Effect of Fcc-Hcp Phase Transformation Produced by

Isothermal Aging on the Corrosion Resistance of a

Co-27Cr-5Mo-0.05C Alloy
C. MONTERO-OCAMPO, R. JUAREZ, and A. SALINAS RODRIGUEZ

The corrosion resistance of two-phase (fcc-hcp) Co-27Cr-5Mo-0.05C alloys produced by isothermal

aging at 800 ⬚C was studied using potentiostatic polarization tests in Ringer’s solution. Critical pitting
potentials were estimated from the potentiostatic polarization curves and were found superior to that
exhibited by the conventional ASTM-F75 cast alloy used for the manufacture of orthopedic implants.
Formation of suitable distributions of hcp embryos (incoherent twin boundaries and stacking faults)
prior to and during the early stages of aging required for isothermal fcc-hcp transformation led to a
relative reduction of the corrosion resistance of two-phase alloys. However, once the transformation
proceeded rapidly, between 4 and 8 hours of aging, the elimination of lattice defects caused a reduction
of the dissolution rates and the breakdown potential became nearly independent of the relative amounts
of fcc and hcp phases present in the microstructure. This behavior was due to the uniform chemical
composition of the two-phase alloys. Concurrent work has shown that the hardness and yield strength
of a 50 pct hcp alloy are increased by at least 30 pct without undue ductility losses. Therefore, the
results of the present article suggest that these materials are excellent candidates for the manufacture
of orthopedic implant devices requiring higher strength than provided by conventional ASTM-F75
materials.

I. INTRODUCTION laboratory have shown that the metastable fcc phase pro-
A MAJOR concern after total hip joint arthroplasty is duced by water quenching from 1150 ⬚C can fully transform
the formation of polyethylene wear debris at the metal- to hcp during isothermal aging at 800 ⬚C.[7] Vander Sande
polyethylene interface of the prosthesis.[1] Co-27Cr-5Mo- et al.[8] showed that aging a wrought Co-26.7Cr-5.5Mo-
0.05C wrought alloy with a low percentage of hard carbide 0.15C alloy at 750 ⬚C promotes a martensitic transformation
particles seems to be an attractive alternative to the cast from fcc to a heavily faulted hcp phase. They found that
ASTM F75 alloy to reduce the formation of such wear the formation of densely packed fcc stacking faults was
debris.[2] Additional abrasive wear reductions may be followed by a very fine carbide precipitation at intrinsic
obtained if the crystal structure of the contact surface of stacking faults. Taylor and Waterhouse[9] found that aging
the metallic implant device is modified to hcp. This Co between 650 ⬚C and 1000 ⬚C a previously solutionized Co
phase exhibits lower coefficients of friction than the fcc alloy resulted in preferential precipitation of M23C6 carbides
phase normally found in conventional cast and wrought at intragranular striations identified as hcp platelets.
alloys.[3] Unfortunately, this structural change is difficult Rajan[10,11] found that the isothermal aging at 750 ⬚C of
to achieve. the solution-treated wrought Co-26.7Cr-5.5Mo-0.15C alloy
Pure Co undergoes a fcc → hcp allotropic transformation induced the fcc metastable phase transformation to the hcp
upon cooling below 417 ⬚C by a martensitic mechanism. phase by a two-stage process. The first was a martensitic
Alloying additions of Cr and Mo (hcp stabilizers) increase nucleation mechanism and the second (hcp2) a recrystalliza-
this temperature.[4] Cong et al.[5] have detected hcp phase tion-type phenomenon, which was associated with the for-
after isothermal aging of a Co-27Cr-5.5Mo-0.25C alloy mation of carbides and hence diffusion. Cong et al.[5]
during 48 hours at 1100 ⬚C, while Weeton and Signorelli attributed the marteniste nucleation kinetics in Co-27Cr-
suggested that the transition temperature is in the range 5.5Mo-0.25C alloy powders to the thermally activated
965 to 1230 ⬚C.[6] This transformation is often incomplete motion of Shockely partial dislocations and their interaction
because of the very low stacking fault energy of Co and with short-range obstacles, e.g., stacking fault intersections
its alloys. This causes the formation of stacking faults and and solute clusters. On the other hand, aging at elevated
twins in the cubic high-temperature phase, which remains temperatures where the stability of the fcc matrix is signifi-
in metastable form at room temperature. These features of cantly improved reduces the formation of hcp intragranular
the microstructure are considered precursors to hcp forma- striations. At these temperatures, the developed striations
tion. The lower the fault energy, the higher are the fault are relatively coarse and discontinuous. This, in turn, affects
and twin densities. Experimental results obtained in our the active carbide precipitation reactions.
In general, Co-Cr-Mo-C alloys exhibit excellent resist-
ance to uniform corrosion under environmental conditions
C. MONTERO-OCAMPO, Research Professor, R. JUAREZ, Student, and simulating those found in the human body. However, their
A. SALINAS RODRIGUEZ, Research Professor, are with CINVESTAV-
IPN, Unidad Saltillo, 25000 Saltillo, Coahuila, México. Contact e-mail:
preferential dissolution is initiated at microstructural heter-
cmontero@saltillo.cinvestav.mx ogeneities.[12] In particular, the differences in volume frac-
Manuscript submitted April 6, 2001. tion and chemical composition of alloy carbides are of

