Progress in Organic Coatings 44 (2002) 217–225

Evaluation of corrosion protection properties of additives

for waterborne epoxy coatings on steel
F. Galliano∗ , D. Landolt
Laboratoire de Métallurgie Chimique, Département des Matériaux, Ecole Polytechnique Fédérale de Lausanne,
CH-1015 Lausanne EPFL, Switzerland
Received 23 September 2001; accepted 8 February 2002

Abstract
The development of environment compatible additives for corrosion inhibition in waterborne coatings requires test methods which yield
significant results on a short time scale. The present study aims at the evaluation of the effect of corrosion inhibiting model additives on the
performance of a waterborne epoxy coating using electrochemical and non-electrochemical methods which measure different properties.
Electrochemical impedance spectroscopy (EIS), linear sweep voltammetry, mechanical pull-off tests and scanning acoustic microscopy
(SAM) in combination with image analysis are used. Two kinds of corrosion inhibiting additives are employed: an organic inhibitor based
on a carboxylic acid neutralized by a polysiloxane base, and ZPA, an inorganic pigment with inhibiting properties. The results obtained
show that corrosion inhibiting additives drastically modify the adhesion, water uptake, blistering behavior and substrate protection of
waterborne epoxy coatings. Both additives improved the dry adhesion and reduced blistering under cathodic polarization conditions. The
experimental approach described in this paper should be useful for additive development and for coating formulation because it yields
a more complete picture than can be obtained by single methods of how a given additive affects the coating performance in a corrosive
environment. © 2002 Elsevier Science B.V. All rights reserved.
Keywords: Waterborne coatings; Corrosion inhibition; Additives

1. Introduction the wetting properties of the binder emulsion with respect

to the substrate and the pigments.
The introduction of strict regulations in the use of volatile The degradation of a polymer-coated metal occurs after
organic compounds (VOCs) has brought about the devel- water penetrates at the coating–substrate interface, where
opment of solvent-free coating technologies like powder failure mechanisms such as osmotic blistering, cathodic de-
coatings, electrocoatings, UV curable coatings and water- lamination and anodic undermining may be initiated [3,4].
borne systems. Waterborne coatings have gained increasing Corrosion inhibiting additives are designed to react at the
importance for many applications. These coatings usually metal surface by improving adhesion and slowing down the
contain different additives such as extenders, dispersion electrochemical reactions. In modern coatings traditional
agents, defoamers, fillers and corrosion inhibitors. The inorganic inhibitors such as chromates, lead oxides, etc.
choice of a functional additive requires a good knowledge are being replaced by less harmful inorganic [5,6] and or-
of its interaction with other additives and pigments, its com- ganic compounds [7,8]. Inorganic pigments like phosphates,
patibility with the binder and the substrate and its influence molybdates, vanadates, silicates and borates are commonly
on the adhesion properties [1]. Additives can interact with used for the protection of solvent-borne and waterborne
waterborne paint system in different ways with correspond- coatings. Environmentally friendly organic inhibitor formu-
ing consequences for the corrosion protective properties. lations specifically designed for waterborne coatings are an
For example, a previous study performed in our laboratory attractive alternative [9].
showed that the selection of a suitable dispersion agent was The development of corrosion inhibiting additives re-
crucial for the performance of coated steel samples sub- quires test methods which yield significant results in a
jected to corrosion tests [2], the reason being that it affects time as short as possible. Salt spray and different cyclic
exposure tests are widely used in the coating industry, but
these tests yield essentially qualitative information. More
∗ Corresponding author. recently, electrochemical methods, especially electrochem-
E-mail address: federico.galliano@epfl.ch (F. Galliano). ical impedance spectroscopy (EIS), have found increasing

0300-9440/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 3 0 0 - 9 4 4 0 ( 0 2 ) 0 0 0 1 6 - 4
218 F. Galliano, D. Landolt / Progress in Organic Coatings 44 (2002) 217–225

