CP620, Shock Compression of Condensed Matter - 2001
edited by M. D. Furnish, N. N. Thadhani, and Y. Horie
© 2002 American Institute of Physics 0-7354-0068-7/02/$ 19.00
DYNAMIC RESPONSE OF TITANIUM CARBIDE-STEEL,
CERAMIC-METAL COMPOSITES.
B. Klein1, N. Frage1, E. Zaretsky2 and M.P. Dariel1
1
Department of Material Engineering,2 Department of Mechanical Engineering,
Ben-Gurion University of the Negev, P.O.Box 653, Beer-Sheva 84105, Israel
Abstract. The dynamic response of a titanium carbide (TiC)-carbon steel, ceramic-metal composite, was
studied in planar impact experiments, using a copper impactor with velocity in the 80 to 450 m/sec range.
The composites were prepared by pressureless infiltration of TiC ceramic preforms with molten steel. The
metallic component had either a pearlitic or a martensitic microstructure, determined by an appropriate
heat treatment. Fully dense composites, consisting of TiC and 1060 steel, in pearlitic and martensitic
states, were used as reference samples. The values of the HEL and of the spall strength were derived from
the VISAR records of the free surface velocity of the impacted samples. The results indicate that the
confining stress, produced by metallic matrix on the TiC particles, changes drastically the dynamic
response of the composite.
INTRODUCTION
MATERIALS
Ceramic-metal composites (cermets) have a
potential as armor plates but require an in-depth
understanding of their dynamic response. The
relevant available information on cermets is very
scarce and the available data are non-systematic. In
particular, the influence of the state of the metallic
component on the dynamic response of cermets was
never studied.
The present work is an attempt to provide some
information regarding any such influence. TiCcarbon steel composite was chosen for the present
study. Although the dynamic properties of dense
TiC ceramics are not high, they present some
definite advantages. Ceramic TiC matrices with
controlled open porosity (preforms) may be
manufactured easily. Moreover, owing to the good
wetting of TiC by molten steel, the porous preforms
can be completely infiltrated by the molten metal.
By varying the heat treatment applied to the cermet,
the state of the metallic component can be changed
while that of the ceramic matrix is kept constant.
The composites were prepared by pressureless
infiltration of TiC ceramic preforms with 0.6%C
molten steel, followed by furnace cooling. The
infiltrated ceramic-metal pieces (presamples)
contained of about 30 vol.% of steel. These
presamples were divided into three groups: The
samples, referred henceforth as (a), were directly
machined from the presamples into 3-4-mm thick,
20-mm diameter disks. A second group of
presamples, (q), was heated to 870°C, waterquenched and machined into disks. A third group of
presamples, (qa), was tempered after the quenching
for one hour at 250°C, furnace cooled, and
machined into the disks. Similar size disks were cut
from a rod of commercial, normalized 1060 steel.
Part of the steel discs underwent the same heat
treatments as the cermet presamples, and are also
referred as (a), (q) and (qa), respectively. Finally,
disk-shape samples made of fully dense TiC were
also prepared and tested.
The various heat-treatments were carried out in
order to determine the effect of the microstructure
1119
of the metallic component on the dynamic response
of the composite material. Prior to the impact
experiments the surfaces of the disk samples were
lapped to better than 0.005 mm parallelism, the
density, p0, determined by liquid displacement in
distilled water and the longitudinal C/ and
transversal Ct sound velocities, measured by the
ultrasonic pulse-echo method. The average results
of these measurements and the corresponding
Poisson's ratio are shown in Table 1.
Since the sample made of dense TiC was
completely destroyed by the compressive stress of
the HEL level (see Fig. 2), a weaker impact
experiment was performed in order to determine the
spall strength of TiC. The detected velocity pullback, Fig.2, yields the TiC spall strength
a™u =0.29 GPa. The weak (below HEL) impact
experiments were also performed with cermet
samples that had undergone different heat
treatments, Fig. Ic. The values of the corresponding
spall strengths, a*sp, are also given in the Tab. 1.
