Progress in Organic Coatings 146 (2020) 105720

Contents lists available at ScienceDirect

Progress in Organic Coatings

journal homepage: www.elsevier.com/locate/porgcoat

Review

Effects of polymer hardness on the abrasive wear resistance of thick organic T

offshore coatings
A.W. Mombera,*, M. Irmerb, T. Marquardta
a
Muehlhan AG, Schlinckstraße 3, 21107 Hamburg, Germany
b
Fraunhofer Institute for Large Structures in Production Engineering IGP, Albert-Einstein-Straße 30, 18059 Rostock, Germany

ARTICLE INFO ABSTRACT

Keywords: For the first time, effects of polymer hardness on the abrasive wear resistance of organic coatings, designed for
Abrasive wear the corrosion protection of offshore wind power structures, are investigated at varying normal forces. The tests
Coatings are performed with a specially designed Taber abrasion machine. The results reveal statistically significant
Hardness effects of the polymer material Vickers hardness on the coating resistance against abrasive wear. With respect to
Offshore
the generic polymer type, the following ranking of the abrasive wear resistance is estimated: epoxy >
Polymers
polysiloxane > polyurethane. Thus, the most frequently applied top coat material (polyurethane) exhibits the
lowest abrasive wear resistance and may not be capable to protect the underlaying epoxy-based intermediate
coats. Polysiloxane would provide a longer protection against abrasive wear. The dominant material removal
modes in all polymers are microcutting and microfracturing, whereby the former mode is dominant at lower
normal forces, and the latter mode is dominant at higher normal forces. The power exponent of the classical
power relationship VA HPn depends on the mixture of the associated removal modes. The authors introduce a
new transition parameter in order to rank the associated individual material removal processes. Based on an
empirical-mathematical model, a two-parameter Weibull distribution function is derived, which links the
transition parameter to the applied normal force. A two-dimensional graph (nomogram) is designed, where the
different material removal modes are situated as functions of normal force and transition parameter.

1. Introduction high values for abrasive loads due to the movement of supply vessels on
boatlanding platforms. Two examples are provided in Fig. 1. Table 1
Abrasive wear notably contributes to the damage to protective summarizes investigations into the abrasive wear of organic coating
coatings of offshore wind power structures during the transport, in- systems. The results reveal that abrasive wear can deteriorate the color
stallation and service periods [1], and it is important to better under- and gloss resistance of top coats as well as the barrier (pore) resistance
stand the performance of organic protection systems under abrasive of intermediate/priming coats. Several references consider effects of
wear. The coating systems usually consist of a priming coat, one or pigments (alumina, glass flakes, MoS2) on the abrasive wear resistance.
more intermediate coats, and a top coat [2]. The latter coat is exposed With respect to offshore wind power coatings, none of the investiga-
to the environment and subject to abrasive wear. For outside areas, the tions focuses on the performance of the generic polymers (binders) of
top coats for marine and offshore structures are based on polyurethane the coatings at varying normal forces. The standard Taber tests cited in
and polysiloxane [2]. The reason is their good resistance to ultraviolet Table 1 are usually run with a normal force of 10 N, and an accurate
light and their superior color and gloss stability. Intermediate coats are and repeatable adaption of lower or higher normal forces is not pos-
almost entirely made of epoxy as the basic polymer. Epoxies provide a sible.
good barrier resistance, but they are not resistant to ultraviolet light Table 2 summarizes investigations into the effects of different
and deteriorate once the protective top coat is damaged or worn. hardness parameters of various polymers on their abrasive wear per-
Therefore, the resistance of top coats against abrasive wear is an im- formance. None of the investigations consider protective coating sys-
portant lifetime parameter for corrosion protection coating systems. tems for offshore wind power structures. Moreover, the results do not
The coats are subject to varying normal forces, ranging from low values indicate any general trends between hardness values and abrasive wear
for rubbing wires during transport of maintenance services to rather target parameters. Reasons for the lack in general trends of hardness

⁎
Corresponding author.
E-mail address: [email protected] (A.W. Momber).

https://doi.org/10.1016/j.porgcoat.2020.105720
Received 18 March 2020; Received in revised form 23 April 2020; Accepted 23 April 2020
Available online 13 May 2020
0300-9440/ © 2020 Elsevier B.V. All rights reserved.
A.W. Momber, et al. Progress in Organic Coatings 146 (2020) 105720

Nomenclature N Number of abrasive wear cycles

VA Volume removed due to abrasive wear
fab Scratch geometry factor VC Volume removed due to microcutting and microploughing
FN Normal force Vf Volume removed due to microfatigue
FNc Threshold normal force for microfracturing VF Volume removed due to microfracture
GIc Specific fracture energy αT Transition parameter
h0 Initial coating thickness β Work hardening factor
HIT Indentation hardness Δh Coating thickness reduction due to abrasive wear
HM Material hardness φ Deformation
hN Coating thickness after abrasive wear λ Weibull distribution scale factor
HP Polymer material Vickers hardness μ Weibull distribution shape factor
KIc Mode-I fracture toughness

effects include the following: in this study are the normal force and the number of cycles (exposure
time), whereas the base body surface parameter considered in this study
• There is no single material deformation/removal mode in abrasive is the hardness of the top coat polymer material. The hardness of ma-
wear processes [13,31,32]; terials is considered an important system parameter for abrasive wear
• Hardness is stochastically distributed over material surfaces, and processes. The effect of material hardness on abrasive wear processes
abrasive wear can be considered a stochastic process [33,34]; can generally be expressed as follows:
• Abrasive wear resistance is not a material parameter, but it is a 1
system parameter (including effects of load, size, temperature, VA
HnM (1)
loading rate, etc.).
• Different abrasive wear testing methods, machines and conditions Here, VA is the volume loss due to abrasive wear, HM is the hardness of
are used. the abraded material, and n is a power exponent. The exponent n de-
pends on the material removal mode (see Table 2). For ductile (plastic)
Abrasive wear resistance is a system parameter. As illustrated in material deformation and removal processes, such as microcutting and
Fig. 2, it depends on loading regime, properties and responses of the ductile ploughing, the exponent can have a value of 1.0 [35,36]:
contributing partners, and on environment. The counter body in this
study is the Taber abrasive wheel, and the base body is the coated steel
VA H M1 (2)
sample. The intermediate zone is determined by the contact between For brittle (elastic) material deformation and removal processes,
the aluminium oxide particles of the abrasion wheel and the surface of such as microcracking, a value of 0.5 was found for the exponent
the top coat of the coated samples. The loading parameters considered [37,38]:

VA H M1/2 (3)

Effects of process and material parameters on the power exponent

have not been investigated yet for polymeric coating materials. The
objectives of this paper are:

- to investigate and to identify the determining material removal

modes during abrasive wear of typical polymer materials at different
applied normal forces;
- to quantify the interactions of varying normal forces and varying
polymer hardness values on the abrasive wear performance of off-
shore wind power top coat polymer materials;
- to establish an empirical-mathematical relationship between
polymer hardness, applied normal force and hardness-power ex-
ponent.

2. Experimental setup

2.1. Investigated coatings

Eight organic coatings were investigated, whereby three generic

polymer materials were considered, namely epoxy, polyurethane and
polysiloxane. The coatings were specified for the corrosion protection
of offshore wind power structures and intended for use in marine tidal
zone and in marine splash zone [39]. Information on the individual
coating materials is provided in Table 3. The application of the coatings
and the surface preparation were carried out according to the specifi-
cations of the coating material manufacturers. The substrate for the
Fig. 1. Damages to offshore wind power coatings due to abrasive wear. a – coatings was rolled, flat, low-carbon high-strength steel (S355J2 +
Damage to transition piece coating from supply boat during service (KEMA AR), typically used for offshore structures, with a plate thickness of
Nederland BV, Arnhem, Netherlands). b – Damage to tower coating during 5.0 mm. All steel substrates were blast-cleaned according to [40] to a
transport (Muehlhan AG, Hamburg, Germany). surface preparation degree of Sa 2½ according to [41]. The roughness

2
A.W. Momber, et al. Progress in Organic Coatings 146 (2020) 105720

Table 1
Review of abrasive wear investigations on organic coatings.
Method Coatings Results Reference

Rotating sand paper disc Acrylic-melamine Weight loss increased with cycles in a square-root mode [3]
600 cycles Alumina particles improved abrasion resistance
Room temperature
Taber test Phenolic, epoxy, novalac Filler material and resin chemistry determined resistance [4]
1000 cycles
Room temperature
Taber test Coil coatings (hot dip galvanized zinc + polyester) Abrasion reduced pore (barrier) resistance [5]
With abrasive sand and electrolyte
600 cycles
Room temperature
Taber test Powder coating (epoxy polyester) Abrasion reduced gloss [6]
With abrasive paste
1000 cycles
Room temperature
Taber test (modified) Powder coating (epoxy polyester) Thickness loss increased linearly with cycle number (small grit particles) [7]
Rubber wheel with abrasive sand Abrasion reduced pore (barrier) resistance
700 cycles
Room temperature
Taber test (modified) Powder coating (epoxy polyester) Thickness loss increased linearly with cycle number [8]
4000 cycles Abrasion increased coating electrical capacity
Room temperature Abrasion reduced pore (barrier) resistance
Taber test Powder coating (epoxy polyester) Abrasion reduced barrier resistance [9]
Rubber wheel with abrasive Coarser grit was more abrasive
particles
2000 cycles Round grit was less abrasive
Room temperature
Taber test Sol-gel films Definition of Abrasion-Index [10]
Room temperature Sol-gel owned high abrasion resistance
Taber test Clearcoats (isocyanate/polyol) Abrasion reduced gloss [11]
Room temperature Nano-silica increases abrasion resistance
Sliding test with alumina ball; 0 °C Polyurethane, epoxy, polysiloxane Polysiloxane had high abrasion resistance [12]
Glass flakes reduced abrasion resistance
Taber test Polyurethane, Epoxy Abrasion resistance decreases with decreasing temperature [13]
2000 cycles Thickness loss increases linearly with cycle number
Low-temperature (0 °C)
Taber test Polyamide-based primer and MoS2/PTFE-modified Abrasion resistance was very high at moderate and low temperatures due [14]
topcoat (30 μm) to MoS2 addition
2000 cycles Thickness loss increased linearly with cycle number
Low-temperature (0 °C)
Taber test Bridge coatings (epoxy, polyurethane) Epoxy performed better than polysiloxane [15]
Total top coat removal
Sliding test with 10-mm steel ball Polymer-ceramic coatings UHMWPE and polymer-ceramic coatings performed better than vinyl- [16]
based coatings
50,000 cycles (wet and dry) Fibrous polymer coatings provided very low abrasion resistance

Table 2
Effects of hardness on the abrasion of polymers.
Test Material/polymer Hardness Target parameter Trend/correlation Reference

