The structure and mechanical properties of thick cobalt electrodeposits

Article abstract of DOI:10.1023/A:1018666804577

Cobalt electrodeposits have been produced in chloride solutions. Depending on the electrolysis parameters applied, two different structures for the electrodeposits, namely α and β cobalt were observed. Their characterization included hydrogen content measurement, the relative volume fraction of the α and β phases determined by X-ray diffraction, X-ray diffraction line profile analysis and microstructural investigation by optical microscopy, scanning electron microscopy and transmission electron microscopy. The influence of structure on mechanical properties was examined. The results showed that the ductility properties of the cobalt electrodeposits were highly sensitive to the structure. A higher β phase content that was measured in some deposits did not improve the ductility due to the existence of trapped hydrogen which always exists in such deposits. However, annealing treatments seem to be a promising route to optimize the ductility of cobalt electrodeposits.

Full text of DOI:10.1023/A:1018666804577

JOURNAL OF MATERIALS SCIENCE 32 (1997) 2637—2646
The structure and mechanical properties
of thick cobalt electrodeposits
J . DILLE, J . CHARLIER, R. W INAN D*
Metallurgie Physique, CP 194/3 and *Metallurgie Electrochim ie, CP 165,
Universite Libre de Bruxelles, Avenue F.D. Roosevelt, 50-B 1050 Bruxelles, Belgium
Cobalt electrodeposits have been produced in chloride solutions. Depending on the
electrolysis param eters applied, two different structures for the electrodeposits, nam ely
a and b cobalt were observed. Their characterization included hydrogen content
m easurem ent, the relative volum e fraction of the a and b phases determ ined by X-ray
diffraction, X-ray diffraction line profile analysis and m icrostructural investigation by optical
m icroscopy, scanning electron m icroscopy and transm ission electron m icroscopy. The
influence of structure on m echanical properties was exam ined. The results showed that the
ductility properties of the cobalt electrodeposits were highly sensitive to the structure.
A higher b phase content that was m easured in som e deposits did not im prove the ductility
due to the existence of trapped hydrogen which always exists in such deposits. However,
annealing treatm ents seem to be a prom ising route to optim ize the ductility of cobalt
electrodeposits.
1. Introduction
[4] to be:
HCP a (1 1 2 0)#FFC b (220) or HCP a (1 0 1 0)
#FCC b (4 2 2)
According to various authors [1—7], the structure of
cobalt electrodeposits strongly depends on the com-
position of the electrolyte and on the electrolysis para-
meters such as the pH of the solution, the temperature
and the current density. Three types of electrolyte are
particularly utilized: sulphate solutions, chloride solu-
tions or Watts type baths.
(1)
On the other hand, Sard et al. [1] have noted the
occasional existence of a ‘‘pseudo random’’ structure
in HCP cobalt deposits. In this context a ‘‘pseudo
random’’ structure is one in which the basal planes
a (0 0 0 2) are almost never parallel to the substrate and
no preferred orientation is shown for the other planes.
Different structural features such as the surface ap-
pearance of deposits or the shape of grains are gener-
ally closely related to the texture.
The principal structural feature of interest is the
type of phases that occur in the deposits. Cobalt exists
in two allotropic modifications, a and b. The a phase
is hexagonal close-packed (HCP) and is thermodyna-
mically stable up to 417 C whereas the b phase is
face-centered cubic (FCC) and is stable at temper-
°
°
atures higher than 417 C. Although the electrolysis
Whilst there is a fairly detailed knowledge of the
inter-relation between the electrolysis parameters and
the structure of the electrodeposit very little work has
been performed to relate the electrodeposit structure
to observed properties of the electrodeposited cobalt.
There are reports in the literature [8—10] on the influ-
ence of the electrodeposit structure on various mag-
netic properties. In addition, Feneau and Breckpot [3]
have reported that cobalt electrodeposits with an
HCP a (0 0 0 2) texture — which is occasionally pro-
duced at high pH values in sulphate baths — are more
brittle than deposits with a HCP a (1 1 2 0) or a (1 0 1 0)
texture.
The purpose of the present work [11] is to study in
more detail the relationships between the structure
and mechanical properties of free standing cobalt elec-
trodeposits. In particular we address the question as
to if an increasing content of FCC b phase improves
the ductility as has been established for cobalt strips
°
temperatures are well below 417 C, the cobalt electro-
deposits often simultaneously contain both phases.
This is especially noticeable at low pH values and low
deposition temperatures and is independant of the
electrolyte used in the electrolysis. The presence of the
face-centered cubic b phase at room temperature re-
sults from the small difference between the free ener-
gies of both modifications.
