfrom
from1010toto300
300nm.
nm.
The
The relative
relative concentration
concentration in
in percentage
percentage of
of each
each oxidation
oxidation state
state (I)
(I) can
can be
be obtained
obtained using
using
Equation (7) [33]:
Equation (7) [33]:
I n+
× 100 = I (n = 0, 1, 2, 3, 4) (7)
IT × = ( = , , , , ) (7)
where In+ is the area of the peak that represents the oxidation state Sin+ , and IT is the total area of
where
the is the
Si2p peak. In area of 7the
Figure peakbethat
it can represents
seen the surface
that for the oxidation state
part Sin+,film
of the andthe is the total
binding area of
energies
the Si2p peak. In Figure 7 it can be seen that for 0the surface
3+ part of the film
present are attributed to the oxidation states of Si and Si , with percentages of 9.9% and 90.1%, the binding energies
present are while
respectively, attributed to the the
in volume oxidation statesdisappear
Si–Si bonds of Si0 andand
Si3+the
, with percentages
oxidation states of
Si1+9.9%
, Si3+and
, and90.1%,
Si4+
respectively,
remain, while in volume
with percentages of 14.4%,the 42.4%,
Si–Si bonds disappear
and 43.2%, and the oxidation states Si , Si , and Si4+
respectively.
1+ 3+
remain,
The with
above percentages
process was of carried
14.4%, 42.4%,
out forand 43.2%,
all the respectively.
spectra on the surface of the film (10 nm), in the
center of the film at a depth of 60 nm, and finally at a depth ofon
The above process was carried out for all the spectra 300the
nm.surface of theare
The results film (10 nm),
shown in the
in Table 5,
center of the film at a depth of 60 nm, and finally at a depth of 300
together with the respective oxidation state attributed to that binding energy. nm. The results are shown in Table
5, together with the respective oxidation state attributed to that binding energy.
Table 5. Binding energies of the decomposed XPS peaks.
Si0 Si1+ Si2+ Si3+ Si4+ Si0 Si1+ Si2+ Si3+ Si4+
A1 99.7
102.5 eV 104.5 eV A2 102.8 eV
10 - - - eV - - -
10.6% 89.4% 10nm 87.7%
nm 12.3%
A1 99.9 101.8 101.2
103.2 eV A2 102.5 eV 103.5 eV
60 eV - eV - - eV -
57.3% 60 nm 40% 47.6%
nm 9.3% 33.4% 12.4%
Table 5. Binding energies of the decomposed XPS peaks.
Si0 Si1+ Si2+ Si3+ Si4+ Si0 Si1+ Si2+ Si3+ Si4+
A1
- - - 102.5 eV 104.5 eV A2 99.7 eV - - 102.8 eV -
10 nm 10.6% 89.4% 10 nm 12.3% 87.7%
A1 99.9 eV - 101.8 eV - 103.2 eV A2
- 101.2 eV - 102.5 eV 103.5 eV
60 nm 9.3% 33.4% 57.3% 60 nm 12.4% 40% 47.6%
A1 99.9 eV - 101.8 eV - 103.2 eV A2
- 101.2 eV - 102.5 eV 103.5 eV
300 nm 9.3% 33.4% 57.3% 300 nm 12.4% 40% 47.6%
A1’
- 101.2 eV 102.4 eV 103.3 eV A2’
- - - 102.6 eV 103.7 eV
10 nm 12.5% 38% 49.5% 10 nm 64% 36%
A1’
- - 102.6 eV 103.7 eV A2’
- - - 102.6 eV 103.7 eV
60 nm 25.8% 74.2% 60 nm 64% 36%
A1’
- 101 eV 102.6 eV 103.7 eV A2’
- - - 102.6 eV 103.7 eV
300 nm 16.5% 48.5% 35% 300 nm 64% 36%
D1 99.5 eV - - 102.3 eV 103.2 eV D2 99.8 eV - - 102.7 eV -
10 nm 7.4% 38.6% 54% 10 nm 9.9% 90.1%
D1
- - 101.5 eV 102.5 eV 103.8 eV D2
- 100.9 eV - 102.6 eV 103.6 eV
60 nm 18.1% 40.1% 40.8% 60 nm 14.4% 42.4% 43.2%
D1
- - 101.5 eV 102.5 eV 103.8 eV D2
- 100.9 eV - 102.6 eV 103.6 eV
300 nm 18.1% 40.1% 40.8% 300 nm 14.4% 42.4% 43.2%
D1’
- 101.2 eV - 102.8 eV 103.8 eV D2’
- - - 102.5 eV 103.5 eV
10 nm 14.8% 45.2% 40% 10 nm 39.3% 60.7%
D1’
- 101.2 eV - 102.8 eV 103.7 eV D2’
- - - 102.5 eV 103.6 eV
60 nm 17.4% 43.8% 38.8% 60 nm 38% 62%
D1’
- - - 102.5 eV 103.8 eV D2’
- - 102.5 eV 103.6 eV
300 nm 43.7% 56.3% 300 nm 38% 62%
From Table 5 it is observed that films without thermal annealing present Si0 states generally on
the surface of the film, except for the A1 film that shows it only in volume, this being also the one with
the highest silicon content. Comparing the films according to the ssd of the deposit, films deposited at
greater ssd without thermal annealing show the oxidation states Si0 and Si3+ on the surface of the film,
while in volume they have oxidation states Si1+ , Si3+ , and Si4+ . When applying thermal annealing to
these films the oxidation states are reduced to Si3+ and Si4+ , both for the surface of the film and for
the volume; that is, the stoichiometry of the films is closer to that of the SiO2 and the films tend to
become uniform in their stoichiometry, in addition the film deposited to greater flow D2’, has a higher
percentage of Si4+ oxidation state than that deposited at 25 sccm (A2’).
As for the films deposited at lower ssd without thermal annealing, the oxidation state Si1+ is
absent while the oxidation state Si2+ is only present in volume and Si3+ is only present on the surface
of the sample with lower flux level (25 sccm), but it exists both on the surface and in volume for the
sample with a higher flux level (100 sccm), and Si4+ exists for both space regions in the two samples.
Additionally, the films show significant changes in the percentages of oxidation states only in the
most superficial layer. On the other hand, the relevant characteristic of the annealed films is that they
present the oxidation states Si3+ and Si4+ in both space regions; however, the oxidation states Si0 and
Si2+ are absent. The oxidation state Si1+ exists on the surface of both samples and in volume but at
different depths.
Once the molecular bonds present in the material were identified using the FTIR technique,
in addition to the oxidation states being quantified using XPS, it was also necessary to know the
structural defects in the SiOx films, for which photoluminescence (PL) measurements were carried out.
