AE MONITORING FROM CVD-DIAMOND FILM SUBJECTED TO
MICRO-INDENTATION AND PULSE LASER SPALLATION
R.Ikeda
Asahi Diamond Industrial Co., Ltd, Chiba, Japan.,
H. Cho and M. Takemoto,
Aoyama Gakuin University, Kanaggwa, Japan.,
K. Ono
UCLA, Los Angeles, USA
ABSTRACT
We monitored AEs during micro-indentation and pulse laser spallation test of CVD diamond
film deposited on sintered SiC, and classified the fracture type by Lamb wave AE analysis.
In the indentation test with a Rockwell sphere indenter, we observed Hertz ring cracks and
simultaneously detected AEs from Mode-I fracture. Crack types were classified from polarity
distribution of the first arrival So-mode Lamb waves. Fracture strength of the diamond film was
estimated as 7.1 GPa by FEM analysis, using the critical indentation force, indentation depth and Hertz
crack diameter,
In the laser spallation test, strong pulse expansion-wave was produced by the pulse laser ablation
of the confined silicone grease on the opposite surface of the diamond film and utilized to cause Mode-I
delamination of diamond film. Critical tensile stress to cause the spallation was calculated as 222 MPa.
We attempted to detect the spallation initiation by monitoring the in-plane motion of So-mode Lamb
wave AE using broad-band pinducer mounted on the edge planes. Waveform change due to the Mode-I
spallation was detected.
KEYWORDS: Diamond film, Indentation test, Ring crack, Fracture strength, Laser spallation,
Interfacial strength
NTRODUCTION
CVD-diamond-coated tools are widely used for machining graphite, ceramics and Al-Si alloys.
273
DGZfP-Proceedings BB 90-CD Lecture 25 EWGAE 2004
Wear performance of diamond film as a cutting tool has been demonstrated to be excellent. However
decohesion (fracture) of the film during machining has been remained as problem. This problem is more
serious than the antifriction problem of the film. In order to study the fracture mechanism, fracture
strength of the film and film/substrate interface must be correctly evaluated.
Indentation fracture test using spherical indenter gives us a lot of useful information about micro-
fracture characteristics. When this test is applied to hard material, Hertz ring crack is caused by the
surface tensile stress given by the Hertz theory [1]. Fracture strength of ceramic coating have been
evaluated using the threshold load to cause the first ring crack and FEM analysis [2,3]. Here, accurate
determination of crack initiation load is critical. We monitored progression of micro-cracks during
loading using AE system. We classified fracture types by analyzing the polarity distribution of the first
arrival So-mode Lamb wave. This AE system is different from the conventional AE system utilized for
indentation [4,5] and scratch test [6,7].
We developed a pulse laser spallation method to measure the absolute adhesion strength of hard
coatings. It utilizes strong shock waves produced by pulse laser breakdown of confined grease or liquid.
For the measurement of adhesive strength, scratching and indentation methods have been adopted so far
[8],[9],[10], but could not measure the absolute adhesion strength. The strength estimated by these
conventional methods changes depending on the testing conditions, such as the surface roughness, shape
and stiffness of the indenter. Furthermore these methods can hardly be applied to diamond film. Contrary
to this, laser spallation technique [11] can estimate the Mode-I adhesion strength when both the out-of-
plane displacement of expansion wave and critical laser energy to cause micro spallation are correctly
measured.
We first conducted Rockwell indentation test to CVD-diamond film to estimate the fracture
strength of the film, then estimated interfacial adhesion strength of the film utilizing the pulse laser
spallation technique. AE was successfully utilized for damage detection and fracture mode classification.
EXPERIMENTAL
INDENTATION TEST
We prepared polycrystalline diamond film with
grain size of 10 to 20 µm deposited on sintered
SiC substrate by hot-filament CVD method of 3 %
methane and 97 % hydrogen gas mixture. The
films were polished using diamond wheel to
obtain a clear indentation and spallation of the
diamond film.