Table 1 additives and pigments as reported in Table 1. The effect

Complete paint compositions of two different inhibitors was evaluated. Inhibitor (A) was
Ingredients a proprietary organic inhibitor (IRGACOR 287, Ciba Spe-
Epoxy binder (Beckopox-Hardener EH 623W)
cialty Chemicals) containing carboxylic acid for corrosion
Dem. water inhibition [18] neutralized with a polysiloxane base for im-
Talc proving solubility and adhesion properties [19]. Inhibitor
Iron oxide (B) was a commercial zinc–aluminum phosphate pigment
Calcium carbonate (Heucophos ZPA, Heucotech LTD) known for its corrosion
Bentone SD2
Borchigel L75
inhibiting properties. The weight percentages were 0.6% for
Additol XL 270 inhibitor (A) and 8.3% for inhibitor (B), respectively. In-
With/without inhibitor hibitor (B) replaced an equivalent mass of calcium carbon-
ate in the base paint, in order to maintain the same weight
proportion of inorganic solids in the coating.
interest for testing of coatings, because these methods are Low carbon steel panels (190 mm × 105 mm, Chemetall
capable to yield mechanistic data which can be quantified Gmbh) were used as substrates. They were cleaned in an
by using suitable models [10,11]. On the other hand, EIS ultrasonic bath by a 1:1 mixture of ethanol and acetone and
data acquired in long term exposure tests, often scatter con- painted using a steel bar coater drawn at a constant speed.
siderably and require, especially in the case of damaged Curing was carried out at 80 ◦ C for 30 min and the panels
coats, parallel experiments on several samples and a statisti- were allowed to completely dry for 24 h at 40 ◦ C. The final
cal evaluation [12,13]. The adhesion properties of coatings average thickness was 45 ± 3 ␮m for the different coatings
are crucial for corrosion protection and durability of a coat- as measured by optical methods. Panels were then cut to
ing system. Corrosion cells can deteriorate the adhesion of obtain 24 mm × 24 mm specimens which were stored in a
organic coatings on steel because the rise in pH at cathodic desiccator.
locations on the surface can lead to cathodic disbonding
and blistering. Several authors have studied the mechanism
2.2. Experimental techniques used
of cathodic delamination and disbonding using different
methods [14–17]. In our laboratory scanning acoustic mi-
For the characterization of the coating performance, elec-
croscopy (SAM) has recently been used for monitoring the
trochemical and non-electrochemical techniques were used.
blister formation due to the loss of adhesion during open
Immersion tests were performed using EIS for the evaluation
circuit immersion tests [2]. For the development of corro-
of the performance of the coated system. The LSV technique
sion inhibiting additives it is important to be able to measure
was employed to study the inhibitive properties of the addi-
by what mechanism a given additive affects the coating
tives on the bare steel surface. The effect of the additives on
performance and on which property of the metal-coating
the adhesion of the coating to the substrate was determined
system has the most effect. The goal of the present study
by pull-off tests performed under dry conditions. Finally, the
is to evaluate the effect of two model corrosion inhibiting
effect of the additives on cathodic delamination and blister-
additives on the performance of a waterborne epoxy coating
ing was studied using in situ SAM of cathodically polarized
using a variety of electrochemical and non-electrochemical
specimens and image analysis. A summary of the different
methods which measure different properties.
properties evaluated by the measurement techniques used is
given in Table 2.
2. Experimental
2.3. EIS experiments
2.1. Materials and samples preparation
All EIS experiments were conducted using a Gamry PC
The fully formulated paint system utilized in this work potentiostat connected to an 8-channel ECM8 multiplexer.
is composed of a waterborne amine-modified epoxy binder, The setup was driven by CMS 300 software which permitted

Table 2
Scheme of the tested properties
Property Measurement Measured or calculated values Coated samples Bare steel
technique
Long term immersion resistance EIS Water uptake, coating resistance, polarization X
resistance, double layer capacitance
Water diffusion barrier EIS Diffusion coefficient X
Additive inhibition LSV Cathodic and anodic currents X
Adhesion Pull-off test Bond strength X
Cathodic delamination resistance SAM Blister growth rate, maximum blistering propagation X
 F. Galliano, D. Landolt / Progress in Organic Coatings 44 (2002) 217–225 219

