Quarterly Progress Report for Period Ending June 30, 1995

HIGH TEMPERATURE MATERIALS TECHNOLOGY RESEARCH FOR ADVANCED THERMIONIC SYSTEMS

Submitted to

U.S. Department of Energy Office of Defense Energy Project

Ralph H. Zee Materials Engineering Program

. M. Frank Rose Space Power Institute

Tel: (334) 844-3320 Fax: (334) 844-3400

DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsi- bility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Refer- ence herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recom- mendation, or favoring by the United States Government or any agency thereof. The Views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

-~ ~ -

-.- . D~STRBUTION OF THIS DOCUMENT fs U K I M J ~ ~ ~ ~

HIGH TEMPERATURE MATERIALS TECHNOLOGY RESEARCH FOR ADVANCED THERMIONIC SYSTEMS

For the period 4/1/95 to 6/30/95

Our effort to understand the creep behavior in tungsten alloys prior to this

report period has been concentrated on the effects of strengthening particles such'as

HfC and ZrC. However, in most commercial alloys, a significant amount of Re (on

the order of several percent) is also added to enhance the low temperature formability

of the materials. During the past three months, the roles of rhenium (Re) on the

' mechanical response of .dispersion strengthened alloys were considered. Results from

this work will be the subject of this progress report and a manuscript is being

pripared based on this work for publication in Scripta Metallurgica.

Introduction

The effects of Re on the mechanical properties of tungsten and its alloys have

been examined by numerous authors [l-41. The addition of Re in tungsten not only

improves the low-temperature ductility of the alloys, but also increases the high

temperature creep 'strength relative to pure tungsten. The combined beneficial effect

of Re on tungsten makes Re an essential element for many particle strengthened

tungsten alloys such as W-Re-HfC, W-Re-ZrC and W-Re-Tho,. In these alloys, the

presence of small amounts of carbide .or oxide leads to dramatic increases in their high

temperature strength. The addition of Re in these alloys has been proven to enhance

1

the fabricability and has-exhibited no detrimental effect on their high temperature

mechanical properties [4]. Consequently, these materials have been proposed for use

'

as structural materials at temperatures above 1650 K. Recently, the creep of particle

strengthened tungsten and tungsten-rhenium alloys has been systematically studied at

temperatures between 1900 K and 2500 K [5-7]. Results from these investigations

provide the database needed to isolate.the effects of solid solution Re on the creep of

particle strengthened tungsten alloys.

Literature Review

The effects of solid solution Re on the creep behavior of particle strengthened

tungsten alloys were reported by Chen [5]. In his work, the steady state creep rates

of W-Re-lThO, alloys were measured between 2000 K and 2500 K with Re contents

ranging from 5 d o to 25 d o ( d o represents atomic percent). Resuiis show that the

stress exponent (n) of the power law creep of the alloy changes from 4.6 to about 3

when the Re content is increased from 5 a/o to 25 a/o at temperatures between 2000 K

and-2200 K. This change in the creep behavior is similar to the case of W-Re solid

solution alloys where the addition of 25 d o Re yields a stress exponent of 3.8 where& /

5 a/o of Re in tungsten only results in a reduction of creep rate while maintaining an n

value of 5 (that of pure tungsten). Thus the effect of Re in thoria strengthened

tungsten alloys can be considered to be similar to that of the simple binary solid

solution tungsten-rhenium alloys. According to previous study of Gao and Zee, the

2

strengthening effect of Re can be described by a shift factor that is a function of

solution atom type and concentration [8]. In solid solution tungsten binary alloys, the

shift factor can be expressed as e~p[-145000neC'~lT] where n is the stress exponent of

creep, e is the solute-solvent atomic size mismatch and C is the solute concentration.

Using this shift factor, the creep rate of the alloy can be determined by multiplying the

creep rate of the pure tungsten with the shift factor (less than unity). This concept has

been proven to give good predictions on the creep rates of several solid solution

tungsten and molybdenum alloys that exhibited the Class 11 creep behavior (creep

exponent n close to 5 indicating that creep is controlled by dislocation climb). In this

study, the concept of shift factor is extended to the particle strengthened tungsten alloy

case so that the strengthening effect of the Re in this class of alloys can also

estimated.