The results obtained of such PL measurements are shown in Figure 8, where the inset in the graph
depicts the spectra obtained from the films without thermal annealing, while the main graph shows
the PL spectra obtained from the annealed films. It is worth noting that the PL intensity from all films
with thermal annealing is augmented approximately up to 4 times with respect to that of the films
without thermal annealing. Additionally, such spectra undergo a shift in the emission wavelength
towards lower energies in such a way that PL intensity lies in the range 650–900 nm. It should be
noted that the strongest PL intensity corresponds to the annealed films deposited at a flow level of
100 sccm, while for films without thermal annealing it corresponds to the films deposited at flow levels
of 100 sccm and 25 sccm, and the emission region is wider, in the range 400–800 nm. It is noticeable
Nanomaterials
Nanomaterials 2020,
2020, 10,
10, x
x FOR
FOR PEER
1415 PEER REVIEW
REVIEW 11
11 of
of 22
21
22
flow
flow levels
levels of of 100
100 sccm
sccm andand 25 25 sccm,
sccm, and and the
the emission
emission regionregion is is wider,
wider, inin the
the range
range 400–800
400–800 nm. nm. It It is
is
that all
noticeable the emission
that all the PL bands
emission are
PL wide,
bands which
are indicates
wide, which that they
indicates are
that
noticeable that all the emission PL bands are wide, which indicates that they are composed of several composed
they are of several
composed emission
of several
peaks
emission
emission thatpeaks
have that
peaks a shape
that havesimilar
have aa shape
shape tosimilar
a Gaussian
similar to one.
to aa Gaussian
Gaussian one.
one.
For SiO x
x , the
the most
most accepted
accepted light
light emission
For SiOx, the most accepted light emission mechanisms which
, emission mechanisms
mechanisms which give
which give rise
give rise to
rise to the
to the visible
visible andand near
near
infrared bands are attributed to the following causes: the
the effect
effect of
of dimensional
dimensional
infrared bands are attributed to the following causes: the effect of dimensional quantum confinement quantum
quantum confinement
confinement
in Si-ncs
in Si-ncsoror
Si-ncs orsilicon
siliconnanoclusters
silicon nanoclusters
nanoclusters (Si-ncl) whose
(Si-ncl)
(Si-ncl) whose
whosereported photoluminescence
reported
reported photoluminescence
photoluminescence lies inlies
the in
lies emission
in the energy
the emission
emission
range from
energy
energy range1.3
range eV 1.3
from
from to 1.7
1.3 eV
eV to eV1.7
to [34–36]
1.7 eV and the
eV [34–36]
[34–36] andeffect
and the related
the effect
effect to defects
related
related to in the
to defects
defects in oxide
in the
the oxide
oxidematrix
matrix
matrix suchsuch
suchas
weak
as
as weak oxygen
weak oxygen bonds
oxygen bonds (WOB)
bonds (WOB) with
(WOB) with a PL
with aa PL peak
PL peak at
peak at 3 eV
at 33 eV [36],
eV [36], neutral
[36], neutral oxygen
neutral oxygen vacancy
oxygen vacancy
vacancy (NOV) (NOV) with
(NOV) with PL PL
ranging 2.9 eV [37], hydrogen-related
ranging from 2.8 eV to 2.9 eV [37], hydrogen-related defects (H) with PL located in the
from 2.8 eV to defects (H) with PL located in the range
range from
from
2.2 eV to 2.5 eV [38], and finally a non-bridging oxygen hole
2.2 eV to 2.5 eV [38], and finally a non-bridging oxygen hole center (NBOHC) with PL from 1.8
eV to 2.5 eV [38], and finally a non-bridging oxygen hole center
center (NBOHC)
(NBOHC) withwith PL PL
fromfrom1.81.8
eVeV
eV to
to
to 2
22 eV eV [39,40].
eV [39,40].
[39,40]. These These
These emission emission
emission mechanisms mechanisms
mechanisms make make
make thethe PL the PL
PL spectra spectra
spectra exhibit exhibit
exhibit aa wide a wide
wide shape,
shape, as shape,
as can
can be as can
be seen
seen inbe
in
seen in 8.
Figure
Figure Figure
8. In
In order
order8. Intotoorder
find to
find allfind
all all possible
possible
possible contributions
contributions
contributions to the to
to the the photoluminescent
photoluminescent
photoluminescent processes
processes
processes in theinSiO
in the thexx
SiO
SiO
films,
x films,
the the deconvolution
deconvolution of of
each each spectrum
spectrum was was performed,
performed, by by
way
films, the deconvolution of each spectrum was performed, by way of example. Figure 9 depicts the way
of of example.
example. FigureFigure
9 9 depicts
depicts the
the deconvolution
deconvolution
deconvolution of
of theof photoluminescent
the the photoluminescent
photoluminescent spectrum
spectrum
spectrum of
of theofD2´
the the film.
D2´ D2’ film.
film. To To perform
To perform
perform the the deconvolution
the deconvolution
deconvolution and
and
and
in in
order order
to to
avoid avoid common
common errors
errors reported
reported in inthethe photoluminescence
photoluminescence
in order to avoid common errors reported in the photoluminescence analysis, the spectra were analysis,
analysis, the
the spectra
spectra were
deconvolved using
deconvolved using the the PLPL intensity-energy
intensity-energy differential
differential versus
differential versus energy
versus energy [41].
energy [41].
[41].
Figure
Figure 8.
8. Photoluminescence
Photoluminescence spectra
spectra measured
measured from
from deposited
deposited SiO
SiOxx films.
films.
Figure
Figure 9. Photoluminescence spectrum
9. Photoluminescence
Photoluminescence spectrum deconvolved
spectrum deconvolved for
deconvolved for the D2’ film.
the D2´
D2´ film.
Table 6 shows the position

