Electrochemical and AFM study of nickel nucleation mechanisms on vitreous carbon from ammonium sulfate solutions

Article abstract of DOI:10.1016/j.electacta.2005.08.017

Reaction and nucleation mechanisms of nickel in ammoniacal solutions have been investigated as a function of nickel concentration, solution pH, deposition potential, temperature and conditioning potential. Electrochemical mechanisms of nickel reduction were found to be pH dependent, while their kinetics was concentration dependent. A surface film formed by anodic oxidation passivates nickel clusters preventing their further oxidation. Nickel nucleation on vitreous carbon, which proceeds according to the progressive nucleation model, shows a large degree of inhibition at both pH 6 and pH 9. Cluster sizes were larger when electrodeposition was carried out from solutions with higher nickel concentrations. The clusters were also larger at more negative deposition potentials and at higher solution pH. Cluster population density increased with the increasing solution temperature. Different activation energies for the nickel-aquo and nickel-ammino complexes calculated from Arrhenius diagram indicate that electroreduction of nickel-ammino complex is energetically more demanding. All electrochemical results were further verified by the atomic force microscopy investigations.

Full text of DOI:10.1016/j.electacta.2005.08.017

Electrochimica Acta 51 (2006) 2678–2690
Electrochemical and AFM study of nickel nucleation mechanisms on
vitreous carbon from ammonium sulfate solutions
∗
Darko Grujicic, Batric Pesic
University of Idaho, College of Engineering, Department of Materials Science and Engineering, McClure Hall, Moscow, ID 83844-3024, USA
Received 22 February 2005; received in revised form 7 May 2005; accepted 7 August 2005
Available online 19 September 2005
Abstract
Reaction and nucleation mechanisms of nickel in ammoniacal solutions have been investigated as a function of nickel concentration,
solution pH, deposition potential, temperature and conditioning potential. Electrochemical mechanisms of nickel reduction were found to
be pH dependent, while their kinetics was concentration dependent. A surface film formed by anodic oxidation passivates nickel clusters
preventing their further oxidation. Nickel nucleation on vitreous carbon, which proceeds according to the progressive nucleation model,
shows a large degree of inhibition at both pH 6 and pH 9. Cluster sizes were larger when electrodeposition was carried out from solutions
with higher nickel concentrations. The clusters were also larger at more negative deposition potentials and at higher solution pH. Cluster
population density increased with the increasing solution temperature. Different activation energies for the nickel–aquo and nickel–ammino
complexes calculated from Arrhenius diagram indicate that electroreduction of nickel–ammino complex is energetically more demanding.
All electrochemical results were further verified by the atomic force microscopy investigations.
©
2005 Elsevier Ltd. All rights reserved.
Keywords: Nickel electrodeposition; Ammonia; Nucleation; Electrode conditioning; Atomic force microscopy—AFM
1
. Introduction
shape [7], of non-uniform size (i.e. progressive nucleation)
9,11]. Progressive nickel nucleation can be inhibited by
[
Nickel coatings are among the earliest commercially elec-
adsorption of nickel hydroxide [12], and by formation of
nickel hydrides due to concurrent reduction of hydrogen ions
[10]. Typical surface coverage by nickel ad-atoms required
to initiate nickel nucleation on vitreous carbon found [8] was
in the range 1–5%.
From the reaction point of view, the most frequently
cited mechanisms [13,14] for nickel electrodeposition from
acidicchloridesolutionscontaininghighnickelconcentration
involve two consecutive one-electron charge-transfer steps,
with adsorption of intermediary cationic Ni complex, reac-
tions (1)–(3):
trodepositedthinmetallicfilms[1,2]. Thefirstnickelbathwas
formulated by Watts [3] in 1916, and this bath is still used
today because of its simplicity to control and low cost to
operate. Although the literature body on nickel electrodepo-
sition is voluminous, the attention will be paid mostly to the
area of nickel nucleation mechanisms. This area of research
is predominantly driven by the application of nickel electro-
chemistry for fuel cell development, resulting in numerous
nickel nucleation studies, mostly on vitreous carbon as the
preferred substrate [4–11].
The mechanisms of nickel deposition, derived from the
chronoamperometric current transients [4–9], indicate that
nickel nucleation on vitreous carbon proceeds through for-
mation of discrete nuclei, of conical [4–6] or hemispherical
Ni2 + OH ↔ NiOH
+
−
+
(1)
(2)
(3)
+
NiOH + e ↔ NiOH
NiOH + e ↔ Ni + OH
ads
−
ads
∗
However, the reaction mechanisms for acidic electrolytes of
the Watts type is not universally accepted [15]. The reac-
Corresponding author. Tel.: +1 208 885 6569; fax: +1 208 885 2855.
E-mail address: pesic@uidaho.edu (B. Pesic).
0
013-4686/$ – see front matter © 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2005.08.017