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 33A, JULY 2002—2229

relevant importance; this has been observed for the ASTM-
F75 as-cast Co-27Cr-5Mo-0.25C alloy,[12] where the effect
of carbon content on the localized alloy corrosion has been
to produce local differences on the rates of dissolution
leading to preferential attack at grain boundaries and inter-
dendritic regions where the carbides precipitate. It was
found that low C content ASTM-F75 as-cast alloys resulted
in a lower release rate of corrosion products.
While the fcc → hcp transformation has been considered
important to reduce the abrasive wear and improve the
mechanical properties of cobalt-based alloys, its role on
the corrosion resistance has never been stated explicitly.
The purpose of this article is to present the results of an
experimental evaluation of the effect of the fcc → hcp
transformation formed by isothermal aging on the corrosion
resistance of Co-27Cr-5Mo-0.05C alloy. The resistance to
localized attack of the alloy is estimated from anodic poten-
tiostatic polarization tests. In this technique, the alloy is
taken through a potential cycle to determine the breakdown
potential and the passive current density.

II. EXPERIMENTAL

Disks of approximately 4-mm thickness were cut from Fig. 1—Effect of isothermal aging time at 800 ⬚C on the X-ray diffraction
a 15.8-mm-diameter wrought Co-27Cr-5Mo-0.05C alloy patterns of Co-27Cr-5Mo-0.05C alloys homogenized at 1150 ⬚C for 1 h
cylindrical rod. They were homogenized at 1150 ⬚C for 1 and quenched into water.
hour and quenched into water. Following this, the alloy
was given isothermal aging treatments at 800 ⬚C for various
times and then water quenched. A tubular furnace provided III. RESULTS AND DISCUSSION
with inert argon gas atmosphere was used in both treat-
ments. The microstructure evolution produced by the fcc → A. Evolution of the Microstructure during Isothermal
Aging
hcp transformation was studied by scanning electron
microscopy, reflected light microscopy, and X-ray diffrac- Figure 1 shows the X-ray patterns obtained after aging
tion. Samples for microscopy were etched by immersion the solution-treated and quenched material. The aging times
in a solution of 90 pct HCl ⫹ 10 pct H2O2 solution. X-ray and corresponding amounts of hcp phase, calculated from
diffraction patterns were measured from 40 to 55 deg of 2␪ the integrated intensities of the (200)fcc and (101̄1)hcp peaks
value using Cu K␣ radiation (␭ ⫽ 1.54 Å). The percentage of using the relationship proposed by Stage and Guillaud,[13]
fcc-to-hcp phase transformation was calculated from the are indicated in the patterns. As can be seen in Figure
integrated areas of the (200)fcc and (101̄1)hcp X-ray diffrac- 1, as aging proceeds, the relative intensities of the X-ray
tion peaks.[13] The sample disks were used for electrode diffraction peaks corresponding to the hcp phase increase,
preparation. The electrodes were mechanically polished while those associated with the fcc phase decrease indicat-
using silicon carbide grinding paper and alumina slurry and ing that a phase transformation has taken place. The effect
then cleaned in alcohol and dried. The corrosion tests were of aging time on the amount of transformed hcp phase is
performed in deaerated Ringer solution at 37 ⬚C. The elec- illustrated in Figure 2. As can be seen, there is good agree-
ment between the amount of hcp estimated from the diffrac-
trolyte was prepared by dissolving 8.6 g NaCl, 0.30 g KCl,
tion patterns and that obtained from quantitative
and 0.33 g CaCl2 in 1000 mL of distilled water. All the
measurements using image analysis on metallographic sam-
electrode potentials were referred to the standard saturated ples. This graph shows that the transformation follows a
calomel electrode. A platinum mesh was used as the auxil- sigmoidal kinetics where the reaction rates for aging times
iary electrode. shorter than 4 hours and larger than 10 hours are very
The rest potential was measured as a function of time slow. In contrast, the transformation rate is very rapid for
from the initial value exhibited after immersion of the elec- intermediate aging times. This type of kinetic behavior
trode into the cell until a steady state was achieved; this depends strongly on the aging temperature[15] and also on
last value was reported as the open-circuit potential (OCP). the solution temperature and time.[16]
Potentiostatic curves were plotted based on stabilization Figure 3 illustrates a series of optical micrographs of the
of the current density measured as a function of time at typical microstructures of the solution-treated alloy (Figure
preselected values of applied potential in the range from 3(a)) and of samples isothermally aged during less than 4
the OCP up to 0.78 V. The corrosion test results of the hours (Figure 3(b)), between 4 and 8 hours (Figure 3(c))
investigated alloys were compared to those obtained in a and more than 10 hours (Figure 3(d)). As can be seen,
typical ASTM F75 as-cast surgical alloy, 0.30 pct C. as the aging time increases, the heavily faulted, equiaxed