Table of the peaksthe
obtained from the deconvolution of the photoluminescent
Table 66 shows
shows thethe position
position ofof the peaks
peaks obtained
obtained fromfrom the
the deconvolution
deconvolution of of the
the
spectra, in addition
photoluminescent to the defects to which the emission is attributed, according to their energy position.
photoluminescent spectra,
spectra, in
in addition
addition to
to the
the defects
defects to
to which
which the
the emission
emission is
is attributed,
attributed, according
according toto
their energy position.
their energy position.
Table 6. PL deconvoluted bands and their proposed defects contributions.
A1 A1´ A2 A2´ D1 D1´ D2 D2´

WOB
Nanomaterials -2020, 10, 1415
3.09 eV 3.17 eV 3.3 eV WOB - 3.15 eV 3.1 eV 3.09 eV21
12 of
NOV 2.79 eV - 2.82 eV - NOV 2.90 eV - 2.82 eV -
H 2.34 eV - 2.53 eV - H 2.40 eV - 2.32 eV -
NBOHC 1.95 eV Table 2.046.eVPL deconvoluted
- bands
1.82 eVand their
NBOHCproposed
2.03defects
eV contributions.
2.04 eV - 1.95 eV
1.51 eV 1.51 eV
1.53 eV 1.48 eV 1.53
Si-ncl - A1 1.58eV A1’ 1.74 -A2 A2’ Si-ncl -D1 D1’
1.59 eV D2
- D2’
1.63 eV eV 1.67 eV
eV 1.76 eV
WOB - 3.09 eV 3.17 eV 3.3 eV WOB - 3.15 eV 3.1 eV 3.09 eV
NOV 2.79 eV - 2.82 eV - NOV 2.90 eV - 2.82 eV -
As can be seen in Table 6, films without thermal annealing show photoluminescence attributed
to NOV and 2.34
H H, ineVaddition- to films
2.53 eV
deposited - a shorter
at 2.40 eV the presence
H ssd showing - 2.32 eV
of NBOHC, -
while
those
NBOHCdeposited at a larger
1.95 eV ssd have- WOB.1.82
2.04 eV ForeVannealed
NBOHCfilms,2.03
photoluminescence
eV 2.04 eV is-mainly1.95
dueeVto
Si-ncl effects and to a lesser
1.51 eVextent to NBOHC and WOB defects. 1.51 eV 1.48 eV
1.53 eV
Si-ncl - 1.58 eV - Si-ncl - 1.59 eV - 1.53 eV
1.63 eV
4. Discussion 1.74 eV 1.76 eV 1.67 eV
Using the calculated values of the BG, the UE parameters, and making a correlation with the
resultsAsobtained
can be seen
usingin XPS,
Tablewe6, films without
identify a clearthermal annealing
dependence showthe
between photoluminescence
stoichiometric ratioattributed
(x) and
to NOV and H, in addition to films deposited at a shorter ssd showing the presence
the BG and UE values. Figures 10 and 11 display how the correlation between the stoichiometric ratio of NBOHC,
while
and BGthose
and deposited
UE evolves. at a larger ssd have WOB. For annealed films, photoluminescence is mainly due
to Si-ncl
Figure 10 shows a aclear
effects and to lesser extent to
tendency inNBOHC and WOB
films in which withdefects.
a higher silicon or lower oxygen content
(lower value of x) the lower BG is obtained. Such a trend is held both with and without annealing,
4. Discussion
and it is relevant to point out that the BG increases significantly due to both thermal annealing and
Using the
the increase calculated
of oxygen values
content. In of the BG,the
addition, theUE
UEdecreases
parameters, and makingina accordance
monotonously correlation with
with thethe
results obtained using XPS, we identify a clear dependence between the stoichiometric
silicon content decreasing with the thermal annealing, as is shown in Figure 11. Special attention ratio (x) and
the BG and
deserves UEpaid
to be values. Figures
to the 10 and
fact that the 11 displayfilms
annealed how with
the correlation
the larger between the stoichiometric
ssd and higher ratio
flow level offer
andhighest
the BG andBG UEandevolves.
stoichiometric ratio and the lowest UE.
Figure10.
Figure Correlationbetween
10.Correlation betweenthe
thestoichiometric
stoichiometricratio
ratioxxand
andthe
thecalculated
calculatedenergy
energyband
bandgap
gapvalues.
values.
Figure 10 shows a clear tendency in films in which with a higher silicon or lower oxygen content
(lower value of x) the lower BG is obtained. Such a trend is held both with and without annealing,
and it is relevant to point out that the BG increases significantly due to both thermal annealing and the
increase of oxygen content. In addition, the UE decreases monotonously in accordance with the silicon
content decreasing with the thermal annealing, as is shown in Figure 11. Special attention deserves to
be paid to the fact that the annealed films with the larger ssd and higher flow level offer the highest BG
and stoichiometric ratio and the lowest UE.
The calculated Urbach energies occupy a wide range of values ranging from 380 meV to 530 meV.
These are considerable energy values, and it indicates in some way the level of structural disorder
that exists in the SiOx films. This disorder is remarkably diminished both with the oxygen increment
and with the thermal annealing. In addition to that the dependence among the stoichiometric ratio,
band gap and Urbach energies is well defined and agrees with the model established by Mott [31].
For values of the stoichiometry ratio x further from that of the stable configuration of silicon dioxide,
the films undergo a greater disorder in the molecular structure of the material, which brings the
increment of aggregate energy levels inside the intrinsic band gap energy of the dioxide matrix, causing
the formation of wider band tails and consequently a smaller band gap energy that tends towards that
of the silicon
Nanomaterials one.
2020, 10, x FOR PEER REVIEW 13 of 22
Figure
Figure11.
11.Correlation
Correlationbetween
betweenthe
thestoichiometric
stoichiometricratio
ratioxxand
andthe
thecalculated
calculatedUrbach
Urbachenergy
energyvalues.
values.
Comparing
The these
calculated results
Urbach with the
energies deposition
occupy a wide parameters,
range of valueswe find a behavior
ranging fromin 380which,
meV with
to 530a
lower These
meV. flow level and shorter ssd,
are considerable filmsvalues,
energy with a and
higher silicon content
it indicates in some(lower
way BG theand higher
level UE) are
of structural
obtained, and therefore, the optical response of the material is substantially modified.
disorder that exists in the SiOx films. This disorder is remarkably diminished both with the oxygen It is worthwhile
to note thatand
increment the trend
with inthewhich the flow
thermal level andInssdaddition
annealing. determine to the
thatdependence
the dependencebetweenamongthe silicon
the
content and BG ratio,
stoichiometric and UE is not
band gapwelland
defined
Urbachfor the case ofisthe
energies films
well withoutand
defined thermal
agrees annealing,
with thecontrary
model
to the case ofby
established theMott
annealed
[31]. films for which
For values the increment
of the stoichiometry of the ssd and
ratio flow level
x further fromdiminish
that of the silicon
stable
content and increase
configuration thedioxide,
of silicon BG and the reduce
filmsthe UE clearly.
undergo Furthermore,
a greater disorder for the molecular
in the films without thermal
structure of
annealing,
the material, thewhich
presence
bringsof hydrogen is observed
the increment in the energy
of aggregate FTIR spectra
levels of thesethe
inside films (see Figure
intrinsic band gap12),
which
energysuggests that hydrogen
of the dioxide plays anthe
matrix, causing important
formation roleofwhich
widercauses a clear
band tails anddependence
consequently between the
a smaller
silicon
band content
gap energyand the
that GB
tendsand UE
towards
Nanomaterials 2020, 10, x FOR PEER REVIEW parameters
that of thenot to exist,
silicon one.as in the case of the annealed films in which
14 of 22
hydrogen is no longer
Comparing these present.
results with the deposition parameters, we find a behavior in which, with a
lower flow level and shorter ssd, films with a higher silicon content (lower BG and higher UE) are
obtained, and therefore, the optical response of the material is substantially modified. It is worthwhile
to note that the trend in which the flow level and ssd determine the dependence between the silicon
content and BG and UE is not well defined for the case of the films without thermal annealing,
contrary to the case of the annealed films for which the increment of the ssd and flow level diminish
the silicon content and increase the BG and reduce the UE clearly. Furthermore, for the films without
thermal annealing, the presence of hydrogen is observed in the FTIR spectra of these films (see Figure
12), which suggests that hydrogen plays an important role which causes a clear dependence between
the silicon content and the GB and UE parameters not to exist, as in the case of the annealed films in
which hydrogen is no longer present.
Figure 12 shows an amplification of the IR spectrum in the region from 2100 cm-1 to 2400 cm-1
for the D2 film; in addition, the deconvolution made from the 2075 cm-1 peak to the 2150 cm-1 one is
shown. Such deconvolutions were performed to observe the absorption peaks attributed to the Si–H
bonds in the H–Si≡Si3, H–Si≡O2 and H–Si≡O1 configurations, while Table 7 lists the peak positions for
all films without thermal annealing, which are the ones that present the Si–H bonds.
Figure 12. FTIR spectrum of the D2 film showing the absorption peaks attributed to the different
Figure 12. FTIR spectrum of the D2 film showing the absorption peaks attributed to the different