D. Grujicic, B. Pesic / Electrochimica Acta 51 (2006) 2678–2690
2679
tion mechanisms for basic solutions, involving another com-
plexing agent, ammonia, are not consistent either. Thus, the
study by Davison and Harrison [16] implies that reduction of
nickel–ammonia complexes proceeds through a single, two-
electron charge-transfer step, which is in contrast to the study
by Philip and Nicol [17], who found that nickel electrodepo-
sition is a two-stage single-electron process.
Because most of the nickel nucleation studies were per-
formed in acidic solutions of Watts type, leaving the ammon-
ical systems relatively unexplored, or incomplete [18], it was
decided to launch a comprehensive nickel nucleation study
in the later, ammoniacal, solutions. The experimental tech-
niques used were cyclic voltammetry and chronoamperom-
etry in combination with the morphological characterization
by atomic force microscopy. Vitrous carbon was selected as
electrode material.
Upon the completion of the experiment, surface mor-
phology of nickel deposits was characterized ex situ by
atomic force microscopy (VEECO, Digital Instruments,
Model Nanoscope IIIa-MultiMode, tapping mode in fluid).
Nickel deposits were characterized ex situ, under the pro-
tective layer of cyclohexane in order to prevent oxidation by
oxygen from air. Distribution–pH diagrams were constructed
by using thermochemical software Stabcal [19].
3. Results and discussion
3.1. Nickel ammonia chemistry
Nickel ion in aqueous solutions has octahedral coordina-
tion, similar to those of copper and cobalt, which allows for
accommodation of six ligands in its first hydration shell. For
each of the solution compositions in this study, the speciation
of nickel in ammonia solution was examined by constructing
the distribution–pH diagrams. Two such diagrams are pre-
sented in Fig. 1a and b.
There are four characteristic pH regions in Fig. 1a and
b. The first region is characterized by the free nickel ion
as the most stable species (pH region 0–1.5). In the second
region, between pH 1.5 and pH 8, the predominant nickel
species is a soluble, negatively charged sulfate complex.
2
. Experimental
Reagent grade NiSO4·5H2O was used to prepare 0.005 M,
0
.01 M and 0.05 M Ni2+ solutions. Solutions were pre-
pared with deionized water (Type I purity water, Barnstead
NANOpureII, 18.2 Mꢀ). Allsolutionscontained1 Mreagent
grade (NH4)2SO4 that served both as a buffer and as a sup-
porting electrolyte. The solutions were adjusted to pH 6 or
pH 9 by addition of dilute H2SO4 or NH4OH, respectively.
Deaeration of the solution in the electrochemical cell (15 ml)
by argon gas purging for 10 min, with the electrode retracted
fromthesolution, precededeachelectrochemicalexperiment.
All experiments were run under quiescent conditions, and
only freshly prepared solutions were used.
The electrochemical setup was a standard three-electrode
cell, with vitreous carbon as a working electrode, plat-
inum foil as a counter electrode, and a silver–silver chlo-
ride reference electrode (Cypress Systems microelectrode,
0
E = +0.222V). The vitreous carbon working electrode
h
2
(
0.442 cm ) was a non-porous disk (Sigri). The working
electrode was prepared by polishing with successively finer
grades of abrasive paper, followed by wet polishing with
1
␮m and 0.05 ␮m alumina on a wet polishing cloth. The final
step was polishing with dry 0.05 ␮m alumina on a coarsely
sanded glass plate. The electrode was then sonicated briefly
in deionized water to remove remaining alumina particles,
and then dried in a stream of nitrogen gas. This method of
surface preparation provides both electrochemically repro-
ducible surface properties (tested by measuring the peak sep-
arationofferrous/ferriccyanideredoxcouple), androughness
(root mean square) close to 1 nm that is suitable for the nucle-
ation studies. Only the dry polishing step was repeated after
every electrochemical experiment, which was found to be
sufficient to electrochemically re-activate the surface, while
reducing the polishing time required to achieve low surface
roughness. Electrochemical experiments were executed by a
potentiostat/galvanostat (Ametek, PAR 273A) controlled by
a PC running the electrochemical software (PAR, M270).
Fig. 1. Distribution–pH diagram for: (a) 0.01 M Ni2+ and (b) 0.05 M Ni2+ in
1
M (NH4)2SO4 solution. Note the appearance of hydroxy–sulfate (shaded)
in (b).