The velocity profile of the stronger shot with fully
dense TiC, Fig. 2, reveals that any compressive
deformation above HEL leads to the complete
fracture of the ceramics (arrow I in Fig. 2). In order
to evaluate the yield stress, YTiC, corresponding to
the fracture of the brittle material in compression,
we used the expression derived by Rosenberg [2]
from Griffith's yield criterion
RESULTS AND DISCUSSION
The prepared samples were studied by planar
impact experiments, using 1-mm thick copper
impactors, accelerated by a 25-mm gas gun to
velocities ranging from 80 to 450 m/sec. The
impactor-sample misalignment did not exceed 0.5
mrad in all the experiments. The velocity w of the
free surface of the samples was continuously
monitored by VISAR [1]. Part of the recorded
velocity profiles is shown in Fig.l. The profiles
corresponding to the materials that underwent heat
treatments are shifted upward for the sake of clarity
and the dimensional-less time is r = f C/ / <5 , where
d is the sample thickness and t is the time after
impact. The estimations of the Hugoniot Elastic
Limit, a//, and the spall strength, <7vpa//, listed in Tab.
1, were obtained from the VISAR records of the
sample free surface velocity w using the formulas
OH =0.5poC/w//£ L and aspan =0.5poC/Aw ¥a //.
1-v
(D
where v is the Poisson's ratio. The data of Tab. 1
yield a value of YTiC = 2.41 GPa for the OH =
=5.87GPa. According to the Griffith's biaxial-stress
l€
is the
yield criterion, YTic = 8o"o |C , where al
material tensile strength under uniaxial stress
tension. By taking the spall strength, osTlllU of the
800
700
\7~i7
600
IqaT
(qa)
T 500
(q)
400
(q)
300
(a)
200
(a)
100
0
1
2
L
(a)
3
4
5
6
nondimensional time after impact
Figure 1. VISAR records of the sample free surface velocity profiles obtained in the shots with steel, a, and cermet, b and c, samples after
different heat treatment (shown near the profiles).
1120
TABLE 1 Static and dynamic properties of the studied materials.
Materials and
treatment
TiC
Composite (a)
Composite (q)
Composite (qa)
Steel 1060 (a)
Steel 1060(q)
Steel 1060(qa)
density,
g/cm3
4.78
5.50
5.43
5.47
7.84
7.76
7.72
c/
km/sec
10.12
8.92
8.76
8.72
5.94
5.87
5.93
cb
km/sec
7.280
6.540
6.370
6.280
4.610
4.590
4.650
Poisson's
ratio
0.216
0.235
0.226
0.217
0.288
0.294
0.296
0_________
1.5
2
2.5
nondimensional time after impact
Figure 2. VISAR records of the sample free surface velocity
profiles obtained in different shots with TiC samples. The arrows
mark the arrival of the plastic wave (I), the unloading wave (II)
and the spall signal (III) to the TiC samples free surface.
YTIC
=
as
a TiC
we
expect
to
waii ,while according to the data of Tab.
Ospalb
®*spall>
GPa
0
0.97
1.04
0.91
3.37
GPa
0.29
0.84
1.48
2.14
-
5.87
1.98
4.47
5.68
2.21
3.64
3.46
2.79
3.16
F,GPa
(Griffith)
2.42
0.73
1.73
2.32
.
-
F,GPa
(von Mises)
4.25
1.37
3.16
4.11
1.32
2.12
2.01
cycle. In the sub-HEL shots, see Fig. Ic, the matrix
thermal treatment results in the increase, with
respect to the over-HEL shots, of the ospaii of the
(q) and (qa) samples, by a factor of 1.42 and 2.35
times, respectively, while the GSpaii of the (a)
sample stays almost unchanged. The increase may
be attributed to some hydrostatic compression p *,
generated within the spheroid ceramic particle by
the steel envelope under tension. This compression
prevents the premature opening of the cracks in the
ceramic particle and, thus, maintains the integrity of
the sample cross section. The 30/70 matrix-toparticles volume ratio results in an average ratio of
the envelope thickness, 8 , to the particle radius,
R , of about 8/R = 0.09. We assume that the excess
spall strength of the ceramic particle is due to this
250
material
OHEL)
GPa
compression, p* = Aaspall = & spall -° Spall • I n tnis
get
1,
case the tension in the steel matrix is
YTIC f°™n = 2.41/0.29 = 8.3 . Within the accuracy
of the experimental data, the agreement allows to
conclude that dense TiC is a typical brittle ceramic.