configurations of the Si-H bonds.
configurations of the Si-H bonds.
Table 7. Si–H stretching vibration bands.
A1 A2 D1 D2
H–Si≡Si 2100 cm-1 2097 cm-1 2100 cm-1 2101 cm-1
H–Si≡O1 2117 cm-1 2108 cm-1 2115 cm-1 2119 cm-1
Figure 12 shows an amplification of the IR spectrum in the region from 2100 cm−1 to 2400 cm−1
for the D2 film; in addition, the deconvolution made from the 2075 cm−1 peak to the 2150 cm−1 one is
shown. Such deconvolutions were performed to observe the absorption peaks attributed to the Si–H
bonds in the H–Si≡Si3 , H–Si≡O2 and H–Si≡O1 configurations, while Table 7 lists the peak positions for
all films without thermal annealing, which are the ones that present the Si–H bonds.
Table 7. Si–H stretching vibration bands.
A1 A2 D1 D2
H–Si≡Si 2100 cm−1 2097 cm−1 2100 cm−1 2101 cm−1
H–Si≡O1 2117 cm−1 2108 cm−1 2115 cm−1 2119 cm−1
H–Si≡O2 2154 cm−1 2150 cm−1 2165 cm−1 2156 cm−1
H–Si≡O3 2258 cm−1 2258 cm−1 2256 cm−1 2256 cm−1
Figure 12 shows the presence of the H–Si≡Si3 , H–Si≡O2 , and H–Si≡O1 configurations,
whose contributions are minimal compared to the one due to the H–Si≡O3 configuration, which is
found at position 2256 cm−1 and shows a significant right shoulder which is ascribed to an effect of the
first neighbors with silicon atoms [30]. When comparing these results with those obtained using XPS,
where it is observed that the predominant oxidation states are Si3+ and Si4+ , t is then confirmed that an
important part of these Si≡O3 bonds are bound to hydrogen atoms.
Regarding the hydrogen content in the films without thermal annealing, it is greater for films
deposited at a larger ssd. This is explained by indicating that a greater deposit distance leads to the
formation of weak silicon-silicon or silicon-oxygen bonds. These bonds, as it is well known, are broken
by the action of hydrogen forming covalent bonds Si–H, which explains the presence of the H–Si≡O3 ,
H–Si≡Si3 , H–Si≡O2 , and H–Si≡O1 configurations in the FTIR spectra.
The results obtained using PL are used in the analysis of the formation of defects in the SiOx films;
however, it is hitherto the most controversial and difficult topic without conclusive reasons. In the case
of the results obtained for films deposited without thermal annealing, the abovementioned result is
shown regarding the presence of weak oxygen bonds, which are shown for films deposited at a greater
ssd. Neutral oxygen vacancies are present for all films without thermal annealing, which is common
for this type of material.
The emission related to defects involving hydrogen is more evident for films deposited at a
larger ssd (A2, D2). In the case of NOVs, it is known that they react with hydrogen to form H–Si≡
configurations, so it is interesting to note that this defect is only present in films without thermal
annealing. It is very probable that films without annealing have converted a part of their NOVs into
H–Si≡O3 configurations.
Since the films without thermal annealing deposited at a shorter ssd have a higher silicon content
and oxidation states Si0 with respect to the other films, we expect that the emission at 1.9 eV is not
mostly due to NBOHC defects that are closely related to the SiO4 , but to another type of defect which
is also reported with this characteristic emission, interstitial oxygen (O− −
2 , O3 ), which would be more
appropriate since NBOHC is not present in the A2 and D2 films that have a greater amount of Si4+
oxidation states; however, it is not discarded that it is a contribution of both emission mechanisms.
In the case of PL in annealing films, they all show the same emission mechanisms, attributed
to WOB, NBOHC, and the greater contribution to PL is attributed to Si-ncl effects, which brings
confinement effects due to a nanocrystal formation producing an increase of the band gap, which is
controlled by the Si nanostructure size [42]; however, it is believed that there is a better explanation for
this purpose, which will be discussed later, in addition to the fact that emission intensity increases
substantially compared to that of films without thermal annealing, which indicates that the structural
rearrangement of the material is closely related to this reported emission. To understand how the
material restructuring occurs, we take into account the results obtained from the deconvolution of the
Si2p peak, as mentioned by Tomozeiu [31]. Based on the Gibbs free energy calculations, it has been
shown that the bonds Si-(Si4 ) and Si-(O4 ) are stable, while the Si-(Sin O4−n ) bonds, with n = 1, 2, 3 are
unstable, being Si-(Si2 O2 ), which corresponds to a state of Si2+ oxidation, the less stable structural
entity. The films deposited at an ssd of 5 mm without thermal annealing present this oxidation
state, which leads to a greater tension or stress of the chemical bond between the central silicon atom
and the oxygen ones, so if the conditions for the migration of an oxygen atom are satisfied, a phase
decomposition will occur following the next reaction:
Si-(Si2 O2 ) + Si-(Si2 O2 ) → Si-(Si1 O3 ) + Si-(Si3 O1 )
Si-(Si2 O2 ) + Si-(Si2 O2 ) → Si-(O4 ) + Si-(Si4 )
The first reaction is observed according to the oxidation states present when applying thermal
annealing to these films (Si1+ and Si3+ ). However, in the films deposited at a larger ssd, the states Si1+
and Si3+ are already present in the films without thermal annealing, indicating that the deposition
of the film under this condition leads to a more stable lattice growth that could be favored as the
results of FTIR indicate. Due to the incorporation of hydrogen and the formation of a greater amount
of H–Si≡O3 bonds, where hydrogen helps to reduce tension in the molecular structure, the presence
of hydrogen could be inferred not only using this technique, but could also be considered through
the photoluminescent emission band (2.2 eV to 2.5 eV) attributed to defects related to hydrogen [40].
When applying thermal annealing to these films as observed using FTIR, a hydrogen desorption occurs
so that the H–Si≡O3 bond can become a Si-(SiO3 ) bond with the addition of a silicon atom. It can be
observed in Table 6 that when the films were annealed, only the oxidation states Si3+ and Si4+ survived,
and such states are present in the A2’ and D2’ films that have a stoichiometry closer to that of SiO2 and
a major BG and a smaller UE.
As previously mentioned, reports that exist regarding the structural arrangements that are
presented when thermal annealing is performing in this type of films suggest, given the stoichiometry,
the formation of isolated nanoclusters or nanocrystals of silicon immersed in a SiO2 matrix, instead of
an arrangement of the type of percolated networks or Si-SiO2 sponges that are formed when there
is an excess of Si greater than 30% [20,43]. One method to characterize these silicon nanocrystals or
nanoclusters is through the XRD technique; however, a difficulty has been reported in obtaining a
diffraction pattern for very small Si nanocrystals dispersed in the amorphous SiOx matrix due to the
fact that they have a short-range crystalline order [44]; therefore, to complement this study, the HRTEM
technique was carried out in order to corroborate the presence of Si-ncs or Si-ncls. The HRTEM
micrographs for films A1 and A2 are shown in Figures 13 and 14, respectively. These figures show
the presence of silicon clusters. In Figure 13, film A1 has silicon clusters of different sizes and has the
largest clusters, which corresponds to what was observed using XPS regarding the presence of Si0
states; in addition, this film has a greater excess of silicon. In the case of Figure 14, which shows the
micrograph of film A2, the presence of silicon clusters can be seen in the same way, but in this case
with a smaller size.
The HRTEM micrographs for films D1 and D2 are shown in Figures 15 and 16, respectively.
These figures show the presence of dark spots that could represent silicon clusters in which the crystalline
orientation was not observed (Si-ncls). They have an average size of 3 nm, which corroborates what
was inferred using XPS with respect to the existence of a greater silicon concentration (Si0 state) in
films without thermal annealing.
Figures 17 and 18 show the HRTEM micrographs of the D1’ fil., Figure 17 shows a large number
of clusters with a diameter <2 nm, while Figure 18 shows an enlarged micrograph showing certain
regions with dimensions between 1 nm and 2 nm that show a certain crystalline orientation. To obtain
the crystallographic orientation, the interplanar distances were measured. The micrographs shown in
Figure 18 were analyzed using a Digital Micrograph program using Fourier transform to obtain the
reciprocal space and thus be able to obtain the interplanar distances. With the interplanar distances,
we proceeded to search the orientation of the nanocrystals by means of the crystallographic data
provided by the database PDF 4. The presence of crystalline silicon with reference code 00-027-1402
is appreciated.
66.28 nm
66.28 nm
15.70 nm
15.70 nm
20 nm Si
Si
20 nm
Figure 13. High

