g..

h 'n
ELSEVIER Thin Sohd Films 307 (1997) 156-162

An energy approach to quantification of adhesion strength from

critical loads in scratch tests
H.-S. Park, D. K w o n *
Division of Materials Science and Engineering, Seoul National Universit); Seou1151-742, South Korea
Received 15 November 1996; accepted 13 March 1997

Abstract

The critical loads measured in scratch tests have been used in assessing adhesion in film/substrate systems. But critical loads are
strongly affected by various parameters such as scratching speed, indenter tip radius, fihn thickness, etc. Therefore, to assess the
interfacial adhesion strength reliably, the stress field developed by a sliding circular indenter was analyzed by considering the effects of
friction and plastic deformation of substrate. On the basis of this stress analysis, the elastic deformation energy of the film was estimated
by strain matching condition at the film/substrate interface. When the elastic deformation energy stored in the film reaches a critical
value, it is released by film failure such as delamination and spallation. Thus, the interfacial adhesion sn'ength can be evaluated in terms
of the work of adhesion by extracting the surface energy generated by spallation from the elastic deformation energy. The work of
adhesion of PECVD-prepared diamond-like carbon (DLC) films on WC-Co substrates was evaluated. In spite of the distinctly different
critical loads measured with the change of scratching speed and film thickness in the same plasma preetched specimen, the comparatively
constant work of adhesion was obtained. © 1997 Published by Elsevier Science S.A.

Keywords: Scratch test: Critical load: Elastic deformation energy: Work of adhesion

1. Introduction coefficient and residual stress in film [2]. To overcome this

restriction, we analyze the stress state ahead of the indenter
The scratch test, which is composed of a device moving and then evaluate the interracial adhesion strength using
the specimen with constant scratching speed and a device the energy approach. We verified this analysis by evaluat-
increasing indenting load, is a method to assess the adhe- ing adhesion strength of DLC films deposited on W C - C o
sion in films deposited on soft or hard substrates. In this substrates using the plasma-enhanced CVD (PECVD)
test, the loading indenter causes increasing elastic and method.
plastic deformation until failure occurs in the surface
region. This failure may be a delamination of the film at
the interface (an adhesive type of failure), and failure can
2. Modeling
also occur within the film itself (a cohesive type of failure)
[1]. The minimum load at which the adhesive failure
occurs has been defined as the critical load L c, which has In scratch tests, delamination at the interface between
been thought to characterize the interfacial adhesion the brittle film and the substrate occurs in the region ahead
strength in the film/substrate system. of the indenter, where the compressive stress develops.
However, the critical load has the restriction that it is Taking this into account, we evaluate the interfacial adhe-
strongly affected by several parameters related to testing sion strength from the measured critical load as follows.
conditions and to film/substrate system: for testing condi- First, the stress field generated by the sliding circular
tions, such parameters are scratching speed, loading rate, indenter is analyzed. From this stress analysis, the elastic
indenter tip radius and indenter wear; for film/substrate deformation energy stored in the film is assessed through
system, they are substrate hardness, film thickness, friction the strain matching condition at the interface. When the
elastic deformation energy reaches the critical value, it is
released by film failure (such as delamination and spalla-
* Corresponding author. tion). Thus, the interracial adhesion strength can be evalu-

0040-6090/97/$17.00 © 1997 Published by Elsevier Science S.A. AII rights reserved.

PII S0040-6090(97)00147- 8
 H.=S. Park, D Kwon / T h i n Sohd Fdms 307 (1997) 1 5 6 - 1 6 2 157

ated in terms of the work of adhesion by extracting the

surface energy generated by spallation from the elastic
deformation energy.

