Electroless nickel-plating on AZ91D magnesium alloy : effect of substrate microstructure and plating parameters

Surface and Coatings Technology 179 (2004) 124–134
Electroless nickel-plating on AZ91D magnesium alloy: effect of substrate
microstructure and plating parameters
Rajan Ambata,*, W. Zhoub
Metallur
a
gy and Materials, School of Engineering, The University of Birmingham, Birmingham B15 2TT, UK
School
b
of Mechanical and Production Engineering, Nanyang Technological University, Singapore 639798, Singapore
Received 17 June 2002; accepted in revised form 14 July 2003
Abstract
Electroless nickel-plating on AZ91D magnesium alloy has been investigated to understand the effect of substrate microstructure
and plating parameters. The initial stage of the deposition was investigated using scanning electron microscopy (SEM) and energy
dispersive X-ray analysis on substrates plated for a very short interval of time. The early stage of growth was strongly influenced
by the substrate microstructure. Plating was initiated on b-phase grains probably due to the galvanic coupling of b and eutectic
a-phase. Once the b-phase was covered with the coating, it then spread onto eutectic a and primary a-phase. The coating
produced with the optimised bath showed 7 wt.% phosphorus with a hardness of approximately 600–700 VHN. The optimum
ligand to metal ion ratio was found to be 1:1.5, while the safe domain for thiourea (TU) was in the range of 0.5–1 mgyl.
Fluoride was found to be an essential component of the bath to plate AZ91D alloy with an optimum value of 7.5 gyl. The
presence of 0.25–0.5 mgyl mercapto-benzo-thiosole (MBT) found to accelerate the plating process.
2003 Elsevier B.V. All rights reserved.
Keywords: Magnesium alloy; Electroless nickel; Microstructure; Plating parameters
1. Introduction
those with high purity, have good resistance to atmos-
pheric corrosion w3x. However, the addition of alloying
Magnesium and its alloys, with one quarter of the
elements modifies the corrosion behaviour in such a
density of steel and only two-thirds that of aluminium
way that it can be beneficial or deleterious. The standard
and a strength to weight ratio that far exceeds either of
electrochemical potential of magnesium is y2.4 V vs.
these, fulfill the role admirably as an ‘ultra light weight’
NHE, even though in aqueous solutions magnesium
alloy. Hence, these alloys have obviously become the
shows a potential of y1.5 V due to the formation of
choice for weight reduction in portable microelectronics,
Mg(OH)2 film w4x. Consequently, magnesium dissolves
telecommunications, aerospace and automobile applica-
rapidly in aqueous solutions by evolving hydrogen
tions etc.
below pH 11.0, the equilibrium pH value for Mg(OH)2
w
The magnesium–aluminium system has been the basis
4x.
of the most widely used magnesium alloys since these
Although the addition of several alloying elements
materials were introduced in Germany during the First
such as aluminium, zinc and rare earths have been
World War. Most of these alloys contain 8–9% alumin-
reported w5–12x to improve the corrosion resistance,
ium with small amounts of zinc w1,2x.
technologically that does not satisfy the requirement for
A serious limitation for the potential use of several
several applications. Hence, the application of a surface
magnesium alloys and AZ91 in particular, is their
engineering technique is the most appropriate method
to further enhance the corrosion resistance. Among the
susceptibility to corrosion. Magnesium alloys, especially
various surface engineering techniques that are available
for this purpose, coating by electroless nickel is of
*Corresponding author. Tel.: q44-121-414-5217; fax: q44-121-
414-5232.
special interest especially in the electronic industry, due
E-mail address: R.Ambat@bham.ac.uk (R. Ambat).
to its conductivity and several other engineering prop-
0257-8972/04/$- see front matter
2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0257-8972(03)00866-1