Modeling Approach

The effects of Re on the creep of particle strengthened tungsten alloys were

.- c,. obtained through a comparison of the creep data in alloys both with and without Re.

Figures 1 and 2 show the creep data of recrystallized W-4Re-0.26HfC alloy [5] and

W-0.37HfC [7]. Prior to creep testing, the W-4Re-0.26HfC samples [9] were given a

recrystallization anneal at 2100 K for one hour to obtain grain sizes of between 30.and

40 pm. To ensure thermal equilibrium, the samples were kept at least two hours at

the test temperature before the introduction of creep load. In the latter study by Gao,

3

the W-0.37HfC samples creep tested at 24vo K anc 2500 K were given a

recrystallization anneal of 2673 K for one hour yielding a grain size of between 50

and 60 pm. For the specimens that were crept at 2000 K to 2300 K, a one-hour

anneal at 2773 K was conducted before the test. Different HfC volume fractions and

different recrystallization temperatures result in different degrees of strengthening

from the HfC carbide addition. The carbide strekgthening effects in this class of

materials have been found to come from two basic contributions. One is called the

direct strengthening where the HfC carbide parkcles pin the moving dislocations

during the creep process. The second is referred to as indirect strengthening where

the worked structure stabilized by the fine dispersed HfC particles provides further

reduction in creep deformation. Using the modified Lagneborg's creep model for the

dispersion strengthened alloy, the creep of W-4Re-O.26HfC, W-4Re-0.33HfC, W-

3.6Re-0.33w/oZrC and W-0.37HfC were expressed in the form [lo, 71

where ,5 is the steady state creep,of a material. The constant A depends on the matrix

properties and the test temperature, CT is the applied stress and KGb/L is the threshold

stress due to particle-dislocation direct interaction during creep. The parameter G is

the shear modulus of the matrix which depends on temperature [l 11, b the Burgers

vector of the material (0.274 nm), and L the interparticle spacing of the carbides.

When the HfC particle volume fraction in an alloy is a constant, the interparticle

spacing L is proportional to the size of the HfC particle. Furthermore, materials with

. higher HfC particle volume fractions possess smaller L values. Applying Equation 1

to all the materials mentioned above, the direct strengthening due to HfC or ZrC can

be expressed explicitly and can be isolated. The A values for all of the materials were

obtained by plotting the applied stress CT versus the one-fourth power of the

corresponding creep rate (S1'"> as shown in Figures 3 to 6 . The x-intercept of each

line is the threshold stress due to particle-dislocation interaction for that particular test

temperature. The slope of the line is the corresponding value of A"4. Thus the A

value of each material at each creep temperature can be determined based on the

experimental data. The constant A can be further expressed as [7]

A = A'D,exp(-Q,/RT)b/G3kT (2)

where A' is a material constant which is independent of temperature, Do and Qc are

the pre-exponential coefficient and the activation energy for the process that controls

creep, such as lattice diffusion or dislocation core diffusion. The parameter R is the

gas constant, k is the Boltzmann constant and T is the absolute temperature. Taking

the logarithm of the Equation 2 and after some rearrangements yield:

l0g(AG3kT/b) = log(A'D0) - (Qc/2.303R)(1/T)

5

(3)

The values of log(AG3kT/b) of each material are plotted against the 1/T in Figure 7.

In this plot, the temperature dependent G values were used [l 13. According to

Equation 3, a straight line can be drawn for each material. The activation energy Qc

for creep of each material can be determined from the slope of the line. The A value

of each material can be directly compared with one another at the same temperature

since G, b and k of all the materials can be assumed to be similar.

Figure 7 provides a comparison of the log(AG3kT/b) values for the materials

with and without Re. For the recrystallized materials, we found that the alloys with

Re all possess lower A values than those without Re. Furthermore, under the same

temperature range, the recrystallized W-4Re-O.26HfC and W-0.37HfC both have an

activation energy that is close to that for the lattice difision in pure tungsten. Thus

the two nearly parallel lines provide a foundation for the comparison of the Re effect.

The reasons for the reduction of the A value of the materials were thus analyzed in

light of the fact that the parameter A may also be influenced by other material factors.

