MICROSTRUCTURAL EVOLUTION DURING

THERMOMECHANICAL. PROCESSING OF ALLOY 625

L. Ferrer, B. Pieraggi and J.F. Uginet *

Laboratoire des Materiaux, ENSCT, 31077 Toulouse, France

* AirForge, 09100 Pamiers, France

,Abstract

Wrought nickel - base 625 Alloy is wildely used because of its specific combination
of fabricability, mechanical properties and corrosion resistance. The primary goal of
the present work was to determine tlhe existing relationships between thermome-
chanical treatments, mechanical properties and corrosion behaviour of Alloy 625.
The precipitation diagram for MC, M6C and M23C6, and 8-Ni3Nb and y’ intermetal-
lit precipitated phases was determined. Dynamic and static recrystallisation kinetics
and grain growth kinetics were determined as a function of temperature and final
strain of thermoechanical treatments. These results permitted to evidence the influ-
ence of microstructure on the properties of wrought Alloy 625, and therefore to de-
fine the forging routes leading to a given set of mechanical properties and corrosion
resistance.

Superalloys 718, 625 and Various Derivatives

Edited by Edward A. L&a
The Minerals, Metals & Materials Society, 1991

217
 introduction

Alloy 625 is an austenitic wrought nickel-base alloy containing substantial amounts

of chromium (20 to 25 wt%), molybdenum (8 to 10 wt%), iron (5 wt %) and niobium
(3.5 to 4.5 wt%) as major addition elements. This superalloy is widely used because
of its specific combination of mechanical properties, fabricability and corrosion resis-
tance; it is one of the best corrosion resistant alloy in various media and industrial
environments. Alloy 625 is not usually considered as an age hardenable superalloy
because of its low content in Al and Ti. However, as a consequence of its content in
Fe and Nb (+Ta), precipitation of the metastable, ordered y’ phase (tetragonal DO,,)
and 6 phase (otthorhombic DO,) occurs during long term aging of solution treated
Alloy 625. Both v and 6 phases are based on the composition Ni3Nb, but with differ-
ing stacking sequence of atomically dense planes and differing atomic arrangement
of Ni and Nb atoms onto these planes. Therefore, the mechanical properties of Alloy
625 mainly results from solid solution hardening induced by MO and strain harden-
ing and thus it is not usually employed at temperature higher than about 650 “C.

Only few studies relative to precipitation in Alloy 625 as a function of time and
temperature are reported in the open literature (l-6) and no detailed studies of ther-
momechanical and recrystallization processes are available. In addition, these stud-
ies (2-5) mainly deal with the nature of precipitates formed after very long term heat
treatments at temperature between 600 and 800 “C and thus are not relevant to
usual and practical thermomechanical treatments.

Thus, the aim of the present work was the determination of the thermomechanical
processing permitting the control of mechanical properties and corrosion resistance
of Alloy 625 forged products for low temperature applications. This goal was partially
reached from the TTT precipitation diagram for carbides and intermetallics which
precipitates during heat-treatments and thermomechanical processing, and from the
recrystalllization processes and kinetics as a function of time, temperature, strain
and strain-rate.

Materials and Exnerimental Methods

The studied alloy was provided by lmphy SA in the form of a forged billet. The aver-
age composition is given in Table I.

Table I Composition of Alloy 625 used in this work

Element Ni Cr MO Fe Nb Co Al Ti Si Mn C

Weight % Bal 20.5 8.2 4.3 3.6 0.1 0.1 0.1 0.1 0.1 0.03
Atomic% Bal 23.4 5.1 4.6 2.3 0.1 0.2 0.1 0.2 0.1 0.15

218
The precipitation was studied on specimens solution treated at 1150 “C for 6 hours
and water quenched. The precipitation temperarure range was between 600-1075
“C and the heat treatment lasted from 5 mn up to 100 hours. The shortest heat treat-
ments (duration smaller than 2 hours) were performed in a fluidized-bed furnace
permitting high heating rate. All heat treatments were performed in air at atmospher-
ic pressure and all specimens were water-quenched after aging.