to automate the experimental procedure. Sequential software. Linear potential sweeps were started at −1.6 V
impedance spectra were acquired at well defined time in- until 1.2 V (relative to a Hg/HgSO4 reference electrode), at
tervals under constant experimental conditions. A specimen a scan rate of 2 mV s−1 . Rotating disc electrodes made from
area of 11 mm diameter was exposed for 28 days to a 5 wt.% a CK 45 steel rod of 5 mm diameter embedded in araldite
NaCl solution at room temperature. Impedance measure- and polished with 400, 600, 1000 grit emery paper were
ments were performed every 20 min for the first 12 h, then used in all the experiments as working electrodes at a rota-
every 2 h for 24 h and then every 12 h until the end of the tion rate of 400 rpm. A platinum coil was used as counter
experiment. When a 20 min interval was used, the analyzed electrode. An aerated 0.12 M NaClO4 solution was chosen
frequency ranged from 10 to 105 Hz, for longer intervals as supporting electrolyte and maintained at a temperature of
from 10−1 to 105 Hz. In the case of premature coating fail- 25 ± 1 ◦ C. Because the inhibitors used have a low solubil-
ure the experiment was stopped. From 3 to 6 samples of ity in aqueous solution all experiments were performed at
each type of coating were tested and evaluated. The classic saturation conditions. For this 2 g l−1 of the inhibitor were
equivalent circuit model for painted metals [10] was re- contacted with the electrolyte for 24 h and then filtered, a
gressed individually to each impedance spectra to calculate procedure suggested by Amirudin et al. [5]. Before each
the following model parameters: Rs electrolyte resistance, experiment, the pH of the test solutions was adjusted to
Rp paint resistance, Cp paint capacitance, Rct charge trans- 7.5 by addition of NaOH or HClO4 . Two samples of each
fer resistance and Cdl double layer capacity (here a constant series were tested to check for reproducibility.
phase element was used). Data collected in the first 12 h,
with an analyzed frequency range 10–105 Hz, were treated 2.5. Pull-off adhesion experiments
with a simplified fitting circuit that included only Rs , Rp and
Cp (Randles circuit). A satisfactory fit could be obtained Samples in dry conditions were sandwiched in an align-
for all data sets. In the following figures, mean values of ment jig between 15 mm diameter aluminum cylinders
the calculated parameters are reported together with the utilizing an epoxy adhesive (Araldite Rapid, Ciba Specialty
respective standard deviations. Chemicals). A 12 h curing was allowed at a pressure of
The water uptake φ was calculated from the measured 30 kPa and the resulting specimens were then subjected
coating capacitance Cp using the Brasher–Kingsbury equa- to tensile testing in a tensile machine (Lloyd 2000R) at
tion a cross-head speed of 2 mm min−1 . Reported adhesion
log(Cp /Cp0 ) strength values are averaged over five measurements.
φ= (1)
log 80
2.6. Cathodic delamination experiments
where Cp0 is the coating capacitance extrapolated for t →
0. The evolution of the fit parameters φ, Rp , Rct , and Cdl , Blistering and cathodic delamination was studied on
with immersion time was used to monitor the coating degra- scribed samples using SAM. The cathodic delamination and
dation. Using the theory of Fickian diffusion for immersed blistering process was accelerated by cathodic polarization
supported coatings, the water uptake φ can be analytically of the specimens [17] using an Amel 5000 potentiostat.
expressed as a function of time φ(t) [20]. By utilizing Eq. (1) The edges and the back of the samples were insulated
the coating capacitance is analogously described as a func- with araldite and the exposed coating surface was scribed
tion of time Cp (t) for a length of 5 mm with a diamond tool. The samples
n=∞
Cp (t) 8
1 were then placed horizontally in a recipient containing a
exp[−(2n+1) π Dt/4l ]
2 2 2
=1− 2 (2) 5 wt.% NaCl solution. A graphite counter electrode and a
Cp sat π (2n + 1) 2
n=0 Ag/AgCl reference electrode (SSE) were inserted in the
where Cp sat represents the coating capacitance value when tank and the samples were cathodically polarized at −1.1 V
water saturation in the coating occurred and l the coating SSE.
thickness. The above equation can be reduced, for short The region adjacent to the scribe was monitored in situ by
immersion time, to a linear relation between the coating SAM, which detects changes in the acoustic impedance at
capacitance and the square root of time (see Eq. (3)) and the the interface between coating and substrate [2,17,22]. The
water diffusion coefficient DW can be easily obtained by a SAM measurements were performed using a Honda Scan-
linear fitting procedure [21] ning Ultrasonic Flaw and Imaging HA-711 in conjunction
  with a focused piezopolymeric (PVDF) transducer with a
Cp (t) ∼ 8 DW t 0.5 center frequency of 80 MHz. At this frequency, the vertical
= 2 (3)
Cp sat π l2 resolution of the acoustic microscope was about 15 ␮m.
Further details on the SAM image acquisition procedure are
2.4. LSV experiments reported elsewhere [2]. SAM images were digitally scanned
and processed by a commercial image analysis software.
The LSV experiments were conducted using a Solartron The total blistered area and the maximum blister propaga-
1287 electrochemical interface controlled by a commercial tion distance from the scribe were determined as a function
220 F. Galliano, D. Landolt / Progress in Organic Coatings 44 (2002) 217–225

