Change of shear modulus and yield stress with pressure and temperature

Yehuda Partom

Citation: AIP Conference Proceedings 1793, 110018 (2017); doi: 10.1063/1.4971681

View online: http://dx.doi.org/10.1063/1.4971681
View Table of Contents: http://aip.scitation.org/toc/apc/1793/1
Published by the American Institute of Physics

Articles you may be interested in

Evolutions of elastic-plastic shock compression waves in different materials
AIP Conference Proceedings 1793, 100030100030 (2017); 10.1063/1.4971655

Shock compression dynamics under a microscope

AIP Conference Proceedings 1793, 020001020001 (2017); 10.1063/1.4971456

A perspective on modeling the multiscale response of energetic materials

AIP Conference Proceedings 1793, 020003020003 (2017); 10.1063/1.4971458

Material response mechanisms are needed to obtain highly accurate experimental shock wave data
AIP Conference Proceedings 1793, 020002020002 (2017); 10.1063/1.4971457

Complete forms of Mie-Gruneisen equation of state

AIP Conference Proceedings 1793, 050001050001 (2017); 10.1063/1.4971535

The behavior of iron under ultrafast shock loading driven by a femtosecond laser
AIP Conference Proceedings 1793, 100035100035 (2017); 10.1063/1.4971660
 Change of Shear Modulus and Yield Stress with Pressure
and Temperature
Yehuda Partom1, a)
1
Retired from RAFAEL, ISRAEL.
a)
ypartom@netvision.net.il

Abstract. It is well known that the shear modulus (G) and the yield stress (Y) of metals increase with pressure (P) and
decrease with temperature (T). Steinberg, in his popular compendium of dynamic material properties, assumes for
Y/Y0(P,T)=G/G0(P,T) linear relations based on derivatives determined experimentally at ambient conditions. But recent
tests with high pressure dynamic loadings of certain metals obtained results that generally deviate from Steinberg’s
predictions. Here we use a different approach to estimate G/G0(P,T). As a first approximation we let G/G0=K/K0, where K
is the isentropic bulk modulus. With this assumption we compute the longitudinal sound speed of tantalum along its
principal Hugoniot and compare to recent measurements. There is a very slight disagreement, which we can correct by
assuming (second approximation) that Poisson ratio decreases slightly with pressure and increases slightly with
temperature. As K=Uc2 is always available in a hydrocode run from the equation of state, so are therefore also G/G 0 and
Y/Y0.

INTRODUCTION
Derivatives of shear modulus (G) with respect to pressure (P) and Temperature (T) at ambient conditions have
been determined for many materials. Steinberg [1] extrapolates these derivatives to high pressure and temperature by:
 13
G § U · 1 § wG · 1 § wG ·
1  A P¨¨ ¸¸  BT  T0  A ¨ ¸ B  ¨ ¸ (1)
G0 © U0 ¹ G0 © wP ¹ 0 G0 © wT ¹ RT
Steinberg, as well as others before him, assumes that:
Y G
P, T  P, T  (2)
Y0 G0
where Y is the yield stress.
In [2] and [3] they determine changes of yield stress and shear modulus of copper and tantalum at high pressure
dynamic loading, and the results generally deviate from the predictions of Steinberg [1].
One way to estimate G(P,T) is by assuming that Poisson ratio (Q) does not change, or changes only slightly, with
pressure and temperature. Assuming (as a first approximation) that Q(P,T)=Q0=Constant, we get:
G 1  2Q
3
2 Const . K Uc 2b (3)
K 1 Q
where K is the isentropic bulk modulus and cb the isentropic sound speed. Equation (3) leads to:
G G0 G K
; (4)
K K0 G0 K0
In any dynamic simulation with a hydrocode, U,cb and therefore K are known for any (x,t). Therefore K/K0, and using
equation (4) also G/G0, can be determined directly, without need to evaluate their separate dependence on P and T.