Another important structural feature of the electro-
deposits is their fibrous texture. In this paper, the
nature of the fibrous texture will be distinguished in
terms of the crystallographic plane which is preferen-
tially oriented parallel to the substrate. In cobalt elec-
trodeposits, the hexagonal close-packed phase often
exhibits an a (1 1 2 0) texture or an a (1 0 1 0) texture
depending on the electrolysis conditions. If both
phases are co-deposited, then a texture correspon-
dence between HCP a and FCC b has been reported
0022—2461 ( 1997 Chapman & Hall
2637

obtained by thermomechanical processes [12—14].
This behaviour could be explained by the fact that the
limited number of slip systems in HCP metals restricts
their ductility whereas the FCC metals with more slip
systems generally exhibit a good ductility.
In order to reach this objective, different cobalt
electrodeposits displaying the most frequently en-
countered structures are required. This can be
achieved by investigating a wide range of pH. The
deposits should be thick enough for their mechanical
properties to essentially depend on the structure
which is in turn controlled by the electrodeposition
parameters. In this case, the role of the substrate
influenced deposit zone can be minimized. The char-
acterization of the deposits will aim at establishing
accurate relations between the electrolysis parameters,
the structure and observed properties.
2. Experimental procedure
2.1. Cobalt electrodeposition
In order to control the hydrodynamics during the
cobalt deposition, a channel cell was used (see Fig. 1).
The commercially pure copper cathode was polished
using 600 grade abrasive paper and then the cathode
was carefully cleaned with ethanol. The anode is
a pure cobalt electrode. An additive free electrolyte
\ꢀ
containing 202 g dm
flowed parallel to the electrodes with a flow rate of
ꢁ>
CoCl · 6H O (0.85 M Co
)
ꢁ
ꢁ
\ꢂ
0.4 m s
The cobalt electrodeposits were obtained at a con-
\ꢁ
.
stant current density of 800 A m . The temperature
°
°
of the electrolyte was kept at 50 C (25 C for some
deposits at pH 1.5 or at pH 4.0). The electrolyte pH
values range from 1.0—5.0 and were adjusted by HCl
additions.
During the electrolysis, the pH of the solution tends
to increase. Some hydrochloric acid was added from
time to time to adjust the pH with an accuracy of
$1%. The deposits were about 200 lm thick. They
were separated from the substrate in order to obtain
free standing foils.
Figure 1 Schematic of the electrolytic cell where (1) is the abc
channel electrolytic cell, (2) holding tank, (3) pH measurement, (4)
centrifugal pump, (5) flowmeter and (6) reference electrode.
2.2. Microscopic characterization
The microstructures of the deposits were studied by
optical microscopy, scanning electron microscopy
(SEM) and transmission electron microscopy (TEM).
The SEM observations were carried out using a Jeol
JSM 820 scanning electron microscope which was
used to examine the surface of the deposits and the
fracture surfaces of tensile test specimens. Thin foil
TEM analyses were carried out using a Philips CM20
ultratwin scanning transmission electron microscope.
The thin foils were obtained by double jet electrolytic
polishing with a solution of 10% perchloric acid in
peaks to enable the determination of crystallite size
and lattice distorsions: the Warren—Averbach method
[15] and the method of De Keijser et al. [16].
The powerful Warren—Averbach method is based
on a rigorous analysis of ‘‘pure’’ peak profiles — cor-
rected for the instrumental peak broadening — in
terms of Fourier coefficients. This method requires
two orders of reflection for each (hkl) plane considered
°
ethanol-2-butoxy at #5 C (30 V).
and leads to an area-weighted crystallite size 1D2 and
!
ꢁ
a mean square (local) strain 1e (¸)2.
The method of De Keijser et al. is a rapid single line
method for the determination of strain and crystal size
by means of the use of a Voigt function to describe the
peak profile shape function. A volume-weighted aver-
2.3. X-ray diffraction analysis
The X-ray diffraction experiments were performed
using a Siemens D 5000 diffractometer. Two methods
were used to analyse the broadening of the diffraction
&
age crystallite size 1D2 is then defined as k/b · cosh
ꢀ
#
2638

&
&
&
and a strain value e' as b /4 tan h where b and b are
2.4. Mechanical properties
'
#
'
respectively the integral width of the Cauchy and the
Gaussian components of the ‘‘pure’’ structurally
broadened Voigt profile and j and h are respectively
Three testing methods were utilized to determine the
mechanical properties of the deposits. Tensile tests
were performed using an Instron machine. Samples
the X-ray Ka wavelength and the angular position.
Averbach and Cohen [17] have developed a direct
°
were strained at 20 C at a crosshead speed of 0.02 cm
ꢂ
per min. Data were measured for the 0.2 per cent offset
comparison method for the quantitative determina-
tion of the volume fractions of various phases in a ran-
domly oriented multiphase polycrystalline aggregate.