Fig. 1. Water uptake (a), paint resistance (b), charge transfer resistance (c) and double layer capacitance (d) as a function of immersion time in 5 wt.%
NaCl solution.
 F. Galliano, D. Landolt / Progress in Organic Coatings 44 (2002) 217–225 221

of time. Two samples of each series were tested to assure

reproducibility.

3. Results

3.1. EIS results

Plots of water uptake, paint resistance, double layer ca-

pacitance and charge transfer resistance for the blank coat,
and the coats containing (A) and (B) inhibitor, respectively,
are reported in Fig. 1. Standard deviation plots are also in-
cluded. In the first 30 h the water uptake values (see Fig. 1a)
increase steeply for all the tested coats, denoting a fast dif-
fusion of water into the coatings. The initial period was fol-
lowed by a slower almost linear increase for the remaining
Fig. 2. Coating capacitance normalized as a function of the square root
test duration. No clear saturation plateau was observed, a
of exposure time. Calculation of the slope allows the determination of
behavior that may be related to the swelling of the poly- the water diffusion coefficient, DW .
mer matrix [21]. The data show the highest water uptake
for the blank coating followed by the coat formulated with
inhibitor (A) and then by the one with (B). At longer im-
mersion time, (B) containing samples exhibited a discon- The water diffusion coefficient DW was calculated from
tinuous trend of the water uptake, corresponding to the the Cp values as function of time using Eq. (3). The sat-
loss of adhesion and onset of blistering. Visual observa- uration capacitance was estimated from Fig. 2 as the ca-
tion of (B) samples at the end of the test revealed some pacitance measured at the transition of water uptake from
blistering while samples (A) were not blistered and still fast to slow linear increase [23]. Since the blank coat and
well adherent. As a general observation one may note that the coat with inhibitor (A) present the same capacitance
the amount of water taken up by these waterborne epoxy behavior at short immersion times, the calculated water
coatings (6–9 vol.%) was always higher than that measured diffusion coefficient is the same within experimental er-
for solvent based epoxy coatings of analogous thickness ror. On the other hand, (B) containing coats showed a
(2–6 vol.%) [23]. lower diffusion coefficient (see Table 3) indicating that the
The coating resistance (see Fig. 1b) presents a very sharp ZPA pigment slowed water diffusion. The overall picture
decrease in the first hours followed by a plateau value that emerging from EIS measurements clearly shows a bene-
is always lower than 107  for all the coats. This indicates ficial effect of both additives on the corrosion properties
limited protective properties of the coatings studied. The of the waterborne epoxy coating system. The use of ZPA
coating resistance is the highest for (B) containing coats fol- inhibitor provided the best corrosion resistance at short
lowed by the (A) containing coats and then by the blank immersion times but at longer exposures a sensitive loss
coats. At longer exposure times, a degradation of the barrier of protective properties was observed. A similar behavior
properties of (B) samples is observed and Rp approaches for zinc phosphate pigments has been described by other
the values measured for (A) samples. The same trend found authors [24].
for Rp is observed for the charge transfer resistance Rct (see
Fig. 1c). At short immersion times (<200 h) the inhibiting 3.2. LSV results
action of (B) is more effective than that conferred by the
(A) and both the inhibitors show higher Rct values than the The experiments were performed in air saturated solutions
blank coat. At longer immersion times, Rct for samples (B) where the reduction of oxygen is the predominant cathodic
drops to values observed for the blank coat, because blister- reaction. The polarization curves measured in the 0.12 M
ing occurs and the metal–paint interface accessible to water NaClO4 solution with and without inhibitor are shown in
sensibly increases. The (A) samples on the other hand ex- Fig. 3. The (A) inhibitor (Fig. 3a) does not appreciably in-
hibit an almost constant value for the entire test duration. An
important scattering in the Rct values characterizes the sharp
transition to blister formation observed on (B) samples. The Table 3
results for the double layer capacitance (see Fig. 1d) are Diffusion coefficients for water as evaluated from coating capacitance data
qualitatively consistent with those of Rct in that the onset of Sample D W × 10−10 (cm2 s−1 )
blistering on the (B) samples is again clearly detected. The
Blank, with (A) 2.7 ± 0.5
data scattering for (B) samples is very large, typical for a With (B) 1.1 ± 0.2
non-uniform coating degradation.
222 F. Galliano, D. Landolt / Progress in Organic Coatings 44 (2002) 217–225