4
Diam ond film
Personal
com puter
1
3
2
Digitizer
SiC substrate
Indenter
60dB pream plifier
Experimental setup for the indentation test is
shown in Fig. 1. We used a Rockwell indenter with
spherical tip (radius: 0.2 mm and angle: 120
degree) . Maximum indentation loads :Fmax were
changed at 98, 196, 294, 392 and 490 N. Loading
Fig. 1 Experimental setup for the indentation
test and AE monitoring.
2
1
4
Sm all A E senser
274
DGZfP-Proceedings BB 90-CD Lecture 25 EWGAE 2004
rate and holding time are controlled to be 0.49 N/s and 10 s, respectively. Four small PZT-type AE
sensors (PAC, Type PICO, center frequency :500 kHz) were mounted on the four edge surfaces of the
substrate. Sensor outputs were amplified by 60 dB using a pre-amplifier (NF9913, NF Circuit Co.),
digitized by a fast A/D converter (Compu Scope, CS12100 Gage Applied Sciences Inc.) and fed to a
personal computer. Sampling interval and sampling points were set as 50 ns and 2048, respectively.
Diamond film of 35 µm thickness on 5 mm thick SiC of 20 mm square was submitted to the test. Thus
the AE system monitors the Lamb (or plate) wave AEs. We classified fracture modes or crack types by
analyzing the polarity distribution or radiation pattern of So-mode Lamb wave. Two types of crack, the
Mode-I fracture with a crack opening vector parallel to the surface corresponding to the ring crack, and
the Mode-II fracture or lateral crack at the film/substrate interface, were classified. Detail of polarity
distribution analysis can be found elsewhere [12]. It is noted that all polarities of So-mode are positive
for the Mode-I crack, but two opposite polarities for the Mode-II crack.
LASER SPALLATION TEST
Fig. 2 Experimental setup for laser spallation and
AE monitoring.
Experimental setup for the
laser spallation is shown in Fig. 2. A
high-energy Q-switched pulse
Nd:YAG laser (New Wave Research,
Tempest300, maximum energy: 300
mJ, pulse duration time: 3-5 ns,
wavelength: 1064 nm) was irradiated
on the opposite surface of the
diamond film. Silicone grease
containing fine MoS2 particles was
painted on the surface as an energy-
absorbing layer. Thickness of the
layer was controlled precisely using a
thickness gage of 20 µm and confined rigidly by a 6-mm-thick silica plate with anti-reflection coating.
Impact wave was excited by the laser breakdown of the grease. The beam diameter was controlled to be
1 mm by changing the defocusing distance. Laser energy was changed from 0.7 to 63 mJ. Diamond film
with 78 µm thickness deposited on 35 mm square and 5.0 mm thick SiC substrate was tested. A
heterodyne-type laser interferometer (He-Ne laser, 532 nm) was used to measure the out-of-plane
displacement of the stress wave at the epicenter of the source. In order to detect the spallation initiation,
small AE sensor (Valpey Fisher, pinducer ,6.0 MHz) was glued on the end surface of the specimen.
Sensor to source distance was kept as 14.5 mm. Leak laser was used as trigger signal for the laser
interferometer and AE system. Signals were digitized and stored in a high-resolution digital oscilloscope
at sampling interval of 0.4 ns with sampling points of 8192. Output of the AE sensor was directly fed to
the digital oscilloscope.
275
Absorbing layer SiC Substrate
Diamond film
Focusing beam
Silica confining layer
Pulsed Laser
1064nm
Laser
interferometer
Pinducer
Oscilloscope
DGZfP-Proceedings BB 90-CD Lecture 25 EWGAE 2004
RESULTS AND DISCUSSION
SPHERICAL INDENTATION
TEST
We observed frequent
ring crack generation during the
indentation test. Number of ring
cracks increased with maximum
indentation load Fmax and expanded
into outward area of the first ring
crack. Figure 3 (a) shows the
indentations produced by the test of
Fmax: 98 N, and (b) the indentation
curve (indentation force F vs.
indentation depth h) with AE
timing. We detected frequent AE
signals from the Mode-II fracture
(open triangles). These AEs were,
however, considered to be the
contact noise. Only one Mode-I AE,
from the ring crack was detected as
indicated by solid triangle at 39.1 N.