Abrasion tribotester PA6, PEEK Shore D hardness Wear depth No correlation (for PEEK only) [17]
Ball impact EP, PU, PMMA, PS, PE, PP Indentation hardness (ball) Wear rate log(H/GIc ) [18]
Scratch test High-performance polymers Martens hardness Scratch depth Depth reduced with increasing [19]
hardness
Abrasive paper Polyamide 6 (filler-dispersed) Microhardness Wear rate 1/H (model) [20]
Microscale abrasion HDPE, PC, PETG, PMMA, PP, PS, Martens hardness Wear rate Linear increase with hardness [21]
PVC
Underwater steel ball and basalt stone Polyurea Shore D hardness Wear depth Non-linear increase with [22]
abrasion hardness
Roller with loose abrasives Polyurethane Shore A hardness Abrasive resistance Slight linear increase with [23]
index hardness
Sand/rubber wheel PEI, PEEK, PES, PA6, PBT Vickers hardness Wear rate No correlation [24]
Taber test Polyurethane coatings Knoop hardness Weight loss Inverse linear correlation [25]
Taber test Polyester (powder coatings) Knoop hardness Weight loss No correlation [26]
Taber test Sol-gel coatings Universal hardness Abrasion resistance No correlation [27]
Taber test Offshore coatings (PU, EP) Mohs hardness (of fillers) Abrasion resistance Linear correlation (high filler [13]
contents)
Wheel test in slurry Acrylate, urethane, Vickers hardness Wear depth Significant negative correlation [28]
polycarbonate
Silicone carbide paper (in water) Different polymer resins Vickers hardness Wear depth No correlation [29]
Scratch test PMMA, PEEK, EP, UHMWPE Scratch hardness Wear volume (via fab) Positive linear correlation [30]

3
A.W. Momber, et al. Progress in Organic Coatings 146 (2020) 105720

Fig. 2. Tribological system for the abrasive wear process. Loading parameters characterize the loading conditions during abrasive wear. The surface structure and
texture parameters characterize the resistance of the coating materials. (The system parameters hardness and normal force are marked with orange arrows.).

Table 3 of the surfaces to be coated was measured with a surface roughness

Investigated top coats (based on information from product data sheets). stylus instrument; it was Rz = 50–75 μm. The coating applications were
Coating Polymer Filler/ Hardener Solids Density performed with airless spray guns, whereby all spray parameters, en-
Pigment [vol.%] [kg/l] vironmental parameters and drying parameters were set according to
the data sheets and application guidelines of the respective coating
1 Polyurethane Zinc Aliphatic 67 1.4
products. The dry film measurements were performed for each layer
2 Polyurethane Zinc Aliphatic 67 1.4
3 Polyurethane Zinc Aliphatic 67 1.4
with a magnetic thickness gauge according to [42] (gauge type: Lep-
4 Epoxy – – 95 – toskop 2042, Karl Deutsch).
5 Polyurethane Polyacrylat/ 57 1.21
Polyisocyanat
6 Polysiloxane – Amine 90 1.4
2.2. Hardness measurements
7 Polysiloxane – Amine 90 1.4
8 Epoxy – Amine 98 – The hardness of the polymer materials was measured by means of
an instrumented nano-indentation testing machine (Fischerscope
HM2000S) according to [43]. The measurements delivered the in-
dentation hardness, which was transferred into Vickers hardness via a
standard protocol [43]:

Fig. 3. Design and set-up for the abrasive wear tests.

4
A.W. Momber, et al. Progress in Organic Coatings 146 (2020) 105720

HC = 94.53 HIT (4) h= h 0 hN (5)

Here, HC is the Vickers hardness of the polymers in MPa, and HIT is the Here, Δh is the loss of coating layer thickness, h0 is the initial coating
indentation hardness of the polymers in GPa. The testing force was layer thickness, and hN is the coating layer thickness after a number of
500 mN, and the contact duration was 20 s. Five measurements were N revolutions. The thickness measurements were carried out after every
performed for each sample. 500 revolutions (N = 500) at eight measuring points with five in-
dividual measurements and allowed the creation of abrasion curves
2.3. Abrasive wear testing device Δh = f(N). The procedure is illustrated in Fig. 4. The volume loss due to
abrasive wear is related to the loss of coating layer thickness as follows:
The abrasive wear resistance was estimated with a rotating abrasive
wheel as recommended in [7] for the testing of offshore coatings. The VA = h (d22 d12)
4 (6)
test configuration is illustrated in Fig. 3. The abrasive wear tests were
performed according to [44]. The tests were carried out in a test rig, Here, VA is the volume loss due to abrasive wear, Δh is the loss of
developed at Fraunhofer IGP, which allowed the precise adjustment of coating layer thickness, and d1 and d2 are the internal and the external
the effective normal force when using two standard rotating abrasive diameter of the abrasive wear path (see Fig. 4).
rubber wheels. Details of the test rig are provided in Fig. 3. By con-
trolling the electric cylinder, a dynamic force control up to a maximum 2.4. SEM inspections
effective normal force of 1.0 kN was possible. In addition, the test bench
offered a significantly increased flexibility in measurements. The tests All abraded surfaces were inspected with a scanning electron
were performed with three different normal forces per rotating wheel, microscope (SEM). The microscope was a type “JSM-IT100”. The
namely 2.5 N, 10 N and 25 N. Standard Taber testing machines are images were taken in the secondary electron mode. The resolution was
usually run with a normal force of 10 N, and an accurate and repeatable 4.0 nm, the acceleration voltage was 5.0 kV, the scanning distance was
adaption of lower or higher normal forces is not possible. Each abrasive 10–12 mm, and the probe current was 35 μA. The samples were
wear test ran over a total of 2000 revolutions. After each 500 revolu- sputtered with gold (10–15 nm).
tions, the wheels were cleaned and re-surfaced. For each coating system
and each normal force level, three specimens were tested. The abrasive 2.5. Statistical investigations
wheel was a type Calibrase CS-10 (rubber matrix, aluminium oxide
particles). The Shore hardness of the wheels was Shore D29 [45]. The In order to evaluate the significance of the contribution of the
wheels were cleaned and resurfaced after every 250 cycles according to coating hardness to the loss of coating thickness, Analysis Of Variance
[44]. Abraded coating debris was vacuumed from the surface via two (ANOVA) was performed. The F-value, p-value and the contribution
nozzles. Usually, the loss due to abrasive wear is determined in milli- were used to evaluate potential significances. The F-value is a para-
grams (weight-loss based). Because the used steel substrate plates had a meter that helps to decide if the means between two populations are
thickness of 5.0 mm and, therefore, a weight of about 590 g plus coating significantly different. It is calculated by dividing the variance of the
system, a sufficiently accurate weighing was not possible. Therefore, group means (here, the means of polymer hardness and the means of
the loss of layer thickness, Δh, was determined by means of dry film coating thickness reduction) by the mean of the within-group variance
thickness measurements. The dry film thickness measurements were (e.g. the means of hardness within one polymer): The higher the value
performed according to [42] (gauge type: Leptoskop 2042, Karl for F, the higher is the significance that the means between two po-
Deutsch). The loss of layer thickness was estimated as follows: pulations are different. The p-value expresses the significance of a

Fig. 4. Coated sample with dimensions and abrasive wear path.