Fig. 3. Linear sweep voltammograms for low carbon steel rotating disc electrode in 0.12 M NaClO4 supporting electrolyte with or without (A) additive
(a) and (B) pigment (b).

fluence the oxygen reduction reaction and the limiting ca- 3.3. Pull-off adhesion results
thodic current is the same as without inhibitor. In absence
of inhibitors the anodic branch of the curve follows a typi- The data reported in Table 4 indicate that both additives
cal active dissolution behavior. The addition of inhibitor (A) provided an important increase of the coat–substrate bond
leads to lowering of the anodic dissolution current and to an strength compared to the blank coat. The pull-off force in
anodic shift of the corrosion potential of about 100 mV. The presence of additive (B) was the highest, but it must be borne
surface of the samples tested in the (A) containing solution in mind that the failure was located within the coating and
presented, after the test, large patches of unattacked metal. not at the interface. The same was observed in presence of
It is concluded that (A) in an anodic inhibitor in agreement additive (A). Thus both additives conferred to the coating
with previous studies performed with similar compounds
[18]. The polarization curves in the inhibitor (B) contain-
Table 4
ing solutions (Fig. 3b) are characterized by a slightly lower Pull-off adhesion test in dry conditions
limiting current for oxygen reduction, while the anodic part
Sample Bond strength (MPa) Detached area
of the curve is not significantly modified with respect to the
pure supporting electrolyte solution. This confirms that (B) Blank 7.7 ± 3.7 60 ± 20
is a cathodic inhibitor reducing the rate of oxygen reduction With (A) 19.7 ± 4.0 0
With (B) 27.1 ± 3.9 0
[25].
 F. Galliano, D. Landolt / Progress in Organic Coatings 44 (2002) 217–225 223