The number of the Mode-I AEs
increased with Fmax. Here we
focused our attentions on the load
to cause the first ring crack in order
to evaluate the tensile strength of
the diamond film. Table 1 shows
the load at which the first Mode-I
AE was detected. The load was
almost the same, except that (127.4
N) for Fmax=196 N. AE signals
from the Mode-I ring crack at the
load of 39.1 N for Fmax=98N are
shown in Fig. 4. Polarities of first
So mode was all positive, and
indicate the initiation of Mode-I
Hertz crack.
Fig. 3 Laser microscopic image of indentation (a) and
indentation curve with AE timings (b), for the test of Fmax: 98N.
Load at First
Mode-I AE [N]
F
,
N
▼：Mode-I
▽：Mode-II
(b)
Ring crack, r = 23m m
r
c
e
100
f
o
t
a
t
i
on
50
den
50µm
I
n
00 2 4 6 8
Indentation depth h, µm
(a)
Table 1 Load at the first Mode-I AE event.
98 196 294 392 490
39.1 127.4 38.1 40.7 41.4
Fmax[N]
1 2 3 4
0 4 8
Time, µs
0.08
ut,
V
0.04
p
0
-0.04
Out
-0.080 4 8
Time, µs
0 4 8
Time, µs
0 4 8
Time, µs
0 4 8
Time, µs
Fig. 4 AE waves from the Mode-I Hertz crack during the
indentation test of Fmax: 98N.
r
20
7.1G P a
7.1G P a
Indenter, 0.2R
r = 23µm
P
a
A
,
G
0
Tension
C om pression
σ
r
B
A: D iam ond film
B: SiC substrate
-200 20 40
Next we calculated the film
stress at the load of 39.1 N by FEM.
Figure 5 (a) shows the axial
r, m m
(a) (b)
Fig. 5 Analysis model for FEM (a) and calculated stress
distribution (b).
276
DGZfP-Proceedings BB 90-CD Lecture 25 EWGAE 2004
symmetric model for the FEM. We calculated the film stress for the indentation depth of 2.7 µm at 39.1
N. Elastic properties of the CVD-diamond measured by laser ultrasonic technique [13] were used.
Figure 5 (b) shows the distribution of radial stress. The maximum tensile stress is computed as 7.1 GPa
to the contact radius of 23 µm. This radius is consistent with the measured radius in Fig. 3. Thus the
tensile fracture strength of the CVD-diamond film was estimated to be 7.1 GPa.
LASER SPALLATION TEST
Preliminary test showed that the
diamond film suffers spallation at laser
energy of around 46 mJ. Figure 6 shows
microscopic images of spallated area.
Spallation occurred above laser energy of
46.1 mJ. We observed very small
delamination of 0.3 mm diameter at 46.1 mJ,
though not clear in the photo. Delaminated
circle expanded with laser energy. In order to
estimate the adhesion strength correctly, we
must determine the delamination initiation. Owing to the transparency of the polished diamond film, we
can detect the spallation by eye inspection when a finite size delamination occurs, however, much
accurate detection technique is needed for non-transparent and rough surface films.
1m m
φ 0.3m m
φ 1.46m m
1m m
(a) 46.1m J
(b) 47.9m J
Fig. 6 Optical microscopic images of spallation.
We are now using two method. One is waveform analysis from the interferometer. Figure 7
compares the outputs of laser interferometer for the case of spallation (b) and no-spallation (a). There
observed very clear waveform difference at first portion of the P-waves. Wave oscillation of 28-33 ns
cycle in (a) is due to the multi-reflection of the P-wave (1400 m/s) in the confined grease layer. This
oscillation disappears
when spallation occurs.
This is because the
reflected P-wave can
not be transmitted to
the delaminated film.
In this case, the film
spallation was induced
by the first arrive P-
wave. We next
estimated the critical
interfacial stress to
cause spallation. We
detected out-of-plane displacement of the stress wave at just epicenter on the substrate. Figure 8 shows
the waveforms as a function of laser power from 4.8 to 47.9 mJ. At low energies from 4.8 to 15 mJ, the
compressive wave or down shooting wave is the major component. The expansion component appears
above 35.5 mJ and increases with laser energy. Tensile stress of the expansion component is given by Eq.