Left: Main sample dimensions and the eight measurement points for dry film thickness measurements; right: Abraded path on a polyurethane top coat with internal
(d1) and external (d2) abrasion path diameter.

5
A.W. Momber, et al. Progress in Organic Coatings 146 (2020) 105720

relationship between parameters. The analysis was based on a sig- modes are dominant at the low normal force, the latter mode becomes
nificance level of p = 0.05 (confidence level of 95 %). Thus, the con- dominant at the high normal force.
tribution of a factor is statistically insignificant for p > 0.05 [46]. In-
formation on the estimation of the statistical parameters is provided in 3.4. Results of the statistical investigations
the Appendix A.
Results of the statistical investigations are summarized in Table 6.
3. Experimental results The estimated significance values are all lower than p = 0.05; thus, the
effect of the polymer hardness on the coating thickness reduction is
3.1. Results of the hardness measurements significant. For the two higher normal forces, the p-values are lower
than a threshold of p = 0.001, which characterizes extreme statistical
Results of the hardness measurements are listed in Table 4. The significance [46]. Thus, polymer hardness statistically significantly in-
hardness values can be ranked according to the generic polymer type as fluences the coating thickness reduction. The high F-values for the
follows: epoxy > polysiloxane > polyurethane. The relative standard higher normal forces indicate that the statistical means of the two
deviations are less than 8% for all materials, indicating a negligible parameters are significantly different, which supports the trend for the
effect of fillers or pigments on the hardness. The estimated hardness p-values. Although significance is proven, coating hardness cannot ex-
values very well correspond to Vickers hardness values reported in [47] plain all effects related to the reduction of coating thickness. As can be
for organic offshore coatings. The ranking of the hardness also agrees seen in Table 6, polymer hardness can explain 8–65% of the effects
with a ranking of polymer hardness values provided in [47], who only, whereby the percentage increases with increasing normal force.
found: epoxy > (polysiloxane, polyurethane) > polyester.
4. Discussion
3.2. Results of the abrasive wear measurements
The presented results reveal that the polymer Vickers hardness has a
Examples of abrasive wear functions for the three generic polymer significant effect on the loss of layer thickness, respectively on the
materials are provided in Fig. 5. All abrasive wear functions follow a abraded volume. Any change in polymer hardness leads to a significant
linear relationship: change in the loss of layer thickness. The qualitative ranking of the
abrasive wear resistance is as follows: epoxy > polysiloxane >
h= a N+ b (7)
polyurethane. With respect to a quantitative ranking, the ratios of the
Here, Δh is the reduction in coating thickness, and N is the number of average abrasive wear resistance values for all measurements at all
revolutions (representing exposure time). The linear trend confirms normal force levels are (epoxy:polysiloxane:polyurethane):
results in [13,14,48] for the Taber abrasion of organic offshore coatings 1.0:0.91:0.58. The most frequently applied top coat material for off-
and results in [7,8] for the Taber abrasion of organic powder coatings shore wind power structures (polyurethane) exhibits by far the lowest
with fine grit particles. The trend in abrasive wear resistance is as fol- abrasive wear resistance and may not be capable to protect the un-
lows: epoxy > polysiloxane > polyurethane, which agrees qualita- derlaying epoxy-based intermediate coatings. Polysiloxane would pro-
tively with the trend for the hardness of the polymer materials. This vide a notably longer protection against abrasive wear. The average
trend also agrees with results in [15], where a higher Taber abrasion abrasive wear resistance ratio (for all normal forces) between poly-
resistance was found for epoxy-based bridge coatings compared to urethane and polysiloxane is 0.52. However, if the individual normal
polysiloxane-based coatings. The relationships between Vickers hard- forces are considered, the picture is more differentiated: the ratio
ness and abraded material for the three normal force levels are illu- polyurethane/polysiloxane is 0.26 for FN = 2.5 N, it is 0.51 for
strated in Fig. 6, and the results of regressions with respect to Eq. (1) are FN = 10 N, and it is 0.77 for FN = 25 N. These results indicate that the
listed in Table 5. differences in the abrasive wear resistance of the two top coat materials
may become less pronounced at high normal forces. This is a topic for
3.3. Results of the SEM inspections subsequent investigations.
Three material removal modes are observed for the polymer mate-
Results of SEM inspections of the three generic polymer matrix rial removal process, whereby microcutting and microcracking are the
materials, abraded at the low normal force (2.5 N) and the high normal dominant processes. With respect to material removal modes during
force (25 N), are provided in Figs. 7–9. Fig. 7 illustrates the abrasive abrasive wear, Zum Gahr [32] distinguishes between four major modes,
wear of a polyurethane matrix (coating 1). At the low normal force namely microcutting, microploughing, microfatigue and micro-
(Fig. 7a), although a mixed-material removal mode can be recognized, cracking. The individual modes can be identified by use of a scratch
the material removal is dominated by ductile microcutting. The two geometry factor, fab. For fab = 0, microploughing dominates; for
wider paths only show traces of microcracking. The situation is similar fab = 1, microcutting dominates; for fab > 1, microcracking dominates
for the high normal force (Fig. 7b), whereas microcracking is more [32]. The sole action of an individual material removal mode during the
severe, particularly in the lower section of the image. Fig. 8 illustrates abrasion process is very unlikely, which is proven in Figs. 7–9. The
the abrasive wear of an epoxy matrix (coating 4). In contrast to the
polyurethane, the material removal process exhibits notable features of Table 4
microcracking at the low normal force (Fig. 8a). This applies as well to Results of hardness measurements.
the situation for the high normal force, but cracking is more pro- Coating Vickers hardness Indentation depth in μm
nounced and dominates the material removal process (Fig. 8b). Fig. 9
illustrates the abrasive wear of a polysiloxane matrix (coating 6). At the Average in MPa Standard deviation in %
low normal force, the material is removed almost entirely due to mi-
1 7.8 4.8 ∼15
crocutting (Fig. 9a), whereas a mixed-mode material removal process 2 10.4 7.9 ∼15
can be recognized for the high normal force, including microploughing 3 11.6 5.6 ∼15
and severe microcracking (Fig. 9b). 4 24.2 4.2 ∼10
In summary, mixed-mode removal processes are identified on all 5 8.4 1.6 ∼17
6 18.5 0.6 ∼12
matrix materials at the two normal force levels, whereby the abrasive
7 16.8 1.8 ∼13
wear process includes the following material removal modes: micro- 8 25.1 2.6 ∼10
cutting, microploughing and microcracking. Whereas the two former