---

R. Ambat, W. Zhou / Surface and Coatings Technology 179 (2004) 124–134
125
Table 1
Chemical composition of the AZ91D alloy (in wt.%)
Al
Mn
Ni
Cu
Zn
Ca
Si
K
Fe
Mg
9.1
0.17
0.001
0.001
0.64
-0.01
-0.01
-0.01
-0.001
Bal
erties. Electroless nickel is well known for its corrosion
polished on a diamond wheel using 6-mm diamond
resistance and hardness w13–18x. However, the nickely
paste. The polished specimens were thoroughly washed
Mg system is a classical example of cathodic coating
with water before passing through the pre-cleaning
on an anodic substrate. Hence, the porosity in the coating
schedule as shown in Table 2. The initial weight and
might influence the corrosion behaviour and service
dimensions of the specimens were measured prior to
lifetime of the electroless nickel-plated magnesium. The
pre-cleaning steps. Immediately after the fluoride acti-
protective ability of electroless nickel on many engi-
vation (last step in the pre-cleaning process), the spec-
neering materials is limited by the porosity in the coating
imen was quickly transferred to the coating bath (500
w19–26x.
ml) in a glass container placed in a constant temperature
Being a highly active metal, electroless plating of
water bath to maintain the required temperature. A fresh
magnesium alloy needs special bath formulations and
bath was used for each experiment to avoid any change
pre-cleaning treatments. Hence, the direct plating of
in concentration of bath species. The bath composition
magnesium is still a challenge for the researchers. The
and other parameters used in the present work are given
process becomes more complicated on AZ91 alloy due
in Table 3. The specimens were coated for the required
to the microstructural heterogeneity owing to the une-
length of time, removed from the bath, washed with
qual distribution of aluminium within the three constit-
water and acetone and air-dried. Final weight of the
uent phases namely primary a, eutectic a and b phases
specimen was determined and the coating rate in mmyh
w26x. Therefore, the substrate material is electrochemi-
was calculated from the weight gain. In this work, the
cally heterogeneous and each constituent behaves differ-
initial weight of the specimen was measured prior to
ently to the plating bath. The available information on
pre-cleaning steps. This is to avoid any surface oxidation
electroless nickel-plating of magnesium alloys is very
due to time delay in transferring the substrate to the
limited. This paper reports the work carried out on
plating bath after fluoride activation step. However, to
electroless nickel-coating of AZ91D magnesium alloy,
estimate the error in the weight gain calculation as a
more specifically, the effect of substrate microstructure
result of dissolution of substrate material during the pre-
on coating nucleation and the effect of various bath
cleaning steps, a control experiment was carried out to
parameters.
measure the weight loss during the pre-cleaning process.
The loss in weight during the pre-cleaning process was
2. Experimental
found to be insignificant, although the weight gain
values reported were corrected for this weight loss.
The substrate material used for the present investiga-
Duplicate experiments were conducted in each case, and
tion was AZ91D ingot-cast alloy. The chemical compo-
the coating rate reported is the average of two experi-
sition of the alloy is given in Table 1. Rectangular
ments. The coated specimens were characterised to
coupons of size 20=40=4 mm were used for the
evaluate the coating performance. Coating morphology
investigation. The surface of the substrate material was
was analysed using optical (Axiolab ZEISS model) and
wet-ground (using water) on 1000 grade SiC paper and
scanning
electron
microscope
(JEOL Model No.
Table 2
Table 3
Optimised pre-cleaning steps
Optimised bath composition and coating parameters
Ultrasonic degreasing using acetone
Bath constituents and parameters
Quantity
x
Rinse in 10% NaOH at 60 8C for 5 min
Basic nickel carbonate
9.7 gyl
x
(NiCO3 2Ni(OH)2 4H O
2
)
Water rinse
Citric acid
5.2 gyl
x
Ammonium bifluoride
7.5 gyl
6% chromic acid–5% nitric acid pickling for 45 s
Hydrofluoric acid
11 mlyl
x
Thiourea (TU)
1 mgyl
Water rinse
Sodium hypophosphite
20 gyl
x
Ammonium hydroxide
To adjust pH
Fluoride activation in HF (250 ml 70% HFyl) for 10 min
Temperature
80 8C
x
Agitation
Mild-
Water rinse
mechanical