-1000 500 1000 1500
100
1st P -w ave
2nd P -w ave
2nd P -w ave
100
m
1st P -w ave
ent
,
n
50
ent
,
n
50
m
m
0
em
0
e
-50
-50
Di
sp
l
a
c
Oscillation
Di
sp
l
a
c
-1000 500 1000 1500
Time, ns Time, ns
(a) 33.9m J
(b) 46.1m J
Fig. 7 Bulk waves detected by laser interferometer at just epicenter
of oil breakdown, (a) without spallation and (b) with spallation.
277
DGZfP-Proceedings BB 90-CD Lecture 25 EWGAE 2004
-50
4.8m J 15.0m J 21.5m J
27.5m J 33.5m J
50
400 450 500 550
50
400 450 500 550
0
0
Expansion
-50
Com pression
C om pression
C om pression
t
,
n
m
50
50
e
n
Expansion Expansion
46.1m J
Expansion
Di
sp
l
ac
e
m
0
0
Spallation
Spallation
C om pression
-50
-50
C om pression
400 450 500 550
Time, ns
400 450 500 550
Fig. 8 Change of first arrival bulk wave from oil breakdown, detected for SiC substrate without
the film.
(1) [14].
∂
-
= )
t
u
Vl ∂
ρ
σ (1)
where ρ is the density of the medium and Vl denotes the velocity of the longitudinal wave. ρ and Vl of
sintered SiC are taken as 3150 kg/m3 and 11803 m/s, respectively. The partial differentiation of
displacement with respect to time defines the particle velocity of the longitudinal wave. Figure 9 shows
the stresses calculated for compression (●) and expansion (○) waves. Here compressive stress was
designated by minus symbol. Tensile stress increased with increasing the laser energy. Data scattering
appears to be due to the difficulty in determination of the particle velocities. The adhesive strength of the
diamond film is estimated as 222 MPa.
In order to determine the spallation initiation accurately, we utilized AE monitoring. We previously
detected the impact-induced cracks in the PMMA hit by
fracture is hidden by strong impact induced AE
and could not be detected by the AE sensor
mounted on the plate surface. We, however,
could monitor the fracture-induced weak AE by
monitoring the in-plane motion of the So-Lamb
waves using the sensors mounted on the edge
plane of the target [15]. We used this technique
and detected Lamb-wave AE from the
spallation using the pinducer mounted on the
edge plane.
Figure 10 sh
flying objects [15]. For such case, weak AE by
(
t
Spallation
Spallation
400
a
P
200
I
nt
er
f
a
ci
al
s
t
r
e
s
s
,
M
222M P a
0
-200
ows detected Lamb wave AE as a
0 20 40 60
Laser energy, m J
function of laser energy. Arrival time of the first
So-mode coincides the calculated arrival time of
So-mode AE from the breakdown of silicone at
Fig. 9 Interfacial Stresses by pulse wave
due to the oil-breakdown.
278
DGZfP-Proceedings BB 90-CD Lecture 25 EWGAE 2004
Timp
Tspal
0.1 21.5m J
15.0m J
27.5m J
0.1
0.05
0.05
0
0
-0.05
-0.05
P ２
P 3
P 4
P 1
-0.1
-0.1
0.1
33.9m J
46.1m J
47.9m J
t
,
V
0.05
Ou
t
p
u
P ２
P 3
P 4
P 1
0
-0.05
-0.1
Spallation
Spallation
Spallation
Spallation
1000 2000
Timp : A rrival tim e of So w ave by laser impact.
Tspal : A rrival tim e of So w ave by spallation.
Time, ns
1000 2000
1000 2000
1000 2000
1000 2000
Fig. 10 First component of So-lamb waves detected by the pinducer mounted on edge place
on SiC with the film.
the sheet velocity of 11,585 m/s. This timing is
shown by a vertical line at Timp. Another vertical
line indicated by Tspal predicts the arrival time of
the first So-mode AE from the spallation.