2
680
D. Grujicic, B. Pesic / Electrochimica Acta 51 (2006) 2678–2690
Fig. 2. Effect of scanning rate on the cyclic voltammograms for nickel–ammonia system at: (a) pH 6 and (b) pH 9. For comparison, dashed lines in (a) and (b)
2
+
are control cv-s containing no nickel ions. Insets represent the ratio of anodic to cathodic charge as a function of the scanning rate. Conditions: 0.005 M Ni
M (NH4)2SO4.
;
1
In the third region, from pH 8 to pH 12, there is a series
of nickel–ammonia complexes with successively increas-
metallic nickel (reaction (6)).
Ni(SO4)22 + e = NiSO4− + SO42−
−
(4)
(5)
(6)
2+
ing number of incorporated ammonia ligands, NiNH3 to
Ni(NH3)62 . In 0.05 M Ni solution, nickel hydroxy–sulfate
precipitates in the third region around pH 8. The fourth
region, beyond pH 12, is characterized by nickel hydroxide
as the only thermodynamically stable species. Representative
solutions for the present study were taken from the second
and third region, i.e. pH 6 was representing a nickel–sulfate
complex, while pH 9 was representing nickel–ammonia
complexes. The instability of 0.05 M Ni2+ solutions due to
precipitation of nickel hydroxy–sulfate has limited the
investigations for this nickel concentration only to cyclic
voltammetry.
+
2+
NiSO4 ⇔ NiSO4(ads)−
−
NiSO4(ads)− + e = Ni + SO42−
According to Davison and Harrison [16], for 0.005 M
2+
Ni concentration the chemical step of adsorption of
nickel–sulfate complex on glassy carbon is fast, i.e. both elec-
trochemical reactions are seen, as the overall rate-controlling
step is the reaction (6). With the increasing nickel ions con-
−
centration, the chemical step of NiSO4 adsorption, reaction
(5), becomes the rate-controlling step, as confirmed by the
studies of nickel electrodeposition from more concentrated
solutions [13–15,17].
3
.2. Cyclic voltammetry
−
The variation in the rate of NiSO4 adsorption is most
Cyclic voltammograms for the investigated solutions were
run from the initial voltage of 0 mV to the vertex potential of
1300 mV, and returned to the initial voltage. A typical set
likely related to the available type of deposition substrate.
Below certain Ni concentration range, the predominant depo-
sition substrate is glassy carbon, and vice versa, deposited
nickel metal becomes the deposition substrate for higher
concentration range. It is known that the concentration of
nickel ad-atoms on glassy carbon increases with the increase
of nickel ions concentration [8]. Low nickel concentrations
result in low glassy carbon surface coverage by nickel ad-
atoms, and consequently sparse nuclei population, i.e. glassy
carbon surface will mostly stay free. Because the adsorption
−
of cyclic voltammograms is presented in Fig. 2a and b.
According to Fig. 2a and b, the cyclic voltammograms for
each pH are characterized by one cathodic and one anodic
peak. The cv-s also include a loop characteristic for elec-
trodeposition [20].
Based on the thermodynamic speciation in Fig. 1a and b
and the results of cyclic voltammetry (Fig. 2a and b), the reac-
tion mechanisms for nickel electrodeposition are categorized
according to the solution pH as described below.
At pH 6, nickel electrodeposition proceeds in three, two
electrochemical and one adsorption, steps. In the first elec-
troreduction step (reaction (4)), nickelous sulfate ions is
formed, followed by its surface adsorption (reaction (5)). The
second electroreduction step is responsible for deposition of
−
rate of NiSO4 on metallic nickel is slow, due to inhibition
2
−
by adsorbed species (H(ads) and SO
) [9,12–14,21],
4(ads)
−
the adsorption and discharge of NiSO4 on glassy carbon
becomes the preferred route. As the result, glassy carbon sur-
face becomes covered by a large number of very small nickel
clusters in the initial stages of nickel electrodeposition. For
high nickel concentrations, the glassy carbon surface cover-

D. Grujicic, B. Pesic / Electrochimica Acta 51 (2006) 2678–2690
2681
age with nickel clusters is extensive, almost complete, thus
−
NiSO4 adsorption and subsequent reduction has to hap-
pen on nickel surface. However, as already stated, because
−
the adsorption rate of NiSO4 on metallic nickel is slow, it
becomes the rate-determining step.
The importance of absorption during nickel deposition
was recognized by Chassaing et al. [13], who discussed the
+
2−
inhibitory effects of H and SO4 ions. The inhibition by
+
H was the result of its reduction to atomic state, subsequent
adsorption of Hads atoms and nickel hydride formation. In
this study, peak Ic was immediately followed by the increase
of cathodic current that corresponds to reduction of hydro-
gen ions and formation of adsorbed hydrogen atoms, which
recombined to form hydrogen gas. The question regarding the
site for hydrogen ion reduction, glassy carbon versus nickel,
was resolved by running control cv-s in nickel-free solutions
(
Fig. 2a and b). Zero current found demonstrates that the
Fig. 3. Cyclic voltammograms of electrodeposited nickel in solutions con-
taining 1 M Na2SO4 at pH 3 (dashed line) and pH 6 (full line). Scanning rate
reduction on glassy carbon can be ruled out, i.e. that hydro-
gen ion reduction takes place only on the surface of newly
formed nickel clusters, supporting the suggested inhibiting
role of adsorbed hydrogen.
−
1
20 mV s
.
The inhibitory presence of SO42 is the result of com-
petitive occupation of adsorption sites on nickel surface. The
electrode potential is essential for adsorption/desorption of
−
centrations reaction (8) becomes the rate-controlling step
[17].
Anodic branches of cyclic voltammograms contain a sin-
gleanodicpeakIa, regardlessofthesolutionpH. Thelowratio
of anodic to cathodic charge (insets in Fig. 2a and b) indi-
cates incomplete nickel oxidation, which was also confirmed
by Philip and Nicol, who suggested that formation of passive
film in ammonical solutions leads to incomplete nickel oxida-
tion [25]. To survey the electrode surface conditions after the
completionofpeakIa, thecyclicvoltammogramswerecycled
three times, in order to build sufficient amount of a reaction
product. It was found that with repeated cycling the onset
of cathodic peak became more positive by 200–400 mV, and
that the magnitude of cathodic and anodic peaks increased.
AFM examination of the electrode was performed for two
characteristic potentials: (1) at the crossover potential of the
first cv cycle and (2) at the beginning of the second cv cycle
(Fig. 4a–d).
2−
SO4 . During electrodeposition by scanning in cathodic
direction, SO42 ions undergo desorption process as the
potential becomes more negative, as verified by the in situ
scanning tunneling microscopy during electrodeposition of
copper (Broekmann et al. [22], Wilms et al. [23] and Yan
et al. [24]). In our recent study on copper electrodeposition,
we were also able to confirm, by means of cyclic voltam-
metry, the sulfate desorption process. Similar approach was
also used in this study. First, nickel was electrodeposited
from nickel–ammonia solution at pH 6 at −1300 mV for 20 s.
After deposition, the nickel–ammonia solution was replaced
with 1 M Na2SO4 solution at either pH 3 or pH 6, and the
cyclic voltammograms recorded in the region −700 mV to
−
−
1000 mV, where sulfate desorption was expected to take
place. The results are presented in Fig. 3.
The peak observed for pH 3 solution in Fig. 3 corresponds
todesorptionofsulfateanions, andtoaconsequentadsorbate-
induced hydrogen reduction reaction [22–24]. Because of the
lower hydrogen ion concentration, the same peak at pH 6
is convoluted with the background current becoming more
difficult to distinguish.
At pH 9, electrodeposition of nickel proceeds according
to the mechanisms proposed by Davison and Harrison [16]
and Philip and Nicol [17] (reactions (7)–(9)):
Nickel clusters appear to be well defined at the crossover
potential for each studied pH (Fig. 4a and c), but their appear-
ance becomes “fuzzy” at the end of anodic branch (Fig. 4b
and d). The fuzzy appearance of nickel clusters, including
the visible horizontal streaks, is characteristic for the AFM
tip–sample interaction with the soft, powdery phase, extri-
cated by the AFM tip, resulting in reduced image quality.
The exact nature of the nickel oxidation products (hydrox-
ide, oxide) or the mechanism of nickel passivation was not
determined, as this was beyond the scope of the present study.
Ni(NH3)n2 + e = Ni(NH3)m+ + (n − m)NH3
+
(7)
(8)
(9)
Ni(NH3)m ⇔ Ni(NH3)m(ads)+
+
3.3. Chronoamperometry and nucleation modeling
+
0
3.3.1. Effect of deposition potential
Ni(NH3)m(ads) + e = Ni + mNH3
The mechanisms of nickel nucleation from ammoniacal
solutions were examined by employing the chronoampero-
metric (ca) technique. The ca-s were initiated from 0 mV to a
Reaction (9) represents the rate-controlling step at lower
nickel concentrations [16], while at higher nickel con-