---

126
R. Ambat, W. Zhou / Surface and Coatings Technology 179 (2004) 124–134
Fig. 1. AZ91 substrate: (a) microstructure and (b) Al and Zn concentration across the b-phase along the line shown in the picture.
5600LV). EDX analysis was used for the determination
In the present investigation, coating formed in optim-
of phosphorus. Microhardness measurements were car-
ised bath composition was found to be amorphous
ried out using Shimadzu hardness testing machine
possibly with microcrystalline areas (Fig. 3a). The broad
(HMV 2000 model) using a load of 100 g. The structure
peak in the diffractogram corresponds to the dominant
of the deposit was determined using X-ray diffraction
peak for nickel. Although not shown, the X-ray diffrac-
technique (Philips, Model No. PW1830).
tion studies of the deposit formed in bath with TU,
For studying the coating nucleation, coupons were
MBT and MBTqTU also revealed amorphousymicro-
deposited for very short internals of time namely 1–5
crystalline structure. Fig. 3b shows the microstructure
min, and were analysed using SEM and EDX.
of the electroless nickel-coating along the thickness
direction. The picture reveals a lamellar structure due to
3. Results
the variation in the composition of phosphorus. The
concentration profiling of phosphorus along the thick-
3.1. Microstructure of substrate material
ness of the coating (along the vertical line in Fig. 3b)
was carried out using EDX, and shown in Fig. 3c. The
Fig. 1a shows microstructure of the substrate material
phosphorus content showed a sinusoidal variation with
AZ91D ingot. The microstructure consisted of primary
average phosphorus content decreasing from interior to
a, eutectic a and b-phases (marked in the figure). b-
exterior. The alternate layers contain high amount of
phase is an intermetallic with the stoichiometric com-
phosphorus with layers in between exhibiting lower
position of Mg17Al12. Coring during solidification
values. The average phosphorus content in the deposit
resulted in considerable variation in the distribution of
was 7–8 wt.% with a hardness value ranging between
aluminium and zinc in the microstructure of AZ91D
650 and 750 VHN depending on the plating parameters.
alloy. Previously, we have reported the variation of
aluminium and zinc concentration adjacent to b-phase
3.3. Coating nucleation and growth
in AZ91 ingot-cast alloy w26x. Fig. 1b gives a typical
concentration profile of aluminium and zinc in AZ91D
alloy along the line shown in the SEM picture (shown
To understand the nucleation and growth of electroless
in the inset) measured using EDX. In general, the
nickel on electrochemically heterogeneous AZ91D sub-
aluminium and zinc concentration at a regions near to
b-phase (eutectic a) was high, which decreased with
increase in distance from the b-phase w26x. The width
of this aluminium rich region ()8%) adjacent to the
b-phase varied at different sites w26x. The concentration
of aluminium typically varied between ;35% at the b-
phase to approximately ;8–6% near or within the
primary a-phase. This microstructural heterogeneity on
the surface of the substrate complicates the process of
electroless deposition.
3.2. Electroless nickel-coating
Fig. 2 shows the variation of coating thickness as a
function of time on AZ91D alloy at constant temperature
and pH. The coating thickness was directly proportional
Fig. 2. Electroless nickel-plating rate on AZ91 substrate as a function
to the plating time.
of time.

---

R. Ambat, W. Zhou / Surface and Coatings Technology 179 (2004) 124–134
127
Fig. 3. Structure of electroless nickel-coating on AZ91 substrate after 4 h: (a) diffractogram, (b) cross-section of the coating showing lamellar
structure and (c) phosphorus content across the lamellar structure.
strate, plating growth was monitored at very short
regions adjacent to b-phase is the eutectic a-phase
intervals of time using SEM and EDX. The studies were
containing a higher amount of aluminium. The etching
carried out on substrates after the specified pre-cleaning
difference between these phases is due to the difference
steps. The effect of various cleaning steps on surface
in aluminium composition as described earlier (Fig. 1b).
features is shown in Fig. 4. Although the alkaline
Electroless nickel nucleated on the AZ91D substrate
cleaning did not etch the surface, the surface after
at different intervals of time is shown in Fig. 5. The
chrome–nitric pickling and fluoride activation was
coating has been preferentially nucleated on b-phase,
etched, so that the different microconstituents could be
which then spread to eutectic a and primary a-phase.
clearly seen. After hydrofluoric acid cleaning, the sur-
The coating on primary a-phase was formed as a
face was found to be smooth and flat (Fig. 4), where
continuation of the deposit on b-phase and eutectic a-
flat b phases can be seen as dark phases. The less dark
phase, after the b-phase and eutectic a-phase was
Fig. 4. Effect of pre-cleaning treatment on substrate microstructure: (a) ultrasonic cleaning, (b) alkaline cleaning in 10% NaOH at 60 8C, (c)
6% chromic acid–5% nitric acid pickling and (d) fluoride activation (250 ml 70% HFyl).