2.1. Analysis of stress and strain ahead of indenter

(a) (b) Cc)
The stress state in scratch test is analyzed in terms of
three contributions to the film failure: (1) a static indenta- Fig. i. Stress generation mechanisms: (a) indentation, (b) fncuon and (c)
residual stress.
tion stress, which is controlled by the indenting load: (2) a
friction stress due to the interaction between the sliding
indenter and the specimen surface; and (3) a residual stress
in film. For this purpose, the contact radius a between the
In Eq. (3), the directions of x and y axis are parallel and
indenter and the specimen must be evaluated. It is consid-
perpendicular, respectively, to the direction of sliding in-
ered that a hard thin film deposited on a substrate is not
denter on the surface of specimen, as shown in Fig. l(a).
deformed by indenting load, but transfers the indenting
When the brittle material is plastically deformed by the
load to the substrate. Thus, this contact radius can be
matelial's compaction beneath the indenter, the stress state
evaluated by Hertzian contact [3], which is formulated on
in Eq. (3) is inappropriate, since Eq. (3) assumes elastic
the assumption of elastic deformation between two bodies
deformation. This plastic deformation generates the addi-
in contact:
tional iocal stress (blister stress), which is superposed on
3 ( I -1~ 2 1 -- Z'ff the preexisting elastic stress (Eq. (3)) [8]. Blister stress is
a3 = 7 LR, ,, Es Ei (1) proportional to 1/r 3, where r is the distance from the
center of the indenter. Thus, by superposition of blister
where L is the indenting load, R i is the radius of indenter, stress on Eq. (3), the appropriate elastic indentation stress
u and E are the Poisson's ratio and the Young's modulus, field can be expressed as
respectively (subscripts s and i represent the substrate and L B
the indenter, respectively). When the plastic deformation cr,~ = 2rra-' (1 - 2v~) 74(2 - z'~) (4)
of the substrate beneath the indenter is dominant, the
contact radius can alternatively be estimated using the L B
relationship between hardness and indenting load: cry,, 2rraZ(1-2v~)+-S.24(1-2vs)
%:=o-~,=cr =o-x=O
a = (2)
where B is a constant depending on the extent of the
plastic deformation and the yield strength of the substrate.
where H s is the hardness of the substrate. When a specimen is indented by a circular indenter, B can
When a specimen is indented by a hard sphere, a be determined through the yield criterion: at the elastic-
symmetric stress distcibution develops, where the three plastic deformation boundary (x = a, y = z = 0), the esti-
principal stresses are compressive beneath the contact re- mated stress state should satisfy the yield criterion [9]. We
gion to a depth of the order of the contact diameter. can then calculate the constant B by adopting the von
Outside the contact region, there exist a radial tensile stress Mises yield criterion as
and a hoop stress, each of which takes on its maximum
value at the edge of the contact region [4]. The surface 16B e 18BE
a6 (7r,~2 - 13v~ + 7) - ~ ( 1 - 2us)(1 - h )
stress and strain fields are independent of the contact stress
distribution generated by the indenting load, assuming the
3L 2
same total indenting load L and a radially symmetric stress -- (1-2u~) z y2=o (5)
distribution inside the contact region [5-7]. Thus, these + 4vr2a 4
surface stresses and strains at the edge of the contact where Ys is the yield strength of the substrate.
region can be obtained as in Eq. (3) in Cartesian coordi- The friction effect generated by a sliding indenter leads
nates: to stress enhancement over the indentation stress field and
L also induces a compressive stress at the )eading edge,
o-uu 2~ra 2 (1 - 2z,~) (3) resulting in the strong compression of the film in front of
the indenter. The friction stress field can be derived by
L regarding the indenting load as a symmetric stress distribu-
%.y = 27ra 2 (1 - 2z~) tion within the contact region [5]. The stress field ahead of
the indenter is obtained by superposition of the stress field
%z = %~ = % . " = %x = 0
corresponding to the friction shear over the contact region
158 H.-S. Park, D. Kwon / Thin Solid Films 307 (1997) 156-162

on the stress field of Eq. (4). Thus, the total elastic stress where pf and Ef are the Poisson's ratio and the Young's
field acting at the leading edge is modulus of the film, respectively, and cG~ is the residual
stress in film.
%x = 2rra 2 (1 - 2us) - (4 4- Ps) - ~T4(2 - us) The adhesion behavior can be modeled in terms of the
strain energy released during film failure. At the critical
(6) load, the film ahead of the indenter releases its elastic
deformation energy by interface delamination and film
O-yy= 2~ra 2 (1 - 2 v ~ ) + 4- ~-~4(1 - 2 v ~ ) spallation. The released energy provides the surface energy
for the film failure. Let us consider a film failure region at
O-zz = % y = O'y: = %x = 0 the critical load as a semicircular area of film with radius c
detached from the substrate. The total stored energy in film
where /~ is the friction coefficient between the indenter can be expressed as
and the specimen.
From this stress analysis, we can derive the strain of the ~( 7rc2t 77-c2
v= - w 2( vr + 2) CYrt (9)
film ahead of the indenter. We assume that the strain in the
-

