Mechanical properties of zirconia thin films deposited by filtered cathodic vacuum arc

Article abstract of DOI:10.1111/j.1551-2916.2005.00420.x

Surface Young's modulus (E), hardness (H), and yield strength (Y) of zirconia films deposited on Si substrate by filtered cathodic vacuum arc (FCVA) under different oxygen flow rate were studied by nano-indentation measurement and finite element modeling.

Full text of DOI:10.1111/j.1551-2916.2005.00420.x

J. Am. Ceram. Soc., 88 [8] 2227–2229 (2005)
DOI: 10.1111/j.1551-2916.2005.00420.x
ournal
J
Mechanical Properties of Zirconia Thin Films Deposited by Filtered
Cathodic Vacuum Arc
w
Zhenghao Gan , Guoqing Yu, Zhiwei Zhao, C. M. Tan, and B. K. Tay
School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore 639798, Singapore
Surface Young’s modulus (E), hardness (H), and yield strength
Y) of zirconia films deposited on Si substrate by filtered ca-
thin film under wear or mechanical contact is easy to be eval-
uated.
(
thodic vacuum arc (FCVA) under different oxygen flow rate
were studied by nano-indentation measurement and finite ele-
ment modeling. Identified by X-ray diffraction, the structure of
the films evolves from amorphous to monoclinic, and then to
amorphous, with oxygen flow rate increasing from 10 to 80
standard cubic centimeters per minute, which affects the me-
chanical properties of the films. It is found that the elastic–per-
fect-plastic constitutive model is successfully applied to the films
with amorphous structure; however inadequate for the mono-
II. Experimental Details
2
The ZrO thin films were deposited using an industrial FCVA
system (Nanofilm Technologies International Pte Ltd, Singa-
pore) described elsewhere. The substrates were Si (100) with
7
4
00 mm average thickness, which were pre-cleaned with acetone,
alcohol, and de-ionized water followed by a nitrogen blow-dry
using a static neutralizing blow-off gun. Prior to deposition, the
substrates were sputter cleaned for 2 min by argon ion beams
clinic ZrO , indicating that dislocation-related plasticity possi-
2
bly occurs during nano-indentation.
(
800 eV and 45 mA), and then transferred into the deposition
ꢁ
6
chamber, which was evacuated to 2.0ꢀ 10 Torr. A Zr cathode
(
99.98% pure) operated at 120 A DC current was ignited
through instant contact between cathode and anode to obtain
the plasma, which was steered out by a toroidal magnetic field
fixed at 40 mT to condense on a substrate. O gas (99.99%) was
2
led into the region near the Zr target. O flow rate chosen here
was varied from 10 to 80 standard cubic centimeters per minute
(sccm).
I. Introduction
IRCONIUM OXIDE (ZrO ) is a refractory material with excel-
2
2
Z
lent chemical and corrosion resistance and low thermal con-
ductivity. Zirconia also has high hardness, making it a potential
wear coating. It is well known that brittle materials such as
1
zirconia are subject to plastic deformation under mechanical
contact, where the amount of elastic deformation is comparable
with that of plastic deformation. It is also known that exper-
The nano-indentation measurement was performed by using
Hysitron TriboScope (Hysitron Inc., Minneapolis, MN). In a
typical indentation process, the Berkovich indenter was forced
into a specimen using a pre-defined test force (400 mN), and a
load/unload curve was recorded for each indentation. For each
specimen, three independent indentations were performed. The
experimental load/unload curve will be compared with that ob-
tained from finite element modeling to derive the mechanical
properties (E, Y, and H) of the thin films. The nominal thickness
of the thin films is 500 nm.
2
imental method itself such as nanoindentation is hard to predict
the complex interaction in the vicinity of the contact zone be-
tween two contact materials. To study the plastic deformation of
brittle materials such as zirconia, a constitutive relationship
should be built. However, the conventional uniaxial tensile test
at room temperature does not give a yield strength (Y) of zir-
conia since it is more likely to fail by fracture rather than plas-
ticity under uniaxial tensile test.