---

128
R. Ambat, W. Zhou / Surface and Coatings Technology 179 (2004) 124–134
Fig. 5. Early growth of electroless nickel on AZ91 as a function of time: (a) and (b) after 1 min, secondary and topographic mode, respectively;
and (c) and (d) after 5 min with (d) at higher magnification showing the morphology of deposit around b-phase.
covered with coating (Fig. 5). However, the coating on
taneously with coating growth. However, this phenom-
the b-phase was discontinuous and grew slowly, which
enon was observed in a very thin layer of coating at the
led to a concentration of defects over this region. The
surface while on the average, as shown in Fig. 3c, the
EDX analysis of the early stage deposits on b-phase at
phosphorus content decreased with coating growth.
different intervals of time showed slightly lower
Fig. 6 shows the surface morphology of the coating
amounts of phosphorus content, which increased simul-
that was attached to the substrate after peeling off from
Fig. 6. SEM surface morphology and results of EDX analysis of the coating surface attached to substrate after peeling off: (a) secondary, (b)
back scattered, (c) magnified picture showing the morphology near the b-phase and (d) results of EDX analysis at locations 1, 2 and 3 in picture
(c).

---

R. Ambat, W. Zhou / Surface and Coatings Technology 179 (2004) 124–134
129
Table 4
Rate, phosphorus content and hardness of coating with different Ligand–nickel (citric acidyNi2q) ratio in the coating bath
Citric acidyNi2q
Coating rate
Phosphorus content
Hardness (VHN)
Ratio
(mmyh)
(wt.%)
0.16
No deposition
0.33
6.00
7.61
665
0.64
5.41
6.86
651
0.95
No deposition
the surface. The position marked 1 in the photograph
3.4.1. Effect of ligand–nickel ratio
shows the areas of coating attached to b-phase. It can
Table 4 gives the coating rate, phosphorus content
be seen that the coating contained b-phases broken
and hardness of the coating as a function of ligand–
while pulling out. The EDX analysis on the surface of
nickel ratio. The optimum ratio of ligand–nickel was
the peeled off coating showed the following results.
between
0.33
(Ligand:Ni2q(1:3)
and
0.64
EDX analysis at locations 1, 2 and 3 shows a variation
(Ligand:Ni2q(1:1.5) (i.e. citrate concentration of 5.2–
of Al, Mg and P content as shown in Fig. 6d. At location
7.8 gyl). Within this range of concentration, the effect
1, the EDX analysis revealed Mg and Al composition
of ligand on coating characteristics was found to be
corresponding to the b-phase, while the phosphorus
similar, although in terms of phosphorus content, the
content was slightly lower compared to coating areas
best ligand–nickel ratio was 0.33 (Ligand:Ni2q(1:3).
attached to primary a and eutectic a-phase. Areas of
Above or below this concentration, the coating rate was
the coating attached to eutectic a-phase showed a higher
drastically reduced or inhibited. The surface morphology
amount of Mg than the primary a-phase. However, both
of the coating did not show significant variation with
primary and eutectic a showed very little amount of
the change in ligand–nickel ratio.
aluminium. The SEM picture shown in Fig. 6c shows
the surface of the substrate material after the coating
3.4.2. Effect of fluoride ion concentration
has been removed from the surface. It can be seen that
The fluoride ion is an essential component for elec-
the b-phase was fractured while pulling out. The frac-
troless nickel-plating bath for magnesium, and its pres-
tured surface is visible on the micrographs (marked 1
ence in the bath was reported to improve adhesion. In
in the picture). Some corrosion attack is evident on
the present investigation, ammonium bifluoride was used
eutectic a-phase (position 2), while the primary a-phase
as the source of fluoride ion. Variation of coating rate,
was essentially unattacked (position 3).
phosphorus content and hardness with concentration of
ammonium bifluoride in the coating bath is given in
3.4. Effect of coating parameters
Table 5. The values show that the optimum concentra-
tion of ammonium bifluoride in the bath should be
Both the pre-cleaning steps and the concentration of
approximately 7.5 gyl, although within the concentration
bath constituents have an influence on electroless nickel-
range of 7.5–15 gyl, the difference was found to be
coating. The effect of these bath parameters on the
insignificant. However, the coating produced at 7.5 gyl
electroless nickel deposition of Mg is not available in
showed a hardness value of ;60 VHN higher than the
the literature. In the present investigation, the effect of
coating produced with 15 gyl of ammonium bifluoride.
complexing agent, fluoride ion concentration, effect of
Fig. 7a and b show the surface topographic pictures of
stabilizer concentration such as thiourea and effect of
the coatings produced from baths with 0 gyl and 7.5 gy
MBT on electroless nickel-plating have been investigat-
l of ammonium bifluoride, respectively. The deposit
ed by monitoring coating rate, surface hardness, phos-
formed at 7.5 gyl was found to be more combat and
phorus content and surface morphology.
defect free. However, the simple tape-test for adhesion
Table 5
Rate, phosphorus content and hardness of coating with different amounts of ammonium bifluoride in the coating bath
Concentration of
Coating rate
Phosphorus content
Hardness (VHN)
Ammonium
(mmyh)
(wt.%)
bifluoride (gyl)
0
1.93
1.93
342
7.5
8.11
8.11
780
15
5.98
5.98
665
22.5
No deposition