2
film may not decrease significantly over the film thickness W = % + y f - %f
since the film is thin enough. And the strains of the film
and the substrate are expected to be the same at the where t is the film thickness, y, and yf are the surface
interface, since the adherence between the film and the energies of the substrate and the film, respectively, and %r
substrate must be maintained at the interface. Then the is the interracial energy between the film and the substrate.
film strain can be expressed as On the right-hand side of Eq. (9), the first term is the
elastic deformation energy stored in the film, the second is
the surface and interfacial energy created by film delami-
G x = ~ - ~ [ o T ~ - vsO'yy]=-~- 2¢ra------~.(1+ v ~ ) ( 1 - 2 G )
nation and the last is the surface energy generated by film

. 3,rr.]
-(4-3vJ--~J-vS(1-v~)(t+n)j
B 1 (7)
spallation. In Eq. (9), the work of adhesion W characteriz-
ing the interfacial adhesion strength is expressed by the
surface and interfacial energy.
1 1[ L If the energy released during film failure is ~eater than
~>'y= "~s [O'yy - P s O ' x x ] = ~ -- 2~ra~( l + ~s) the increase of the interfacial and surface energy due to the
delamination and spallation, the film failure will be contin-
ued. The above condition can be expressed through the
× (1 - 2v~) + 7 4 ( 1 - v~)(1 + v~) Griffith energy-balance approach:
8
There also exists the strain component in the z direction, Oc ( E c Z ) T r c t - W~rc - 2(~r+ 2 ) y f t = 0 (10)
but it is too minimal to stretch the film compared with
strains in the x and y directions. Thus, its contribution in From Eq. (10), we can derive the work of adhesion W:
evaluating the elastic deformation energy stored in the film ~( 2(7r + 2) yf
is neglected. W= 2e2)t t (11)
"/7" C

2.2. Evahtation o f interfacial adhesion strength as w o r k o f In Eq. (11), the value of " y J c can be estimated by
adhesion assessing the elastic deformation energy stored in a unit
volume of film with the change of film thickness under the
We can evaluate the elastic deformation energy stored same interracial adhesion strength:
in a unit volume of film using the strain field in Eq. (7). n 2 ( ~ + 2) yf
However, we must take the film residual stress into consid- ~~f ~ S 2 = - -t + ~r c (12)
eration. This residual stress effect can be evaluated by
superposition of the residual strain on Eq. (7). Thus, the In other words, through plotting the elastic deformation
elastic deformation energy can be given as energy stored per unit volume of film as a function of 1/t,
we can obtain the value of y ¢ / c by extrapolating this plot
to 1 / t = O, and the work of adhesion can be evaluated
from Eq. (11) using the value of y r / c .
1 oh, ( 1 - v~.)
~xx = -- "" Ef
3. E x p e r i m e n t a l details and results
1 %~(1 - uf)
8yy Cryy psO'xx] 4- Ef In this work, the WC-Co K20 substrates were mechani-
cally polished with diamond grinder and then with dia-
 H.-S. Part:, D. Kwon/Thin Sohd FzIms 307 (1997) 156-162 159

Table I
Experimental results obtained at indenter tip radii 200 and 800 `am and theoretically evaluated work of adhesion (IV) for DLC film deposited on WC-Co
Deposition Film Indenter tip radius
time finn) thickness
200 ,am 800 `am
(tam)
Critical Friction Contact radius Work of adhesion Critical Friction Contact radius Work of adhesion
load (N) coefficient (,am) (J/m:) load (N) coefficien{ (`am) (JTm2)
I5 0.78 33.32 0.0797 25.6 17.87 68.85 0.0615 36.8 8.81_
30 1.31 30.70 0.0552 24.6 17.81 50.73 0.0424 31.6 4.85
45 2.32 12.50 0.0425 15.7 17.90 23.89 0.0532 21.7 11.52

Loading rate: 100 N/min; scratching speed: 10 mm/min.