13. High Resolution
Resolution Transmition Electronic Microscopy
Transmition micrograph of the A1 SiOx film.
Figure Figure 13. High Resolution TransmitionElectronic Microscopy
Electronic Microscopy micrograph
micrograph of theofA1
the A1
SiO SiOx film.
x film.
6.87 nm
6.87 nm
19.68 nm
19.68 nm
12.94 nm
12.94 nm
12.25 nm
nm 50 nm
50 nm
12.25
Figure 14. HRTEM micrograph of the A2 SiOx film.

Figure 14. HRTEMmicrograph
Nanomaterials 2020, 10, x FORFigure 14. HRTEM
PEER REVIEW
micrograph ofofthe A2A2
the SiOSiO
x film.
film. 17 of 22
x
The HRTEM
The HRTEM micrographs
micrographs forfor films
films D1
D1 and
and D2
D2 are
are shown
shown inin Figure
Figure 15
15 and
and Figure
Figure 16,
16,
respectively. These figures show the presence of dark spots that could represent silicon clusters
respectively. These figures show the presence of dark spots that could represent silicon clusters in in
which the crystalline orientation was not observed (Si-ncls). They have an average size of 3 nm,
which the crystalline orientation was not observed (Si-ncls). They have an average size of 3 nm, whichwhich
corroborates what
corroborates what was was inferred
inferred using
using XPS
XPS with
with respect
respect toto the
the existence
existence of
of aa greater
greater silicon
silicon
concentration (Si
concentration (Si00 state)
state) in
in films
films without
without thermal
thermal annealing.
annealing.
Figure
15. 15. HRTEMmicrograph
HRTEM micrograph ofofthe
theD1SiOx film.
D1SiO
Figure x film.
Figure 15. HRTEM micrograph of the D1SiOx film.
Figure
Figure
Nanomaterials 2020, 10, x FOR PEER 16.HRTEM
16.
REVIEW HRTEM micrograph
micrograph of
of the
the D2
D2SiO
SiOx xfilm.
film. 18 of 22
Figures 17 and 18 show the HRTEM micrographs of the D1’ fil., Figure 17 shows a large number
of clusters with a diameter < 2 nm, while Figure 18 shows an enlarged micrograph showing certain
regions with dimensions between 1 nm and 2 nm that show a certain crystalline orientation. To obtain
the crystallographic orientation, the interplanar distances were measured. The micrographs shown
in Figure 18 were analyzed using a Digital Micrograph program using Fourier transform to obtain
1.80 nm
the reciprocal space and thus be able to obtain the interplanar distances. With the interplanar
1.71 nm 1.80 nm
distances, we proceeded to search the orientation of the nanocrystals by means of the crystallographic
1.71 nm
data provided by the database PDF 4. The presence of crystalline silicon with reference code 00-027-
1402 is appreciated.
1.31nm
1.31 nm
20
20nm
nm
Figure
Figure 17.
17.
Figure HRTEM
17.HRTEM micrograph
HRTEMmicrograph of the
ofthe
micrograph of D1´
theD1´ SiO
D1’SiO
SiO xfilm.
film.
x xfilm.
5 nm
5 nm nm
5 nm
20.511/nm
Si-nc 20.51
Si-nc 1/nm
1.83 nm
1.83 nm (4,4,0) d=0.097
d=0.097 nm
nm 1.83 nm
1.83 nm
(4,4,0)
1.67nm
1.67 nm 1.67
1.67nm
nm
1.94
1.94 nmnm 1.94
1.94nm
nm
1.90 nm 1.90 nm
1.90 nm 1.90 nm
Si-nc
d=0.191 nm
Si-nc Si-nc
d=0.191 nm d=0.110 nm
(2,2,0) Si-nc d=0.110 nm
d=0.167 nm (4,2,2)
d=0.167 nm Si-nc (2,2,0) (4,2,2) 18.10 1/nm
(3,1,1)
(3,1,1) Si-nc
11.96 1/nm 10.43 1/nm 18.10 1/nm
11.96 1/nm 10.43 1/nm
Figure
18. 18.
HRTEMHRTEM micrographofofthe
micrograph theD1’
D1´ SiO
SiOX film with crystallographic orientations.
Figure X film with crystallographic orientations.
Figure 18. HRTEM micrograph of the D1´ SiOX film with crystallographic orientations.
Correlating the results obtained from the various characterizations performed and the
Correlatingshown,
micrographs the results obtained
it is important from thethat
to highlight various characterizations
the presence performed
of the Si0 state in all theandfilmsthe
micrographs shown,
without thermal it is important
annealing to formation
is due to the highlight ofthat the with
Si-ncls presence of the Si
sizes greater 0 state
than in In
10 nm. allthe
thecase
films
of the film
without A1 which
thermal is the is
annealing one
duewith
to the highest excess
formation of silicon
of Si-ncls corresponding
with sizes greatertothan
the one that In
10 nm. reached
the case
highest
of the film A1 temperature
which is the of onethewith
deposit and hasexcess
the highest the lowest flow, corresponding
of silicon it also presents Si-ncls of greater
to the one size
that reached
as shown in Figure 13. The existence of the state Si 0 in the SiOx films, as can be seen in the phase
the highest temperature of the deposit and has the lowest flow, it also presents Si-ncls of greater size
as decomposition,
shown in Figure generally
13. Theoccurs afterof
existence samples areSi
the state subjected to ax thermal
0 in the SiO films, asannealing
can be seenabovein 1000 °C
the phase
Correlating the results obtained from the various characterizations performed and the micrographs
shown, it is important to highlight that the presence of the Si0 state in all the films without thermal
annealing is due to the formation of Si-ncls with sizes greater than 10 nm. In the case of the film A1
which is the one with the highest excess of silicon corresponding to the one that reached the highest
temperature of the deposit and has the lowest flow, it also presents Si-ncls of greater size as shown in
Figure 13. The existence of the state Si0 in the SiOx films, as can be seen in the phase decomposition,
generally occurs after samples are subjected to a thermal annealing above 1000 ◦ C [31]. This thermal
process provokes the formation of Si-ncs in a SiO2 matrix. With this process the material reaches a
stable configuration; however, in this case, when applying thermal annealing, the state Si0 disappears.
This may be due to the resolution reached from the XPS measurement in depth, which may be less
than what is required to detect Si-ncs of sizes less than 10 nm.
As previously mentioned, the band of PL emission in the range from 1.3 eV to 1.7 eV is attributed
by several authors to the effect of quantum confinement in Si-ncs or silicon Si-ncls [34–36]; however,
in this theory, the size of the nanocrystal determines the color of the emission. In larger nanocrystals,
the emission tends to lower energies, and as we can see from the results, the Si-ncs and Si-ncls are
of a size that would correspond to a blue emission. These results make us take into account what is
reported by other authors [35,45] who suggest that the intense emission around 1.5 eV is caused by
small, oxidized silicon grains (1 nm or 2 nm). According to a model of molecular type emitters at the
Si/SiO2 interface in the work of Guerra and Ossicini [45], the importance of oxidation in nanocrystals
for emission is shown, and it is corroborated that those mostly oxidized have a high photoluminescence.
This could be compared with that obtained for porous silicon oxidized by Wolking et al. [46]. In their
report it is corroborated that for clusters greater than 3 nm the behavior attributed to quantum
confinement (QC) is fulfilled and that when the cluster size is smaller the emission intensity results