2
682
D. Grujicic, B. Pesic / Electrochimica Acta 51 (2006) 2678–2690
Fig. 4. AFM of the electrode surface at two characteristic cv potentials: pH 6—(a) crossover of the first cv (−960 mV); (b) onset of the second cv (0 mV); pH
2
+
−1
9
1
—(c) crossover of the first cv (−980 mV); (d) onset of the second cv (0 mV). Conditions: 0.01 M Ni ; 1 M (NH4)2SO4; scanning rate 20 mV s , scale bar
␮m.
ꢆ
ꢇ
ꢈꢉ
range of deposition potentials, chosen from the cv-s in Fig. 2a
and b to be more negative than the peak Ic. Recorded current
transients are presented in Fig. 5a–d.
ꢂ
ꢃ2
2
2
i
1.2254
t
=
1 − exp −2.3367
(11)
2
t
i
tm
m
tm
The mechanisms of nickel nucleation were compared to
the theoretical models derived by Scharifker and Hills [26].
The models utilize the coordinates of the ca peaks in order
to distinguish between the two limiting nucleation mecha-
nisms: instantaneous and progressive. Instantaneous nucle-
ation corresponds to immediate activation of all nucleation
sites, and the rate of further nuclei formation is negligible in
the time frame of the experiment. During progressive nucle-
ation, the rate of new nuclei formation in the time frame of
the experiment is not negligible. Instantaneous and progres-
sive nucleation models are represented by Eqs. (10) and (11),
respectively:
where im and tm are the current and the time, as respec-
tive peak coordinates. Data from Fig. 5a–d are replotted
in Fig. 6a–d with reduced current-time coordinates, along
with the theoretical models for instantaneous and progres-
sive nucleation.
According to Fig. 6a–d, nickel nucleation follows the pro-
gressive nucleation model, regardless of the solution pH or
nickel concentration. Negative deviation of chronoampero-
grams from the models at around t/tmax = 2, was also observed
by others [11], although on gold substrate and from ammonia-
free solutions. The observed negative deviation from the
model is most likely related to reduction of hydrogen ions
and hydrogen evolution. The simultaneous increase in levels
of current plateaus with more cathodic deposition potentials
under all experimental conditions in Fig. 5a–d is a clear indi-
ꢀ
ꢁ
ꢂ
ꢃꢄꢅ
2
2
i
1.9542
t
=
1 − exp −1.2564
(10)
2
t
t
i
tm
m
m

D. Grujicic, B. Pesic / Electrochimica Acta 51 (2006) 2678–2690
2683
Fig. 5. Chronoamperometric deposition of Ni as a function of pH and Ni concentration in 1 M (NH4)2SO4: (a and b) pH 6 and (c and d) pH 9.
cator of the additional current due to hydrogen ions reduction
that causes the observed deviation from the models. This
invalidates the interpretation of the results in Fig. 6, because
the model given by Eqs. (10) and (11) cannot distinguish
between the currents contributed by the nickel ion and hydro-
gen ion reduction.
Scharifker–Hills model also allows the type of nucleation
to be determined by analyzing the ascending parts of the
chronoamperograms, prior to the overlap of nuclei diffusion
zones. At this stage, the reduction of hydrogen ions is not so
prominent, and majority of current corresponds to reduction
of nickel ions. The ascending portions of the current–time
Fig. 6. (a–d) Chronoamperograms from Fig. 5a–d presented in reduced current–reduced time coordinates. Theoretical models for instantaneous and progressive
nucleation are represented by solid and dashed lines, respectively.