mond paste (particle size 3 /~m). They were ultrasonically min with a plasma etching time of 15 min. Scratch tests
cleaned in trichloroethylene, acetone, methanol and ethanol were performed on each specimen. The Rockwell C dia-
sequentially. The DLC films were deposited by plasma-en- mond indenters (conical angle: 120°; hemispherical tip of
hanced CVD (PECVD). Before deposition, the chamber 200 and 800 # m radii) of the test apparatus were drawn
was evacuated to a pressure of 7.0 × 10 -~ Torr and then over the sample surface with continually increasing normal
backfilled with argon gas to a pressure of 0.15 Ton" by load. The loading rate was 100 N / m i n and the scratching
flowing the argon gas at 14 sccm. Under this condition, the speed was 10 m m / m i n . The indenting load, tangential
Ar plasma was generated with the application of H.F. friction force, friction coefficient and AE signal from the
(high frequency: 250 kHz) power and plasma preetching of diamond tip were continually monitored. The detailed
the substrate was performed by exposing the substrate experimental results are shown in Table 1. Then, we
surface to the generated Az plasma. Then the composition deposited the DLC films for 30 min with plasma etching
of gas was changed in a few seconds from Ar 14 sccm and times of 5, 15 and 30 min. Scratch tests were carried out
C H 4 0 sccm to Ar 0 sccm and C H 4 14 sccm. After this with scratching speeds of 5, 10, 15 and 20 m m / m i n , under
procedure, DLC films were deposited under an initial the constant loading rate of 100 N / m i n and the indenter
pressure of CH 4 0.3 Torr. To evaluate the residual stress in tip radius of 200 /xm. Table 2 shows the experimental
the film, we deposited the DLC film on the {001} flat results under these conditions.
silicon wafer of thickness 2 0 0 / x m and aspect ratio at least Fig. 2 shows the variation of critical load measured in
13. After the deposition, the curvature of substrate was specimens with different deposition times. The measured
measured by laser deflection method and then the residual thicknesses of films were 0.78, 1.31 and 2.32 /zm with the
stress in the film was estimated [10]. The film thickness increase~ffleposition time, and the compressive residual
was measured using an a-step profilometer. All the critical stresses in films were estimated as 813, 669 and 632 MPa,
load measurements were performed by the scratch tester respectively. As seen in Fig. 2, the critical loads decrease
equipped with an acoustic emission (AE) detector. significantly with the increase of film thick_hess and in-
The detailed experimental conditions and results are as crease with the change of tip radius from 200 ttm to 800
follows. We deposited the DLC films for 15, 30 and 45 tzm. Fig. 3 illustrates the variation of critical load with

Table 2
Experimental results obtained at plasma etching times of 5, 15 and 30 rain and theoretically evaluated work of adhesion (IV) for DLC film deposited on
WC-Co
Plasma etching Scratching speed Film thickness Critical load Friction coefficient Contact radius Work of adhesion
time (min) (mm/min) (`am) (N) (`am) (J/m 2)
5 5 1.16 37.2i 0.0622 27.0 19.50
10 22.08 0.0593 20.8 17.60
15 I3.63 0.0693 16.4 I9.21
20 13.77 0.059i 16.5 16.08
15 5 1.31 38.67 0.0606 27.6 20.O7
i0 30.70 0.0552 24.6 17.78
15 12.13 0.0493 15.4 12.85
20 11.8i 0.0513 15.2 13.39
30 5 1.30 66.83 0.0825 36.2 27.16
10 51.86 0.0832 3-1.9 27.17
15 33.89 0.0787 25.8 24.88
20 28.01 0.0715 23.5 21.91

Indenter tip radius: 200 `am, loading rate: 100 N/rain.

160 H.-S. Park, D. Kwon / Thin Solid Fihns 307 (1997) t56-152

, . , . , --
7o 60
E
.-.}
v

50.
50. • ©
E
95 UJ
O 40. =
._1 E: 40.
O
"~ 30. ~ =

• -- indenter 200/am
30.
O 2o.
- - . - - indenter 8 0 0 g m

0,5 1.0 1.5 20 2 5 .-- 20 , ' L , i , i

0.2 0,4 0.6 0.8 1.0 1,2 1.4

Thickness (gm)
[]
1/Thickness (ll~tm)
Fig. 2. Variation of critical load Lc with film thickness (t) for indenter tip
radii of 200 and 800/xm (loading rate: 100 N/min, scratching speed: 10 Fig. 4. Variation of elastic deformation energy stored in a unit volume of
ram/mAn). film with film thickness ( l / t ) for indenter tip radii of 200 and 800 #m
(loading rate: I00 N/min, scratching speed: 10 ram/rain).

