Wear 271 (2011) 1341–1348

Contents lists available at ScienceDirect

Wear
journal homepage: www.elsevier.com/locate/wear

Droplet erosion performance of composite materials electroplated with a hard

metal layer
P. Lammel a,b,c , A.H. Whitehead a,∗ , H. Simunkova a , O. Rohr b , B. Gollas a,d
a
CEST Centre for Electrochemical Surface Technology GmbH, Wiener Neustadt, Austria
b
EADS Innovation Works, Munich, Germany
c
University of Paderborn, Institute for Technical and Macromolecular Chemistry, Paderborn, Germany
d
Graz University of Technology, Institute for Chemistry and Technology of Materials, Austria

a r t i c l e i n f o a b s t r a c t

Article history: The effectiveness of resistance to water droplet erosion of various electro-deposited metal layers on glass
Received 31 August 2010 fibre reinforced polymer (GFRP) was investigated. The number of impacts by 2 mm droplets at 225 m s−1
Received in revised form that could be applied before visible damage to the GFRP occurred was significantly higher with the
22 December 2010
metallic coatings than without. The impact resistance was found to increase in the order: Cu-only < Cr on
Accepted 22 December 2010
Cu ∼ Ni–SiC on Cu < Ni on Cu. The best coating increased the impact resistance of the GFRP by more than
20 times compared to the resistance of the uncoated material.
© 2011 Elsevier B.V. All rights reserved.
Keywords:
Liquid impact erosion
Polymer-matrix composite
Metal-matrix composite
Electrochemistry
Erosion testing

1. Introduction
Composite materials of polymers with glass or carbon fibre rein-
forcement (GFRP and CFRP, respectively) are finding increasing use
in aerospace applications [1]. They typically offer high strength to
weight ratios, but rather poor erosion resistance [2,3]. In order to
improve the erosion resistance of these materials electrodeposited
protective metallic layers have been considered [2,4–6]. Similarly
GFRP blades of large wind turbines are prone to environmental ero-
sion and require protective coatings [7]. Good erosion protection
is offered by hard chromium in many applications and has been
advised for protection against cavitation erosion [8]. However, due
to the carcinogenic nature of Cr(VI)-based electrolytes required to
electrodeposit hard chromium, less toxic alternatives have been
sought [9–14]. One candidate for chromium replacement is Ni con-
taining finely dispersed hard particles, such as SiC. Ni and Ni alloys
have proved particular popular matrices for fine ceramic dispersion
coatings because of their relatively low cost, ease of electrodeposi-
tion and reasonable atmospheric corrosion resistance [15,16]. The
ceramic particles improve the hardness of the metal matrix through
dispersion hardening. Generally hard coatings are considered to
∗ Corresponding author. Tel.: +43 2622 222 66 32.
E-mail address: adam.whitehead@cest.at (A.H. Whitehead).
have better wear properties [17], with encouraging results having
been found for sol–gel coatings with incorporated ceramic particles
[18–20]. However, increased hardness does not automatically infer
higher water droplet erosion resistance because various erosion
mechanisms can operate on different materials [21].
Electrodeposited Ni was examined to determine the effect on
erosion resistance of SiC particles in the matrix. Additionally Cu
was present on the GFRP as supplied and provided a convenient
thin conducting layer on which to electrodeposit the more erosion
resistant materials.
2. Experimental
2.1. Sample preparation
1.5 mm thick epoxy-based GFRP clad with ∼35 ␮m of Cu on one
side (Conrad Electronics, Cu on FR4 epoxy-based GFRP, #528412)
was used as supplied and also as a substrate for electrodeposition
of erosion resistant layers. Although it is recognised that, typi-
cally CFRP, materials used for aeronautical applications will differ
in terms of chemistry and mechanical properties; this substrate
was chosen as an example of a fibre-reinforced polymer-matrix
composite of low cost, available in large quantities with certified
uniformity and already possessing a well-adhered, electrically con-
ducting layer, suitable for reinforcement by electrodeposition. The

0043-1648/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.wear.2010.12.034
1342 P. Lammel et al. / Wear 271 (2011) 1341–1348

Table 1
Bath composition for Ni and Ni–SiC electrodeposition.