Shock Compression of Condensed Matter - 2015

AIP Conf. Proc. 1793, 110018-1–110018-5; doi: 10.1063/1.4971681
Published by AIP Publishing. 978-0-7354-1457-0/$30.00

110018-1
 Generally, it is known that Poisson ratio decreases slightly with pressure and increases slightly with temperature.
For our second approximation we therefore assume that:
Q Q 0  Q P P  Q T T  T0  (5)
where QP and QT are quite small. Equation (5) is a reasonable approximation only as long as the state (P,T) is away
from melting. When (P,T) is close to melting, we can use the relation:
Q 0  Q P P  Q T T  T0   Q m PT  T0 
2
Q (6)
where Qm(P) is determined from:
0.5  Q P P  Q T >Tm P   T0 @
Q m P  (7)
>Tm P   T0 @2
and where Tm(P) is the melting curve. Here we use equation (5), as we compare to test results only up to 50GPa, which
is far away from melting.
One way to determine QP and QT for a metal is through planar impact tests in which the longitudinal sound speed
cL is measured [4,5]. The longitudinal sound speed cL is given by:
c 2L c 2b  43 c s2 c s2 GU (8)
where the bulk sound speed cb is known from the equation of state, and where cs is the shear wave speed. Measuring
cL yields cs, and therefore also G.
In what follows we check (calibrate) our approach against data for c L along the principal Hugoniot of tantalum
measured in [5]. We do this up to 50GPa, because there is a break in the cL curve there.

G(P) FOR TANTALUM ALONG THE PRINCIPAL HUGONIOT

In Fig. 1 we show cL(P) data from [5] up to 50GPa and our curve fit to it. (At 50GPa there is a kink in the data
curve which the authors of [5] attribute to a phase change) We see from Fig. 1 that on this scale there is some scatter
in the data. From the definition of cL and from the shock equations it follows that the relation cL(P) along the Hugoniot
should be somewhat less than linear. We therefore fit the cL(P) data with a parabola that starts at cL(0)=4.14km/s
(cb(0)=C=3.41km/s, G0=69.0GPa, U0=16.69g/cc). The equation of this parabola is cL=4.14+0.0269P-0.000131P2
(GPa, km/s), and it is only valid for interpolation within the range 0 to 50 GPa.
For computing the Hugoniot curve we use the usual Uu straight line. For tantalum we use the parameters from [1]:
U c 0  Su c 0 3.4 km s ; S 1.2 (9)
where U is the shock velocity and u is the particle velocity.

FIGURE 1. Data of longitudinal sound speed in tantalum up FIGURE 2. Longitudinal sound speed in tantalum up to 50
to 50 GPa along the Hugoniot, from [5]. GPa. Points: curve fit to data from [5]. Full curve: calculated
with a constant Poisson ratio.

The Hugoniot curve in the PV plane is then:

110018-2
 H V
Ph U 0 c 02 H 1 V 1U U0 16.69 gr cc (10)
1  SH  2
V0
where U is density and V is specific volume. We use the usual Gruneisen equation of state referred to the Hugoniot
with constant U*, where * is the Gruneisen parameter. The equation of state is:
V0
EP, V  E h V   >P  Ph V@ Eh 1
2 Ph V0  V  (11)
*0
where E is the specific internal energy, and the index h refers to the Hugoniot curve. It is easy to verify that:
P  wE wV
c 2b V2 (12)
wE wP
and we do not detail here the expressions for the partial derivatives of equation (12) along the Hugoniot.
We first calculate cL(P) along the Hugoniot assuming that Q=Q0=Const.. We show the result in Fig. 2, and we
compare it there with the data curve fit given in Fig. 1. We see from Fig. 2 that even for a constant Poisson ratio the
agreement is quite good, especially in view of the scatter in the data. In Fig. 3 we show the result for G/G0=K/K0
obtained from the calculation of Fig. 2. We compare this result to Steinberg’s curve from equation (1), with the
coefficients given in [1]: A=0.0145/GPa, B=1.3e-4/K.

FIGURE 3. Full curve: G/G0=K/K0 along the Hugoniot from FIGURE 4. Longitudinal sound speed in tantalum along the
the calculation of Fig. 2. Broken curve: Steinberg’s curve [1]. Hugoniot up to 50 GPa. Curve for QP=-1.e-4/GPa, compared
with the curve with Q=Const..