This method is based on the measurement of the
integrated intensity of an isolated diffraction peak for
each existing phase. Sage and Guillaud [18] applied
this method to the quantitative determination of the
proportion of a HCP and b FCC phases in cobalt
powders. Using the most suitable a (1 0 1 1) and
b (2 0 0) diffraction peaks they found the following
relationship
yield strength, (R
and per cent elongation at fracture (A).
) ultimate tensile strength (R )
%ꢄꢅꢁ%
ꢃ
The ductility measurements were made by means of
a work-hardening bend test. This test consists of ap-
plying cantilever bending to a strip piece clamped at
one end; the specimen is bent backwards and forwards
around two mandrels with a bend radius of 5 or 2 mm.
This bending is repeated until fracture occurs. The
bending ductility is defined as the number of cycles
before fracture. In addition, Vickers micro hardness
tests were performed with a 200 g load.
x
Ib(2 0 0)
"k ·
(2)
1!x
Ia(1 0 1 1)
2.5. Hydrogen content determ ination
The hydrogen content of the electrodeposits were
determined using a Leco RH 1 inert gas fusion ap-
paratus.
in which Ib(2 0 0) and Ia(1 0 1 1) are the measured integrated
intensities, k is a constant depending on the radiation
wavelength and x the volume fraction of the b FCC
phase.
Preferred orientations, frequently observed in elec-
trodeposits, introduce errors in the volume fraction
determination. Thus it is necessary to replace the
integrated intensity of one diffraction peak by the total
intensity obtained from the corresponding complete
pole figure [19]. Owing to the rotational symmetry in
an electrodeposits texture, a plot of pole density versus
3. Results and discussion
3.1. Electrochem ical results
The influence of the pH of the solution on the current
efficiency is shown in Fig. 2. The current efficiency,
always greater than 85%, regularly increased as the
pH increased and reached more than 99% at pH 3.5.
This behaviour was due to hydrogen co-deposition at
low pH and temperature.
Fig. 3 shows the cathodic potential as a function of
the pH for deposits obtained at 50 C. A sudden dis-
continuity is observed at a pH between 3.0—4.0. These
results are in agreement with those reported for cobalt
electrodeposition from chloride solutions [6].
°
angle ꢀ between 0—90 gives a complete description of
the texture; ꢀ is the angle between the symmetry axis
normal to the deposit surface and the normal towards
the reflecting plane. The pole figure is then represented
by the distribution curve of pole density Iꢀ versus
angle ꢀ. The calculation of total intensity from a pole
figure should include the length of every parallel cor-
responding to Iꢀ. One obtains [20]:
°
pꢃꢁ
3.2. Hydrogen content
I
"2p
Iꢀ · sinꢀ dꢀ
(3)
2ꢁꢂ
ꢀ
The hydrogen contents shown in Fig. 4 were measured
about 30 days after electrodeposition. The hydrogen
content of the deposits increased when the pH or the
electrolysis temperature decreased. It should be noted
ꢄ
For the experimental determination of the pole den-
sity Iꢀ, it is necessary to use both a reflection and
a transmission X-ray diffraction method.
The reflection method covers the inner part of the
°
°
pole figure from ꢀ"0 to about 70 whereas the
transmission method covers the outer part of the pole
°
figure from ꢀ values of about 50—90 . However, Scoyer
et al. [19] have established with the help of crystallo-
graphic considerations that the data obtained by re-
flection are sufficient for the calculation of the total
intensity from the pole figure for both a HCP
and b FCC cobalt phases when rotational symmetry
occurs.
In this work, the reflection method of Schulz [21]
was used to obtain experimental a (1 0 1 1) and b (2 0 0)
pole figures. Then, the corresponding total intensities
were determined using Equation 3 and applying the
arguments of Scoyer et al. [19]. Finally, these total
intensities were introduced into Equation 2 in order to
obtain the volume fraction of the b FCC phase.
Figure 2 Current efficiency r versus pH (᭹) at 50 °C and (ᮀ) at
25°C.
#
2639

Figure 3 Cathodic potential º versus pH (᭹) at 50 °C and (ᮀ) at
E
25°C.
Figure 5 Hydrogen content measured after annealing for 1 h versus
annealing temperature. (᭹) deposits obtained at pH 1, (;) deposits
obtained at pH 2, (ᮀ) deposits obtained at pH 4.