The phases of the thin films were identified by XRD (X-ray
diffractometer, D5005, Siemens, Madison, WI) using CuKa ra-
Nanoindentation is a well-established technique for measur-
ing films’ mechanical properties. Both Young’s modulus (E) and
hardness (H) of the thin films are readily obtained by analyzing
˚
diation (wavelength of 1.54 A) at 40 kV and 40 mA with a thin
film goniometer (Rigaku, Tokyo, Japan). The incident angle of
the detailed shape of the unloading curve during nanoindenta-
tion. However, the report on yield strength of ZrO thin film is
the X-ray is 11.
3
2
limited since it could not be derived directly from the unloading
curve of the nano-indentation measurement. For a metallic ma-
terial under conditions of fully plastic indent, the yield strength
III. Finite Element Approach
4
Some assumptions are made as follows to simplify the simula-
tion: (1) The root-mean-square (RMS) roughness of the surface
of the thin films is less than 0.1 nm. Thus, the surface roughness
is ignored in the model; (2) Residual stress (i.e., pre-existing
stress after deposition) is not included in the model; (3) Silicon
is believed to be 1/3 of the hardness. However, in ceramic ma-
2
terials like ZrO , yield occurs by bond cleavage, and the above
relationship does not apply. The H/Y value is always lower than
5
.
3
Under nanoindentation, high hydrostatic stress level is
around the contact zone, and an equivalent yield strength
EYS) could be defined to replace the usual yield strength giv-
2
substrate and ZrO are assumed to be isotropic and behave as
elastic perfect plastic. That is, only two parameters (i.e., E and
Y) are needed to decide the constitutive model. This is a good
(
en by the uniaxial test. In this work, with the aid of finite element
simulation, the EYS of zirconia films are studied, which were
deposited on Si (100) by filtered cathodic vacuum arc (FCVA) .
8–10
approximation widely accepted
since they are brittle materi-
6
als, where yield usually happens by bond cleavage and work
hardening is not usually observed. The diamond indenter is as-
sumed to be isotropic and to deform purely elastically, with
With the EYS at hand, the in-progress deformation of zirconia
1
1
E 5 1140 GPa and n 5 0.07; (4) The friction between the ind-
enter and the specimen under test is ignored; (5) Film-substrate
is assumed to be mechanically bonded perfectly.
D. Marshall—contributing editor
The indentation with a Berkovich indenter was modeled as a
two-dimensional axisymmetric contact problem. This in one
way could save a lot of computational time and in another
Manuscript No. 10733. Received December 14, 2003; approved February 14, 2005.
Author to whom correspondence should be addressed. e-mail: ezhgan@ntu.edu.sg
w
2
227

2
228
Journal of the American Ceramic Society—Gan et al.
Vol. 88, No. 8
8
way produce a reasonable result. Finite element package AN-
SYS (version 5.6) was employed for the indenter-specimen con-
tact analysis in this work. The proposed finite element model
500
400
300
Experimental
Simulation with E=220GPa and Y=6.0GPa
was verified by employing nano-indentation measurement on
the standard bulk fused silica provided by Hysitron (Minnea-
s
1
polis, MN). A detailed verification process was elaborated
1
(a) O flowing rate: 20 sccm
2
1
2
elsewhere.