We see from Fig. 3 that Steinberg’s curve is below our curve by about 25% in G/G 0. In Figs. 4 and 5 we show the
influence of the pressure dependence coefficient QP. From Fig. 4 we see that for QP<0 (which is the usual case), when
going up the Hugoniot Q decreases slightly and cL increases slightly. From Fig. 5 we see that going up the Hugoniot
G/G0 increases above K/K0.
Next we check the influence of having QT>0. To this end we need first to calculate the increase of temperature
along the Hugoniot. We do this by using the first TdS equation of thermodynamics together with the first law:
*0
TdS dE  PdV TdS C V dT  TdV (13)
V0
Equating and applying to the Hugoniot we get:
dT dP * dP dPh 1  SH V
CV  12 P  12 V0  V   0 U02c02 H 1 (14)
dV dV V0T dV dV 1  SH 2 V0
Integrating (with a standard ODE solver) and using for CV (constant volume heat capacity) 0.15J/gr/K, we get the
temperature increase shown in Fig. 6. In Fig. 7 we see that for QT>0 (which is the usual case), when going up the
Hugoniot Q increases slightly and cL decreases slightly. From Fig. 8 we see that going up the Hugoniot G/G0 decreases
below K/K0.

110018-3
 FIGURE 5. G/G0 and K/K0 for tantalum along the Hugoniot FIGURE 6. Temperature along the Hugoniot of tantalum up
up to 50 GPa for QP=-1.e-4/GPa. to 50 GPa.

FIGURE 7. Longitudinal sound speed in tantalum along the FIGURE 8. G/G0 and K/K0 for tantalum along the Hugoniot
Hugoniot up to 50 GPa. Curve for QT=5e-5/K, compared to the up to 50 GPa for QT=5e-5/K.
curve with Q=Const..

SUMMARY
It is well known that the shear modulus and yield stress of metals increase with pressure and decrease with
temperature. Steinberg [1] extrapolates pressure and temperature derivatives of the shear modulus at ambient
conditions to high pressure and temperature. But indirect determination of strength in high pressure dynamic
experiments shows deviations from such an extrapolation.
Here we use data from [5] for the longitudinal sound speed of tantalum along the Hugoniot to 50 GPa to calibrate
the change of the shear modulus (and thereby also of the yield stress) with pressure and temperature. As a first order
approximation we assume that Poisson ratio stays constant, which leads to G/G0=K/K0, where K is the isentropic bulk
modulus. With this assumption the data along the Hugoniot of tantalum are reproduces almost exactly. As a second
order approximation we let Poisson ratio decrease slightly with pressure and increase slightly with temperature, and
show the sensitivity to these changes.
One advantage of our approach is that the bulk modulus follows directly from the equation of state, and there is
no need to calculate pressure and temperature changes separately.

110018-4
 REFERENCES
1. D. J. Steinberg, Equation of State and Strength Properties of Selected Materials, LLNL Report, UCRL-MA-
106439 (1996).
2. H. S. Park et al. (17 authors), Experimental Results of Tantalum Material Strength at High Pressure and High
Strain Rate, in Shock Compression of Condensed Matter-2011, AIP Conf. Proc. 1426, 1371-1374 (2012).
3. S. D. Rothman, R. J. Edwards, T. J. Vogler and M. D. Furnish, Inferring the High Pressure Strength of Copper
by Measurement of Longitudinal Sound Speed in a Symmetric Impact and Release Experiment, in Shock
Compression of Condensed Matter-2011, AIP Conf. Proc. 1426, 104-107 (2012).
4. P. A. Rigg, R. J. Scharff and R. S. Hixon, Sound Speed Measurements in Tantalum using the Front Surface
Impact Technique, 18th APS-SCCM and 24th AIRAPT, J. of Phys. Conf. Series 500 (2014).
5. J. Hu et al. (9 authors), Sound Velocity Measurements of Tantalum under Shock Compression in the 10-110 GPa
Range, J. Appl. Phys. 111, 033511 (2012).

110018-5