---

130
R. Ambat, W. Zhou / Surface and Coatings Technology 179 (2004) 124–134
Fig. 7. Effect of ammonium bifluoride on the morphology of the coating on AZ91: (a) 7.5 gyl and (b) without fluoride.
indicated higher adhesion for the coating produced with
concentration of MBT in the presence of TU. Fig. 9
15 gyl bifluoride in the bath. Absence of ammonium
shows the effect of TUqMBT at different concentra-
bifluoride or a concentration above 15 gyl produced
tions on coating properties. A concentration of 0.5 mgy
very inferior coating (Fig. 7b) or almost inhibited the
l TUq0.25 mgyl MBT was found to be the optimum
coating.
value, so that it produced increased coating rate, high
phosphorus content and hardness.
3.4.3. Effect of stabilizer (thiourea) concentration
4. Discussion
The effect of thiourea concentration on coating rate,
phosphorus content and hardness of the coating is given
4.1. Early stage growth and coating morphology
in Table 6. It can be seen that the safe range for thiourea
stabiliser is within the range of 0.5–1 mgyl. Within this
Direct plating of magnesium alloys with electroless
range, the phosphorus content in the coating remained
nickel is still a challenge. Only a limited amount of
the same. Below 0.5 mgyl, the bath decomposes spon-
literature is available w28,29x on the electroless nickel-
taneously, correspondingly a higher amount of phospho-
plating of magnesium alloys and applications. The pro-
rus content can be seen in the coating, possibly due to
cess is more complicated when the substrate contains
the formation of nickel phosphide. However, above a
second phase particles as for AZ91, which makes the
concentration of 1 mgyl, the coating was totally
alloy
electrochemically
heterogeneous.
The
three
inhibited.
microstructural constituents in AZ91 alloy (Fig. 1)
namely b, eutectic a and primary a-phase have different
3.4.4. Effect of MBT on electroless nickel-plating
electrochemical potentials w10x. This is attributed to the
Heterocyclic organic compounds are widely used as
different levels of aluminium in these phases. As shown
accelerators in electroless plating processes. MBT is one
in Fig. 1b, aluminium concentration typically varies
among them, which is found useful in many electroless
from ;35% in the b-phase to ;12 and ;5 wt.% in
plating processes. In the present investigation, the effect
the eutectic and primary a-phase, respectively.
of MBT was studied with and without the presence of
Electrochemically b-phase is reported to be more
thiourea in the bath. The coating properties as a function
cathodic (more noble) to eutectic a and primary a w10x.
of MBT concentration is shown in Fig. 8. Addition of
Lunder et al. w10x reported that the corrosion potential
0.5 mgyl of MBT gave the best deposition with maxi-
of b-phase in 5% NaCl saturated with Mg(OH)2 is
mum hardness, although the effect was similar in the
approximately y1.2 V vs. SCE, whereas the values for
concentration range of 0.25–0.5 mgyl. However, addi-
pure Mg and AZ91 are y1.66 V and y1.62 V vs. SCE,
tion of MBT without thiourea had destabilised the bath
respectively. The lower aluminium containing primary
in the long run. So it was necessary to optimise the
a-phase is expected to have most active (more negative)
Table 6
Phosphorus content and hardness of coating with different amounts of thiourea in the coating bath
Concentration of
Coating rate
Phosphorus content in
Hardness (VHN)
thiourea (mgyl)
(mmyh)
the deposit (wt.%)
0
8.15
25.00
527
0.5
6.70
8.00
593
1.0
6.77
8.00
620
2.0
1.63
–
–