Thermomechanical treatments were performed on cylindrical bars solutionnized at

1150 “C for 2 hours. The deformation temperature varied between 1050 and 1175
“C and the final true strain E between 0.25 and 1.5. The ram speed was kept con-
stant and equal to 14 mms-1 , which roughly corresponds to a logarithmic variation
of E with time.
_-
4

(a) (b)
Figure 1 Simulation of the deformatilon of a cylindrical bar parallel to its long axis.
- a) Grid before deformation
- b) Same grid after deformation (&m = 0.63)

The cylindrical bars were deformed parallel to their longitudinal axis, which mini-
mizes the extent of dead zone in contact with the dies, offers a better reproducibility
and leads to a relatively large central zone of homogeneous deformation permitting
easy machining of specimens for mechanical testing. Figure 1 a-b shows an exam-
ple of simulated deformation map used to determine the extent of the central speci-
men zone homogeneously strained. Local deformations were computed from
Vulcain computer program (7). In sorne case, static recrystallization was studied di-
rectly after deformation at temperature between 900 and 1050 “C.

Metallographic examinations were performed on all heat-treated and thermome-

chanically processed specimens. Specific etching were used to reveal grain bound-
aries and intergranular carbide precipitates. The mean grain size was determined
from the Jeffries’ Intercept Procedure (ASTM El 12) and the percentage of recristal-
lized phase by planimetry. These examinations were correlated with TEM examina-
tions of thin foils and extraction replicas. Extraction replicas permitted to observe the
real shape and morphology of grain boundary carbides and also a semi-quantitative
determination of their composition by means of EDX. These analysis were per-
formed on a JEOL 200 CX transmiission electron microscope equipped with an
EDAX energy-dispersive analyser.

219
 Intergranular corrosion tests were performed according to Method A of ASTM G28
test procedure consisting in the measurement of weight loss resulting from the im-
mersion of a test piece of 25x20~3 mm3 in a boiling solution of concentrated sulfuric
acid and ferric sulfate. The measured weight loss after immersion for 120 hours is
converted in a uniform corrosion rate expressed in mm per year.
. . . .
tprt&on m Alloy 675

Precipitates formed after solution heat-treatment and annealing between 600 and
1075 “C were characterized by classical metallography, TEM examinations and
EDX analysis of thin foils and extraction replicas.

As usually observed in various austenitic stainless steel and Ni - Base superalloys

(8-lo), primary carbides (Ti,Nb)C and (Ti,Nb)(C,N) carbonitrides are present after
the solution treatment. TEM examinations of extraction replicas permitted to distin-
guish diamond-shaped and irregularly rounded carbides, but despite their high con-
tent inTi and Nb, these diamond-shaped carbides are secondary MC carbides
formed during annealing treatments.

lnteraranular Secondary Carbides.

Secondary intergranular carbides are formed during annealing treatments. The for-
mation of two different carbides, M23C6 and M6C, is observed as a function of the
annealing temperature. Because of their similar f.c.c cristallographic structure of very
close parameters (a = 10.6-l 0.9 A for M23C6 and a = 11 .1-l 1.3 A for M6C), the pre-
cipitation domain of these two intergranular carbides was determined by classical
metallographic examinations from the different response of grain boundaries to met-
allographic etching (Figures 2a and b). Indeed, strong preferential grain boundary
dissolution was observed for intergranular carbides formed at temperatures lower
than 950 “C while carbides precipitated between 950 and 1075 “C did not induce
such a preferential dissolution of grain boundaries.

(a) (b)
Figure 2 - Microstructure of heat treated Alloy 625 : Specimens annealed during 1
hour at : - a) 900 “C (M23C6 major carbide), - b) 1000 “C (M6C major carbide)

220
MC carbide precipitates were observed at temperatures between 900 and 950 “C
with a dendritic morphology or a diamond shape (Figures 3a and 3b). M3C carbides
were detected after precipitation treatment at temperatures between 1075 and 800
“C, while the precipitation of M23C3 (Figures 3c and 3d) occured at temperatures
between 950 and 700 “C. Table II reports some results of EDX analysis performed
on differing intergranular carbides ; MO is the main metallic constituant of MgC car-
bides while Cr and Nb are the major elements detected in M23C6 carbides.

