8729
is extremely toxic [32]. Hence, to perform a simultaneous thermal
oxidation and nitridation on metallic Zr, N2O gas is more appropri-
ate and preferable due to its non-toxic property [32]. However, up
to date, there is no report on the effects of either N2O or NO gas on
a sputtered Zr thin film on Si substrate. Therefore, this is the main
objective of this manuscript to report on the effects of thermal oxi-
dation and nitridation in N2O ambient of a sputtered Zr/n-type Si
system on their structural and chemical properties. Based on these
results, a possible model related to the oxidation and nitridation
mechanisms has been proposed.
elements. The sensitivity factor is dependent on XPS system. In this
work, the sensitivity factors for Zr 3d, Si 2p, O 1s, N 1s, and C 1s were
2.576, 0.328, 0.78, 0.47, and 0.278, respectively.
Energy filtered transmission electron microscopy (EFTEM)
(Zeiss Libra 200) was employed to analyze cross-sectional images
of the films. Prior to this, a protective layer of resist was deposited
on the top most layer of a sample. Then it was ion-milled by
a focused ion beam system. The zone axis used for the imaging
was [1 1 0] and the reciprocal lattice vectors were [0 0 1], [0 2 0],
¯
and [0
1 0]. Interplanar spacing, d, of the polycrystalline struc-
ture was measured from the EFTEM images with the assistance
of ImageJ software, which gives the accuracy of measurements in
three decimal places. Crystallinity of the films were characterized
by X-ray diffraction (XRD) system (P8 Advan-Bruker) in a scan range
of 2ꢃ = 20–40◦ with step time of 71.6 s and step size of 0.0342◦.
Copper (Cu K␣) with wavelength (ꢄ) of 1.5406 nm was used as X-
ray source. Raman spectroscopy measurements (Jobin Yvon HR 800
UV) using an Ar+ incident beam with wavelength of 514.5 nm, were
conducted to ascertain the stability of chemical bonding upon ther-
mal oxidation and nitridation process. Fourier Transform infrared
(FTIR) (Pelkin Elmer Spectrum GX) analysis was performed to ana-
lyze the chemical functional groups of the films.
2. Experimental details
n-Type Si substrates (ꢀ1 0 0ꢁ oriented, 1–10 ꢁcm) were first sub-
jected to an ultrasonic and Radio Corporation of America (RCA)
cleaning then followed by a HF dipping (1 HF: 50 H2O) process
to remove native SiO2 from the surface. A 5-nm thick Zr film
was deposited on the cleaned Si substrates by a RF sputtering
system (Edwards Auto 500). During the sputtering process, work-
ing pressure, RF power, and inert Ar gas flow were configured at
1.2 × 10−7 Torr, 170 W, and 20 cm3/min, respectively. The depo-
´
˚
sition rate was set at 2 A/s. The initial thickness of Zr film was
monitored by an in situ monitoring system in the RF sputtering
system. The thickness was also re-measured by ellipsometer, to
confirm the thickness. The thickness of Zr film was 5.00 0.05 nm.
After the deposition, samples were placed into a horizontal tube
furnace and heated up from room temperature to 700 ◦C in an Ar
flow ambient and the heating rate was set at 10 ◦C/min. Once the
set temperature was achieved, N2O gas was purged in with a flow
rate of 150 mL/min for a set of durations (5, 10, 15, and 20 min).
The samples were withdrawn from the furnace after the furnace
was cooled down to room temperature in an Ar ambient. A sim-
ple qualitative cellophane tape adhesion test has been carried out
to examine the adhesion quality of ZrO2 on Si substrate and there
was no peeling off issue arisen.
Evaluations of structural and chemical properties of the samples
were carried out by various characterization techniques. Composi-
tional and depth profile measurements of the films were conducted
by an X-ray photoelectron spectroscopy (XPS) (Kratos Axis Ultra-
DLD) with a monochromatic Al-K␣ X-ray source (hꢂ = 1486.69 eV)
operated at 150 W and a take-off angle of 0◦ with respect to surface
normal. In order to perform a depth profiling analysis, a 5-kV Ar ion
etching was used. The pressure of Ar in the analysis chamber was
3 × 10−7 Torr, while the area of analysis was 220 × 220 m2. The
chemical compositions of the films were obtained from a combi-
nation of wide and narrow scans. A wide scan was first performed
with a passing energy of 160 eV for 9 min to determine the elemen-
tal chemical states. The core-level spectra that had been detected
were Zr 3d, Si 2p, O 1s, N 1s, and C 1s. Subsequently, for a narrow
scan, a passing energy of 20 eV for 5 min was used to scan through
the binding-energy range of interest. The recorded C 1s peak due
to the adventitious carbon-based contaminant on the surface, with
respect to the literature value of 284.6 eV [33,34], was used as a
help of CasaXPS software (version 2.3.15) before deconvolution of
the XPS spectra was performed. The total concentration of an ele-
ment (Cx) available in the investigated films was calculated based
on the following equation [35–37]
3. Results and discussion
3.1. XPS measurements
Based on wide scan of XPS, core-level spectra of Zr 3d, Si 2p,
O 1s, and N 1s have been detected in all oxidized samples. Then
narrow scan was performed for each element. Depth profiles of
these elements as a function of etching time for each sample were
obtained (Fig. 1(a)–(d)). It is observed that at earlier etching time,
Zr and O are the dominant elements with atomic ratio of about
1:2. This indicates a stoichiometric ZrO2 has been formed. Based
on this justification, a layer of ZrO2 located at the outmost layer
of each samples has been identified and indicated in Fig. 1. Atomic
percentage of both Zr and O elements are decreasing as the etching
time is extended and their ratio is deviated from 1:2, until they are
totally disappeared. This is the boundary of oxide and Si. Beyond
this boundary, atomic percentage of Si is increased significantly;
indicating the appearance of Si substrate. In between Si and ZrO2
boundaries, an interfacial layer (IL) with a mixture of Zr, O, Si, and
N have been detected. Detail analysis of the chemical compound
of the IL was performed by narrow scan of XPS and it will be dis-
cussed in the subsequent paragraphs. The thickness of ZrO2 and
IL are depending on oxidation time. Based on the figure, thickness
of ZrO2 is marginally increased as the oxidation time increases. A
minute variation has been detected from EFTEM images and these
will be presented and discussed in later paragraphs. Distribution
of nitrogen in terms of broadness, maximum atomic percent, and
its location vary with samples. It is seen that sample oxidized for
5 min reveals the highest atomic percent (18.90 at%) of nitrogen.
The nitrogen concentration is in the range of 18–19% and it is
consistently detected even more than 3000s of etching time (not
shown). With this observation, one may hypothesize that nitrogen
is embedded deep into the Si substrate. However, this type of dis-
tribution does not observe for other oxidized samples (10–20 min).
In these samples, nitrogen is accumulated at the interfacial layer
with concentration of 1.98–2.20 at%, which is much lower than
the concentration detected in 5-min oxidized sample. Among the
three oxidized samples (10, 15, and 20 min), 15-min oxidized sam-
ple demonstrates the highest concentration of nitrogen (2.20 at%).
For 10- and 20-min oxidized samples, concentration of nitrogen is
1.98 and 2.08 at%, respectively. This is in line with the trend of IL
thickness with 15-min oxidized sample showing the thinnest layer.
Ix/Sx
ꢀ
Cx
=
× 100
(1)
iIi/Si
where, Ix and Ii were peak intensity of the evaluated element and
all other detected elements, respectively. Sx and Si were sensitivity
factor of the respective evaluated element and all other detectable
Wong, Yew Hoong