IV. Results and Discussion
1) Structure of the Thin Films
2
The XRD patterns for ZrO films deposited at different O flow
2
1
00
00
(
2
rate are shown in Fig. 1. Amorphous structure is observed at 10
sccm (Fig. 1(a)), indicated by a wide peak centered at 2y 5 32.41,
which is related to the oxygen-deficient Zr–O solid solution. As
6
5
4
3
00
0
O flow rate increases to 20 sccm (Fig. 1(b)), one strong peak at
2
Experimental
Simulation with E=280GPa and Y=7.5GPa
2
y 5 33.71 corresponds to (200) plane of monoclinic ZrO (de-
2
noted as m(200)), accompanied by appearance of other weak
peaks from monoclinic phase. At 35 sccm, in addition to more
weak peaks, m(200) peak becomes weak while m(ꢁ202) peak
gets strong. At 50 sccm, the strongest peak is m(ꢁ111) and more
00
00
(
b) O flowing rate: 50 sccm
2
2
monoclinic peaks present. However, when O flow rate rises to
6
5 sccm or above, the film transforms from polycrystalline into
amorphous structure, evidenced by a wide peak at 2y 5 31.31
Fig. 1(e)). It is seen that the structure of the zirconia thin films
evolves from amorphous through monoclinic to amorphous,
with the increase of O flow rate
200
(
1
00
2
(
2) Mechanical Properties Versus Oxygen Flow Rate
Figures 2(a)–(c) shows the experimental loading–unloading
curves and their simulation fittings for O flow rate of 20, 50,
0
5
4
00
Experimental
Simulation with E=230GPa and Y=6.5GPa
2
and 80 sccm, respectively. In Figs 2(a) and (c), the simulation
shows excellent agreement with the experimental results for both
loading and unloading curves. However, for O flow rate of 50
00
2
(c) O flowing rate: 80 sccm
2
sccm (Fig. 2(b)), one may see the discrepancy in the latter part of
the unloading curves: the experimental residual plastic defor-
mation is much less than that of simulation, which may be at-
300
200
tributed to the indentation-induced dislocation activity in ZrO
2
1
3
crystalline, as reported by Li et al. Dislocation activity typi-
cally leads to plastic deformation, and thus the work hardening
effect.
1
00
0
A qualitative explanation is given here for the larger predicted
plasticity in the modeling (Fig. 2(b)). Figure 3 is a schematic
comparison of the two hypothetical constitutive models of plas-
0
5
10
15
20
25
30
35
tic (case 1) and perfect plastic (case 2). Note that work harden-
0
Depth (nm)
ing is not included in case 2 (i.e., tangent modulus E
0
1
5 0), but
be included in case 1 (i.e. E 40). W and W are the total works
Fig. 2. Experimental loading-unloading curves and their simulation fit-
tings for oxygen flow rate of: (a) 20; (b) 50, and (c) 80 sccm.
1
1
2
of indentation of cases 1 and 2, respectively. Subscripts e and p
denote elastic and plastic portion, respectively. The total work
of indentation should be equal for the two hypothetical cases
(e)
(a) 10 sccm
(b) 20 sccm
(c) 35 sccm
(d) 50 sccm
(e) 65 sccm
(d)
(c)
(b)
(a)
Fig. 3. Schematic constitutive models for plastic (case 1) and perfect
2
0
30
40
50
Theta
60
70
plastic (case 2). W
and 2, respectively. Subscripts e and p denote elastic and plastic portion,
respectively. It is clear that the residual strain in case 2 (OC ) is larger
than that in case 1 (OC ).
1 2
and W are the total works of indentation of cases 1
2
Fig. 1. X-ray diffraction patterns for the ZrO
ferent O flow rate.
2
films deposited at dif-
2
2
1

August 2005
Mechanical Properties of Zirconia Thin Films Deposited by FCVA
2229
2
0
were then derived by nano-indentation measurement and finite
element modeling. Identified by XRD, the structure of the zir-
conia thin films evolves from amorphous, through pure mono-
3
3
2
2
2
30
00
70
40
10
(
a)
16
H
E
2
clinic, to amorphous again, with the increase of O flow rate
1
8
2
from 10 to 80 sccm. It is found that the elastic-perfect plastic
constitutive model is successfully applied to the thin films with
amorphous structure; it is however not adequate for the mono-
(
b)
8
7
6
5
4
3
clinic ZrO , indicating that dislocation related plasticity possibly
2
occurs during nano-indentation examination of the monoclinic
ZrO2. This information indicates that a suitable constitutive
model should be adopted for the simulation, providing a guide-
line for future work.
0
20
40
60
80
100
Oxygen flow rate (sccm)
References
1
Fig. 4. (a) Young’s modulus and hardness and (b) yield strength of the
ZrO thin films vary with the O flow rate.
M. J. Paterson, P. J. Paterson, and B. Ben-Nissan, ‘‘Dependence of Structural
2
2
and Mechanical Properties on Film Thickness in Sol-Gel Zirconia Films,’’ J. Ma-
ter. Res., 13, 388–95 (1998).