Chemical compound or condition Ni deposition Ni–SiC deposition

−1
NiSO4 ·6H2 O 1.14 mol l 1.14 mol l−1
NiCl2 ·6H2 O 0.17 mol l−1 0.17 mol l−1
H3 BO3 0.65 mol l−1 0.65 mol l−1
Saccharin 92 mg l−1 –
2-Butyne-1,4-diol 43 mg l−1 –
SiC (Ø 1 ␮m, Alfa Aesar) – 50 g l−1 Fig. 1. Schematic illustration of the pulsed jet erosion test rig (PJET).
Temperature 328 K 328 K
pH 4.2a 4.2a
a
The pH was adjusted with 0.1 M NaOH or 10 wt% H2 SO4 .

manufacturer’s information (Shanghai Nanya Copper Clad Lami-

nate Co. Ltd.) gave the peel strength of the Cu layer as 85–89 kPa.
The electrodeposition of Ni and Ni–SiC dispersion coatings was
carried out galvanostatically using a reverse pulse power supply
(Plating Electronic GmbH, Germany). Although sulfamate baths
generally give lower stress deposits, Watt’s bath compositions were
used in this work because of the lower perceived hazard (nickel
sulfamate is carcinogenic). The plating bath compositions and con-
ditions were as shown in Table 1.
800 ml of electrolyte were used in both instances with mechan-
ical agitation via a magnetic stirrer having at a rotation speed
of 1100 rpm. The anode consisted of electrolytic Ni spheres held
within a Ti basket. The substrate was masked off to a defined area
of 0.4 dm2 . Degreasing of the Cu clad FRP substrate was performed
in a commercial alkaline rinse solution (Enprep, Enthone) at 60 ◦ C
for 600 s, followed by rinsing in flowing water and drying with
dried compressed air. Before plating, the substrate was immersed
in 20 vol% H2 SO4 for 30 s, rinsed with deionised water and fixed
into the plating bath. Plating commenced 30 s after immersion of
the substrate in the plating bath.
Low roughness was required for optimal aerodynamic perfor-
mance of the coated FRP. Saccharin and 2-butyne-1,4-diol were
added to the Ni bath to improve the smoothness of the deposit
[22,23]. In this respect they failed with the measured roughness
(Ra = 0.27 ␮m) not differing significantly from the deposit without
additives, however, they were found to increase the hardness from
HV0.05 ∼ 220 to HV0.05 ∼ 500.
For the dispersion coating the suspension was placed in an ultra-
sonic bath for 600 s immediately prior to use, in order to break down
large SiC agglomerates. Directly after plating, the dispersion coat-
ing sample was cleaned by ultrasound in deionised water for 10 s
to remove loose particles from the surface.
Cr was deposited from an acidic plating solution containing the
equivalent of 3 mol l−1 CrO3 and 0.03 mol l−1 H2 SO4 . A Pb plate
was used as the anode. A continuous current was applied at either
50 A dm−2 or 30 A dm−2 at 55–70 ◦ C and with magnetic stirring at
700 rpm. Despite the difference in current densities the plating rate
was found to be 0.6–0.7 ␮m min−1 in both cases, implying that
the deposition at 50 A dm−2 was significantly less efficient than at
30 A dm−2 .
2.2. Droplet erosion testing
Rain erosion experiments were carried out using a pulsating jet
erosion test rig (PJET) [20,24]. A standard equivalent water droplet
size of 2 mm diameter [25–27] was distributed by a disc with two
orifices rotating at a frequency of 20 Hz, which corresponds to an
impact frequency of 40 s−1 . The nozzle diameter was 0.8 mm and a
pressure of 32.5 MPa was chosen to receive the required velocity.
The PJET apparatus allows precise alignment of the sample with
respect to the nozzle. After a given number of droplet impacts at a
preset position the sample is moved and the droplets then repeat-
edly impact the next site. In this manner an array of spots with a set
Fig. 2. Typical droplet impact erosion test matrix applied by the PJET.
number of impacts can be defined. For the samples tested in this
study an array of spots with between 5 × 102 and 2 × 105 impacts
were used. Up to 5 repetitions of each impact number were made
within each row, with the number of impacts increasing from one
row to the next. The impact sites were regularly spaced at inter-
vals of typically 5 or 10 mm, as shown in Figs. 1 and 2. Testing was
prematurely terminated if the substrate became severely damaged.
The correlation of erosion testing in lab with in-flight condi-
tions is very difficult. Déom et al. developed a correlation for a
similar erosion test, the multiple impact jet apparatus (MIJA), in
which water jets are sent randomly onto the sample [28]. These test
conditions can be compared to aircraft flying time through natural
rainfall [28,29]. Flying for 1 h at 225 m s−1 through rain of 25 mm
h−1 intensity corresponds to 820 droplet impacts (taking an impact
area of 1.2 mm diameter and power velocity dependence of 3 [30]).
Therefore, a rough idea of the coating durability may be obtained by
this method in a relatively short time (20 s of testing corresponded
to roughly 3600 s flying through rain). These results for the tests
with MIJA should only be taken as an order of magnitude estimate
for the performed tests with PJET. An uncertainty of just ±0.1 mm
in the impact area diameter would lead to an uncertainty of ±140
droplets for 1 h equivalent flying time. Additionally, as rainfall is
unusual above 5 km [31] and civilian jet aeroplanes typically fly at
cruising altitudes in excess of this, the proportion of overall flying
time spent at around 225 m s−1 velocity in rain will be relatively
low.
The sample was fixed on a holder with a set distance of 60 mm
to the nozzle of the water jet. In addition, a dry air jet was directed
at the sample with a pressure of 0.3 MPa to remove excessive sur-
face water, which has previously been found to reduce the erosive
influence of the droplets.
The PJET provides a fast, inexpensive method for estimating the
rain erosion resistance of materials. The testing area is, however,
small compared to the overall sample and therefore some effects
 P. Lammel et al. / Wear 271 (2011) 1341–1348 1343