an adhesive strength exceeding its cohesive strength. On the formed in the vicinity of the scribe on a coat containing
other hand, for the blank coat the rupture occurred primar- inhibitor (A) after 53 h of cathodic polarization.
ily at the coating–metal interface as a result of a deadhesive The occurrence of the blistering was always preceded by
failure. Polysiloxane additives are well known as adhesion a delay time where presumably an initial disbonding was
promoters in the organic coating technology [19] and the re- going on, as previously described by Crossen et al. [22].
action mechanism has been extensively studied in the past Once the first blisters appeared, it was possible to follow
years [26]. On the other hand, to the authors’ knowledge, the their growth and the formation of new blisters. To quantify
marked increase of coating adhesion caused by the presence the SAM observations image analysis was utilized. In Fig. 5
of ZPA has not been investigated in detail so far and fur- the total blister area and the maximum blistering propaga-
ther studies are needed to elucidate the interfacial reactions tion away from the scribe (an average of the two opposite
involved. In summary, the adhesion tests prove that both ad- directions) are plotted as a function of time for the three dif-
ditives (A) and (B) significantly improved the adhesion of ferent coats tested. A linear relationship between blistered
the coating to the substrate. area and time is observed. From the slope the growth rate of
the blister area was calculated and results are also shown in
3.4. SAM results the figure. Assuming that, during the cathodic delamination
process, water and ions must diffuse at the coating–metal in-
Observation by SAM of the metal–polymer interface terface from the scribe to the blister, one would expect from
permitted to follow the formation of blisters with time simple diffusion theory that the maximum distance from the
in the vicinity of a previously made scribe. Cathodic po- scribe to blisters increases linearly with the square root of
larization of the scribed samples accelerated the cathodic time. Such a relationship has been found by Leng et al. [16]
disbonding and blistering process. The cathodic reaction who explained the behavior in terms of a mechanism in-
provides hydroxyl ions and short-lived radical intermediates volving cation transport at the metal–polymer interface and
which weaken, by hydrolization and oxidation, the adhesive limiting the delamination kinetic. The data of Fig. 5b for the
strength at the metal–polymer interface [16]. This permits blank coat and for the coat with inhibitor (A) yield the ex-
advancement of the delamination front and penetration of pected relationship, whereas the behavior of the coating with
solution. As a result blisters are formed where the interface inhibitor (B) differs in that the maximum distance from the
adhesion is weakest. Unfortunately, the limited vertical res- scribe at which blisters form is much smaller and increases
olution of our microscope (∼ =15 ␮m) did not allow us to very slowly during the test duration. This coating also ex-
detect the penetration of liquid at the interface before blis- hibits a much longer initiation time before blistering occurs.
ters were formed. As an illustration, Fig. 4 shows blisters The inhibitor additives significantly affected the blister
formation kinetics. Blank samples, after 14 h, exhibit large
blisters which spread rapidly far from the scribe. After 87 h
the surface of the samples was completely covered by blis-
ters and the test was stopped. On (A) containing samples
blisters were detected only after 22 h and the size of blis-
ters remained small although the blister front advanced quite
rapidly. The blister growth rate of the blank was about 3.7
times larger than that of the (A) containing sample (0.75
and 0.20 mm2 h−1 , respectively), but the maximum blister-
ing distance from the scribe was only 1.7 times larger (1.37
and 0.81 mm h−0.5 , respectively). For the (B) containing
samples the first blisters were detected after 76 h and devel-
oped at a rate of 0.08 mm2 h−1 . The maximum distance for
blistering stabilized after about 164 h. The different results
show that inhibitor (B) effectively slowed down blistering
under cathodic polarization conditions. Inhibitor (A) also
yielded a reduced blister area compared to the blank coat
but it was less effective in slowing the spreading of blister
formation. Several factors may contribute to the beneficial
effect of (B) on blister formation and growth. The adhesion
test showed that (B) generally improves the adhesion of the
coating and the LSV experiments indicated that (B) inhibits
the cathodic partial reaction. Finally, the solubility of the (B)
pigment apparently increases with increasing pH [27] per-
Fig. 4. SAM image of the coat formulated with (A) after 53 h of cathodic mitting a higher concentration at the cathodic delamination
polarization. At the center of the image there is the scribe mark. front, where the hydroxyl ions are produced.
224 F. Galliano, D. Landolt / Progress in Organic Coatings 44 (2002) 217–225

Fig. 5. Blistered area (a) and maximum blistering propagation from the scribe mark (b) as a function of time during cathodic delamination experiments
at −1.1 V SSE in a 5 wt.% NaCl solution.

4. Conclusions Two kinds of corrosion inhibiting additives were em-

ployed: an organic inhibitor based on a carboxylic acid neu-
The present study demonstrates the usefulness of apply- tralized by a polysiloxane base (A), and ZPA (B) which is an
ing a variety of experimental methods for characterizing the inorganic pigment. EIS showed that (A) was beneficial over
effects of corrosion inhibiting additives on the performance an immersion time of 1 month in a saline solution, whereas
of waterborne coating systems. The obtained data show that the inhibitor (B) exhibited the best performance at short im-
corrosion inhibiting additives drastically modify properties mersion times. Polarization curves of low carbon steel sub-
such as adhesion, water uptake or blistering, and they also strates showed that (A) inhibits the anodic partial reaction
provide evidence that different additives affect these prop- while (B) is a cathodic inhibitor inhibiting oxygen reduction.
erties differently. The experimental approach described here Both inhibitors greatly improved the adhesion of the
allows one to identify specific effects of coating additives coating under dry conditions. Using SAM in conjunction
which should be useful for their optimization and for coat- with image analysis blister formation in saline solution
ing formulation. under cathodic polarization conditions could be quantified.
 F. Galliano, D. Landolt / Progress in Organic Coatings 44 (2002) 217–225 225