---

R. Ambat, W. Zhou / Surface and Coatings Technology 179 (2004) 124–134
131
Fig. 8. Effect of MBT on electroless nickel-coating on AZ91: (a) coating rate, phosphorus content and hardness; and typical surface morphology
of coating at concentrations (b) 1 mgyl, (c) 0.5 mgyl and (d) 0.75 mgyl.
potential in AZ91 w26x. In such a situation, one would
electroless
nickel
on
b-phase.
The
presence
of
expect preferential nucleation of electroless nickel on
Mg(OH)2 over eutectic a-phase and less phosphorus in
the a-phase. However, the present study shows prefer-
early nickel deposits might support this argument. The
ential nucleation of nickel deposit on b-phase (Fig. 5).
above explanation could be represented by the chemical
Possible explanation is that the initial stage of deposition
equation as follows:
has been influenced by a galvanic coupling between b-
At the a-phase, magnesium dissolved as Mg2q ions
phase and adjacent eutectic a-phase. The electrons
according to the reaction:
produced by the anodic dissolution of magnesium from
2q
y
MgsMg
q2e
(1)
the a-phase are consumed by the cathodic deposition of
Fig. 9. Effect of MBTqTU on electroless nickel-coating on AZ91: (a) coating rate, phosphorus content and hardness; and surface morphology
of coating at concentrations (b) 0.5 mgyl TUq0.25 mgyl MBT, (c) 0.5 mgyl TUq0.5 mgyl MBT and (d) 1 mgyl TU.

---

132
R. Ambat, W. Zhou / Surface and Coatings Technology 179 (2004) 124–134
At the b-phase, the electrons produced from the Eq.
above which the rate decreases (Table 4). Several
(1) are consumed for the reduction of nickel.
theories have been proposed to explain the variation in
coating rate with ligand concentration. De Minjer and
2q
y
o
Ni
q2e sNi
(2)
Brenner w34x argued that the maximum is the result of
the low adsorption of ligand on the catalytic surface at
However, more electrochemical evaluation is needed
low concentrations, which accelerates the reaction. At
to prove the above mechanism in order to understand
higher concentrations, there is a high adsorption of
the early stage growth behaviour of electroless nickel
ligand on the surface, which poisons the reaction.
on AZ91 alloy. Xiang et al. w30x also have reported
A more plausible explanation suggests that the
preferential nucleation of electroless nickel on second
increase in rate is due to the buffering action of the
phase particles in magnesium alloys during plating, and
complexing agents w35x. Maximum rates occur while
no phosphorus was found in the initial deposited nickel.
there are uncoordinated (solvated) sites remaining on
The lamellar structure of the (Fig. 3b) deposit has
the nickel ions, i.e. nickel ions are only partially com-
been attributed to the compositional variation of phos-
plexed or chelated. Partially complexed nickel ions
phorus within the deposit as observed in Fig. 3c. The
retain some of the properties of free, solvated nickel
cause of the compositional variation has been explained
ions. Hence, for ligand concentrations up to values
in terms of periodic fluctuations in the pH of the plating
where the maximum plating rates occur, buffering is the
solution adjacent to the deposit surface w31,32x. The
dominant function of the various ligands.
overall decrease in phosphorus content with coating
An explanation based on pH is inadequate to explain
thickness might be attributed to the reduction in the
the decrease in rate with further increase in ligand
concentration of hypophosphite as coating proceeds. The
concentration beyond maximum. The pH of the plating
structure of the deposit observed is in agreement with
bath changes very little with continued plating. More-
earlier observation that the structure of electroless nickel
over, the buffer capacity of the complexing agent does
with 4–7 wt.% phosphorus is generally amorphous w33x.
not parallel the plating rates obtained with them.
However, a decrease in plating rate with increase in
4.2. Effect of bath composition
ligand concentration (or ligand–nickel ratio), is most
likely due to the co-ordination of the remaining solvated
The nickel source in the present investigation was
sites with ligand atoms w35x. The concentration of free
basic nickel carbonate (NiCO3 2Ni(OH)2 4H2O) and
nickel ions decrease and the metal ions take on the
the reducing agent was sodium hypophosphite. Citric
characteristics of complexed or chelated nickel ions.
acid monohydrate was added as the complexing agent
Fluoride ion is an essential component in electroless
that provides a quadridentate anion for chelation with
nickel-plating of magnesium alloys. The initial dip in
nickel cation. To activate the magnesium surface, fluo-
hydrofluoric acid activates the substrate surface, and
ride was added in the form of hydrofluoric acid and
forms a fluoride film over the surface. Several research-
ammonium bifluoride, which also increases the adhesion
ers studied the reactions between fluoride ion and
of the coating. The presence of citric acid and ammo-
magnesium w36,37x in relation to corrosion protection.
nium bifluoride also serves the purpose of buffers and
Fluoride ion is reported to be a good corrosion inhibitor
accelerators. The pH of the solution was controlled
for magnesium and its alloys. Further, good corrosion
within the range of 7–8 using ammonium hydroxide.
resistance of magnesium base materials in severe envi-
The autocatalytic electroless nickel deposition was ini-
ronments such as F2 gas and aqueous and gaseous HF
tiated by catalytic dehydrogenation of the reducing agent
has been reported w36,37x. The protection is due to the
with the release of hydride ion, which then supplied
formation of a MgF2 film on the surface of the metal.
electrons for the reduction of nickel ion w27x.
However, the protective effect in aqueous solutions
Electroless deposited nickel usually contains 3–15
seems to depend on solution pH and fluoride concentra-
wt.% of co-deposited phosphorus originating from hypo-
tion. As a result, fluorides are constituents of several
phosphite ion. The quantity of phosphorus in the coating
common anodic coating baths, known as DOW17 and
determines several engineering properties of electroless
HAE w37–39x. In electroless nickel-plating, formation
nickel.
of magnesium fluoride film over the substrate surface
The role of the complexing agent on electroless
during the pre-cleaning step prevents the oxidation of
nickel-coating on AZ91 alloy can be explained on the
otherwise reactive magnesium during the transfer of
basis of three aspects i.e. (a) a reduction in the concen-
material to plating bath. In the plating bath, the deposi-
tration of free nickel ions, (b) preventing the precipita-
tion of electroless nickel takes place by replacement of
tion of basic nickel salts and nickel phosphate, and (c)
the fluoride film. The presence of fluoride in the bath
exerting a buffering action.
stabilizes the film on the surface, however, an increase
The coating rate on AZ91 alloy increased initially
in concentration of fluoride above a value makes the
with increase in ligand–nickel ratio upto a ratio of 0.33,
removal of fluoride film for nickel deposition impossible