Fig. 3. (a) Photograph of a 50 mm wide GFRP after droplet erosion testing from the uncoated side with 5 × 102 to 3 × 104 droplet impacts at 225 m s−1 , each row consisted
of three impact sites of equal number of impacts (indicated on the right) (b) the same sample but viewed from the side as indicated by the black arrow in (a) to show detail
of the delamination and damage caused by repeated water droplet impacts.

may not be observed (e.g. if localised loss of adhesion occurs at
an internal interface the outer layer may remain attached via the
surrounding material).
2.3. Sample characterisation
Samples were characterised by visual observation, optical and
scanning electron microscopy (SEM) before and after erosion test-
ing. SEM was conducted with a Philips XL30 ESEM, typically at
20 keV with a working distance of 10 mm in environmental (ca.
0.8 mbar gas pressure) or high vacuum modes.
The average roughness (Ra ) was determined using a profilome-
ter (Mahr Pertometer S2, sensor MFW-250), with profiles measured
over 6 mm. However, for average surface roughness (SRa) measure-
ments, especially when the stylus profilometer was found to have
insufficient sensitivity an Olympus confocal laser scanning micro-
scope (Lext OLS3100, 10 nm height resolution) with 408 nm laser
illumination was employed. A 50× magnification objective was
used for roughness measurements and 10× objective for profile
measurements.
Vickers’ hardness measurements were made by micro-
indentation of a cross-sectioned, polished surface with a force of
0.49 N (HV0.05 ), for 10 s. Cross-sections were examined rather than

Fig. 4. Scanning electron micrographs of a Cu surface (500× magnification) taken at the base of the erosion scar after: (a) 0, (b) 1 × 104 and (c) 3 × 104 droplet impacts at
225 m s−1 .
1344 P. Lammel et al. / Wear 271 (2011) 1341–1348

the original surface to reduce the influence of the underlying lay- 4

ers. The indentations were examined by optical microscopy. In each
3
case the reported values are the average of at least 5 separate mea-
surements. 2

Height [µm]
1
3. Results
0
3.1. Bare GFRP erosion resistance
-1
Erosion testing was performed on the reverse face (i.e. that not
-2
covered by the Cu layer) of the GFRP substrate. 3 spots of equal
impact were performed in each row with the number of impacts -3
increasing from 5 × 102 to 3 × 104 (at which point the test was
stopped due to extensive delamination of the test piece). -4
0 1000 2000 3000 4000 5000 6000
Very little damage of the bare GFRP was visible to the naked eye
Distance [µm]
up to 5 × 103 impacts, Fig. 3. Although delamination occurred in
the area of these spots, this followed subsequently, during higher Fig. 5. Profile of a Cu-clad GFRP substrate, determined by stylus profilometry, after
impact testing of the adjacent spots. The delaminated area is much spot erosion from 2 × 104 water droplet impacts in the region marked by the bar
paler than the surrounding dark GFRP. (centred around 3300 ␮m).
At higher impact numbers significant amounts of GFRP were
removed in roughly rectangular sections. The shape of the frac-
ture reflected the weave direction (i.e. the alignment of the glass
fibres). At the end of testing severe delamination had occurred, with
6 layers clearly separated. The sharp transition in damage between
5 × 103 and 1 × 104 impacts allows a critical impact limit for bare
GFRP to be considered to be ca. 1 × 104 impacts. The impact limit,
according to the Déom model, would be equivalent to 12 h of flying
time through rain.