Results showed that addition of additive (A) led to a re- [5] A. Amirudin, C. Barreau, R. Helouni, D. Thierry, Prog. Org. Coat.
duction in blister size and growth rate, but the maximum 25 (1995) 339.
distance from the scribe where blisters formed was not sig- [6] R.C. MacQueen, R.R. Miron, R.D. Granata, J. Coat. Technol. 68
(1996) 75.
nificantly affected. Additive (B) dramatically reduced the [7] G. Reinhard, P. Simon, U. Rammelt, Prog. Org. Coat. 20 (1992) 383.
progress of the blistered front and the blister growth rate. The [8] P. Agarwal, D. Landolt, Corr. Sci. 40 (1998) 673.
enhanced adhesion and the cathodic inhibition are thought [9] A. Braig, Prog. Org. Coat. 34 (1998) 13.
to be responsible for the effectiveness of additive (B) for re- [10] F. Mansfeld, J. Appl. Electrochem. 25 (1995) 187.
ducing blistering. The different data show that both additives [11] P.L. Bonora, F. Deflorian, L. Fedrizzi, Electrochem. Acta 41 (1996)
1073.
(A) and (B) tested here improved coating performance, but [12] W.S. Tait, J. Coat. Technol. 66 (1994) 59.
their relative effectiveness depends on the system properties [13] J.A. Grandle, S.R. Taylor, Corrosion 53 (1997) 347.
studied. [14] H. Leidheiser Jr., J.W. Wang, L. Igetoft, Prog. Org. Coat. 11 (1983)
19.
[15] D.H. van der Weijde, E.P.M. van Westing, J.H.W. De Wit, Corr. Sci.
Acknowledgements 36 (1994) 643.
[16] A. Leng, H. Streckel, K. Hofmann, M. Stratmann, Corr. Sci. 41
(1999) 579.
The authors thank A. Braig, D. Renoux and M. Frey from [17] M. Kendig, R. Addison, S. Jeanjaquet, J. Electrochem. Soc. 137
Ciba Specialty Chemicals Inc., Basel Switzerland, for pro- (1990) 2690.
viding the paint systems and the additives. Dr. A. Kulik [18] G.T. Hefter, N.A. North, S.H. Tan, Corrosion 53 (1997) 657.
kindly permitted the use of the SAM at the Institut de Génie [19] P. Walker, J. Adhes. Sci. Technol. 5 (1991) 279.
[20] J. Crank, G.S. Park, Diffusion in Polymers, Academic Press, London,
Atomique, EPFL Lausanne. Financial support from CTI 1968.
Bern is gratefully acknowledged. [21] E.P.M. van Westing, G.M. Ferrari, J.H.W. de Wit, Corr. Sci. 36
(1994) 957.
[22] J.D. Crossen, J.M. Sykes, T. Zhai, G.A.D. Briggs, Faraday Discuss.
References 107 (1997) 417.
[23] F. Geenen, Ph.D. Thesis, Technische Universiteit Delft, The
[1] K. Dören, W. Freitag, D. Stoye, Waterborne Coatings, Hanser/ Netherlands, 1991.
Gardner, Cincinnati, 1994. [24] R. Romagnoli, V.F. Vetere, Corrosion 51 (1995) 116.
[2] V. Guillaumin, D. Landolt, Corr. Sci. 44 (2001) 179. [25] D. Labouche, M. Traisnel, Surf. Coat. Int. 10 (1994) 404.
[3] H. Leidheiser Jr., Corrosion 38 (1982) 374. [26] C. Reinartz, W. Fürbeth, M. Stratmann, Fresenius J. Anal. Chem.
[4] T. Nguyen, J.B. Hubbard, J.M. Pommersheim, J. Coat. Technol. 68 353 (1995) 657.
(1996) 45. [27] T. Schauer, W.M. Liu, L. Dulog, Eur. Coat. J. 3 (1997) 292.