According to Chen [5] and Tsao et al. [9], W-4Re-0.26HfC become fully

recrystallized after annealing at 2100 K for one hour. ..The creep tests of the W-4Re-

0.26HfC from 1955 K to 2500 K were all conducted after such a recrystallization

process. The initial particle radius was about 20 nm. For W-0.37HfC;the creep tests

were conducted after the material has been recrystallized from 2673 K to 2773 K for

one hour. The starting average particle sizes were estimated at between 60 nm and 77 I

nm according to the prediction of Ozaki [12]:

6

r(in meter) = [2.8~1O~'~(t/T)exp(-472340/RT) +r;]ln

where r is average HfC particle radius obtained after exposure at temperature T for t

seconds and r, is the initial HfC particle radius. Equation 4 was obtained using the

average of three particle coarsening equations obtained from three independent studies

on the HfC particle coarsening observation using transmission electron microscopy

[13,5,12]. The larger particle size resulted from the higher temperature anneal is not

the governing reason .for [he higher A values observed in W-0.37HfC. As shown in

Figure 7, the log(AG3kT/b) value for the as-extruded W-0.37HfC tested at a high

temperature of 2500 K is larger than the corresponding value for the recrystallized

material due to the dynamic recrystallization of the as-extruded alloy during creep

testing. This value is even larger than the value of log(AG3kT/b) for the W-4Re-

0.26HfC tested in the recrystallized condition. According to Ozaki [12], the average

particle radius in the as-extruded W-0.37HfC alloy used in this study was 20 nm.

However, this smaller particle size did not lead to smaller A value compared to the

pre-recrystallized W-4Re-0.26HfC samples wherit was tested at 2500 K. In fact, the

A value for the as-extruded W-0.37HfC at 2500 K with a worked structure and a high

particle volume fraction is over 3 times as large as that for the pre-recrystallized W-

4Re-0.26HfC crept under the same temperature. It can therefore be concluded that

particle size is not directly responsible for the reduction in A observed.

The A value was further found to be relatively independent of grain size as long

7

as the full recrystallization is achieved [7]. The .final grain size for the as-exkuded

W-0.37HfC specimen crept at 2500 K is between 50-60 pm due to dynamic

recrystallization [7] while the grain size of the recrystallized W-4Re-0.26HfC

specimen is 30-40 pm. After one hour annealing at 2500 K, the latter grain size was

found to increase to 100 pm [9]. If the parameter A were affected by the grain size

(d) in the same manner as in pure tungsten (k a do."), the A value for the W-4Re-

0.26HfC would only be 30% larger than that for the W-0.37HfC [14]. Therefore the

main difference between the two materials is directly due to the presence of 4 a/o Re

in the W-4Re-0.26HfC.

In order to estimate the solid solution strengthening effect of Re, the concept of

shift factor was used [8]. One of the basic assumptions of this shift factor is that

solute atoms provide an extra energy barrier for the dislocations to'climb during

creep. Since creep is a thermally activated process, the shift factor has the form of

exp([-f(e,C)/@T)]. The energy barrier f(e,C) is a function of solute-solvent atomic

size mismatch, e (e= I rsolUk-rsolvent I / rsolvent, where r is the atomic radius) and the solute

mole fraction C. The shift factor-for the solid solution tungsten is exp(-

145000neCVI'), where n is the stress exponent of the creep equation of pure

tungsten. Since the creep of W-(4Re)-HfC alloys is controlled by the climb of

dislocations over the HfC particles, the shift factor concept is still valid. Following

the idea that dilute solid solution only shift creep rate to lower values, the

corresponding shift factor can be written as:

8

S,,,, = e~p(-580000eC’~/T)

Thus for a 4 d o Re matrix strengthened with HfC particles with an e value of 0.029

for Re [l5] and C of 0.04, the shift factor was found to depend on temperature

through the term exp(-3364/T). Dividing the value of log(AG3kT/b) obtained for. W-

4Re-0.26HfC at each temperature by exp(-3364/T), the solid solution effect of Re can

be isolatedl and accounted for. The resultant line for the recrystallized W-4Re-

0.26HK1 alloy overlaps the line of the W-0.37HfC where no Re is present in the

matrix, as shown in Figure 8. .From the shift factor due to 4 a/o Re, the extra energy

barrier is estimated at 28 kJ/mole. On the other hand, an activation energy of 585.6

W/mole was obtained for the W-4Re-0.26 HfC alloy after the Re effect has been

separated. This activation energy is very close to the activation energy for creep

observed in W-0.37HfC (575.2 kJ/mole). The activation energy values for the W-

0.37HfC and W-4Re-0.26HfC obtained after eliminating the Re effect are in good

agreement with the reported lattice diffusion activation energy of 585.2 kJ/mole [16].