2
684
D. Grujicic, B. Pesic / Electrochimica Acta 51 (2006) 2678–2690
Fig. 7. Ascending parts of the current time transients (symbols) in: (a) i–t1/2 and (b) i–t3/2 domain. Nickel concentration 0.01 M.
1
/2
the type of nucleation mechanisms are P2 and P3:
transients should exhibit linear dependence in either the i–t
or i–t3 domains, for instantaneous and progressive nucle-
ation mechanisms, respectively [26] (Fig. 7a and b).
The results in Fig. 7a and b indicate that the initial portions
/2
P2 = N0πkD
P3 = A
(13)
(14)
of chronoamprograms can better be linearized in the i–t1
/2
domain, which corresponds to the instantaneous nucleation
regime. It should be noted that the linearization is not ideal
in either of the two domains.
where N0 is the density of nucleation sites, A the rate of nucle-
ation, k = (8πc0/ρ)0 , c0 the concentration of metal ion, ρ the
depositdensityandDisthediffusioncoefficientofmetalions.
The comparison between the two experimental chronoamper-
ograms and the current–time transients that were generated
by the non-linear Marquardt–Levenberg fitting algorithm is
presented in Fig. 8a and b.
.5
Because of the apparent discrepancy in the assigned nickel
nucleation mechanisms, which change depending on the
employed diagnostic criteria, a newly proposed nucleation
model by Palomar-Pardave et al. which takes into account
the contribution from the concurrent hydrogen ion reduction,
was also used [27]. The model, given by Eq. (12) represents
the sum of the metal ion and hydrogen ion current contribu-
tions:
∗
Parameters P , P2, P3 and P4 were varied during the curve
1
fitting, and are presented in Table 1 for some of the chronoam-
perograms from Fig. 5a–d.
AccordingtoTable1, bothAandtheN0 increasewithmore
cathodic deposition potentials. The rate of new nickel nuclei
∗
−1/2
)
i
(t) = (P + P4t
total
1
−
5
−4
ꢂ
ꢀ
ꢁ
ꢄꢅꢃ
formation per active site is slow, with only 10 to 10
1
− exp(−P3t)
P3
−1 −1
new sites activated per second, compared to the 10
s
to
×
1 − exp −P2 t −
−
2
−1
10
s
when cobalt is deposited [27]. The density of nucle-
(12)
10
−2
ation sites on the other hand is on the order of 10 cm
to 1011 cm , which is higher than that of cobalt deposited
−2
∗
Complete definition of the parameters P , P2, P3 and P4 is
on the same substrate, where the theoretical density of the
1
6
−2
given in Appendix A. The two parameters most relevant for
nucleation sites was on the order of 10 cm [27], while
Fig. 8. Comparison between the experimental chronoamperograms (symbols) at: (a) pH 6 and (b) pH 9 and the theoretical current transients (solid lines)
2
+
obtained by fitting the expression (12) to the experimental data. Conditions: 0.005 M Ni ; Edep = −1300 mV.