Cd)
Figure 3 - TEM examination of intergranular carbides :
- a) Dendritic MC (900 OC, 1hr.) - b) Diamond-shaped MC (900°C, 1 hr.)
- c) M23C6 (800 OC, 100 hrs.) - d) M6C (1025 OC, 1 hr.).

221
 Table II EDX analysis of extracted carbides (concentrations in atomic percent)

Heat Treatment Ni Cr Nb MO Si Carbide

700 “C-100 hrs. 10.0 73.0 ---- 17.0 ---- M23C6

800 “C-100 hrs. 8.0 82.0 ---- 10.0 ---- M23C6
800 “C-100 hrs. 44.0 21.4 9.0 25.0 0.5 M6C
850 “C-15 min. ---- 60.0 33.0 7.0 ---- M23C6
900 “C-1 hr. ---- 6.0 75.0 19.0 ---- MC
940 “C-100 hrs. 10.0 6.0 75.0 4.0 ---- MC
1025 “C-1 hr. 26.0 15.0 13.5 40.0 5.5 M6C

From these observations, the precipitation diagram of carbides was determined as

shown by Figure 4. This precipitation diagram is in good agreement with the previ-
ous determinations (2-5).

MC
M6C
M23C6
gamma ”
delta

Figure 4 - Precipitation diagram of MC, M6C and M23C6 carbides, and &Ni3Nb and
y” intermetallic phases for the studied Alloy 625.

Intermetallic Precioitates.

In addition to intergranular carbides, precipitates of 6 Ni3Nb and y” phases were ob-

served after long term aging at temperatures between 650 and 820 “C (Figure 4).
The 6 phase mainly precipitates in the form of needle, near Nb rich zones often char-
acterized by an important amount of primary NbC carbide (Figure 5a). The 6-Ni3Nb
needles seem to growth from particles of NbC carbide partially dissolved and subse-
quent diffusion of Nb along the y/NbC interface (Figure 5b). The precipitation of y’ is
usually heterogeneous and occurs preferentially on dislocations (Figure 5~).

222
 (a> (b) @>
Figure 5 - Examinations of long term aged specimens of Alloy 625
- a) 6 precipitates formed after aging at 820 ‘33 for 200 hrs.
- b) 6 needles growing from a NbC precipitate.
- c) y’ precipitates formed after aging at 675 “C for 100 hrs.

The amount of precipitated y is small and insufficient to consider the studied Alloy
625 as an age-hardenable alloy as shown by the variation of the yield strength and
ultimate tensile stress as a function of temperature for aging times of 10 and 100
hours ( Figure 6a). The variation of elongation and area reduction (Figure 6b) is in-

Mpa-i
duced by intergranular carbide precipitation.

Q A%after1Oh
* Z%after1Oh
700 * A%after100h
+ Rp.21Oh

n
* Z%afterlOOh
+ Rm10h
Q Rp.21OOh
-+- Rm100h
540

I
300
500
I

700Temperature

(a)
1

(“C)goo 500
I
600
m I
700
- I ’
800
(W
Temperature
I -
900
(“C)

1000

Figure 6 - Variation of - a) Yield strength (Rp.2)and Ultimate tensile stress (Rm),

- b) Elongation (A%) and area reduction (Z%), of heat treated Alloy 625
as a function of aging temperature.

223
 Recrvstallization of Alloy 625

Dvnamic recrvstallization

Figure 7 shows the variation of the fraction of recrystallization as a function of the

true strain E and the deformation temperature. The recrystallization is approximately
complete at any true strain for temperatures exceeding 1175 “C. These sets of ex-
perimental data can be described by an Avrami equation of the form :

R%= lOO[ 1 -exp(K@)] (1)

where R% is the percentage of recristallized phase, K and n are constants deter-

mined by means of least-square fitting. The continuous curves were drawn from the
so-calculated K and n values. The variation of K and n with the temperature is
shown on Figure 8.