Similar results were also obtained in all the creep data of W-4Re-0.33HfC within the

same temperature range (Figure 8). This material has been recrystallized for 1.5

hours at 2438 K [17]. The solid solution strengthening effect of Re in the W-4Re-HfC

alloys was successfully represented by the shift factor, exp(-145000neC1”/T) while

maintaining an n value of 4.

Since the creep of both W-0.37HfC and W-4Re-0.26HfC was found to be

9

controlled by lattice diffusion, the parameter Do in the Equation 3 will be the same

and possessing the value of 5 . 6 ~ 1 0 ~ m2/s [16]. The resultant A' values for both

materials were found to be 2 . 4 ~ 1 0 ' ~ for W-0.37HfC and 4.4~10" for W-4Re-

0.26HfC. The average of the value is 3 . 4 ~ 1 0 ' ~ which could be used as a universal .

value for the recrystallized W-(4Re)-HfC alloys. Therefore the creep W-Re-HfC alloy

can be written as:

i: = 3.4x10'0(b/G3kT)DLexp(-580000eC"Z/T)(a-)4 (1955 KITS2500 K) (6)

where e and C are the atomic mismatch and mole fraction of Re respectively. The

parameters G and CT are both in unit of MPa. This result expresses explicitly the solid

solution strengthening effect of Re in the creep equation for W-4Re-O.26HfC.

Conclusion

The concept of shift factor was used successfully to develop a model which

described thcrole of solid solution atoms in dispersion strengthened tungsten alloys.

This shift factor separates the solid solution strengthening effect of Re in the creep of

W-Re-HfC materials. The creep of the alloys is expressed by the modified

Lagneborg's creep model in the following form:

i: = A'(b/G3kT)DLexp(-580000eC'R/T)(~-up)4 (1955 KSTS2500 IS) (7)

10

where DL is the coefficient of lattice diffusion of tungsten.

Reference$

1.

2.

3.

4.

5.

6.

7.

8.

9.

Raffo P.L., W.D. Klopp and W.R. Witzke, "Mechanical Properties of Arc- Melted and Electron-Beam-Melted Tungsten-Base Alloys" , NASA TN D-256 1 , Lewis Research Center, Cleveland, Ohio (1965).

Raffo F.L. and W.D. Klopp, "Mechanical Properties of Solid Solution and Carbide-Strengthened Arc-Melted Tungsten Alloys", NASA TN D-3248, Lewis Research Center, Cleveland, Ohio, (1 966).

Vandervoort R.R., "The Creep Behavior of W-SRe", Metall Trans. l(1970) 857- 864.

Vandervoort R.R. and W. L. Barmore, "Elevated Temperature Deformation and Electron Microscope Study of Polycrystalline Tungsten and Tungsten-Rhegum Alloys, I' Proceedings of 6th Plansee Seminar, Metallwerk Plansee Ag. , Reutte- Tyrol, (1969) 108-137.

Chen B.L., "High-Temperature Creep Behavior of Second-Phase Particle- Strengthened Tungsten Rhenium Alloys, I' Dissertation,. Arizona State University, 1990.

Luo A., D.L. Jacobson, and Kwang S . Shin, "Particle Strengthened Tungsten for Space Power Applications, " Proceedings 26th Intersociety Energy Conversion Engineering Conference, Boston, Massachusetts, August 4-9, (1991), 142-147.

Gao.H, "A Study of Creep Behavior in Tungsten Alloys for Thermionic' Emitter, " PhD Dissertation, Auburn University, 1995.

.

Gao H. and R.H. Zee, "A Semi-Mechanistic Creep Model for Tungsten and Molybdenum based Solid Solution Alloys," Scripta Metallurgica et Materialia 32 (10) (1995) 1665-1670.

Tsao B.H., M.L. Ramalingam, D. Tang, B.L. Chen, and D.L. Jacobson, T h e Effect of Tho, and HfC on the Recrystallization of W-Re Alloys," Tungsten and Tungsten Alloys, Recent Advances, Andrew Crowson and Edward S. Chen Ed. 1991.