D. Grujicic, B. Pesic / Electrochimica Acta 51 (2006) 2678–2690
2685
Table 1
Values of the parameters P , P2, P3 and P4, obtained by non-linear fitting of the experimental chronoamperograms from Fig. 5a–d with Eq. (12) [27]
∗
1
∗
−2
−1
−1
P4 (mA cm s0.5
−2
−1
−2
N0 (cm
[
Ni2+] (M)
.005
pH 6
Edep (mV)
P
(mA cm
)
P2 (s
)
P3 (s
)
)
A (s
)
)
1
0
−
5
4
−5
−4
11
11
−1250
1.03
3.31
11858
28907
3 × 10
4.34
3.85
3 × 10
1.00 × 10
−
−1300
1 × 10
1 × 10
2.45 × 10
−
−
5
5
−5
−5
10
11
pH 9
−1250
0.04
1.89
7519
28826
2 × 10
8 × 10
3.22
1.36
2 × 10
8 × 10
6.37 × 10
2.44 × 10
−1300
0
.01
pH 6
−
−
5
5
−5
−5
10
11
−1250
1300
3.27
5.64
9823
18975
6 × 10
6.53
7.22
6 × 10
5.88 × 10
−
9 × 10
9 × 10
1.14 × 10
−
5
−5
10
11
pH 9
−1250
0.92
3.57
9193
29630
4 × 10
4.67
3.18
4 × 10
5.50 × 10
1.77 × 10
−
4
−4
−1300
1.7 × 10
1.7 × 10
the experimentally determined cluster population densities
According to Fig. 10a and b, nickel clusters are of uneven
sizes in the early stages of deposition, a known character-
istic for progressive nucleation mechanisms. The clusters
size is very small, particularly at pH 6 where the average
height of a nucleus is only several nanometers. Despite their
small size, the overlapping clusters can be still visualized. At
pH 9, the numerous overlapped (fused) clusters can readily
be seen.
9
−2
10
−2
were on the order of 10 cm to 10 cm [28]. It follows
that the initial nickel nucleation on a very large number of
active nucleation sites exhausts their further availability, thus
limiting the growth of nickel deposit to the already formed
nuclei. This scenario corresponds to the instantaneous nucle-
ation regime, and is in agreement with the results in Fig. 7a
and b.
The effect of deposition potential on morphology of nickel
clusters is presented in Fig. 9a–f.
The morphology of nickel clusters electrodeposited from
solutions with higher Ni concentration (0.01 M) is presented
in Fig. 11a and b.
According to analysis of Fig. 11a, at pH 6, the increase
of nickel concentration from 0.005 M to 0.01 M resulted in
more than a three-fold increase of clusters population density
According to Fig. 9a–f, the clusters population density at
pH 6 is much higher than at pH 9. This may likely be the
result of higher surface coverage by adsorbed nickel–sulfate
complex at pH 6, than by nickel–ammino complex at pH 9,
as discussed above.
1
0
−2
10
−2
N, from 1.5 × 10 cm to 5.5 × 10 cm , respectively.
2
+
At pH 6 (Fig. 9a–c), there is an increase in the clusters pop-
ulation density between −1200 mV and −1250 mV, with uni-
form surface coverage. However, the last image in this series
shows coarsening of nickel clusters, with regions of barren
glassy carbon still visible. The evidence for coarsening of the
clusters was previously found in the studies of nickel elec-
trodeposition on pyrolytic graphite [13] and in the studies of
cobalt electrodeposition on glassy carbon [28]. Both studies
suggest that hydrogen evolution is responsible for the aggre-
gation of small clusters into larger ones during the course
of electrodeposition, presumably by physical interaction
between hydrogen bubbles and nickel clusters. Another con-
sidered explanation, the aggregation of nickel clusters due to
magnetostatic interactions, was unlikely because the average
diameters of nickel clusters found in this study were below
the calculated superparamagnetic–ferromagnetic limit of the
face centered cubic nickel spherical particles (78 nm) [29].
At pH 9, deposition at more cathodic potentials resulted in
the increase of nickel clusters population density (Fig. 9d–f),
without the coarsening effect observed at pH 6.
At pH 9, Ni-clusters size increased for 0.01 M Ni , but
the clusters population density stayed constant to around
6 × 10 cm , irrespective of nickel concentration. The clus-
ter population densities are well in agreement with the den-
sities of active sites determined from the parameters of Eq.
(12).
9
−2
Qualitative interpretation of the degree of nucleation inhi-
−2
bition can be obtained from the log N versus η plots, as
the saturation clusters population density is proportional to
−
2
nucleation overvoltage, η [30]. Nucleation overvoltage,
η, was determined as the difference between the deposition
potential and the crossover between the cathodic and anodic
2+
current traces (the redox Ni /Ni potential [20]). The values
of redox potentials were −940 mV and −970 mV at pH 6 and
pH 9, respectively.
According to the results in Fig. 12, the log N versus
lines are characterized with very small slopes for all
−
2
η
investigated conditions, an indication of the highly inhibited
nucleation mechanisms. A high degree of inhibition is also
consistent with the results obtained in the study of nickel
deposition from ammonia-free solutions [10–15], and with
the low rates on nucleation determined from Eq. (12).
In order to better visualize the morphology of Ni-clusters
in the earlier stages of development, the AFM examination
was performed after two seconds of deposition, right after
the formation of a chronoamperometric peak, i.e. at the time
when diffusional zones overlapped. The AFM images, for
solutions at pH 6 and pH 9, are presented in Fig. 10a and b.
3.3.2. Effect of solution temperature
The effect of solution temperature was investigated in the
◦
range 5–55 C. Nickel electrodeposition was carried out at

2
686
D. Grujicic, B. Pesic / Electrochimica Acta 51 (2006) 2678–2690
Fig. 9. AFM height images—effect of deposition potential on the morphology of electrodeposited nickel: pH 6—(a) −1200 mV, (b) −1250 mV and (c)
2
+
−
1300 mV; pH 9—(d) −1200 mV, (e) −1250 mV and (f) −1300 mV. Conditions: 0.005 M Ni ; 1 M (NH4)2SO4; scale bar 400 nm; vertical height scale range
0
–75 nm.
−
1250 mV for 10 s. from both pH 6 and pH 9 solutions.
where N is the number of clusters, A the constant, Ea the acti-
vation energy, R the gas constant and T is the temperature.
Increase in the clusters population densities with the increas-
ing temperature is most likely related to the thermal activation
of the glassy carbon electrode, and the consequent increase in
the number of available active nucleation sites, and increasing
mobility of nickel ions at elevated temperatures.
Different activation energies at pH 6 and pH 9, 15 kJ
and 23 kJ, respectively, could be related to different desol-
vation pathways for ammino and aquo solvated nickel ions,
Reduced current transients (not presented here) indicate
that nickel electrodeposition mechanisms remain progres-
sive, regardless of the solution temperature for both pH 6
and pH 9. Clusters population densities as a function of tem-
perature are presented in Fig. 13.
According to Fig. 13, the clusters population density fol-
lows the Arrhenius type of temperature dependence, Eq. (15):
N = A × e−
Ea/RT
(15)