The value of strain exponent n of Eq. (1) is comprised between 1 and 2, and the
constant K between -6 and -2. The n value is in agreement with the ones usually re-
ported for a grain growth control of recrystallization process. This exponent would be
approximately equal to 4 in case of a nucleation control. However, these possible
theoritical values of exponent n associated to a given recrystallization mechanism
must not be considered as a real proof of the occurence of the corresponding mech-
anism because of the oversimplified hypothesis introduced in the calculations.

-2
n

Temperature
(“C)
I m I . I ' I '

1040 1060 1080 1100 1120 1140

Figure 7 - Variation of the percentage of Figure 8 - Variation of the coefficients K

recrystallization as a function of true and n of Avrami equation describing dy-
strain & and temperature. namic recrystallization of 625.

The variation of the average grain size of recristallized material has been deter-
mined as a function of deformation temperature and final deformation. The results
are reported on Figure 9 for two different final strains of 0.25 and 0.8. The observed
variation is approximately linear within the considered temperature range and for

224
the two reported deformations, which may be interpreted as a control of recrystal-
lization kinetics by the dynamic recrystallization itself.

60

40

20

Temperature
(“C)
0
1025 1075 1125 1175

Figure 9 - Variation of the average size! of Figure 10 - Variation of the percentage

recrystallized grains as a function of tem- of recrystallization R% as a function of
perature and final true strain. heat treatment temperature T.

Static recrvstallization

Static recrystallization has been studiled as a function of temperature as shown by

Figure IO for an initial deformation E : 0.4 at room temperature and a one hour an-
q

nealing. This figure show that, under these specific conditions, static recrystallization
of deformed specimen of Alloy 625 is observed between 900 and 980 “C. As shown
by the presence of some large grains, secondary recrystallization occurs at anneal-
ing temperature higher than 1100 “C. Static recrystallization and secondary carbide
precipitation interfer as they occur within the same temperature range and are inter-
correlated. However, it appears that static recrystallization precedes secondary car-
bides precipitation, which means that the intergranular carbides could not help to
control the grain size of Alloy 625 duriing thermomechanical processing.

Grain arowth

Grain growth was studied between 850 and 1100 “C from the evolution of grain size
of specimen annealed for 0.5 and 2 hrs. The initial microstructure of specimens was
equiaxed with a mean grain size of 15 - 20 pm ( ASTM index 9-l 0). Figure 11 shows
the variation of the mean grain size of annealed specimens as a function of anneal-
ing temperature. For the shorter annealing time a sharp increase of grain size oc-
curs at temperatures between 1050 and 1075 “C while the grain size increases over
a larger temperature range (940 - 1050 OC) for specimens annealed during 2 hrs.
Plotting log (d2-do2) as a function of l/T gives a straight line of same slope for the
two annealing times (Figure 12), which indicates that grain growth kinetics are
parabolic and thermally activated. The average apparent activation energy is found
equal to 225 kJ.mole-l.
225
 20.
I II II IOIO
II
Temperabe
(“C) 10000/T
(OK)
O- I
1200 7,2 7,4 7,6 7,8 8,0 8,2 8,4
Figure 11 - Variation of mean grain size Figure 12 - Variation of log(d2-d02) as
as a function of annealing temperature. a function of i/T(“K).

Mechanical Properties and corrosion Behaviour

The primary goal of the present work was to determine the existing relationships be-
tween thermomechanical treatments and mechanical properties and corrosion be-
haviour of Alloy 625.

The relation between microstructure and mechanical properties is well illustrated by

Figures 13 a-b showing the linear relationships existing between yield strength and
ultimate tensile strength, and the reciproqual square root of recrystallized grain size.
These two curves were drawn from data collected on specimens totally recrystal-
lized but submitted to different sequences of thermomechanical treatment.

YS0.2(MPa)

800

(a)

Figure 13 : Variation as a function of d-li2 of a) Yield strength and b) Ultimate

tensile stress

226
This d-li2 variation of yield strength and UTS shows the importance of grain size
control in thermomechanical processing of Alloy 625. This importance of grain size
means that intergranular carbide precipitation are of secondary importance in the
control of mechanical properties of forged and recrystallized Alloy 625. Indeed, at
approximately constant grain size, different thermomechanical treatments do not in-
duce significant variations of mechanical properties of completely recristallized
specimens, most of the observed variations are not significant .