2230—VOLUME 33A, JULY 2002 METALLURGICAL AND MATERIALS TRANSACTIONS A

microstructure of the solution-treated material is replaced lamellae with different orientations within a given prior fcc
by a pearlitic-type microstructure with colonies of very fine grain. Figure 4 shows examples of the typical morphology
of the hcp lamellae formed at the early stages and at the
end of the transformation. It is noteworthy that the new
hcp phase nucleates at intersections of fcc twins, stacking
faults, and grain boundaries. This observation is consistent
with Olson and Cohen’s model for hcp martensite nucle-
ation[17] and the observations of Rajan[10] concerning the
formation of hcp embryos at structural defects, such as
stacking faults and incoherent twin boundaries in Co-Cr-
Mo-C alloys. It is noteworthy that energy dispersive X-ray
spectrum (EDXS) microarea characterization of regions
with colonies of hcp platelets, about 1 by 1 ␮m in size,
did no show significant differences with similar spectra
obtained in regions free from hcp platelets in the same
sample or in samples characterized prior to aging.
Previous work[16] has shown that the defect structure
(twins and stacking faults) formed on quenching after the
Fig. 2—Hcp transformation-time curves for the isothermal aging at 800
⬚C of Co-27Cr-5Mo-0.05C alloy homogenized at 1150 ⬚C for 1 h and
solution heat treatment affects strongly the kinetics of the
quenched into water. Data obtained by X-ray diffraction and quantitative transformation during aging. If the metastable fcc phase
metallography. transforms to hcp as a result of quenching, then the kinetics

Fig. 3—Typical microstructures (a) of Co-27Cr-5Mo-0.05C alloy heat treated at 1150 ⬚C for 1 h and water quenched and of samples isothermally aged
during (b) less than 4 h, (c) 7 h, and more than (d ) 10 h at 800 ⬚C.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 33A, JULY 2002—2231

 (a)

Fig. 4—SEM photomicrographs illustrating the typical morphology of the

hcp phase formed in homogenized at 1150 ⬚C for 1 h and water-quenched
Co-27Cr-5Mo-0.05C alloy during isothermal aging at 800 ⬚C (a) at the
early stages (⬍4 hours) and (b) at the end (⬎10 h) of the hcp transformation.