Nevertheless as previously mentioned no change in
hydrogen concentration was found in this study. This
fact means that the detected hydrogen cannot diffuse
out of the cobalt sheet at room temperature. To ex-
plain this, various possibilities have to be considered
for the origin of this measured hydrogen: such possi-
bilities include atomic or molecular hydrogen in
microstructural trapping sites or an occluded chem-
ical compound.
In order to investigate this question, annealing
treatments were made on the cobalt deposits for 1 h at
different temperatures. Fig. 5 shows the measured hy-
drogen contents in the deposits after annealing. For
deposits obtained at pH 1 and pH 2, the hydrogen
concentration decreased as the annealing temperature
increased. This phenomenon can be explained most
satisfactorily in terms of hydrogen detrapping during
annealing. Hydrogen then escapes out of the deposit.
No information is available in the literature on poten-
tial trap sites in the cobalt structure but they are
probably the same as in iron or nickel where the main
trapping sites have been determined to be dislocations
and grain boundaries [24].
The larger hydrogen contents found in deposits ob-
tained at pH 1 should be related to the more important
hydrogen co-deposition observed at this pH value. On
the other hand, the measured hydrogen content for
deposits obtained at pH 4 remained unchanged with
increasing annealing temperature. In this case the de-
tected hydrogen probably comes from the decomposi-
tion of some occluded cobalt compound such as
Figure 4 Hydrogen content measured 30 days after electrodeposi-
tion versus pH. (᭹) deposits obtained at 50 °C, (ᮀ) deposits ob-
tained at 25 °C.
that for an individual deposit the hydrogen content
did not change between 30 min and 100 days after the
end of the electrodeposition.
The diffusion equations can be applied to the de-
gassing phenomenon of a sheet of thickness h to en-
able the calculation of the average concentration c of
hydrogen remaining inside the deposit at a time t after
the end of the electrolysis.
The solution of the diffusion equation can be writ-
ten as [22]
8 ꢆ
+
Hꢇꢄ
1
(2j#1)n ꢁ
c/c "
exp !
ꢁ
· D · t
ꢄ
ꢁ
p
(2j#1)
h
A C D B
(4)
Co(OH) . The occlusion of this compound during co-
ꢁ
balt electrodeposition at high pH values has already
where c "hydrogen concentration into the deposit
been mentioned in the literature [6].
ꢄ
at the end of the electrolysis
h"deposit thickness (200 lm, in our study)
D"diffusion coefficient of hydrogen in cobalt
3.3. X-ray diffraction patterns
The X-ray diffraction patterns measured on the de-
From the data obtained by Caskey et al. [23],
the diffusion coefficient of hydrogen in HCP cobalt.
extrapolated to room temperature is 3.5;
°
posits obtained at 50 C, at pH 1 to 3, revealed no
significant preferred orientation (Fig. 6a) but, a too
\ꢂꢁ ꢁ \ꢂ
10 cm s . Using this value, Equation (4) gives
weak
a
(0 0 0 2) peak intensity was noticed;
[a (0 0 0 2)#b (1 1 1), if both phases coexist]. For the
deposit obtained at pH 1, a weak b (2 0 0) extra reflec-
tion was detected.
c/c "0.95 for t "30 min and c/c "0.37 for
ꢄ
ꢂ
ꢄ
t "100 days.
ꢁ
2640

Figure 6 X-ray diffraction patterns, (a) from a deposit obtained at pH 1.5 and T"50 °C; (b) from a deposit obtained at pH 4 and T"50 °C.
TABLE I b (FCC) phase volume fraction x in the deposits
3.5. X-ray diffraction line profile analysis
Tables II and III respectively list the average crystal-
Electrolysis conditions
x
lite sizes 1D2 and the strain values e obtained by the
ꢀ
method of De Keijser et al. [16] from different
pH 5; 50 °C
pH 4.5; 50 °C
pH 4; 50 °C
pH 3.5; 50 °C
pH 3; 50 °C
pH 2; 50 °C
pH 1.5; 50 °C
pH 1; 50 °C
pH"4; 25°C
pH"1.5; 25 °C
41%
41%
a (HCP) diffraction peaks for the non-significantly
41%
textured deposits. Table IV contains 1D2 and e' cal-
ꢀ
"1.5%
"1.5%
"5.5%
"7%
culated in the same way from the single a (1 1 2 0)
line profile for the almost perfectly (1 1 2 0) textured
deposits.
"12%
"14.5%
"22%
Table V summarizes the results obtained by the
Warren—Averbach method. This method cannot be
applied to perfectly textured deposits because the sec-
ond order a (2 2 4 0) peak is not available. From these
results, three main considerations have to be made.