---

R. Ambat, W. Zhou / Surface and Coatings Technology 179 (2004) 124–134
133
(Table 5). Xiang et al. w30x have reported that for
2. A strong influence of substrate microstructure was
electroless nickel-plating of magnesium alloys, different
found. Initially, the coating was nucleated preferen-
fluorideyoxide ratio in the surface film caused different
tially on b-phase. The coating spread over to primary
deposition rates.
a-phase, once the b-phase and eutectic a-phases were
Although the two heterogeneous partial reactions
covered with electroless nickel.
require a catalytic surface on which to occur, Gutzeit
3. Optimum ligand–nickel ratio was found to be 0.33
w40,41x found that they could also easily happen on the
(Ligand:Ni2q(1:3), whereas the best concentration
surfaces of solid particles of colloidal dimensions present
of thiourea was 1 mgyl. Fluoride ions were essential
in the solution. As the formation of these colloidal
to plate electroless nickel on AZ91D alloy with an
particles is uncontrollable and the nickel deposition
optimum concentration of 7.5 gyl of ammonium
process is autocatalytic, working electroless nickel baths
bifluoride.
may decompose spontaneously at any time by triggering
4. The presence of 0.5 mgyl of MBT in the plating bath
a self-accelerating chain reaction on those colloidal
doubled the plating rate. However, for a stable bath,
particles. So in practice, all working electroless nickel
the optimum concentration of thiourea and MBT was
bath contain some kind of stabiliser (catalytic inhibitor)
0.5 mgyl of former with 0.25 mgyl of the latter.
to prevent such spontaneous decomposition. The strong
adsorption of stabiliser ions on a catalytic surface block
the catalytic sites from being used for the adsorption of
References
reactants, thus impedes the deposition process. When
the surface density of adsorbed stabiliser ions reaches a
w1x D. Magers, Magnesium alloys and their applications, in: B.L.
certain level, the deposition process could be completely
Mordike, K.U. Kainer (Eds.), Proceedings of the Conference
on Magnesium Alloys and their Applications, Wekstoff-infor-
inhibited (Table 6). As the concentration of stabiliser in
mationsgesellschaft, 1998, p. 105.
bulk solution is usually in the ppm range, unlike the
w2x S. Schumann, F. Friedrich, (in Ref. D. Magers), Magnesium
adsorption of reactants, which is kinetically controlled,
alloys and their applications, in: B.L. Mordike, K.U. Kainer
the adsorption of stabiliser is diffusion limited.
(Eds.), Proceedings of the Conference on Magnesium Alloys
Although complexing agent is essential in electroless
and
their
Applications,
Wekstoff-informationsgesellschaft,
1988, p. 3.
nickel-plating baths, its presence sometimes reduces the
w3x L. Whitby, in: F.L. LaQue, H.R. Copson (Eds.), Corrosion
speed of the plating reaction. To overcome this, accel-
Resistance of Metals and Alloys, 2nd ed, Reinhold, New York,
erators can be added to electroless nickel-plating solu-
1963, p. 169.
tion. For AZ91, presence of MBT alone in the bath
w4x G.L. Makar, J. Kruger, Int. Mater. Rev. 38 (1993) 138.
without thiourea, doubled the plating rate. The presence