of the isothermal transformation during aging is severely (b)

slowed down. In the present work, the solution temperature Fig. 5—Effect of applied potentials (a) 0.650, 0.670, and 0.700 V and (b)
and time were selected so as to inhibit as much as possible 0.740, 0.760, and 0.780 V on the dissolution of the homogenized at 1150 ⬚C
the athermal formation of the hcp phase. This is why the for 1 h and water-quenched Co-27Cr-5Mo-0.05C alloy in Ringer’s solution.
X-ray diffraction pattern for the solution-treated alloy does
not exhibit any of the hcp diffraction peaks (Figure 1).
Thus, it can be argued that the slow rates of transformation HS21 Co-based alloys. These researchers also observed,
observed during the early stages of aging (aging times less using transmission electron microscopy, the precipitation
than 4 hours) are associated with the creation of appropriate of very fine carbide particles at intrinsic fcc stacking faults.
lattice defect distributions required for the nucleation of This process was observed concurrently with the formation
the hcp phase during aging. Once this has taken place, the of the hcp2 phase at large aging times.
transformation proceeds by continued nucleation of hcp
lamellae and a reduction of the density of lattice defects
in the metaestable fcc phase. As will be show later, this B. Effect of Microstructure on Corrosion Behavior
process has profound effects on the corrosion behavior of Figures 5 and 6 shows the effect of the magnitude of
the material. the applied potential on the current density-time curves of
The morphology of the hcp lamellae found in the present solution-treated (Figure 5) and isothermally aged Co-27Cr-
aged specimens is consistent with the morphology of the 5Mo-0.05C alloys to produce microstructures with 10 pct
fault-free martensitic, hcp2 observed by Rajan[10] in Co- (Figure 6(a)), 55 pct (Figure 6(b)), and 100 pct hcp phase
26.7Cr-5.5Mo-0.15C alloy aged 15 h at 700 ⬚C. Vander (Figure 6(c)). These current density-time curves should be
Sande et al.[8] found similar results in cast and wrought compare with that shown in Figure 6(d) obtained for the

2232—VOLUME 33A, JULY 2002 METALLURGICAL AND MATERIALS TRANSACTIONS A

 (a) (b)

(c)
(d )
Fig. 6—Effect of applied potentials (0.650, 0.670, and 0.700 V) on the dissolution in Ringer’s solution of the Co-27Cr-5Mo-0.05C alloy isothermally aged
to produce microstructures with (a) 10 pct, (b) 55 pct, and (c) 100 pct hcp phase after a homogenization treatment at 1150 ⬚C for 1 h followed by water
quenching. (d ) These curves are compared with those of the as-cast ASTM F75 alloy.

as-cast Co-27Cr-5Mo-0.30C ASTM-F75 alloy. In general, class 1 behavior for applied potentials of up to 0.700 V
the current density-time behavior was characterized by two (Figure 5(a)). At more noble potentials, e.g., 0.740 V, this
regions, the extent of which depended on the applied poten- alloy continued to exhibit a class 1 behavior (Figure 5(b)),
tial and alloy microstructure. All samples showed an initial until at 0.780 V repassivation no longer occurred and a
decrease in current density up to about 50 seconds. After gradual current density increase was observed after an incu-
this, the material exhibited a typical passive behavior (class bation period of 70 seconds. Solution-treated and aged Co-
1) or a corrosion behavior (class 2), which showed a depen- 27Cr-5Mo-0.05C alloys exhibited a class 1 behavior for all
dence on alloy microstructure and applied potential. Class applied potentials investigated up to 0.7 V (Figures 6(a)
1 behavior was associated with a continuous decrease in through (c)). In contrast, the as-cast alloy exhibited a class 2
current density due to a dynamic breakdown-repassivation behavior for applied potentials higher than 0.630 V (Figure
process. In contrast, the continuously increasing current 6(d)) indicating that the breakdown of the passive film is
density observed in samples exhibiting class 2 behavior experienced at relatively more active potentials.
was attributed to continued dissolution. Figure 7 shows the potentiostatic polarization curves for
Solution-treated Co-27Cr-5Mo-0.05C alloy exhibited a Co-27Cr-5Mo-0.05C alloy with different distributions of

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 33A, JULY 2002—2233

 Fig. 8—Effect of the hcp transformation on the breakdown and open-circuit
potentials of the Co-27Cr-5Mo-0.05C alloy isothermally aged at 800 ⬚C
after a heat treatment at 1150 ⬚C during 1 h followed by water quenching.