(i) First, the 1D2 values calculated from (h k i l)
diffraction peaks with h!k"3n$1 and lO0 [e.g.,
ꢀ
°
For the deposits obtained at 25 C (pH 1.5 and 4)
°
and at 50 C (pH'3), only one peak [a (1 1 2 0) or, if
both phases coexist, a (1 1 2 0)#b (2 2 0)] was ob-
served (Fig. 6b). Such deposits are perfectly textured.
TABLE II Average crystallite sizes 1D2 (in nm) obtained by the
ꢀ
method of De Keijser et al. [16] from various a (HCP) diffraction
peaks for the non-significantly textured deposits
3.4. Determ ination of the volum e fractions
of the a (HCP) and b (FCC) phases
Diffraction peak
pH 1
pH 1.5
pH 2
pH 3
Table I summarizes the b (FCC) phase volume frac-
tions in the deposits. These volume fractions were
determined by means of the above described proced-
ure using pole figures obtained by the reflection
method of Schulz.
(1 0 1 0)
(1 0 1 1)
(1 1 2 0)
(1 0 1 3)
(2 0 2 0)
(2 0 2 1)
53.2
11.7
40.8
8.0
51.4
20.1
151.8
18.5
101.3
14.5
90.7
27.6
241.3
33.7
172.7
28.9
142.0
65.7
181.0
45.3
103.2
39.5
120.7
61.1
Comparing Table I to Fig. 4, it is apparent that the
b (FCC) content depends on the hydrogen concentra-
tion in the deposit. A larger concentration of incorpor-
ated hydrogen leads to a larger b (FCC) volume
fraction. This observation confirms the dominant in-
fluence of incorporated hydrogen on b face centered
cubic phase formation as proposed by various authors
[6, 7, 25].
TABLE III Strain values e' obtained by the method of De Keijser
et al. [16] from various a (HCP) diffraction peaks for the non-
significantly textured deposits
Diffraction peak
pH 1
pH 1.5
pH 2
pH 3
It should also be noted that the b (FCC) modifica-
tion can exist in the two different types of deposits
nearly random oriented or strongly textured.
Finally, the pole figures definitely confirm the
observations from the measured X-ray diffraction
patterns.
(1 0 10)
(1 0 1 1)
(1 1 2 0)
(1 0 1 3)
(2 0 2 0)
(2 0 2 1)
0.0017
0.0014
0.0019
0.0006
0.0011
0.0022
0.0008
0.0012
0.0011
0.0007
0.0008
0.0009
0.0006
0.0009
0.0005
0.0003
0.0006
0.0011
0.0008
0
0.0009
0
0.0006
0.001
2641

TABLE IV 1D2 and e' obtained by the method of De Keijser
ꢀ
et al. [16] for the single a (1 1 2 0) line profile for the almost perfectly
a (1 1 2 0) textured deposits
Electrolysis conditions
1D2
(lm)
e'
ꢀ
pH 5; 50 °C
pH 4,5; 50 °C
pH 4; 50 °C
pH 3.5; 50 °C
pH 4; 25 °C
pH 1.5; 25 °C
'1
0.0004
0
0.0004
0
0.001
0.0053
'1
'1
'1
0.3982
0.0733
TABLE V Results of the Warren—Averbach [15] line profile
analysis
Electrolysis conditions
1D2
(nm)
1eꢁ(L)2ꢂꢃꢁ
L"5.0 nm
!
pH 1; 50 °C
pH 1,5; 50 °C
pH 2; 50 °C
pH 3; 50 °C
31.0
83.0
104.0
72.0
0.00338
0.00293
0.00148
0.00227
Figure 7 Cross-sectional SEM image of a cobalt electrodeposit
obtained at pH 5 and T"50 °C. Deposit growth direction: from left
to right.
a (1 0 1 0)] are smaller than those obtained from the
other peaks [e.g., a (1 0 1 1")].
This behaviour could be due to the existence of
stacking faults in the electrodeposits. Indeed the gen-
eral theory of X-ray diffraction from faulted HCP
structures indicates a broadening due to stacking
faults for diffraction lines with h!k"3n$1 and
1O0. This line broadening does not affect the strain
values but leads to a calculated ‘‘effective’’ size smaller
than the true crystallite size. The occurrence of numer-
ous stacking faults in the deposits is confirmed by the
microscopy investigations that will be described in the
following section.
(ii) Comparing the calculated microstrain values to
hydrogen contents shown on Fig. 4, it seems that the
hydrogen co-deposition results in the formation of
crystalline defects that are the cause of the obtained
microstrains.
(iii) Both the Warren—Averbach method and the
De Keijser et al. method indicate that the crystallite
size for electrodeposits obtained at pH 1 and T"
°
50 C is significantly smaller than for electrodeposits
obtained at other experimental conditions. This is also
confirmed by microscopy observations and suggests
°
that, during the electrodeposition at pH 1 (50 C), the
nucleation rate is higher.