different plasma etching times. Film thicknesses were 1.16, 30

1.31 and 1.30 ,am with the increase of etching time, and
c'4
the compressive residual stresses in films were 741, 733 E 20-
and 7 3 5 MPa, r e s p e c t i v e l y . A s s h o w n in Fig. 3, the critical i-- --i .i
c"
loads decrease distinctly with increasing scratching speed o
ffl 10.
and increase with increasing etching time. Although there t-- "~.,~
is a little deviation in critical loads measured at scratching "0
<
speeds greater than 15 mm/min, the general variation of "5 --,-- indenter 200gin
critical load is in good agreement with the above explana- --.-- indenter 800rtrn
tion. -10
O5 110 I'5 2'0 25
By assuming the Poisson's ratio of the film as 0.25, a Thickness (I-zm)
typical value for the brittle material, and its Young's
Fig. 5. Variation of work of adhesion with film thickness (t) for indenter
modulus as 2 0 0 G P a [11,12], w e e s t i m a t e d the elastic
tip radii of 200 and 800 p~m (loading rate: i00 N/min, scratching speed:
deformation energy in a unit volume of film with the 10 mm/min).
change of film thickness. These values are shown in Fig. 4.
The values of 7f/c calculated from Eq. (12) were 8.79 and
10.96 J / m 3 with indenter tip radii 200 and 800 ,am, ,am, we evaluated the work of adhesion for different
respectively. Using these values, we evaluated the work of plasma etching times, as listed in Table 2 and shown in
adhesion (see Table 1). The evaluated values of the work Fig. 6. The comparison of Fig. 6 with Fig. 3 shows that in
of adhesion using tip radius 200 ,am were 17.87, 17.82 spite of the distinctly different critical loads with scratch-
and 17.90 J/re'-, nearly constant regardless of the change ing speed, the evaluated values of the work of adhesion are
in f i l m thickness; but those o f the w o r k o f a d h e s i o n using relatively constant except at scratching speed 15 and 20
tip radius 800 ,am were 8.81, 4.85 and 11.52 J/re", as mm/min. The values increase: 18.6, 18.9 and 27.2 J / m 2
shown in Fig. 5. with plasma etching times of 5, 15 and 30 min, respec-
Using the value of 7Jc obtained with tip radius 200 tively.

70 , , , ~ , , , ;~ , ~ ;-- T -,-- T 50

6o --,-- 5rain etching

N" --.-- 5rain etching
--,--15m~n etching E 40.
A- ~ - - . - - 3 0 m m
-'-- --.-- 15min etching
Z 50 etching
E --,,-- 30min etching
O 30 -

_J
3o

o
20j
1o t
4 6 8 10 12 I4 16 18 20 22 4 6 8 10 12 14 16 18 20 22

Scratching S p e e d (mmtmin) Scratching Speed (mmtmin)

Fig. 3. Variation of critical load Lo with scratching speed for plasma Fig. 6. Variation of work of adhesion with scratching speed for plasma
etching times of 5, 15 and 30 min (loading rate: 100 N/mAn, indenter tip etching times of 5, 15 and 30 min (loading rate: 100 N/mAn, indenter tip
radius: 200 p,m). radius: 200 /zm).
 H.-S. Par~ D. Kwo~t / T h i n Solid Films 307 (1997) 156-162 161