6
A.W. Momber, et al. Progress in Organic Coatings 146 (2020) 105720

Fig. 5. Coating thickness loss due to abrasive wear versus number of revolu-
tions (exposure time) for three generic polymer materials (applied normal
force: 25 N); solid lines express linear regressions.

material removal process in polymers is rather the result of a combi-

nation of material removal modes [13,31,49]. When microcutting
(plastic response) and microcracking (elastic response) are considered
the two dominant modes, the following relationship for a mixed-mode
material removal process can be established (adapted from [50,51]):

VA = (1 T) VC + T VF (8)

Here, VA is the volume removed due to abrasive wear, VC is the volume

removed due to microcutting, VF is the volume removed due to mi-
crocracking, and αT is a balancing transition parameter, which de-
termines the probability for an individual material removal mode to
occur. For αT = 0, microcutting is the sole abrasive wear mode; for
αT = 1, abrasive wear occurs solely due to microcracking; for
0 > αT > 1, mixed-mode abrasive wear occurs. It shall be emphasized
that the transition parameter αT has some analogy to Zum Gahr’s [32]
parameter fab1, whereby higher fab-values correspond to higher αT-va-
lues. Ikramov and Machkamov [49], too, modeled the abrasive wear of
materials to be the sum of two simultaneously occurring material re-
moval mechanisms. Based on the results listed in Table 5, and when
Eqs. (2) and (3) are combined, Eq. (7) can be further specified as fol-
lows:

VA = [(1 T) f1 HP 1.0] + [ T f2 HP 1/2] (9)

Here, VA is the volume removed due to abrasive wear, and HP is the

polymer material Vickers hardness. The terms f1/2 are functions of the
tribological system, and they cover effects of other process and material
parameters. From αT = 0 follows: VA HP 1, and from αT = 1 follows:
VA HP 1/2 . The results listed in Table 5 can be used to establish at least
a qualitative relationship between αT and the normal force. For the
highest normal force (FN = 25 N), the power exponent is n = 0.53,
which is close to the exponent for brittle material removal due to mi-
crocracking (αT→1). The power exponents for the two lower normal
forces (FN = 2.5 N and 10 N), in contrast, are notably higher (n = 0.71
for 10 N), whereby the lowest normal force leads to a power exponent Fig. 6. Effects of polymer Vickers hardness and normal force on the loss of
of 0.92, which is close to the exponent for microcutting (αT→0). coating layer thickness after abrasive wear; solid lines represent regressions
according to Eq. (1). a – Applied normal force FN = 2.5 N; b – Applied normal
force FN = 10 N; c – Applied normal force FN = 25 N.
1
For example, Zum Gahr [[32]] suggested: VA = fab VP + (1 fab) Vf
Ploughing + Cutting Fatigue

7
A.W. Momber, et al. Progress in Organic Coatings 146 (2020) 105720

Table 5
Regression parameters for Eq. (1).
Normal force in N Power exponent n Radj²

2.5 0.92 0.44

10 0.71 0.69
25 0.53 0.70

Fig. 8. Abraded surface sections of coating 4 (epoxy at varying normal force

after 2000 revolutions. a – Low applied normal force (FN = 2.5 N); the domi-
nant material removal modes are microcutting and microploughing, and some
microcracking. b – High applied normal force (FN = 25 N), mixed material re-
moval modes, including microcutting, microploughing and microcracking.

[53] introduced transition maps and showed that material removal

modes turned from plastic deformation to lateral fracture if a given
Fig. 7. Abraded surface sections of coating 1 (polyurethane) at varying normal threshold normal load (force) was exceeded. The following relationship
force after 2000 revolutions. a – Low applied normal force (FN = 2.5 N); the can be assumed between the material hardness and a threshold normal
dominant material removal modes are microcutting and microploughing, slight force for introducing cracking during abrasive wear [54] (Fig. 10):
microcracking. b – High applied normal force (FN = 25 N), mixed material re- 1
moval modes, including microcutting, microploughing and microcracking. FNc
HC (10)
This equation suggests that the polymer matrix with a higher
Although the power exponents do not reach unity (for a plain micro-
hardness experiences a higher susceptibility for cracking; the corre-
cutting process at αT = 0), the higher power exponents for the two
sponding ranking would be: epoxy > polysiloxane > polyurethane.
lower normal forces indicate a mainly microcutting-driven material
The meaning of the threshold normal force is illustrated in Fig. 10.
removal process (αT→0). This agrees with the results of the SEM in-
Because all investigated materials exhibit features of cracking at all
spections of the abraded surfaces at different normal forces (Figs. 7–9).
normal force levels, the threshold normal force is FNc < 2.5 N for the
None of the power exponents reaches the theoretical values for the
conditions in this study. Results of the SEM inspections, provided in
plain material removal modes, and this is an expected result because of
Figs. 7–9, confirm this approach. The response of the epoxy matrix is
the simultaneous (mixed) action of microcutting and microcracking
characterized through notable microcracking at the low normal force
(see Figs. 7–9).
(Fig. 8a), whereas the material removal was dominated by microcutting
More supporting evidence is provided in [52,53]. Briscoe et al. [52]
for polyurethane (Fig. 7a) and polysiloxane (Fig. 9a) at the low normal
introduced scratch deformation maps for polymers (polycarbonate,
force.
polyethylene), where normal forces were plotted versus the cone angles
For the three frame conditions: (i) αT = 0 for microcutting; (ii)
of the scratching tools. For cone angles less than 60°, the material re-
αT = 1 for microcracking; (iii) inflection/transition point between the
moval mode changed from ductile ploughing and plastic deformation at
abrasion modes, and for FN > FNc, the relationship between αT and FN
low contact forces to ductile ploughing, cracking and chip formation at
can be approximated with a two-parameter Weibull distribution func-
higher normal forces. For abrasion and erosion processes, Hutchings
tion [50,54]:

8
A.W. Momber, et al. Progress in Organic Coatings 146 (2020) 105720

Fig. 10. Suggested relationship between the derived transition parameter αT

and the applied normal force according to Eq. (11); the micro-abrasion illus-
trations in the graph are taken from [18]; elastic response = microcracking,
plastic response = microcutting; the left section represents entirely plastic re-
sponse for FN < FNc.

an issue for further investigations.

The transition parameter αT shall also depend on the material
properties of the partners in the tribological system. Zum Gahr [32]
found the transition parameter fab to depend, among others, on Young’s
modulus and work-hardening degree of the abraded materials. For
small-scale, impact-based material removal processes, Momber [55]
found that the transition between plastic and elastic material removal
mechanisms was governed by a ratio KIcm/HM2·m/3, whereby the ex-
ponent m depended on the particular loading mode. This ratio is in
Fig. 9. Abraded surface sections of coating 9 (polysiloxane) at varying normal agreement with results of ball impact erosion tests on polymers re-
force after 2000 revolutions. a – Low applied normal force (FN = 2.5 N); the ported in [18], who found that the wear resistance of numerous poly-
dominant material removal modes are ductile microcutting; slight micro- mers was related to the ratio HM/GIc (see Table 2). However, systematic
cracking. b – High applied normal force (FN = 25 N); the dominant material investigations are required in order to further explore these issues and
removal mode is microcracking.
to establish valid relationships.
The results in Table 6 reveal that polymer hardness, although it
Table 6 exhibits significant, or even extremely significant, correlations, cannot
Results of ANOVA calculations. explain all effects on the loss of coating thickness. Therefore, other
Normal force ANOVA material parameters shall be considered. According to standard
polymer abrasion models, these parameters include elongation (strain)
Contribution F-value p-valuea to rupture, Young’s modulus, and fracture toughness. With respect to
coating composition, additional effects particularly include filler/pig-
2.5 N 8.1 % 4.13 0.048
10 N 52.6 % 53.37 0.000
ment type and properties [13,31]. These effects seem to be less im-
25 N 64.9 % 92.50 0.000 portant at the higher normal forces. These results require further con-
sideration.
p < 0.001 = extremely significant [46].
a
p > 0.05 = insignificant; p < 0.01= very significant.
5. Summary and conclusions
T (FN) =1 exp [ ( FN ) µ ] (11)
For the first time, effects of polymer hardness on the abrasive wear
In that equation, αT is the transition parameter, FN is the normal resistance of organic top coats, designed for the corrosion protection of
force, λ is the scale parameter, and μ is the shape parameter of the offshore wind power structures, are investigated at varying normal
distribution function. This approach has an analogy to a suggestion forces with a specially designed Taber abrasion machine. The following
made by Zum Gahr [32], who related fab to the normal force, ranking of the abrasive wear resistance is estimated: epoxy >
fab = f(FN) , via a Weibull distribution function2 . The general structure polysiloxane > polyurethane. Thus, the most frequently applied top
of Eq. (11) is illustrated in Fig. 10, where the two abrasion material coat material (polyurethane) exhibits the lowest abrasive wear re-
removal modes are separated at the critical normal force. For FN≤FNc→ sistance and may not be capable to protect the underlaying epoxy-based
αT = 0. The detailed investigation of the two distribution parameters is intermediate coatings. Polysiloxane would provide a longer protection
against abrasive wear. This is a result with an important practical re-
2 levance. The dominant material removal modes in all polymers are
Zum Gahr [[32]] suggested: fab = 1 , with φS=f(FN)
2 S
exp ln
S,lim microcutting and microfracturing, whereby the former mode is

9
A.W. Momber, et al. Progress in Organic Coatings 146 (2020) 105720

dominant at lower normal forces, and the latter mode is dominant at influence the work reported in this paper.
higher normal forces. Based on an empirical-mathematical model, a
two-parameter Weibull distribution function is derived, which links a Acknowledgements
newly introduced transition parameter to the applied normal force.
Based on a two-dimensional nomogram, the different material removal The investigations were funded by the German Federal Ministry of
modes can be rated as functions of normal force and transition para- Education and Research (BMBF) in the frame of the innovation in-
meter. itiative „Wachstumskerne - Unternehmen Region“, sub-program: „OWS-
MV: Offshore Wind Solutions-Mecklenburg-Vorpommern“. Thank is
Declaration of Competing Interest addressed to Nieradzik, Fraunhofer IFAM, Bremen, Germany, where the
hardness measurements were performed. Thank is also addressed to
The authors declare that they have no known competing financial Kathrin Hasche of Fraunhofer IGP, Rostock, Germany, who conducted
interests or personal relationships that could have appeared to the SEM inspections.