D. Grujicic, B. Pesic / Electrochimica Acta 51 (2006) 2678–2690
2687
Fig. 10. Nickel clusters formed after 2 s of electrodeposition from: (a) pH 6 and (b) pH 9 solutions. Conditions: electrodeposition performed from 0.005 M
2
+
Ni + 1 M (NH4)2SO4 solution at −1250 mV. Note that for clarity, the images are binary coded, black representing nickel clusters and white representing
glassy carbon surface. Scale bar 200 nm.
Fig. 11. AFM height images of nickel clusters deposited from 0.01 M Ni2+ + 1 M (NH4)2SO4 solutions at: (a) pH 6 and (b) pH 9. Conditions: deposition
2+
potential −1200 mV; tdep = 10 s; height = 75 nm; scale bar 400 nm (compare with morphology in Fig. 7a and d for deposition of Ni from 0.005 M Ni ).
similar to those calculated for cobalt [31]. Ab initio study
by Mendoza-Huizar et al. [31] shows that desolvation of
cobalt–ammino complex is less energetically demanding pro-
cess than the desolvation of cobalt–aquo complex, which
was experimentally confirmed by the lower electrocrystal-
lization overpotential of cobalt–ammino complex [28,32]. In
the present study, the electrocrystallization overpotential of
nickel–ammino complex is higher than that of nickel–aquo
complex (Fig. 2a and b). Higher overpotential, together with
higher activation energy found from the slope in Fig. 13
may indicate that the desolvation of nickel–ammino com-
plex represents energetically more demanding process. This
conclusion could further be supported by an ab initio study
of nickel–aquo–ammino complex desolvation and electrore-
duction, similar to that for cobalt [31].
3.3.3. Effect of conditioning potential
Theglassycarbonelectrodewassubjectedtoelectrochem-
ical conditioning at either +1 V or −1 V for 2 min in the
Fig. 12. Measured clusters population density as a function of deposition
overvoltage.

2
688
D. Grujicic, B. Pesic / Electrochimica Acta 51 (2006) 2678–2690
nickel-free solutions containing 1 M (NH4)2SO4 at pH 6 or
pH 9. Nickel was then deposited on the conditioned elec-
trode chronoamperometrically for 10 s, at −1250 mV, from
2+
pH matching solutions containing 0.005 M Ni . The mor-
phology of electrodeposited nickel was then examined by the
AFM. The electrode surface images, after the conditioning-
deposition sequence, are presented in Fig. 14a–d.
According to Fig. 14a and c, conditioning at negative
potentials results in reduction of clusters population den-
sity in both pH solutions. Electrode conditioning at positive
potentials, on the other hand, results in a significant increase
of clusters population density (Fig. 12b and d). Similar con-
ditioning effect was observed earlier for copper electrode-
posited from ammoniacal solutions [33]. The explanation
for the observed effect is difficult to provide, because of
the complex nature of glassy carbon surface with respect
to the presence of oxide groups serving as catalysts for the
electron transfer [34], and the unknown content of surface
impurities.
Fig. 13. Ni-clusters population density, determined from the AFM image
analysis, as a function of temperature. Full lines are linear fits for data at pH
6
and pH 9. Conditions: Edep = −1250 mV; deposition time 10 s; 0.005 M
2+
Ni ; 1 M (NH4)2SO4.
Fig. 14. The effect of electrochemical modification of the glassy carbon electrode on the size and population density of electrodeposited nickel clusters: pH
—(a) Econd = −1 V and (b) Econd = + 1 V; pH 9—(c) Econd = −1 V and (d) Econd = + 1 V. Conditions: Edep = −1250 mV; scale bar 500 nm.
6