Corrosion behaviour lead to completely differing observations, Figure 14 shows that

the variation of the uniform corrosion rate calculated from the weight loss measure-
ments, present a maximun for annealing treatments performed at temperatures be-
tween 700 and 850 “C. But, the results of G28 ASTM tests would also depend on the
specimen grain size. Indeed, very high value of corrosion rate would be observed as
soon as the depth of intergranular attack exceeds the grain size, condition which is
more easily fulfilled for fine-grained specimens. Furthermore, a small corrosion rate
may correspond to a very deep grain boundary corrosion, particularly in the case of
specimens with large grain size.

-g 30
$

E20
s
!Y!!
5 10
yjj
2
B
0
500 600 700 800 900 1000 1101
temperature (“C)
Figure 14 : Variation of corrosion rate deduced from ASTM G28 Test as a function of
annealing temperature after thermomechanical processing.

Qnclusions

Relative to the thermomechanical processing of Alloy 625, the present work evi-
denced the importance of the final grain on the mechanical propeties of recrystal-
lized alloy. This effect of final grain size is even more important in the case of the
studied alloy because of its low conlent in Al en Ti. Therefore, precipitation of y’
could not change or improve signific(antly the mechanical properties of Alloy 625.
Similarly, the small amount and the small size of intergranular carbides and 6 -
Ni3Nb precipitates do not permit to colntrol the final grain size of forged products.

227
 Furthermore, heat treatments in the temperature range 650 -900 “C has a very detri-
mental influence on the corrosion behaviour. Surprisingly, intergranular precipitates
formed after long term aging at temperature between 800 and 900 “C are not
M23C6 carbides, which means that the resistance of Alloy 625 to intergranular cor-
rosion is not only dependent on the precipitation of Cr - rich carbides. Other ele-
ments such as MO might have an important effect on corrosion behaviour.

The temperature of the final solution treatment is a very important parameter relative
to the mechanical properties and corrosion behaviour of Alloy 625, as shown by the
relationships existing between heat-treatment, specimen microstructure and corro-
sion behaviour, and the grain size dependence of mechanical properties.

References

1) G.P. Sabol and R. Sticker, “Microstructure of Ni - base superalloys”, Phys. Stat.

Sol., & 11, (1969), 11-52

2) F. Garzarolli, A. Gerscha, and K.P. Francke, “Untersuchungen iiber das

Ausscheidungsverhalten und die mechanischen Eigenschaften der Legierung
lnconel 625”, Z. Metallkde, m, 8, (1969), 643 - 652

3) H. Bohm, K. Ehrlich, and K.H. Kramer, “Das Ausscheidungsverhalten der

Nickellegierung lnconel 625”, Metall. und Technik, 24, 2, (1970), 139 - 144

4) E. Schnabel, H.J. Schtiller, and P. Schwaab, “The Precipitation and

Recrystallization Behaviour of Ni - Base - Alloy lnconel 625”, Practical
Metallography, (1971), 521 - 527

5) M. Schirra, “Einfluss des Vorbehandlungszustandes auf das Zeitstandfestigkeits

und Kriechverhalten einer matrixhtirtenden Ni - Basis - Legierung”, Metall. und
Technik, s, 4, (1982), 394 - 401

6) M. Sundararaman, P Mukhopadhyay, and S. Banerjee, “Precipitation of the 6-

Ni3Nb phase in two nickel base superalloys”, Met. Trans., m, (1988), 453 - 465

7) J.C. Boyer, F. Chevet, “Calcul par elements finis des ecoulements en forgeage “,
Mem. Scient. Rev. Met., (1988) 337 - 348

8) M. Dalhen, ‘Carbide precipitation in superalloys”, Superalloys 1984, ed. by M.

Gell et al., (Warrendale, PA : The Metallurgical Society, 1984), 449 - 454

9) L.k. Singhal, J.W. Martin, “The nucleation and growth of Widmannstatten M23C6
type carbide in an austenitic stainless steel”, Acta Metall., s, 9, (1968), 1159 - 1165

10) P Lacombe, B. Baroux, G. B&anger, “ Les Aciers lnoxydables “, Editions de

Physique, (1990)

228