Appendix A. Statistical parameters

(Manual “Minitab®”, Minitab 17, Minitab Inc., State College, Pennsylvania, USA, 2016)
The F-value is defined as follows:
MS
F=
MSerror (A1)
Here, MS is the mean square, and MSerror is the mean square for error. The mean square for each term is the sum of squares divided by the degrees of
freedom:
SS
MS=
DF (A2)
The mean square for error is an estimate of the variance in the data left over after differences in the means have been accounted for. The sum of
squares is a measure of deviation from the mean. The sum of squares for a set X of N items is defined as follows:
N
SS= (Xi X)2 (A3)
i= 1
th
In the equation, SS is the sum of squares, N is the number of items, Xi is the i item of the set, and X is the mean of all items in the set.
The parameter “contribution” is defined as follows:
SS 100
Contribution =
SStotal (A4)
Here SStotal is the total sum of squares. The total sum of squares is the sum of squares for the factor and the sum of squares for error.
The p-value is a tabulated parameter, which can be read as a function of F-value and degree of freedom.

References 348–349 (2016) 166–180.

[14] A.W. Momber, M. Irmer, N. Glück, Investigation into the performance of a dual-
layer thin-film organic coating during accelerated low-temperature offshore testing,
[1] A.W. Momber, Quantitative performance assessment of corrosion protection sys- ASME J. Offshore Mech. Arctic Engng. 139 (2017) 041402 1–9.
tems for offshore wind power transmission platforms, Renew. Energy 94 (2016) [15] E.A. Toubia, S. Emami, Experimental evaluation of structural steel coating systems,
314–327. ASCE J. Mater. Civ. Eng. 28 (12) (2016) 04016147.
[2] A.W. Momber, T. Marquardt, Protective coatings for offshore wind energy devices [16] P.V. Vijay, P.R. Soti, D.A. Banerjee, K.A. Sierros, Abrasion resistance of polymer and
(OWEAs): a review, J. Coat. Technol. Res. 15 (1) (2018) 13–40. polymer–ceramic composite coatings for steel hydraulic structures, J. Coat.
[3] D. Hosseinpour, J.T. Guthrie, J.C. Berg, The effect of α-alumina filler/acrylic-mel- Technol. Res. (2019), https://doi.org/10.1007/s11998-019-00286-3.
amine polymer interfacial interactions on the abrasion resistance of an automotive [17] L. Zsidai, L. Katai, Abrasive wear and abrasion testing of PA6 and PEEK composites
topcoat layer, Progr. Organ. Coat. 62 (2008) 214–218. in small-scale model system, Acta Polytech. Hungarica 13 (6) (2016) 197–214.
[4] R.S. Lauer, Advancements in the abrasion resistance of internal plastic coatings, [18] K. Friedrich, Erosive wear of polymer surfaces by steel ball blasting, J. Mater. Sci.
Mater. Perform. 53 (2015) 52–55. 21 (1986) 3317–3332.
[5] S. Rossi, F. Deflorian, J. Fiorenza, Environmental influences on the abrasion re- [19] K. Friedrich, H.J. Sue, P. Liu, A.A. Almajid, Scratch resistance of high performance
sistance of a coil coating system, Surf. Coat. Technol. 201 (2007) 7416–7424. polymers, Tribol. Int. 44 (2011) 1032–1046.
[6] S. Rossi, F. Deflorian, E. Scrinzi, Reduction of aesthetical properties of organic [20] Y. Morioka, Y. Tsuchiya, M. Shioya, Correlations between the abrasive wear, fa-
coatings caused by mechanical damage, Mater. Design 30 (2009) 1511–1517. tigue, and tensile properties of filler-dispersed polyamide 6, Wear 338–339 (2015)
[7] A. Cambruzzi, S. Rossi, F. Deflorian, Reduction on protective properties of organic 297–306.
coatings produced by abrasive particles, Wear 258 (2005) 1696–1705. [21] P.H. Shipway, N.K. Ngao, Microscale abrasive wear of polymeric materials, Wear
[8] S. Rossi, F. Deflorian, M. Risatti, Modified Taber apparatus and new test geometry 255 (2003) 742–750.
to evaluate the reduction of organic coatings corrosion protective properties in- [22] X. Wang, S.Z. Luo, G.S. Liu, L.C. Zhang, Y. Wang, Abrasion of flexible protective
duced by abrasive particles, Surf. Coat. Technol. 201 (2006) 1173–1179. materials on hydraulic structure, Water Sci. Engng. 7 (1) (2014) 106–116.
[9] S. Rossi, F. Deflorian, M. Risatti, Reduction of corrosion protection properties of [23] D. Capandis, The influence of hardness of polyurethane on its abrasive wear re-
organic coatings due to abrasive damage produced by natural sands, Wear 261 sistance, Tribolog. 4 (2016) 29–39.
(2006) 922–931. [24] A.P. Harsha, An investigation on low stress abrasive wear characteristics of high
[10] H. Yahyaei, M. Mosheni, Use of nanoindentation and nanoscratch experiments to performance engineering thermoplastic polymers, Wear 271 (2011) 942–951.
reveal the mechanical behavior of sol-gel prepared nanocomposite films on poly- [25] H. Aliman, M. Hinnawi, P. Pradhan, Y. Alkassar, ANN & ANFIS models for pre-
carbonate, Tribol. Int. 57 (2013) 147–155. diction of abrasive wear of 3105 aluminium alloy with polyurethane coating,
[11] E. Scrinzi, S. Rossi, P. Kamarchik, F. Deflorian, Evaluation of durability of nano- Tribol. Industr. 38 (2) (2016) 221–228.
silica containing clear coats for automotive applications, Progr. Organ. Coat. 71 [26] R. Barbato, R. Boi, R. Ragazzini, Determination of micro-indentation hardness of
(2011) 384–390. organic coatings, VDI Berichte (1685) (2002).
[12] A. Bjoergum, O.O. Knudsen, A.K. Kvernbraten, N.I. Nilsen, Corrosion protecting [27] J. Wen, V.J. Vasudevan, G.L. Wilkes, Abrasion resistant inorganic/organic coating
coating systems in arctic areas, Eurocorr 2011, European Corrosion Congress, materials prepared by the Sol-Gel method, J. Solgel Sci. Technol. 5 (1995) 115–126.
Stockholm, Sweden, 2011 4–8 September. [28] M.N. Mandikos, G.P. McGivney, E. Davis, P.J. Bush, J.M. Carter, A comparison of
[13] A.W. Momber, M. Irmer, N. Glück, P. Plagemann, Abrasion testing of organic cor- the wear resistance and hardness of indirect composite resins, J. Prosthet. Dent. 85
rosion protection coating systems with a rotating abrasive rubber wheel, Wear (4) (2011) 386–395.