11

10. Lagneborg R., "Bypassing of Dislocation Past Particles by a Climb Mechanism", Scripta Metall. 7 (1973) 605-613.

11. Lowrie R. and A.M. Gonas-, J. Appl. Phys., 36 (1965) 2189-2192. I

12. Ozaki Y., "Thermal Properties of Refractory Metals for Advanced Energy Conversion Systems, " PhD Dissertation, Auburn University, 1994.

13. Klopp W.D. and W.R. Witzke, "Mechanical Properties of Arc-Melted Tungsten- Rhenium-Hafnium-Carbon Alloys, " NASA TN D-5348, Lewis Research Center, Cleveland, Ohio, (1969).

14. Klopp W.D., W.R. Witzke and P.L. Raffo, "Effects of Grain Size on tensile .

creep Properties of Arc-Melted and Electron-Beam-Melted Tungsten at 2250°F to 4140"F," Trans. Met. SOC. AIME, 233 (1965) 1860-1866.

15. Laves F., "Crystal Structure and Atomic Size," Theory of Alloy Phases, American Society of Metals, (1956) 124-193.

16. Robinson S.L. and O.D. Sherby, "Mechanical Behavior of Polycrystalline Tungsten at Elevated Temperature," Acta Metall. 17 (1969), 109-125.

17. Chen B.L., A. Luo, K.S. Shin and D.L. Jacobson, I' High-Temperature Mechanical Properties of W-Re-HfC Alloys, I' Refractory Metals: State-of-the-Art 1988, P. Kumqr and R.L. Ammom Ed. TMS 1989.

12

' W I- Q I- v,

I W-.4.ORe-O.Z6HfC - - -

>- Q w I- in

n

10-4

I o - ~

I o-6

I o-7

1 0-8

I o-'

0 2500K Chen (1990)

2400K p n=5.2-5.4

0

A A

2300K 2200K 2190K 2100K 2070K 1955K

IO

STRESS (MPa) ---

I00

Figure 1. Steady state creep of recrystallized W-4Re-O.26HfC.

13

IO-^

IO-^

lo-6

IO-'

IO-^

10-'O

V 2500 K . 1 .

10

Stress (MPa)

100

Figure 2. Steady State creep of Recrystallized W-0.37 HE.

14

0.1 4

0.1 2

0.10 \ n

v1 0.08 \ - 0.06 4

I 1 I 1 I I I / I I !/ I 0 2500 K 0 2400 K w-4.

t v 2300 K / /

0.02

0.00 50 I 0 0

Stress (MPa) . . ,.

Figure 3. Modeling of creep of W-4Re-0.26HfC using the modified Lagneborg’s

approach.

15

0.14

0.12

0.10

0.08

0.06

0.04

0.02

0.00

I I I I 1 I 1 1 1 I 1 0 2400

2200 K 2100 K 2000 K 1900 K

W-3.6Re- 0133ZrC

0 50 100

Stress (MPa)

Figure 4. Same as figure 3 but for W-3.6Re-O.33ZrC.

16

0.14 - I I I I

W-0.37HfC 1 I I I I I 1

As received - -

0.10 o-12 1 O 0.08 1 0.06

0.04 1 0.02. .

2300 K 1

1 I 1 I I I

u.uu - ' ' 0 20 40 60 80 I00 120

Stress (MPa)

Figure 5. ' Same as figure 3 but for as-received W-O.37HfC.

17

- . . . .

0.1 4

0.1 2

0.1 0

0.08

0.06

0.04

0.02

0.00

I 1 - W-O.d7HfC I 1 1 1 I

-

0 20 40 60 80 1.00

Stress (MPa)

Figure 6 . Same as figure 3 but for recrystallized W-0.37HfC.

18

M 0 - IO-’

1 0-l1

1 0-l2 3.5 4.0 4.5 5.0 . 5.5

Figure 7. Comparison of the parameter A for different materials.

19

rl u 0 0 0

u3 g I

X Q)

P \

1 o-8

1 o-’

1 0-l2 4.0 4.5 5.0 5.5

1 O ~ / T ( 1 o ~ / K )

-

3.5

Figure 8. Same as figure 7 .but normalized for the Re effect using the shift factor

approach.

20