3.2. Cu-coated GFRP erosion resistance

From the previous sample, Fig. 3, it is also interesting to note

that even at the maximum number of impacts the Cu layer on
the reverse side of the test piece remained well adhered to the
glass–fibre composite.
A droplet impact matrix was used to test the Cu-clad side of
the GFRP. 3 samples were tested with different maximum impact
numbers from 3 × 104 to 1.5 × 105 impacts, Fig. 4. When examined
visually and microscopically the erosion features were similar for
any given pair of spots with the same number of impacts. Therefore,
the erosion was not significantly affected by the location on the
surface.
For the first sample circular erosion features first became visible
at 2 × 103 impacts and increased in diameter with greater impact
numbers. However, even at 3 × 104 impacts there was no penetra-
tion of the Cu layer or delamination of the underlying GFRP.
The Cu surface underwent significant plastic deformation,
including the formation of 10 ␮m diameter tunnels (not shown) as
have previously been reported after droplet erosion of Al–Cu alloys
by focused erosion in relatively soft, ductile material [32,33].
Stylus profilometry across the eroded spots revealed that the
surface had become considerably roughened, especially with high
numbers of droplet impacts, Fig. 5. However, integration of the pro-
file over the impact areas revealed that very little material had
been removed, e.g. the average depth of the eroded spots, calcu-
lated over the entire spot diameter, was <1 ␮m, even after 3 × 104
impacts. The diameter of the roughened area was 0.9 ± 0.1 mm at
>1 × 104 impacts, in good agreement with the impact area diameter
of 1.2 mm.
A periodic undulation in the non-eroded surface area with char-
acteristic length of ca. 600 ␮m and peak-to-peak amplitude of ca.
0.8 ␮m was due to the weave pattern in the GFRP.
The stylus technique is limited by the probe dimensions and
hence cannot accurately measure narrow features of high aspect
ratio (e.g. tunnels). Laser profilometry was therefore used and
Fig. 6. Photograph of Cu-clad GFRP after droplet erosion testing using the PJET
apparatus with 1 × 104 to 4 × 104 impacts per spot.
showed greater average maximum height of the profile (Rz) of
up to 29 ± 4 ␮m at 2 × 104 impacts (initially 1.3 ␮m) compared to
4 ± 1 ␮m by stylus profilometry (Rz is defined as the average of the
distance between maximum to minimum height in 5 equal length
segments of the measured profile).
For the other two samples similar trends were observed,
although the Cu layer was removed at 2.5 × 104 and 3 × 104
impacts, Fig. 6. At higher impacts the GFRP was delaminated and
rectangular sections eroded.
Therefore, it would appear that the Cu underwent plastic defor-
mation, roughening and work hardening, with little erosion until
catastrophic failure at around 3 × 104 impacts.
3.3. Cr-coated GFRP erosion resistance
Three samples of Cr electrodeposited on Cu-clad GFRP were pre-
pared. Characteristics of the samples are tabulated below (Table 2).
In each case the Cr layers were slightly smoother than the Cu
layer on which they were deposited (Ra = 0.33 ␮m). The samples
were subjected to similar testing as previously but with greater
maximum impacts. Erosion scars were not visible on Cr-1 after test-
 P. Lammel et al. / Wear 271 (2011) 1341–1348 1345

Fig. 7. Scanning electron micrographs of Cr-2, electrodeposited Cr layer on Cu-clad GFRP (a) as plated, at 100×, (b) and 2000× magnification, (c) after 2 × 104 water droplet
impacts, at 100×, (d) and 2000× (right) magnification revealing very rough Cu where a Cr fragment has been removed.

Table 2
Selected properties of Cr-plated Cu-clad GFRP test pieces for examination by droplet erosion.