are greater. In this, case the emission process is attributed to the recombination of the free excitons.
For clusters between 2 and 3 nm, the previous behavior is still satisfied but not to the same rate in
which it is fulfilled in QC. Here, the recombination is related to a free hole and a trapped electron,
and finally, for clusters less than 2 nm, an emission redshift occurs and the recombination processes are
attributed to trapped excitons.
Given the size observed in Figures 17 and 18, it could be suggested that the thermal annealing in
the films induces a phase separation that further reduces the size of the Si-ncl. In addition, by virtue of
the presence of interstitial oxygen in the films without annealing, it is possible that an oxidation process
occurs of these small, silicon nanocrystals in the Si-nc/SiO2 interface. Such an event may explain why
the PL emission shifts towards red. In addition to the other experimental results obtained regarding
the increase in the amount of oxygen in the films and the rearrangement of the molecular network,
this would corroborate the theory presented in the literature [35,45].
5. Conclusions
Using the spectroscopic ellipsometry technique, the transmittance, reflectance, and absorption
coefficient of the films were calculated, with which it was subsequently possible to calculate the values
of BG and UE. Such parameters show a tendency by which, at a higher oxygen content, the value
of BG increases, while the value of the UE decreases. Similarly, the films with thermal annealing
present a decrease in the UE, which suggests a decrease in the molecular structure disorder and the
level of defects. By means of the XPS technique, it was possible to obtain an approximate value of
the stoichiometric ratio for deposited SiOx films whose values varied between 1.2 and 1.61, as well
as the oxidation states of silicon. It was observed that films deposited at a higher ssd show a greater
value of x in the stoichiometric ratio, possibly due to a greater incorporation of hydrogen, which was
sustained according to what was observed using FTIR. From the results of PL, the bands attributed
to the main defects observed in the SiOx were found. Thus, in the case of films without thermal
annealing, the violet band is ascribed to WOB defects, the blue band to NOV defects, the green band
to H, and the orange band to NBOHC, while in the case of the annealed films, they exhibited the
emission attributed to WOB, NBOHC, and quantum confinement effects (with emission in red), due to
silicon nanocrystals. This contradicts the results obtained using XPS where the concentration of silicon
decreases with thermal annealing, which leads us to suggest the presence of small, oxidized silicon
grains (less than 2 nm) according to the results obtained, which were corroborated using HRTEM.
In addition, Si-ncls were obtained without applying thermal annealing to the deposited films, while the
annealing induces a phase separation that dramatically increases the PL, which could be an advantage
for the numerous potential applications where silicon agglomerates with small diameters enrich the
luminescent properties of the material.
Author Contributions: Conceptualization, J.A.L.L.; data curation, Z.J.H.S., J.A.L.L., and A.D.H.d.l.L.; formal
analysis, Z.J.H.S., and J.A.L.L.; investigation, Z.J.H.S., A.D.H.d.l.L., S.A.P.G., A.B.L., G.G.S., J.C.L., G.O.M.C.,
and H.P.M.H.; methodology, J.A.L.L.; project administration, J.A.L.L.; resources, J.A.L.L.; supervision, J.A.L.L.;
validation, J.A.L.L.; writing—original draft, Z.J.H.S.; writing—review and editing, Z.J.H.S., J.A.L.L., A.D.H.d.l.L.,
S.A.P.G., A.B.L., G.G.S., J.C.L., G.O.M.C., and H.P.M.H. All authors have read and agreed to the published version
of the manuscript.
Funding: This research received no external funding.
Acknowledgments: This work has been partially supported by CONACyT-CB-255062 and VIEP-LULJ-EXC-2019
projects. The authors acknowledge CIDS, CIMAV-Monterrey, INAOE, and IFUAP laboratories for their help in the
characterizations of the samples.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Aceves-Mijares, M.; González-Fernández, A.A.; López-Estopier, R.; Luna-López, A.; Berman-Mendoza, D.;
Morales, A.; Murphy-Arteaga, R. On the origin of light emission in silicon rich oxide obtained by low-pressure
chemical vapor deposition. J. Nanomater. 2012, 5. [CrossRef]
2. Sopinskyy, M.; Khomchenko, V. Electroluminescence in SiOx films and SiOx -film-based systems. Curr. Opin.
Solid State Mater. Sci. 2003, 7, 97–109. [CrossRef]
3. Zhang, W.L.; Zhang, S.; Yang, M.; Liu, Z.; Cen, Z.; Chen, T.; Liu, D. Electroluminescence of as-sputtered
silicon-rich SiOx films. Vacuum 2010, 84, 1043–1048. [CrossRef]
4. Vivaldo, I.; Carrillo, J.; López, O.; Jiménez, S.; Martínez, J.; Murias, D.; López, J.A. Study of the photon
down-conversion effect produced by thin silicon-rich oxide films on silicon solar cells. Int. J. Energy Res.
2017, 41, 410–416. [CrossRef]
5. Yuan, Z.; Pucker, G.; Marconi, A.; Sgrignuoli, F.; Anopchenko, A.; Jestin, Y.; Pavesi, L. Silicon nanocrystals as
a photoluminescence down shifter for solar cells. Sol. Energy Mater. Sol. Cells 2011, 95, 1224–1227. [CrossRef]
6. Jung, K.; Kim, K.; Song, S.; Park, K. Effects of double SiOx (x < 2) ultra-thin layers on bipolar resistive
switching of ZnO-based nonvolatile memory. Microelectron. Eng. 2019, 216, 111015.
7. Zhou, F.; Chang, Y.F.; Fowler, B.; Byun, K.; Lee, J.C. Stabilization of multiple resistance levels by current-sweep
in SiOx-based resistive switching memory. Appl. Phys. Lett. 2015, 106, 063508. [CrossRef]
8. Chang, Y.F.; Ji, L.; Wu, Z.J.; Zhou, F.; Wang, Y.; Xue, F.; Lee, J.C. Oxygen-induced bi-modal failure phenomenon
in SiOx-based resistive switching memory. Appl. Phys. Lett. 2013, 103, 033521. [CrossRef]
9. Wang, Y.; Fowler, B.; Zhou, F.; Chang, Y.F.; Chen, Y.T.; Xue, F.; Lee, J.C. Effects of sidewall etching on electrical
properties of SiOx resistive random access memory. Appl. Phys. Lett. 2013, 103, 213505. [CrossRef]
10. Miyazaki, H.; Goto, T. SiOx films prepared using RF magnetron sputtering with a SiO target. J. Non-Cryst.
Solids 2006, 352, 329–333. [CrossRef]
11. Alarcón-Salazar, J.; Zaldívar-Huerta, I.E.; Aceves-Mijares, M. Electrical and electroluminescent characterization
of nanometric multilayers of SiOx /SiOy obtained by LPCVD including non-normal emission. J. Appl. Phys.
2016, 119, 215101. [CrossRef]
12. Hegemann, D.; Vohrer, U.; Oehr, C.; Riedel, R. Deposition of SiOx films from O2 /HMDSO plasmas.
Surf. Coat. Technol. 1999, 116, 1033–1036. [CrossRef]
13. He, L.; Inokuma, T.; Kurata, Y.; Hasegawa, S. Vibrational properties of SiO and SiH in amorphous SiOx :
H films (0 ≤ x ≤ 2.0) prepared by plasma-enhanced chemical vapor deposition. J. Non-Cryst. Solids 1995, 185,
249–261. [CrossRef]
14. López, J.L.; Salgado, G.G.; Pedraza, A.P.; Valerdi, D.V.; López, J.C.; Sánchez, A.M.; Santiesteban, H.J. Structural