increase in lattice defects due to (a) the defects created on

the fcc phase as a result of quenching from the solution
Fig. 7—Potentiostatic polarization curves of homogenized at 1150 ⬚C for temperature and (b) the defects introduced by the initial
1 h and water-quenched Co-27Cr-5Mo-0.05C alloy in Ringer’s solution formation of the hcp phase during the early stages of aging.
(ST) and of samples isothermally aged to produce microstructures with 10, As a result, the overall corrosion resistance of the alloy
55, and 100 pct hcp phase. Photomicrographs illustrate the appearance of
the electrode surface with 100 pct hcp transformation at selected potentials
decreases and this is observed for weight fractions of hcp
up to 0.7 V (left) and of the as-cast F75 alloy (right). phase less than 10 pct produced by short aging times. As
discussed earlier, these defects are required to form the
distribution of hcp embryos (incoherent twin boundaries
fcc and hcp phases. From these data measured between and stacking faults), which act as nucleation sites for the
open-circuit potential and 0.78 V, passive current density isothermal transformation. However, once the transforma-
and breakdown potential were obtained. The effect of the tion is well under way, the density of lattice defects must
dissolution process on the appearance of the surface micro- decrease. Since the chemical composition of both the fcc
structure of the electrodes was investigated at one point of and hcp phases is the same, the measured potentials become
the polarization curve: at 0.7 V. The micrograph at the top less sensitive to the hcp-fcc phase distribution and remain
left of Figure 7 shows the type of corrosion attack exhibited approximately constant. Because of the precipitation of fine
by the solution-treated Co-27Cr-5Mo-0.05C alloy aged to carbide particles, which starts near the end of the fcc-
produce a microstructure with 100 pct hcp phase. As can be hcp transition, the breakdown measured potentials decrease
seen, the microstructure was revealed by a weak preferential very slightly at the end of the transformation (Figure 8).
attack at grain boundaries and carbides precipitated after The as-cast F75 alloy, on the other hand, experienced a
long periods of aging. It is noteworthy that the surface of more active breakdown and OCPs (0.650 and ⫺0.480 V,
the electrodes solution treated and after aging to produce respectively) than any of the single-phase or two-phase Co-
microstructures with 10 pct and 55 pct hcp phase did not 27Cr-5Mo-0.05C alloys investigated.
show any corrosion attack. In contrast, the surface appear- In the passive region, solution-treated and aged Co-27Cr-
ance of the as-cast Co-27Cr-5Mo-0.30C ASTM-F75 alloy, 5Mo-0.05C alloys exhibited an enhanced corrosion resist-
illustrated in the micrograph at the top right side of Figure ance, and a low current density flowed at a negligible rate,
7, showed a highly localized corrosion with the grain ⬃1 ␮A⭈cm⫺2. In contrast, the as-cast alloy exhibited a
boundaries and carbide particles being attacked preferen- diminished corrosion resistance in this region, since a
tially. The electrode, clearly under nonuniform attack, higher passive current density (⬃4 ␮A⭈cm⫺2) was
revealed the dendritic solidification microstructure. The Cr- observed.
rich interdendritic regions exhibited a slower rate of disso- The experimental results presented previously indicate
lution than the Co-rich dendrites. that the preferential dissolution of aged Co-27Cr-5Mo-
The effect of the amount of hcp phase in the microstruc- 0.05C alloy is initiated at the most energetic features of
ture on the open-circuit and breakdown potentials of solu- the microstructure such as fcc twins and stacking faults.
tion-treated and aged Co-27Cr-5Mo-0.05C alloys is shown Therefore, the corrosion resistance of solution-treated and
in Figure 8. As can be seen, the measured potentials decay aged Co-27Cr-5Mo-0.05C alloys decreases only in material
rapidly at low values of the amount of hcp phase in the aged during short periods where there is a maximum density
microstructure, then remain approximately constant until of lattice defects (hcp embryos). The chemical composition
the microstructure of the alloy is fully hcp. The first poten- of the solution-treated material does not change signifi-
tial drop observed in Figure 8 can be associated with an cantly as a result of the phase transition. Thus, once the