4. Discussion

Interfacial adhesion strength in scratch tests has been

described as either critical shear stress generated at the
interface [13,14] or the surface energy created by delami-
nation [15,16]. However, complete film removal does not
normally occur abruptly; in general, a greater indenting
load is required for complete film removal than is required
to exceed the critical shear stress of the interface [17]. In
addition, film removal ahead of the indenter occurs by
compressive stress [15]. These observations suggest that
the process of film failure may be better described in terms
of an energy balance criterion, and this is adopted here. l(b)
The elastic deformation energy stored in the film in-
creases with increasing indenting load. When this energy
reaches the critical value, it is released to provide the
surface energy created by film failure and the energy of
crack-tip plasticity. However, the failure at the interface is
brittle failure, since both the substrate and the film are hard
and brittle. We can therefore neglect the crack-tip plastic-
ity effect and evaluate the interfacial adhesion strength as
the work of adhesion by using a Griffith energy-balance
criterion.
The determination of the constant B in Eq. (4) can be Fig. 7. (a) Optical micrograph of DLC fiIm failure ahead of indenter
made by adopting a yield criterion at an elastic-plastic radius 200 #m. (b) Optical micrograph of DLC film failure beside
boundary. A material indented with a spherical indenter, indenter radius 800 /xm.
first deforms elastically. As the indenting load increases,
plastic deformation occurs first on the centeriine 0.5a 2. As the contact region betxyeen the indenter and the
below the surface of specimen where the yield criterion is specimens increases with the increase of tip radius, a
satisfied. The plastically deformed region expands with greater indenting load is needed to achieve the same stress
increasing indenting load and reaches the radius (x = a, state; thus, the critical loads increase with the tip radius.
y = z = 0) of the contact region in an indenting load range Fig. 7(a,b) show the film failure regions using tip radii 200
on the order of 10 N [18]. Therefore, the constant B can be and 800 /,m in the specimen of 2.32 # m film thickness.
determined through yon Mises yield criterion (Eq. (5)) by With tip radius 800 /,m. film failure occurred not ahead of
assuming the elastic-plastic boundary as the contact radius but beside the indenter. The reason for the different failure
(x=a, y=z=O). mode-with the tip radius change is probably that the
Fig. 2 shows the variation of critical load with the indenting load !s !ransferred to t)e f i!m at thfl side of the
change of film thickness for indenter tip radii 200 and 800 indenter as tip radius increases, so that the film failure
/,m. In general, if we assume that the adhesion is the same occurs beside tee indenter, not ahead of it.
and the critical load is determined by the extent of defor- We thus used the indenter tip radius 200 /xm instead of
mation, an increased film thickness requires an increased 800 # m to measure the critical load with varying scratch-
indenting load to obtain the same deformation, so that the ing speed. The critical loads measured for different etching
critical load increases with the film thickness [19,20]. But, times are shown in Fig. 3 with scratching speeds of 5. 10,
considering the residual stress effect in the film, the varia- 15 and 20 mm/min. As the figure shows, the critical load
tion of critical load with film thickness may be described decreases with greater scratching speed and increases with
in another way. In our experiment, the estimated residual longer etching time. The Iatter effect is mainly due to
stresses are 813, 669 and 632 MPa for film thicknesses of better cleanness of the surface as the etching time in-
0.78, 1.31 and 2.32 /.Lm, respectively. These residual creases. The former effect is probably because, when
stresses decrease as film thickness increases, but the elastic scratching speed increases under constant loading rate, the
deformation energies stored in the film (oc VCrr~t) due to probability of encountering defects in the film within a
the only residual stresses increase. Thus the critical loads certain load range increases, so that critical load decreases.
decrease with increasing film thickness. At the scratching speeds above 15 mm/min, however,
To determine the effect of the indenter tip radius on the critical load apparently did not decrease with increasing
critical load, the specimens described above were used. scratching speed. It is likely that the film may wear at
The critical loads measured by a diamond indenter with higher scratching speeds. In general, when wear occurs
different tip radii 200 and 800 /,m are also shown in Fig. between materials in contact, the softer material is de-
162 H.-S. Park, D. K~von/Thin Solid Fihns 307 (1997) 156-162