and optical properties of SiOx films deposited by HFCVD. Procedia Eng. 2011, 25, 304–308. [CrossRef]
15. Luna-López, J.A.; García-Salgado, G.; Díaz-Becerril, T.; López, J.C.; Vázquez-Valerdi, D.E.; Juárez-Santiesteban, H.;
Coyopol, A. FTIR, AFM and PL properties of thin SiOx films deposited by HFCVD. Mater. Sci. Eng. B 2010,
174, 88–92. [CrossRef]
16. López, J.A.L.; López, J.C.; Valerdi, D.V.; Salgado, G.G.; Díaz-Becerril, T.; Pedraza, A.P.; Gracia, F.F.
Morphological, compositional, structural, and optical properties of Si-nc embedded in SiOx films.
Nanoscale Res. Lett. 2012, 7, 604. [CrossRef] [PubMed]
17. Espinosa-Torres, N.D.; Hernández-de La Luz, D.; Flores-Gracia, J.F.J.; Luna-López, J.A.; Martínez-Juárez, J.;
Vázquez-Valerdi, D.E. Evaluation of optical and electronic properties of silicon nano-agglomerates embedded
in SRO: Applying density functional theory. Nanoscale Res. Lett. 2014, 9, 507. [CrossRef] [PubMed]
18. Iacona, F.; Bongiorno, C.; Spinella, C.; Boninelli, S.; & Priolo, F. Formation and evolution of luminescent Si
nanoclusters produced by thermal annealing of SiOx films. J. Appl. Phys. 2004, 95, 3723–3732. [CrossRef]
19. Fang, Y.C.; Zhang, Z.J.; Lu, M. Room temperature photoluminescence mechanism of SiOx film after annealing
at different temperatures. J. Lumin. 2007, 126, 145–148. [CrossRef]
20. Schumann, E.; Hübner, R.; Grenzer, J.; Gemming, S.; Krause, M. Percolated Si: SiO2 Nanocomposites: Oven-vs.
Millisecond Laser-Induced Crystallization of SiOx Thin Films. Nanomaterials 2018, 8, 525. [CrossRef]
21. Liu, Y.; Qiu, J.; Liu, L. Applicability of the effective medium approximation in the ellipsometry of randomly
micro-rough solid surfaces. Opt. Express 2018, 26, 16560–16571. [CrossRef] [PubMed]
22. Fujiwara, H. Spectroscopic Ellipsometry: Principles and Applications; John Wiley & Sons: West Sussex, UK, 2007;
pp. 43–48.
23. Mirabella, S.; Agosta, R.; Franzò, G.; Crupi, I.; Miritello, M.; Lo Savio, R.; Terrasi, A. Light absorption in
silicon quantum dots embedded in silica. J. Appl. Phys. 2009, 106, 103505. [CrossRef]
24. Nikitin, T.; Khriachtchev, L. Optical and structural properties of Si nanocrystals in SiO2 films. Nanomaterials
2015, 5, 614–655. [CrossRef] [PubMed]
25. Espinosa-Torres, N.D.; Luna-López, J.A.; Flores-Gracia, J.F.; de la Luz, A.D.H.; Martínez-Juárez, J. Theoretical
survey of luminescence observed in nanostructured silicon rich oxide films attributed to annealing processes.
Adv. Mater. Lett. 2016, 7, 480–484. [CrossRef]
26. Tauc, J.; Grigorovici, R.; Vancu, A. Optical properties and electronic structure of amorphous germanium.
Phys. Status Solidi (B) 1966, 15, 627–637. [CrossRef]
27. Saxena, N.; Agarwal, A.; Phase, D.M.; Choudhary, R.J.; Kanjilal, D. Controlled formation of silicon nanocrystals
by dense electronic excitation in PLD grown SiOx films. Phys. E: Low-Dimens. Syst. Nanostructures 2010, 42,
2190–2196. [CrossRef]
28. Van Duy, N.; Jung, S.; Nga, N.T.; Son, D.N.; Cho, J.; Lee, S.; Yi, J. The investigation of an amorphous SiOx
system for charge storage applications in nonvolatile memory at low temperature process. Mater. Sci. Eng. B
2010, 175, 176–180. [CrossRef]
29. Hassanien, A.S.; Akl, A.A. Effect of Se addition on optical and electrical properties of chalcogenide CdSSe
thin films. Superlattices Microstruct. 2016, 89, 153–169. [CrossRef]
30. San Andrés, E.; del Prado, A.; Mártil, I.; González-Dıaz, G.; Bravo, D.; López, F.J.; Sieber, I. Bonding
configuration and density of defects of SiOx H y thin films deposited by the electron cyclotron resonance
plasma method. J. Appl. Phys. 2003, 94, 7462–7469. [CrossRef]
31. Tomozeiu, N. Silicon oxide (SiOx , 0 < x < 2): A challenging material for optoelectronics. Optoelectron. Mater. Tech.
2011, 55–98. [CrossRef]
32. Chae, K.H.; Son, J.H.; Chang, G.S.; Kim, H.B.; Jeong, J.Y.; Im, S.; Whang, C.N. Visible photoluminescence in
ion beam mixed SiO2 /Si/SiO2 layers. Nanostructured Mater. 1999, 11, 1239–1243. [CrossRef]
33. Zhang, W.; Zhang, S.; Liu, Y.; Chen, T. Evolution of Si suboxides into Si nanocrystals during rapid thermal
annealing as revealed by XPS and Raman studies. J. Cryst. Growth 2009, 311, 1296–1301. [CrossRef]
34. Lin, J.H.; Chiu, H.C.; Lin, Y.R.; Wen, T.K.; Patil, R.A.; Devan, R.S.; Chen, C.H.; Shiu, H.W.; Liou, Y.;
Ma, Y.R. Electrical and chemical characteristics of probe-induced two-dimensional SiOx protrusion layers.
Appl. Phys. Lett. 2013, 102, 031603. [CrossRef]
35. Nikitin, T.; Velagapudi, R.; Sainio, J.; Lahtinen, J.; Räsänen, M.; Novikov, S.; Khriachtchev, L. Optical and
structural properties of SiOx films grown by molecular beam deposition: Effect of the Si concentration and
annealing temperature. J. Appl. Phys. 2012, 112, 094316. [CrossRef]
36. Takagahara, T.; Takeda, K. Theory of the quantum confinement effect on excitons in quantum dots of
indirect-gap materials. Phys. Rev. B 1992, 46, 15578. [CrossRef] [PubMed]
37. Ledoux, G.; Gong, J.; Huisken, F.; Guillois, O.; Reynaud, C. Photoluminescence of size-separated silicon
nanocrystals: Confirmation of quantum confinement. Appl. Phys. Lett. 2002, 80, 4834–4836. [CrossRef]
38. Morales-Sánchez, A.; Leyva, K.M.; Aceves, M.; Barreto, J.; Domínguez, C.; Luna-López, J.A.; Pedraza, J.
Photoluminescence enhancement through silicon implantation on SRO-LPCVD films. Mater. Sci. Eng. B
2010, 174, 119–122. [CrossRef]
39. Ding, L.; Chen, T.P.; Liu, Y.; Ng, C.Y.; Yang, M.; Wong, J.I.; Huang, Y. Evolution of photoluminescence
mechanisms of Si+-implanted SiO2 films with thermal annealing. J. Nanosci. Nanotechnol. 2008, 8, 3555–3560.
[CrossRef]
40. Yu, Z.; Aceves, M.; Carrillo, J.; Flores, F.; Falcony, C.; Domínguez, C.; Llobera, A.; Morales, A. Photoluminescence
in off-stoichiometric silicon oxide compounds. Superf. Vacío 2004, 17, 1–6.
41. Wang, Y.; Townsend, P.D. Common mistakes in luminescence analysis. J. Phys. Conf. Series 2012, 398, 012003.
[CrossRef]
42. Teixeira, F.; Berjoan, R.; Peraudeau, G.; Perarnau, D. Solar preparation of SiOx (x≈1) nanopowders from
silicon vaporisation on a ZrO2 pellet. XPS and photoluminescence characterisation. Sol. Energy 2005, 78,
763–771. [CrossRef]
43. Friedrich, D.; Schmidt, B.; Heinig, K.H.; Liedke, B.; Mücklich, A.; Hübner, R.; Mikolajick, T. Sponge-like
Si-SiO2 nanocomposite—Morphology studies of spinodally decomposed silicon-rich oxide. Appl. Phys. Lett.
2013, 103, 131911.
44. Park, E.; Yoo, H.; Lee, J.; Park, M.S.; Kim, Y.J.; Kim, H. Dual-size silicon nanocrystal-embedded SiOx
nanocomposite as a high-capacity lithium storage material. ACS Nano 2015, 9, 7690–7696. [PubMed]
45. Guerra, R.; Ossicini, S. High luminescence in small Si/SiO2 nanocrystals: A theoretical study. Phys. Rev. B
2010, 81, 245307. [CrossRef]
46. Wolkin, M.V.; Jorne, J.; Fauchet, P.M.; Allan, G.; Delerue, C. Electronic states and luminescence in porous
silicon quantum dots: The role of oxygen. Phys. Rev. Lett. 1999, 82. [CrossRef]
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Nanomaterials 10 01415