Sample name Layer thickness (␮m) Surface finish Initial HV0.05 Initial Ra (␮m) Cr deposition current
density (A dm−2 )

GFRP Cu Cr

Cr-1 1500 35 35 Matt 869 ± 11 0.28 50

Cr-2 1500 35 39 Half-bright 863 ± 7 0.23 50
Cr-3 1500 35 39 Half-bright 951 ± 18 0.29 30

ing up to 3 × 104 impacts (the maximum used) and there was no trodeposited Cr layers, Fig. 7a [34]. The cracks were only apparent
obvious damage to the underlying GFRP. under microscopic inspection and not visible to the unaided eye.
The as deposited Cr-2 layer (similar to the other deposits), was Cr-2 was tested to higher impact numbers, which resulted in no
covered by a network of fine cracks, that are hard to avoid in elec- visible damage prior to 2 × 104 impacts. At 2 × 104 impacts frac-

Fig. 8. Photograph of Cr-3, Cr-plated Cu-clad GFRP, after (a) 3 × 104 water droplet impacts and (b) 7 × 104 water droplet impacts.
1346 P. Lammel et al. / Wear 271 (2011) 1341–1348

Fig. 9. Schematic illustration of the stages of droplet impact erosion of a Cr-layer on Cu-clad GFRP: (a) initial erosion behaviour by incident water droplets and (b) later stage
erosion where later water jets impinging on freshly exposed Cr–Cu surfaces cause spalling of larger fragments.