The spall strength of the cermets in the stronger
(over-HEL) shots is almost independent of the state
of the metallic matrix. The spall strength of the
cermets varies between 0.91 and 1.04 GPa, while
the spall strength of the corresponding steel samples
changes from 2.79 to 3.37 GPa. The cermet-to-steel
spall strength ratio varies between 29 to 35% and is
close to the volume fraction of the metal within the
cermet. During the tensile path of the over-HEL
shots, the effective cross-section of the cermet is
actually the cross section of the metallic matrix,
because the ceramic component has been destroyed
in the course of the compressive path of the loading
OT = Sp*/R
and is equal to 0.10 and 0.16 GPa for (q) and (qa)
materials, respectively. Since the response of the
steel matrix under such a low tensile stress is
purely elastic, it stays unchanged during the
compression. The fracture of the sample starts
when this excess pressure is cancelled in tension.
The ceramic particles in the un-treated, (a), cermet
sample seems free of this excess pressure.
The presence of the compression p* may
explain the striking difference revealed between
the values of <JHEL of the steel and cermet samples,
tempered after quenching. The <JHEL of both the
steel (q) and cermet (q) samples increased with
respect to the OHEL of the untreated materials (a).
The tempering (qa) of the cermet sample results in
a substantial increase of the OHEL while the
1121
of the steel sample decreased by a few percent
after tempering, see Tab. 1.
It was shown that the yield behavior of pure
dense TiC follows the Griffith's criterion, Eq. (1).
Metals, and, in particular, steel follow the von
Mises yield criterion, which, in the case of 1-D
compression, can be written:
1-v _ _
GHEL -~——-* (l-2vY
Notice that (5) is the larger root of Eq. (4) and
expression (5) coincides with (1) when p* = 0.
The dependence (5) of the <JHEL on the
hydrostatic pressure p* is shown in Fig. (3) for
different values of the yield stress Yc.
cd
CU
(2)
O
The values of the yield stress, calculated for steel
samples according to Eq. (2), and for cermet
samples, according to both Eqs. (1) and (2), are
listed in the two right columns of Tab. 1. It is
apparent from the data in these columns, that for
the same value of OHEL, yielding according to Eq.
(1) occurs at a lower stress, and, consequently,
yielding of the cermets starts in their ceramic
component. Since three cermets were made from
similar TiC performs and differed only in the state
of the metal component, the difference in the
values of the OHEL of the cermets is due to the
mechanical interaction of the steel matrix with TiC
inclusions. In the presence of the hydrostatic
compression/?*, in the Griffith criterion
(<J\ -o- 2 ) 2 = Y(a\ +<7 2 )
o
0.5
1
1.5
hydrostatic pressure, GPa
2
Figure 3. Hugoniot Elastic Limit of cermet samples that had
undergone different heat treatments as function of the excess
pressure p* (squares), and pressure dependence of the HEL
according to Eq. (5) calculated for different values of the
constant yield stress Yc.
The experimentally measured values of a HEL of
the cermets that had undergone different heat
treatments are shown in the same figure as
function of the calculated pressure surplus.
Accounting in the error margins of the
experimentally obtained figures and the
assumptions made (v = const, Yc = const), the
agreement seems reasonable.
The results obtained allow us to conclude that
the dynamic response of the ceramic-metal
composite is governed by the mechanical
interaction between the metallic matrix and the
ceramic particles and, thus, may be controlled by
choosing a proper thermal treatment of the cermet.
(3)
the stress a\ and a 2 have to be replaced by
<7i+/?* and 0"2+/?*, respectively. Accounting
in that for 1-D strain conditions 02 = <J\v l(l-v)
the Eq. (3) the Eq. (3) yields
(4)
The yield stress of the ceramic Yc is assumed to
be constant. Equation (4) yields the dependence of
& HEL on the excess pressure /?*:
REFERENCES
2(1 -2v) 2
1. Barker, L. M., and Hollenbach, R. E., J.Appl.Phys.,
45,4872, (1974).
2. Rosenberg, Z,, J, Appl. Phys. 74, 752, (1993)
(5)
^
2(1 -2v)2
1122