Figure 8 TEM micrograph of a deposit obtained at pH 5 and
T"50 °C. Section parallel to the substrate.
3.6. Microscopy investigations
Two different characteristic structures were observed
in the deposits. These structures are in fact those most
commonly encountered in cobalt electrodeposits.
Each structure is exactly related to one type of deposit
determined by the X-ray diffraction analysis namely
an almost perfectly textured deposit or a nearly ran-
dom oriented deposit.
corresponds to a F T (field oriented texture) type in
Fisher’s classification [26]. The columnar aspect of
this structure is revealed in Fig. 7 which shows an
SEM cross-sectional picture. This cross-section had
previously been polished and etched.
Fig. 8 is a TEM micrograph of a thin foil parallel to
the substrate for a deposit obtained under the same
conditions as those for Fig. 7. For such a section the
grain structure is equiaxed.
The first structure was found in all strongly tex-
°
°
tured deposits (50 C, pH'3; 25 C, pH 1.5 and 4). It
2642

Figure 10 Cross-sectional SEM image of a cobalt electrodeposit
obtained at pH 2 and T"50 °C. Deposit growth direction: from left
to right.
Figure 11 TEM micrograph from a deposit obtained at pH 1.5 and
T"50 °C. Section parallel to the substrate.
(0 0 0 2) with the plane of the foil and are due to
stacking faults.
°
Nearly random oriented cobalt deposits (50 C, pH
1 to 3) exhibit a totally different structure. Their struc-
ture consists of an assembly of dihedrons. A SEM
cross-sectional image (Fig. 10) reveals three dimen-
sional nucleation. Figs. 11 and 12 are TEM micro-
graphs of thin foils parallel to the substrate. These
figures show the dihedral shape of the crystallites. The
curvature of the ridge between part A and part B for
the dihedron seen in Fig. 12 results from the fact that
each part of the dihedron is not a single crystal but
consists of a series of subgrains. The electron diffrac-
tion patterns indicate a weak misorientation between
Figure 9 Electron diffraction pattern from crystal 1 on Fig. 8, zone
axis [0 1 1 0] (spots (0 0 0 n), n odd, are double diffraction spots).
All the grains have electron diffraction patterns
similar to the one represented by Fig. 9. Its zone axis is
[0 1 1 0]. This fact confirms the strong a (1 1 2 0) tex-
ture. The electron diffraction patterns also reveal that
in many cases two adjacent grains are slightly mis-
°
oriented (3 for example between grain 3 and grain 4
in Fig. 8) around the zone axis. In other cases, the
°
°
misorientation is greater (15 between grain 1 and
two adjoining subgrains (about 2 ). The electron dif-
grain 2). The closely spaced parallel lines observed in
each grain represent the intersection of the basal plane
fraction analysis (Fig. 13) moreover reveals that both
parts of a dihedron have a twin relationship to one
2643

another. The twin plane (TT) lies perpendicular to the
substrate and contains the ridge of the dihedron.
The observed structure in nearly random oriented
cobalt electrodeposits is identical to the ‘‘pseudo ran-
dom’’ structure mentioned by Sard et al. [1] Similar
dihedral shapes have been reported in (2 1 1) tex-
tured FCC nickel electrodeposits by various authors
[27, 28].
Thevenin [29] and Atanassov et al. [30] have pro-
posed that the (2 1 1) texture in nickel electrodeposits
is initiated by the development of paracrystalline de-
cahedral and icosahedral clusters. The authors of this
paper think that a further elucidation of the crystalline
structures forming in cobalt electrodeposits, which is
beyond the scope of this study should take into ac-
count the similarity with nickel electrodeposits.
A characteristic surface morphology corresponds to
each type of deposit microstructure. Almost perfectly
textured a(1 1 2 0) cobalt electrodeposits exhibit
a smooth surface (Fig. 14) with a satin like appearance
while nearly random oriented deposits exhibit curved
dihedral shapes on their surfaces (Fig. 15) with a dull
dark grey or black appearance. The surface morpho-
logy shown in Fig. 15 is commonly mentioned in the
investigations of cobalt electrodeposits [1, 6, 31].
3.7. Mechanical properties of the deposits
Conventional Vickers microhardness tests do not re-
veal significant differences between the electrodeposits
°
obtained at 50 C (Fig. 16). On the other hand, the
°
microhardness of the electrodeposits obtained at 25 C
is clearly higher. This fact can be explained by the
higher microstrain values, e', for these deposits.
Figure 13 (a) Electron diffraction pattern of crystal A (on Fig. 12),
(b) electron diffraction pattern of crystal B (on Fig. 12), (c) crystallo-
graphic relationship between (;) crystal A and (⅙) crystal B.