turing of the Cr layer into fragments of roughly 50 ␮m diameter foreign material remaining on the other. This would explain the
was just visible, Fig. 7c. Cracking of hard, brittle materials by water irregular outline of the eroded Cr layer and rectangular fracture of
droplets, followed by dislocation of fragments is a well known ero- the underlying GFRP, which broke along lines parallel to the glass
sion mechanism for glasses and ceramics [35]. fibres.
Some Cu remained attached to the dislocated Cr chips, leaving a
very rough Cu layer remained in contact with the GFRP. At 2000× 3.4. Ni–SiC coated GFRP erosion resistance
magnification striations, typical of fatigue fracture, can be observed.
At 5 × 104 some spalling of the Cr layer together with the Cu was A detailed examination of the erosion of a Ni–SiC composite
observed, leaving a virtually metal-free GFRP surface exposed. At layer at low water droplet impacts (2 × 101 to 3 × 103 ) has already
1 × 105 and 2 × 105 there was complete perforation of the GFRP, been submitted to this journal for publication by the authors. In
with extensive delamination of the polymer-matrix composite. brief, a depth-profiling EDX method was used to examine the rela-
Therefore, the critical impact limit lay between 5 × 104 and 1 × 105 tive intensities of X-rays generated by Ni and Si by electrons under
impacts: a significant improvement on the bare GFRP. different accelerating voltages. Because the penetration depth is
Cr-3 was initially harder than Cr-2 and of the same thickness, dependent on the kinetic energy of the electrons the sampled vol-
therefore it may have been expected to show greater erosion resis- umes are dependent on the accelerating voltage and by means of
tance. However, the first signs of damage were similarly visible Monte Carlo simulations a simple model of spatial elemental distri-
at 3 × 104 impacts, where the Cr layer had started to flake off the bution was proposed. It was found that SiC particles that extended
underlying GFRP. By 5 × 104 impacts the GFRP was further exposed
and by 7 × 104 impacts rectangular sections of the GFRP had been
eroded, Fig. 8. The critical impact limit of Cr-2 was therefore similar
to Cr-3 at 5–7 × 104 impacts.
It should be noted that the exposed GFRP area was much larger
than the primary impact area of the droplets (ca. 1.2 mm diameter)
and was likely caused by lateral water jets from collapse of the
impinging liquid [36].
In contrast to the Cu-only layer the Cr showed no evidence of
deformation or surface roughening prior to brittle fracture. Fur-
thermore, visual examination of the sample after erosion testing
revealed many fine cracks in the Cr layer around the eroded region
that were not initially apparent. In places the macroscopic cracks
joined to form irregular, angular frames around sections of the
metallic layer adjoining the exposed GFRP. The flakes so formed
were significantly larger (several mm) than those formed in the
initial erosion stages (ca. 50 ␮m). Gentle teasing with a pair of
tweezers revealed that in some cases the Cu had become detached
from the GFRP but remained well adhered to the Cr, although the
Cr flake itself was otherwise disconnected from the surrounding
Cr layer. Thus a plausible scheme for the erosion of the Cr elec-
troplated Cu-clad GFRP is shown schematically in Fig. 9. It may be
assumed that the Cu layer then effectively formed a hinge for the
stiff Cr layer, until repeated movement under the impact of water
droplets (or in this case tweezers) caused it to work harden and
eventually fracture along the same lines as the overlying Cr. Typ-
Fig. 10. Photograph of a Ni–SiC composite layer on Cu-clad GFRP after droplet ero-
ically the break between Cu and GFRP was rather clean with little sion testing using the PJET apparatus with 1 × 104 to 1.5 × 105 impacts per site.
 P. Lammel et al. / Wear 271 (2011) 1341–1348
2.0E+05
1.5E+05
Critical impact limit 1.0E+05
5.0E+04
0.0E+00
none
Fig. 11. Summary of the critical water droplet impacts that could be applied to various coatings on GFRP before damage to the GFRP became evident.
beyond the plain of the surface were selectively removed, even after
rather few impacts, presumably by the action of high-speed lateral
jets from the impinging water droplets. SRa also increased within
the first 1 × 103 impacts but by less than 0.3 ␮m above an initial
roughness of ca. 1.1 ␮m.
In this work the maximum number of impacts was extended
above that in the initial study from 3 × 104 to 1.5 × 105 . From which
it was observed that the metal surface exhibited slight depres-
sions after 1 × 104 to 5 × 104 impacts. By 7 × 104 impacts the Ni–SiC
and underlying Cu had fractured off the GFRP, but the polymer
composite was without visible damage. At 1 × 105 impacts the
typically rectangular erosion marks were formed in the GFRP and
by 1.5 × 105 impacts the substrate was perforated with extensive