K. Tanaka, M. Kanari, and N. Matsui, ‘‘Continuum Dislocation Model of
2
(
i.e., W
1
5 W ), which is also equal to the experimental one.
2
Vickers Indentation on a Zirconia,’’ Acta Mater., 47, 2243–57 (1999).
W. C. Oliver and G. M. Pharr, ‘‘Improved Technique for Determining Hard-
3
However, it is seen in Fig. 3 that the elastic recovery W 4W
because of work hardening (i.e., area B C D 4B C D ). There-
2
fore, the residual plastic strain for perfect plastic (OC ) is larger
than that for plastic model (OC ). In other words, the elastic–
1
perfect plastic model assumed for the amorphous structure of
the thin films (Figs. 2(a) and (c)) would be inappropriate for the
monoclinic ZrO (Fig. 2(b)).
Figure 4 presents the E, Y, and H of the zirconia thin films,
exhibiting the same trend as the O flow rate increases. The val-
ues of the properties are the largest in the middle gas flow region
(i.e., in-between 35 and 50 sccm), which is corresponding to the
monoclinic ZrO . The future work will be intended to study the
structure variations controlling the changes in yield strength or
hardness. The plastic model (case 1) considering work hardening
will be employed to simulate the indentation behavior of the
monoclinic ZrO2.
e1
e2
ness and Elastic Modulus Using Load and Displacement Sensing Indentation Ex-
periments,’’ J. Mater. Res., 7, 1564–80 (1992).
1
1
1
2
2
2
4
D. Tabor, The Hardness of Metal. London, Oxford, 1951.
S. S. Chiang, D. B. Marshall, and A. G. Evans, ‘‘Response of Solids to Elastic/
5
Plastic Indentation-Stresses and Residual Stresses,’’ J. Appl. Phys., 53, 298–311
(1982).
G. Q. Yu, B. K. Tay, and Z. W. Zhao, ‘‘Structure and Properties of Zirconium
6
2
Oxide Thin Films Prepared by Filtered Cathodic Vacuum Arc,’’ Appl. Phys. A:
Mater. Sci. Process., 81, 405–11 (2005).
S. Xu, B. K. Tay, H. S. Tan, L. Zhong, Y. Q. Tu, S. R. P. Silva, and W. I.
7
2
Milne, ‘‘Properties of Carbon Ion Deposited Tetrahedral Amorphous Carbon
Films as a Function of Ion Energy,’’ J. Appl. Phys., 79, 7234–40 (1996).
8
J. A. Knapp, D. M. Follstaedt, and S. M. Myers, ‘‘Precipitate-Hardened
Aluminum Alloys Formed Using Pulsed Laser Deposition,’’ J. Appl. Phys., 79,
2
1
116–22 (1996).
J. A. Knapp, D. M. Follstaedt, S. M. Myers, J. C. Barbour, and T. A. Fried-
9
mann, ‘‘Finite-Element Modeling of Nanoindentation,’’ J. Appl. Phys., 85, 1460–
74 (1999).
10
D. J. Ma, K. W. Xu, and J. W. He, ‘‘Numerical Simulation for Determining
the Mechanical Properties of Thin Metal Films Using Depth-Sensing Indentation
Technique,’’ Thin Solid Films, 323, 183–7 (1998).
1
1
TriboScope User Manual. Hysitron Inc., Minneapolis, MN, USA, 2002.
12
Z. H. Gan, Y. B. Zhang, G. Q. Yu, C. M. Tan, S. P. Lau, and B. K. Tay,
V. Conclusions
‘
by Filtered Cathodic Vacuum Arc,’’ J. Appl. Phys., 95, 3509–15 (2004).
W. B. Li, J. L. Henshall, R. M. Hooper, and K. E. Easterling, ‘‘Mechanisms of
Indentation Creep,’’ Acta Metall. Mater., 39, 3099–110 (1991).
‘Intrinsic Mechanical Properties of Diamond-Like Carbon Thin Films Deposited
2
ZrO thin films were deposited on Si substrate by FCVA with
different oxygen flow rate. Their mechanical properties (includ-
ing E, H, and yield strength (Y)) without the substrate effect
13
&