w5x G. Song, A. Atrens, M. Dargusch, Corros. Sci. 41 (1999) 249.
w
of MBT and thiourea together in the bath also gave
6x O. Khaselev, J. Yahalom, Corros. Sci. 40 (1998) 1149.
w7x D. Daloz, P. Steinmetz, G. Michot, Corrosion 53 (1997) 944.
significant increase in plating rate. These accelerators
w8x T.J. Warner, N.A. Thorne, G. Nussbaum, W.M. Stobbs, Surf.
are thought to function by loosening the bond between
Interface Anal. 19 (1992) 386.
hydrogen and phosphorus atoms in the hypophosphite
w9x F. Hehmann, R.G.J. Edyvean, H. Jones, F. Sommer, in: Pro-
molecule, allowing it to be more easily removed and
ceedings of the International Conference on Powder Metallurgy
adsorbed on to the substrate surface. The most suitable
and Aerospace Materials, Lucerne, Switzerland, November 2–
compounds for this purpose are heterocyclic organic
4, 1987, p. 46.
w10x O. Lunder, J.E. Lein, T.Kr. Aune, K. Nisancioglu, Corrosion
compounds with oxygen or nitrogen atoms as these have
45 (1989) 741.
a lone pair of electrons to participate in the bonding
w11x C.F. Chang, S.K. Das, D. Raybould, A. Brown, Met. Powder
process. The delocalisation of electrons possible in these
rep. 41 (1986) 302.
molecules makes them suitable to use as accelerators in
w12x C.F. Chang, S.K. Das, D. Raybould, Rapidly Solidified Mate-
electroless nickel-plating process. The MBT heterocyclic
rials, American Society for Metals, Metals Park, OH, 1986,
ring consists of two sulfur and a nitrogen atom. These
pp. 129–135.
w13x M. Bayes, I. Sinitskaya, K. Schell, R. House, T I Met. Finish.
atoms have lone pairs of electrons to participate in the
69 (1991) 140.
electroless nickel deposition process.
w14x R.N. Duncan, Plat. Surf. Finish. 83 (1996) 65.
w15x K.L. Lin, J.W. Hwang, Mater. Chem. Phys. 76 (2002) 204.
w
5. Conclusions
16x M.I. Jafar, M. Broadhurst, S.A. Ashton, Br. Corros. J. 31
(1996) 239.
w17x M.A. Sanchez, L.A. Parra, O.A. Perez, O. De Rincon, Corros.
1. The electroless nickel-coating deposited on AZ91D
Rev. 19 (2001) 105.
alloy in optimised bath showed amorphous structure
w18x G.O. Mallory, J.B. Hajdu (Eds.), Electroless Plating: Funda-
with 7 wt.% P and a hardness value of 600–700
mentals and Applications, AESF, Florida, 1990, p. 1.
w19x L. Das, D.T. Chin, R.L. Zeller, G.L. Evarts, Plat. Surf. Finish.
VHN. The microstructure of the coating in the trans-
82 (1995) 56.
verse direction showed lamellar structure with a
w20x C. Kerr, D. Barker, F.C. Walsh, T I Met. Finish. 74 (1996)
phosphorus content varying in a sinusoidal manner.
214.

---

Document Outline

• Electroless nickel-plating on AZ91D magnesium alloy: effect of substrate microstructure and plating parameters
  • Introduction
  • Experimental
  • Results
    • Microstructure of substrate material
    • Electroless nickel-coating
    • Coating nucleation and growth
    • Effect of coating parameters
      • Effect of ligand-nickel ratio
      • Effect of fluoride ion concentration
      • Effect of stabilizer (thiourea) concentration
      • Effect of MBT on electroless nickel-plating
  • Discussion
    • Early stage growth and coating morphology
    • Effect of bath composition
  • Conclusions
  • References

---