delamination, Fig. 10. From the cracks in the metal layer and the
angular fracture edges it is clear that the Ni–SiC layer has undergone
brittle fracture.
On a second sample, prepared under the same conditions, at
6 × 104 impacts Ra was 1.6 ± 0.4 ␮m; within experimental error of
the value measured initially. Thus the behaviour resembled that
of Cr, in that there was negligible surface roughening or erosion
before brittle fracture. Equally the Ni–SiC to Cu interface remained
largely intact after fracture. It is likely that the Ni–SiC on Cu fails
by a similar mechanism to that sketched in Fig. 9, at least for the
later-stage erosion.
3.5. Ni-coated GFRP erosion resistance
2 pieces of Cu-clad GFRP were electroplated with 30 ␮m of Ni.
The layers as deposited were intermediate in hardness between Cu
and Cr (HV0.05 = 493 ± 24) as expected. The droplet erosion charac-
teristics were similar for both Ni samples, with erosion marks just
visible at 5 × 104 impacts but no GFRP visible or obviously damaged
after 2 × 105 impacts (the maximum number used).
Post erosion investigation by laser profilometry revealed that
2 × 105 droplet impacts caused a lip-free, curved depression of
roughly 0.8 mm diameter had been formed, with a maximum aver-
age depth of ca. 6 ␮m. SRa barely increased from an initial value
of 85 ± 10 nm to 115 ± 10 nm after 2 × 105 impacts. Therefore, the
critical impact limit was >2 × 105 and possibly significantly higher
than this value given that coating still retained 80% of its original
thickness after this number of impacts. Therefore, the critical limit
had been extended by more than a 20 times in comparison to bare
GFRP and was at least thrice that of a Cr coating of comparable
thickness.
1347
Did not fail
Cu only Cr on Cu Ni-SiC on Cu Ni on Cu
Coating on GFRP
It would be of interest in future work to relate the internal
stress of the deposits with the critical impact threshold, because
low stress deposits may be expected to withstand more impacts
prior to failure.
4. Conclusions
In line with current wisdom GFRP was found to be eroded rather
rapidly by water droplets at 225 m s−1 showing obvious visible
damage after 1 × 104 impacts. All of the metallic coatings increased
the erosion resistance of the GFRP, Fig. 11.
Cu underwent plastic deformation, with considerable roughen-
ing but very little erosion until ca. 3 × 104 impacts, whereupon it
failed catastrophically. Ni on Cu showed little deformation, negli-
gible roughening and little erosion after 2 × 105 impacts. Cr on Cu
in contrast underwent brittle fracture of the Cr to form flakes of
roughly 50 ␮m diameter. On shedding of the Cr flakes a rough Cu
surface was revealed. Finally it is proposed that lateral jets from the
impinging water droplets caused extensive detachment of Cr–Cu
flakes after work hardening of the Cu. As erosion progressed metal-
free GFRP was exposed after (3–5) × 104 impacts and significantly
damaged after (5–7) × 104 impacts.
Ni with dispersed 1 ␮m SiC particles failed after approximately
the same number of impacts as Cr. The layer was more brittle than
that of pure Ni.
The thickness of the Cr and Ni layers were rather similar, as
were the roughness of the layers. It is proposed that the difference
in erosion protection offered by these two electrodeposited layers
lies in the tendency of electrodeposited Cr layers to undergo brittle
fracturing. Similarly when micron-sized SiC particles were incor-
porated into Ni it became more brittle, and failed more readily than
pure Ni.
Acknowledgements
This work was funded by the COMET programme of the Austrian
Research Advancement Agency (Österreichische Forschungs-
förderungsgesellschaft FFG) and the government of Lower Austria
and supported additionally by Atotech Deutschland GmbH and
EADS Deutschland.
References
[1] R. Stewart, Carbon fibre composites poised for dramatic growth, Reinf. Plast.
53 (2009) 16–21.
1348 P. Lammel et al. / Wear 271 (2011) 1341–1348
[2] N.L. Hancox, The erosion of carbon fibre reinforced plastic by repeated liquid
impact, Wear 23 (1973) 71–81.
[3] S. Arjula, A.P. Harsha, Study of erosion efficiency of polymers and polymer
composites, Polym. Test 25 (2006) 188–196.
[4] J. Zahavi, S. Nadiv, G.F. Schmitt Jr., Indirect damage in composite materials due
to raindrop impact, Wear 72 (1981) 305–313.
[5] F.J. Gammel, O. Tuscher, R. Suchentrunk, Galvanisieren kohlenfaserverstärkter
Kunstoffe, Galvanotechnik 75 (1984) 988–993.
[6] M. Menningen, H. Weiss, U. Fischer, Metallic wear resistant coatings
for carbon fibre epoxy composite rolls, Surf. Coat. Technol. 71 (1995)
208–214.
[7] N. Dalili, A. Edrisy, R. Carriveau, A review of surface engineering issues
critical to wind turbine performance, Renew. Sust. Energ. Rev. 13 (2009)
428–438.
[8] Y. Iwai, T. Okada, T. Fujieda, K. Awazu, Effects of hard chromium plating on
cavitation erosion, Wear 128 (1988) 189–200.