10
A.W. Momber, et al. Progress in Organic Coatings 146 (2020) 105720

[29] A. Harrison, R.A. Draughn, Abrasive wear, tensile strength, and hardness of dental ISO, Geneva, Switzerland, 2016.
composite resins - is there a relationship? J. Prosthet. Dent. 36 (4) (1976) 395–398. [44] ISO 7784-2, Paints and Varnishes - Determination of Resistance to Abrasion - Part 2:
[30] S.K. Sinha, W.L. Chong, S.C. Lim, Scratching of polymers - modeling abrasive wear, Method With Abrasive Rubber Wheels and Rotating Test Specimen, ISO, Geneva,
Wear 262 (2007) 1038–1047. Switzerland, 2016.
[31] J.O. Bello, R.J. Wood, Micro-abrasion of filled and unfilled polyamide 11 coatings, [45] ISO 7619-1, Rubber, Vulcanized or Thermoplastic: Determination of Indentation
Wear 258 (2005) 294–302. Hardness - Part 1: Durometer Method (Shore Hardness), ISO, Geneva, Switzerland,
[32] K.H. Zum Gahr, Microstructure and Wear of Materials, Elsevier, Amsterdam, 1987. 2010.
[33] G. Beckmann, I. Kleis, Abtragsverschleiß von Metallen, VEB Deutscher Verlag für [46] A. Garcia-Diaz, D.T. Phillips, Principles of Experimental Design and Analysis,
Grundstoffindustrie, Leipzig, 1983. Chapman & Hall, London, 1994.
[34] W. Wuttke, Tribophysik, Carl Hanser Verlag, München-Wien, 1987. [47] A. Bjoergum, O.O. Knudsen, A.K. Kvernbraten, N. Nilsen, Protective coatings in
[35] I.M. Hutchings, P. Shipway, Tribology, Friction and Wear of Engineering Materials, arctic environments, Eurocorr 2012: European Coating Congr. Istanbul, Turkey,
2nd ed., Butterworth-Heinemann, Oxford, 2017. 2012 9–13 September.
[36] E. Rabinowicz, Friction and Wear of Materials, Wiley, New York, 1966. [48] A.W. Momber, M. Irmer, N. Glück, Performance characteristics of protective coat-
[37] A.H. Evans, Abrasive Wear in Ceramics: an Assessment, Report LBL-8608, Lawrence ings under low-temperature offshore conditions. Part 2: surface status, hoarfrost
Berkeley Laboratory, Univ. of California, Berkeley, CA, USA, 1979 January. accretion, and mechanical properties, Cold Regions Sci. Technol. 127 (2016)
[38] M.A. Moore, F.S. King, Abrasive wear of brittle solids, Wear 60 (1980) 123–140. 109–114.
[39] ISO 12944-12949, Paints and Varnishes: Performance Requirements for Protective [49] U. Ikramov, K.C. Machkamov, Machkamov, Berechnung und Bewertung des abra-
Paint Systems for Offshore and Related Structures, ISO, Geneva, Switzerland, 2018. siven Verschleißes, VEB Verlag Technik, Berlin, 1987.
[40] ISO 8504-2, Preparation of Steel Substrates Before Application of Paints and Related [50] A.W. Momber, A transition function for the solid particle erosion of rocks, Wear
Products - Surface Preparation Methods - Part 2: Abrasive Blast-cleaning, ISO, 328–329 (2015) 348–355.
Geneva, Switzerland, 2019. [51] A.W. Momber, The response of geo-materials to high-speed liquid drop impact, Int.
[41] ISO 8501-1, Preparation of Steel Substrates Before Application of Paints and Related J. Impact Engng. 89 (2016) 83–101.
Products - Visual Assessment of Surface Cleanliness - Part 1: Rust Grades and [52] B. Briscoe, E. Pelillo, S.K. Sinha, Scratch hardness and deformation maps for
Preparation Grades of Uncoated Steel Substrates and of Steel Substrates After polycarbonate and polyethylene, Polymer Engng. Sci. 38 (24) (1996) 2996–3005.
Overall Removal of Previous Coatings, ISO, Geneva, Switzerland, 2007. [53] I.M. Hutchings, Ductile-brittle transitions and wear maps for the erosion and
[42] ISO 2178, Non-magnetic Coatings on Magnetic Substrates - Measurement of Coating abrasion of brittle materials, J. Phys. D Appl. Phys. 25 (1992) A212–A118.
Thickness - Magnetic Method, ISO, Geneva, Switzerland, 2016. [54] K.H. Zum Gahr, Wear by hard particles, Tribol. Int. 31 (10) (1998) 587–596.
[43] ISO 14577-1, Metallic Materials - Instrumented Indentation Test for Hardness and [55] A.W. Momber, Brittle-ductile transition index for small-contact damages to geo-
Materials Parameters - Part 4: Test Method for Metallic and Non-metallic Coatings, materials, ASCE J. Mater. Civil Engng. 28 (10) (2016) 060160111–9.

11