D. Grujicic, B. Pesic / Electrochimica Acta 51 (2006) 2678–2690
2689
4
. Conclusions
1/2
2
FD c0
P4 =
1/2
π
The mechanisms of electrochemical deposition of nickel
from ammoniacal solutions, on glassy carbon, were investi-
gated by cyclic voltammetry and chronoamperometry. Corre-
sponding nickel morphology was examined by atomic force
microscopy. The following conclusions can be made:
c0 is the bulk concentration of metal ions, M the molar mass
of deposit, ρ the density of deposit, zPRF the molar charge
transferred during the proton reduction process, kPR the rate
constant of the proton reduction reaction, N0 the density of
nucleation sites, k = (8πc0/␳)0 , D the diffusion coefficient
of metal ions and A is the rate of nucleation.
.5
1
. Deposition involves two electrochemical and one chem-
ical (adsorption) steps, i.e. nickel nucleates according to
the ECE mechanisms. At pH 6, and low nickel concentra-
−
tion (0.005 M), adsorption of Ni(SO4) is fast. The rate of
−
Ni(SO4) adsorption decreases with the increasing nickel
References
concentration. At pH 9, the mechanisms are similar except
+
that the adsorbing species are Ni(NH3)m , i.e. deposition
[1] W.H. Remington, U.S. Patent 82,877 (1868).
[2] I. Adams, U.S. Patent 93,157 (1869).
happens from Ni–ammine complexes.
. Electrodeposition of nickel is inhibited by SO42− and
[3] O.P. Watts, Trans. Am. Electrochem. Soc. 23 (1916) 395.
2
[4] M. Abyaneh, Electrochim. Acta 27 (1982) 1329.
[5] M. Abyaneh, M. Fleischmann, Electrochim. Acta 27 (1982) 1513.
[6] M. Abyaneh, W. Visser, E. Barendrecht, Electrochim. Acta 28 (1983)
285.
hydrogen adsorption. According to the cyclovoltammet-
ric and morphological studies, the degree of inhibition is
higher at pH 6 than at pH 9.
[
[
[
7] J. Amblard, M. Froment, G. Maurin, N. Spyrellis, E. Trevisan-
Souteyrand, Electrochim. Acta 28 (1983) 909.
8] Chr. Bozhkov, Chr. Tzvetkova, St. Rashkov, J. Electroanal. Chem.
3
. Chronoamperometric study indicates that nickel nucle-
ation conforms to the instantaneous nucleation mecha-
nism, under all investigated conditions. The statistical
analysis of AFM images showed higher clusters popula-
tion density when nickel was deposited from sulfate (pH
296 (1990) 453.
9] E. Gomez, C. Muller, R. Pollina, M. Sarret, E. Valles, J. Electroanal.
Chem. 333 (1992) 47.
[
[
[
[
10] M. Fleischmann, A. Saraby-Reintjes, Electrochim. Acta 29 (1984)
9.
11] A.N. Correia, S.A.S. Machado, L.A. Avaca, J. Electroanal. Chem.
88 (2000) 110.
12] A.G. Munoz, D.R. Salinas, J.B. Bessone, Thin Solid Films 429
(2003) 119.
6
) than from amino complexing solutions (pH 9). Cluster
6
population densities were on the same order as the theo-
retically calculated number of active sites.
4
4
. Nickel clusters population density increased with the
increase of solution temperature, following the Arrhe-
nius type of dependence. The difference in calculated
activation energies may be related to different energetic
requirements for nickel–ammino and nickel–aquo ions
desolvation. An ab initio study may be required for further
confirmation.
13] E. Chassaing, M. Joussellin, R.J. Wiart, J. Electroanal. Chem. 157
(1983) 75.
[14] R. Wiart, Electrochim. Acta 35 (1990) 35.
[15] A. Saraby-Reintjes, M. Fleischmann, Electrochim. Acta 29 (1984)
557.
[16] W. Davison, J.A. Harrison, J. Electroanal. Chem. 36 (1972) 399.
[
17] H.I. Philip, M.J. Nicol, The Electrodeposition of Nickel from Ammo-
niacal Solutions, Project Report No. 1804, National Institute for
Metallurgy, Randburg, South Africa, April 1976.
Acknowledgements
[
[
18] M.P. Zach, R.M. Penner, Adv. Mater. 12 (2000) 878.
19] H.-H. Huang, STABCAL—Stability Calculation for Aqueous Sys-
tems, Montana Tech of the University of Montana, Butte, MT.
20] Southampton Electrochemistry Group, in: T.J. Kemp (Ed.), Instru-
mental Methods in Electrochemistry, Ellis Horwood Ltd., Chichester,
UK, 1985.
This work has been sponsored by the U.S. Department of
Defense, Department of the Navy, Office of Naval Research,
Grant Number N000140110829, and The National Science
Foundation RII Award (EPS 0132626).
[
[
21] A. Lachenwitzer, O.M. Magnussen, J. Phys. Chem. B 104 (2000)
7
424.
[22] P. Broekmann, M. Wilms, K. Wandelt, Surf. Rev. Lett. 6 (5) (1999)
07.
[23] M. Wilms, P. Broekmann, C. Stuhlmann, K. Wandelt, Surf. Sci. 416
1998) 121.
Appendix A
9
∗
This section contains definitions of parameters P , P2, P3
and P4 in Eq. (12). For further explanation refer to [27].
1
(
[24] J.W. Yan, J.M. Wu, Q. Wu, Z.X. Xie, B.W. Mao, Langmuir 19
ꢂ
ꢃ
1/2
(2003) 7948.
2
c0M
∗
P = P1
[25] M.J. Nicol, W.M. Stapleton, A.N.E. Balaes, An Electrochem-
ical Investiagtion of the Dissolution of Copper, Nickel and
Copper–Nickel Alloys in Ammonium Carbonate Solutions, Project
Report No. 1630, National Institute for Metallurgy, Randburg, South
Africa, May 1974.
1
πρ
P1 = zPRFkPR
P2 = N0πkD
P3 = A
[
[
26] B. Scharifker, G. Hills, Electrochim. Acta 28 (7) (1983) 879.
27] M. Palomar-Pardave, B.R. Scharifker, E.M. Arce, M. Romero-Romo,
Electrochim. Acta 50 (2005) 4736.
[28] D. Grujicic, B. Pesic, Electrochim. Acta 49 (2004) 4719.

2
690
D. Grujicic, B. Pesic / Electrochimica Acta 51 (2006) 2678–2690
[
[
[
29] B.D. Cullity, Introduction to Magnetic Materials, Pearson Addison
Wesley, Boston, MA, 1972, p. 614.
30] A. Serruya, J. Mostany, B.R. Scharifker, J. Electroanal. Chem. 446
[32] M. Palomar-Pardave, I. Gonzales, A.B. Soto, E.M. Arce, J. Elec-
troanal. Chem. 443 (1998) 125.
[33] D. Grujicic, B. Pesic, Electrochim. Acta 50 (2005) 4426.
[34] P. Chen, R.L. McCreery, Anal. Chem. 68 (1996) 3958.
(1999) 39.
31] L.H. Mendoza-Huizar, M. Palomar-Pardave, J. Robles, Electrochim.
Acta 46 (2001) 2749.