[9] B.R. Mayers, S.C. Lynn, E. Jang, Case study—alternatives to the use of chromium
plating and conversion coatings at McClellan Air Force Base, California, in: Pro-
ceedings of the 9th Annual Aerospace Hazardous Materials Management Conf.,
Denver, CO, 1994.
[10] C. Steffani, M. Meltzer, Electrodeposited Tungsten–Nickel–Boron—A Replace-
ment for Hexavalent Chromium, in Report UCRL-ID-120611, Lawrence
Livermore National Laboratory, 1995.
[11] L Roberts, A. Galanis, Non-chromated technology works on aircraft, Plat. Surf.
Finish. 90 (2003) 8–11.
[12] E.W. Brooman, Compliant electrodeposited and electroless nano-structured
and nano-composite coatings to replace chromium coatings. Part 1, Galvan-
otechnik 96 (2005) 2843–2853.
[13] E.W Brooman, Compliant electrodeposited and electroless nano-structured and
nano-composite coatings to replace chromium coatings. Part 2, Galvanotechnik
97 (2006) 58–66.
[14] A.S.M.A. Haseeb, U. Albers, K. Bade, Friction and wear characteristics of elec-
trodeposited nanocrystalline nickel–tungsten alloy films, Wear 264 (2008)
106–112.
[15] V. Medeliene, The influence of B4 C and SiC additions on the morphological,
physical, chemical and corrosion properties of Ni coatings, Surf. Coat. Technol.
154 (2002) 104–111.
[16] H. Simunkova, Investigation of the mechanism of particle incorporation in
nickel-based dispersion layers and its influence on hardness and tribological
behaviour, Ph.D. Thesis, Institut für Chemische Technologien und Analytik, TU
Vienna, 2009.
[17] W. Wang, F. Hou, H. Wang, T. Guo, Fabrication and characterization of Ni–ZrO2
composite nano-coatings by pulse electrodeposition, Scripta Mater. 53 (2005)
613–618.
[18] M. Grundwürmer, O. Nuyken, M. Meyer, J. Wehr, N. Schupp, Sol–gel derived
erosion protection coatings against damage caused by liquid impact, Wear 263
(2007) 318–329.
[19] M. Grundwürmer, N. Schupp, M. Meyer, J. Wehr, Protection of erosion-stressed
aerospace structures through nanoparticle-enhanced inorganic-organic hybrid
coatings, in: patent WO 2008052510, EADS Deutschland GmbH, 2006.
[20] M. Grundwürmer, Development and characterisation of liquid impact resistant
sol–gel coatings, Ph.D. Thesis, Technical University of Munich, 2008, p. 171.
[21] C. Westmark, G.W. Lawless, A discussion of rain erosion testing at the United
States Air Force rain erosion test facility, Wear 186–187 (1995) 384–387.
[22] V. Darrort, M. Troyon, J. Ebothé, C. Bissieux, C. Nicollin, Quantitative study by
atomic force microscopy and spectrophotometry of the roughness and bright-
ness of electrodeposited nickel in the presence of additives, Thin Solid Films
265 (1995) 52–57.
[23] D. Mockute, G. Bernotiene, The interaction of additives with the cathode in
a mixture of saccharin, 2-butyne-1,4-diol and phthalimide during nickel elec-
trodeposition in a Watts-type electrolyte, Surf. Coat. Technol. 135 (2000) 42–47.
[24] M. Grundwürmer, N. Schupp, J. Wehr, Neuartige Prüfmethode zur Simulation
von Regenerosion, in: patent DE 102007060511, EADS Deutschland GmbH,
2007.
[25] J.V. Hackworth, A mechanistic investigation of the rain erosion of infrared
transmitting materials at velocities to Mach 2, in: Proceedings of the Fifth Inter-
national Conference on Rain Erosion and Associated Phenomena, Cambridge,
England, 1979, pp. 10.11–10.12.
[26] N.S. Corney, J.S. Pippet, The rain erosion resistance of some radome and IRdome
materials, in: Proceedings of the Sixth International Conference on Rain Erosion
and Associated Phenomena, Cambridge, England, 1983, pp. 24.21–24.27.
[27] A. Déom, D.L. Balagaes, Rain erosion resistance of IR materials: comparison of
rotating arms and MIJA measurements, in: Proceedings of the 8th European
Structures Conference, Nottingham, UK, 1994.
[28] A. Déom, R. Gouyon, C. Berne, Rain erosion resistance characterizations link
between on-ground experiments and in-flight specifications, Wear 258 (2005)
545–551.
[29] Ministry of Defence, Environmental Handbook for Defence Materiel, Part 4.
Natural Environments, DEF STAN 00-35, Chapter 6-02. The Effects of Rain, 2006,
pp. 181–206.
[30] Y.I. Oka, H. Miyata, Erosion behaviour of ceramic bulk and coating materials
caused by water droplet impingement, Wear 267 (2009) 1804–1810.
[31] B. Natarajakumar, Estimation of rain height from rain rate using regression-
based statistical model: application to SeaWinds on ADEOS-II, M.Sc. Thesis,
Dept. of Electrical Engineering, University of Kansas, 2004, p. 129.
[32] C.M. Preece, J.H. Brunton, A comparison of liquid impact erosion and cavitation
erosion, Wear 60 (1980) 269–284.
[33] C.M. Preece, J.H. Brunton, R.F. Lawn, Liquid impact erosion of Al-Mg and Al–Cu
alloys, Wear 60 (1980) 285–304.
[34] J.K. Dennis, T.E. Such, Nickel and Chromium Plating, 3rd ed., Woodhead Pub-
lishing, Ltd., Somerset, 1993.
[35] W.F. Adler, S.V. Hooker, Rain erosion mechanisms in brittle materials, Wear 50
(1978) 11–38.
[36] W.F. Adler, Waterdrop impact modeling, Wear 186–187 (1995) 341–351.