Figure 12 TEM micrograph of a deposit obtained at pH 2 and
T"50 °C. Section parallel to the substrate.
Figure 14 SEM image of the surface of a deposit obtained at pH
4 and T"50 °C.
2644

TABLE VI Summary of tensile test results (deposits obtained at
¹"50 °C)
Electrolysis conditions
A
(%)
R
(MPa)
R
ꢃ
(MPa)
%;ꢄꢅꢁ%
pH 5
pH 4
pH 3
pH 2
pH 1.5
pH 1
0.5
0.8
5.9
5.1
2.8
2.1
430
435
335
315
385
450
455
480
565
540
550
620
TABLE VII Summary of work-hardening bend test results
(deposits obtained at ¹"50 °C)
Electrolysis
conditions
Number of bends before fracture
bend radius 5 mm
bend radius"2 mm
pH 5
pH 4
pH 3
pH 2
pH 1.5
pH 1
0
0
53
28
26
13
0
0
23
15
6
2
Figure 15 SEM image of the surface of a deposit obtained at pH 1.5
and T"50 °C.
The mechanical properties measured by tensile test
are much more sensitive to microstructure than
microhardness as is shown in Table VI. For the nearly
random oriented cobalt electrodeposits, the ductility
decreases when the pH value of the electrolyte
decreases.
Consequently, the higher b (FCC) phase content
produced in deposits at low pH does not correspond
to an increase of ductility as was expected. This is
probably due to the presence of trapped hydrogen in
this kind of deposit; severe embrittlement can indeed
be produced in FCC and HCP metals by very small
amounts of hydrogen [32]. It should also be noted
that the smaller grain size for deposits obtained at pH 1
leads to a higher yield strength and tensile strength.
The reason for the total lack of ductility in strongly
textured electrodeposits is not fully understood.
Figure 16 Vickers microhardness versus pH. (᭹) deposits obtained
at 50 °C, (ᮀ) deposits obtained at 25 °C.
°
Nevertheless, an annealing treatment at 400 C for
TABLE VIII Summary of main observations
Type of deposits
1a
1b
2
Electrolysis conditions
¹"50°C;
pH'3
¹"25°C;
pH 1.5 or 4
¹"50 °C;
p43
structure
texture
columnar structure
columnar structure
assembly of dihedrons
very strong
a (1 12 0)
very strong
a (1 12 0)
nearly random
oriented
texture
low
#b (220) texture
high
hydrogen content
varying; increases
if pH decreases
b (FCC) phase
content
nil
rather high
(up to 22%)
varying; increases
if pH decreases
ductility
very weak
extremely weak
rather weak:
varying; decreases
if pH decreases
surface
satin like or grey pale;
relatively smooth
satin like or grey pale;
relatively smooth
dull dark grey or black;
rather rough
2645

these deposits indicates that their prismatic grain mor-
phology and pronounced preferred crystallographic
orientation are not the main structural features limit-
ing the ductility. Indeed, after a 1 h annealing at
treatment on the structure and mechanical properties
of cobalt electrodeposits will be studied in a further
publication.
°
400 C the elongation A for a cobalt deposit at pH 4 or
pH 5 increases from less than 1% up to about 8%
while the grain morphology and the strong texture
remain. Furthermore, this result shows that it would
be very interesting to systematically study the effect of
annealing in order to maximize the ductility of cobalt
electrodeposited foils.
References
1. R. SARD, C. D. SCWARTZ and R. WEIL, J. Electrochem.
Soc. 113 (1966) 424.
2. M. FROMENT and G. MAURIN, C.R. Acad Sc. Paris-Se&rie
C 266 (1968) 1017.
3. C. FENEAU and R. BRECKPOT, A.¹.B. Me&tallurgie 9 (1969)
115.
Owing to their brittleness, no tensile test specimen
4. S. VITKOVA, S. ARMIANOV and N. PANGAROV, Elec-
trodepos. Surf. ¹reat. 3 (1975) 225.
5. J. SCOYER and R. WINAND, Surf. ¹echn. 5 (1977) 169.
6. S. NAKAHARA and S. MAHAJAN, J. Electrochem. Soc. 127
(1980) 283.
°
could be made from deposits obtained at 25 C. This
brittleness was probably due to both the structure and
the high hydrogen content of such deposits.
The work-hardening bend tests confirm the tensile
test results. They allow the distinction between the
non ductile strongly textured and the nearly random
oriented types of deposits to be made (see Table VII).
7. I. POVETKIN and I. KOVENSKII, Sov. Electrochem. 22
(1986) 1101.