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nanomaterials 10 01415

Uploaded by

Copyright:

Available Formats

nanomaterials

Nanomaterials 2020, 10, 1415; doi:10.3390/nano10071415 www.mdpi.com/journal/nanomaterials

2. Materials and Methods

Table 1. Description of the labeling of the eight SiOx films obtained.

Hydrogen Flow Level: 25 sccm Hydrogen Flow Level: 100 sccm

Figure 2. Reflectance and transmittance of the A1, A1’

Figure 3. The absorption coefficients of the SiOx films without

Figure 4. Energy band gap calculated for D1’

Table 2. Energy band gaps and Urbach energies.

at 2100 cm−1, 2156

A1A1 A2A2 D1D1 D2

Binding energies Si2p

Binding energies 99.5 eV 100 eV 101 eV 101.5 eV 102.5 eV 103.5 eV

Table 5. Binding energies of the decomposed XPS peaks.

Table 5. Binding energies of the decomposed XPS peaks.

Table 6 shows the position

Table 6. PL deconvoluted bands and their proposed defects contributions.

A1 A1´ A2 A2´ D1 D1´ D2 D2´

Table 7. Si–H stretching vibration bands.

Table 7. Si–H stretching vibration bands.

Si-(Si2 O2 ) + Si-(Si2 O2 ) → Si-(Si1 O3 ) + Si-(Si3 O1 )

Si-(Si2 O2 ) + Si-(Si2 O2 ) → Si-(O4 ) + Si-(Si4 )

Figure 13. High

Figure 14. HRTEM micrograph of the A2 SiOx film.

You might also like