formed by debris on contact surface, or the asperities of strain. In addition, by analyzing the strain in the film, we
the surface are deformed. Indenting load is consumed in could evaluate the work of adhesion through the Griffith
these deformations. So the measured critical load will be energy-balance approach.
greater than is expected. To avoid this wear, we measured (3) The critical load L c in the same preetched specimen
the critical load at scratching speeds 10 r a m / r a i n in increased with increasing indenter tip radius, and de-
differently deposited specimens. creased with increasing film thickness and scratching speed.
Fig. 4 illustrates the variation in elastic deformation This indicates that the critical load can just be quantita-
energy stored in a unit volume of film with varying film tively used in a given system and with external parameters
thickness ( t / t ) . By tinear regression of data, we obtained fixed. However, the values of the work of adhesion calcu-
the typical line for tip radius 200 gin, indicating that the lated in current analysis were relatively constant irrespec-
film failure region c is constant regardless of film thick- tive of external parameters such as film thickness and
ness. But for tip radius 800 /xm, the elastic deformation scratching speed.
energy for film thickness 2.32 # m is distinctly different (4) The work of adhesion in differently preetched speci-
from that for tip radius 200 /,zm. This discrepancy is mens increased with increasing plasma etching time: 18.6,
mainly due to the fact that the film failure induced by tip 18.9 and 27.2 J / m 2. This shows that the work of adhesion
radius 800 /zm occurred beside the indenter, not ahead of based on micromechanical analysis of indentation stress
the indenter, as described above. field successfully reflects the tendency of the increase in
Using the above y f / c value, we evaluated the work of adhesion by plasma etching; so, current analysis provides a
adhesion in differently preetched specimens with varying reliable tool for quantitative evaluation of the interracial
scratching speed; these values are shown in Fig. 6. The adhesion strength.
values of the work of adhesion were evaluated constantly
except at scratching speeds 15 and 20 ram/rain. This
discrepancy may result from the wear described above. Acknowledgements
Therefore, when considering the scratching speed range in
which the wear is not severe, we could evaluate the quite
This work was supported by the Korea Science and
constant work of adhesion not influenced by any effect of
Engineering Foundation (961-0801-011-2).
scratching speed. Furthermore, the constant work of adhe-
sion could be assessed eliminating the film thickness effect
in the current stress analysis by extracting the surface
energy generated by film spallation. References
These experimental results mean that interfacial adhe-
sion strength can be analytically evaluated as the work of [1] H.K. Pulk, A.L Perry, Surf. Technol. 14 (1981) 25.
adhesion from the critical load, which is dependent on [2] H.E. Hintermann, J. Vae. Sci. Technoi. B 2 (1984) 816.
[3] S. TLmoshenko,J.N. Goodier, Theory of Elasticity, McGraw-Hill,
various parameters besides adhesion. This evaluation gives New York (1951) p. 414.
us reliable data on interfacial adhesion strength that ex- [4] K. Zeng, K. Breder, D.J. Rowliffe, Aeta MetalL Mater. 40 (10)
clude the effects of external experimental factors. (1992) 2595.
[5] G.M. Hamilton, L.E. Goodman, J. Appt. Mech. 33 (1966) 371.
[6] M.J. Boussinesq, Mere. Soc. Sci. Agfic. 13 (1885) 99.
[7] S. Way, E. Pittsburgh, J. Appl. Mech. 7 (1940) 147.
5. Conclusions [8] E.H. Yoffe, Phil. Mag. A 46 (4) (1982) 617.
[9] A.K. Ghosal, S.K. Biswas, Phil. Mag. B 67 (3) (1993) 371.
We have evaluated the interracial adhesion strength as [10] G.G. Stoney, Proc. R. Soc., London, A 82 (1909) 172.
the work of adhesion through stress analysis in the region [11] X. Jiang, K. Reichelt, B. Stritzker, J. Appl. Phys. 68 (1990) 1018.
ahead of the indenter in DLC film deposited on W C - C o . [12] N. Savvides, T.J. Bell, J. Appl. Phys, 72 (t992) 2791.
[13] C. Weaver, J. Vac. Sei. TechnoI. 12 (1975) 18.
The conclusions are as follows.
[14] B. Oliver, A. Matthews, J. Adhesion Sci. Technol. 8 (6) (1994) 651.
(1) In a scratch test, the stress state in the material [15] M.T. Laugier, J. Mater. Sci. 21 (1986) 2269.
ahead of the spherical indenter can be described by consid- [16] S. Venkataraman, D,L. Kohlstedt, W.W. Gerberich, J. Mater. Res. 7
ering the blister stress caused by the plastic deformation in (5) (1992) 1126.
brittle material and the friction stress. [17] M.T. Laugier, Thin Solid Films 76 (198t) 289.
[18] K. Sadeghipour, W. Chen, G. Barant, J, Phys. D 27 (1994) 1300.
(2) Through stress analysis, the strain of the film can be
[19] P.A. Steinmann, Y. Tardy, H.E. Hintermann, Thin Solid Films 154
described by using the strain-matching criterion at the (1987) 333.
film/substrate interface and by considering the residual [20] J. Valli, J. Vac. Sci. Technol. A 4 (1986) 3007.