2234—VOLUME 33A, JULY 2002 METALLURGICAL AND MATERIALS TRANSACTIONS A

density of lattice defects is reduced as a result of the forma- orthopedic implant devices requiring higher strength than
tion of the new phase, the corrosion resistance of the alloy provided by conventional ASTM-F75 materials.
is no longer affected by the phase transformation. This
conclusion was drawn from the breakdown potentials. ACKNOWLEDGMENTS
C. Montero-Ocampo thanks the National Science and
IV. CONCLUSIONS Technology Council of Mexico (CONACYT) for the finan-
cial support of this work through Project Nos. 26389A
The results of the present experimental investigation show and 38467U.
that the corrosion resistance of two-phase (fcc-hcp) Co-
27Cr-5Mo-0.05C alloys produced by isothermal aging at
REFERENCES
800 ⬚C is superior to that exhibited by the conventional
ASTM-F75 cast alloy used for the manufacture of orthope- 1. D.C. Mears: Int. Met. Rev., 1977, Rev. 218, pp. 119-54.
dic implants. 2. J.B. Medley, J.J. Krygier, J.D. Bobyn, F.W. Chan, and M. Tanzer: 21st
Annual Meeting of the Society for Biomaterials, San Francisco, CA,
The isothermal fcc → hcp transformation requires forma- Mar. 18–22, 1995, p. 47.
tion of suitable distributions of hcp embryos (incoherent 3. D.H. Bucleky and R.L. Johnson: Trans. ASLE, 1965, vol. 8, p. 123.
twin boundaries and stacking faults) prior to and during the 4. C.T. Sims: JOM, 1969, Dec., pp. 27-42.
early stages of aging. This requirement leads to a relative 5. N. Cong Dahn, D. Morphy, and K. Rajan: Acta Metall., 1984, vol.
reduction of the corrosion resistance of two-phase alloys, as 32, pp. 1317-22.
6. J.W. Weeton and R.A. Signorelli: Trans. Am. Soc. Met., 1955, vol.
inferred by the faster dissolution rates observed in material 47, p. 815.
aged during less than 4 hours. However, once the transforma- 7. A. Salinas Rodriguez and A. Mani Medrano: TMS Annual Meeting,
tion proceeds rapidly, between 4 and 8 hours of aging, the New Orleans, LA, Feb. 12–15, 2001, TMS, Warrendale, PA, p. 174.
elimination of lattice defects causes a reduction of the disso- 8. J.B. Vander Sande, J.R. Coke, and J. Wulff : Metall. Trans. A, 1976,
vol. 7A, pp. 389-97.
lution rates and the breakdown potential remains constant. 9. R.N.J. Taylor and R.B. Waterhouse: J. Mater. Sci., 1983, vol. 18, pp.
This behavior was attributed to the uniform chemical compo- 3265-80.
sition of the two-phase alloys. As a result, the corrosion 10. K. Rajan: Metall. Trans. A, 1982, vol. 13A, pp. 1161-66.
behavior of the material becomes nearly independent of the 11. K. Rajan: Metall. Trans. A, 1984, vol. 15A, pp. 1335-38.
12. C. Montero-Ocampo and A. Salinas Rodriguez: J. Biomed. Mater.
relative amounts of fcc and hcp phases present in the
Res., 1995, vol. 29, pp. 441-53.
microstructure. 13. M. Sage and C. Guillaud: Rev. Met., 1950, vol. 49, pp. 139-45.
Concurrent work on the mechanical behavior of these 14. A.H. Graham and J.L. Youngblood: Metall. Trans., 1970, vol. 1, pp.
two phase Co-27Cr-5Mo-0.05C alloys,[7] has shown that the 423-30.
hardness and yield strength of a 50 pct hcp alloy can be 15. A. de J. Saldivar Garcia: Metall. Mater. Trans. A, 1999, vol. 30A, pp.
1177-84.
increased by at least 30 pct without undue ductility losses. 16. A. de J. Saldivar Garcia: Scripta Mater., 1999, vol. 40 (6), pp. 717-22.
Therefore, the results of the present article suggest that these 17. G.B. Olson and M. Cohen: Metall. Trans. A, 1976, vol. 7A, pp.
materials are excellent candidates for the manufacture of 1905-14.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 33A, JULY 2002—2235