8. M. F. QUINN and I. M. CROLL, Adv. X-Ray Anal 4 (1961)
151.
9. R. D. FISHER, J. Electrochem. Soc. 109 (1962) 479.
10. S. ARMYANOV and S. VITKOVA, Surf. ¹echn. 7 (1978) 319.
11. J. DILLE, PhD thesis, Universite Libre de Bruxelles (1994).
12. R. W. FRASER, D. J. L. EVANS and V. N. MACKIW, Cobalt
25 (1964) 171.
13. M. BECKERS, L. FONTAINAS, B. TOUGARINOFF and
L. HABRAKEN, ibid 25 (1964) 171.
14. E. DIDERICH, J. M. DRAPIER, D. COUTSOURADIS and
L. HABRAKEN, ¸e Cobalt 1 (1975) 7.
15. B. E. WARREN and B. L. AVERBACH, J. Appl. Phys. 23
(1952) 497.
16. Th. H. DE KEIJSER, J. L. LANGFORD, E. J. MITTE-
MEIER and A. B. P. VOGELS, J. Appl. Cryst. 15 (1982) 308.
17. B. L. AVERBACH and M. COHEN, ¹rans. Met. Soc. AIME
176 (1948) 104.
18. M. SAGE and CH. GUILLAUD, Rev. Met. 47 (1950) 139.
19. J. SCOYER, R. WINAND and J. CHARLIER, A¹B Me&tal-
lurgie 15 (1975) 222.
4. Conclusions
The main observations of this investigation are listed
in Table VIII. Depending on the electrolysis para-
meters, two types of cobalt electrodeposits were ob-
tained in chloride solutions. Their structures were
completely different in that:
(1) the first type of deposit was a FT (field oriented
texture) type. Its columnar structure consisted of nar-
row prismatic crystallites with a strong aHCP (1 1 2 0)
[b FCC (220)] preferred orientation; the surface of
these satin like or pale grey deposits was relatively
smooth, this type of deposit was obtained under the
°
following electrolysis conditions: T " 50 C, pH '3;
20. B. D. CULLITY, ‘‘Elements of X-Ray diffraction’’, 2nd Edn
(Addison-Wesley, Reading, MA, 1978).
°
T"25 C, pH 1.5 or 4.
(2) the second type of deposit consists of an assem-
bly of dihedrons with nearly random crystallographic
orientation; the surface of these dull dark grey or
black deposits was rather rough. This type of deposit
was obtained under the following electrolysis condi-
21. L. G. SCHULTZ, J. Appl. Phys. 20 (1949) 1030.
22. P. SHEWMON, ‘‘Diffusion in solids’’, 2nd Edn (The Minerals
Metals and Materials Society, Warrendale, PA 1989).
23. G. R. CASKEY, R. G. DERRICK and M. R. LOUTHAN,
Scipta Metall. 8 (1974) 481.
24. SUNG MAN LEE and JAI YOUNG LEE, Met. ¹rans. A 17
(1986) 181.
25. G. MAURIN, Thesis, Faculte des Sciences de Paris (1970).
26. H. FISCHER, ‘‘Electrolytische Abscheidung und Elektrokris-
tallisation von Metallen’’ (Springer Verlag, Berlin, 1954).
°
tions: T"50 C, pH43.
A sudden discontinuity in the cathodic potential
appeared exactly at the transition between these two
different structures.
27. H. SCHLO® TTERER, Z. Kristall 119 (1964) 321.
The ductility of each kind of deposit was signifi-
cantly different. The F.T. type of deposit had no duc-
tility compared to the nearly random oriented cobalt
electrodeposits. On the other hand, the higher b(FCC)
phase content in deposits obtained al low pH values
did not correspond to an increase of ductility as ex-
pected. This seems to be due to the trapped hydrogen
that always exists in these deposits.
28. M. FROMENT and G. MAURIN, J. Microsc. Spectrosc.
Electron 12 (1987) 379.
29. J. THEVENIN, ibid 1 (1976) 7.
30. N. ATANASSOV, S. VITKOVA and S. RASHKOV, Surf.
¹echnol 14 (1981) 215.
31. M. ROJAS, C. L. FAN, H. J. MIAO and D. L. PIRON,
J. Appl. Electroch. 22 (1992) 1135.
32. G. DIETER, ‘‘Mechanical Metallurgy’’ (3rd Edn) (McGraw
Hill, New York, 1986).
On the other hand, an annealing treatment clearly
improves the ductility of the cobalt electrodeposits.
This seems to be the most promising means to opti-
mize the malleability of cobalt foils obtained by elec-
trodeposition. Therefore, the effects of an annealing
Received 25 March
and accepted 2 July 1996
.
2646