Arrival time of first three peaks, P1, P2 and P3,
do not change for any laser energies, but their
amplitude and polarity changes with laser
energy. This means that the kinetics of oil
breakdown does not change. A small spallation
of 0.3 mm diameter occurred at the energy of
46.1 mJ. Waveform at 47.9mJ is so much
different form another waves. Both the polarity
and amplitude of P3 and P4 wave significantly
changed at 47.9 mJ. Plus polarity of P4 wave
indicates the Mode-I spallation. Cross-
correlation factor of these waves is shown in Fig.
11 as a function of laser energy. It shows a sharp change at 47.9 mJ which produce large spallation.
2
l
at
i
on
c
oef
f
i
ci
ent
Spallation
1.5
Spallation
1
0.5
0
-0.5
r
r
e
-1
Co
-1.5
0 20 40 60
Laser energy, m J
Fig. 11 Cross correlation coefficients of
waveform at in Fig. 10.
C
ONCLUSION
We utilized AE for detecting and classifying the film fractures during Rockwell micro-
indentatio
n and laser spallation test of CVD-diamond films deposited on the SiC plate. Results are
279
DGZfP-Proceedings BB 90-CD Lecture 25 EWGAE 2004
summarized below:
1) Mode-I Hertz ring crack during Rockwell indentation was detected by advanced AE system utilizing
2) e absolute interfacial strength of CVD-
3) plane motion of the So-mode Lamb wave
CKNOWLEDGEMENT
his research was conducted as part of the High-Technology Research Center Program and the
C
EFERENCES
] H. E. Powell and F. W. Preston, J. Amer. Cerami. Soc., 28, 6, (1945) 145.
ech. Eng. A60, 570 (1994),
[3] K. Hayashi and Y. Fang, Trans. J. Soc. Mech. Eng. A60, 617 (1998), 30.
. Kishi, Sur. Coat. Technol.,
[5] Y roda, A. Sato and H. Yamaguchi, Thin Solid Films, 213 (1992), 220.
1) 1.
1.
shed.
-93.
ction, 53, 4 (2004), 230.
four AE sensors around the indenter. Critical indentation load, indentation depth and ring crack
diameter to cause the first Hertz crack were determined and submitted to FEM analysis. Tensile
strength of the diamond film was estimated to be 7.1 GPa.
Laser spallation technique makes it possible to determine th
diamond film on SiC substrate by measuring the fast out-of-plane displacement of the expansion
stress wave. Adhesion strength was estimated as 222 MPa.
In order to detect the spallation initiation, we monitored in-
by broad band small sensor (pinducer) mounted on the distal planes. The sensor detected the AE from
large spallation.
A
T
enter of Excellence (COE) Program, funded by the Ministry of Education, Culture, Sports, Science and
Technology of Japan.
R
[1
[2] H. Kagawa, M. Ichikawa, T. Takamatsu and H. Kuwano, Trans. J. Soc. M
396.
[4] M. Shiwa, E. R. Weppelmann, A. Bendeli, M. V. Swain, D. Munz and T
68/69 (1994), 598.
. Tsukamoto, H. Ku
[6] P. A. Steinmann and H. E. Hintermann, J. Vac. Sci. Technol., A3(6) (1985), 2394.
[7] C. Julia-Schmutz and H. E. Hintermann, Sur. Coat. Technol., 48 (1991) 1.
[8] P. J. Burnett and D. S. Rickerby: Thin Solid Films 157 (1998) 233.
[9] C. Julia-Schmutz and H. E. Hintermann: Sur & Coat. Tech. 48 (199
[10] P. C. Jindal, D. T. Quinto and G. J. Wolfe: Thin Solid Films 154 (1987) 36
[11] R. Ikeda, S. Tasaka, H. Cho and M. Takemoto, Jpn. J. Appl. Phys, to be publi
[12] A.Yonezu, T.Ogawa and M.Takemoto, Progress in Acoustic Emission XI (2002)86
[13] R.Ikeda, Y.Hayashi and M.Takemoto, Progress in Acoustic Emission XI (2002), 53.
[14] V. Gupta, J. Yuan and D. Martinez, Surf. Eng. Struc. Ceram. 76 (1993) 305.
[15] H. Tanabe, Y. Mizutani and M. Takemoto, J. Jpn. Soc. Non destructive inspe
280
| 
760 KB