Progress in Organic Coatings 46 (2003) 49–54

Effect of environmental temperature variations on

protective properties of organic coatings
A. Miszczyk∗ , K. Darowicki
Department of Anticorrosion Technology, Faculty of Chemistry, Technical University of Gdansk,
11/12 Narutowicza St., 80-952 Gdansk, Poland
Received 13 September 2001; received in revised form 10 October 2002; accepted 17 October 2002

Abstract
Effect of environmental temperature variations on mechanical integrity and protective properties of coated metal systems has been
discussed based on stress development and the cumulative damage model. The effect of temperature cycling between 20 and 55 ◦ C, and
−20 and 20 ◦ C on the durability of different paint coating systems on steel was investigated. Using the impedance spectroscopy technique
changes of parameters of coatings submitted to thermal treatment have been monitored. Results of impedance measurements show a
negative influence of temperature variations on coating protective properties.
© 2002 Elsevier Science B.V. All rights reserved.
Keywords: Environmental; Impedance; Coating

1. Introduction rial due to its properties has the ability to respond to stress by
molecular rearrangement. Temperature- and time-dependent
Organic coatings have been used for a long time to pro- behaviour of polymeric material systems subjected to
tect metals against corrosion in atmospheric conditions. The thermal changes is well known [6,7]. Below glass transition
fundamental property of natural environment conditions is temperature (Tg ) spontaneous changes occur in polymer
variability, mainly connected with sun operation and mete- amorphous materials. These processes are ascribed to con-
orological phenomena. Coating systems on metal undergo formational changes of the polymer backbone and are related
cyclic thermal loading connected with temperature fluctu- to physical ageing. As a result the density and elastic modu-
ations in the 24 h cycle and seasonal cycle (Fig. 1) [1]. lus of the polymer increase, the impact strength, fracture en-
This aspect does not reflect conditions of accelerated weath- ergy decrease; creep and stress relaxation rates decrease [8].
ering tests or formulation of new coating products. Usu- A number of phenomenological theories have been proposed
ally, testing of coating system is carried out at constant to describe the time-dependent changes of physical proper-
temperature or at two levels of temperature in cyclic tests ties [6,8,9]. Loss of coating ability to relax stress is the effect
[2,3]. of physical ageing and chemical degradation. As a result,
Environmental temperature changes affect coated metal the mechanical damage to the coating occurs (microcracks)
in two ways: by inducing thermal stresses in the system and as well as delamination from the metallic base [10–12].
by changing the mechanical properties of the viscoelastic Subsequently, during cyclic temperature changes propa-
coating. Thermal stresses result from mismatch of the co- gation of defects and delamination on the metal/coating
efficients of thermal expansion of the coating and substrate interface occur. In this way protective properties of the
(e.g., for an epoxy resin from 45 × 10−6 to 65 × 10−6 ◦ C−1 , coating connected with the barrier function and adhesion
for alkyds 55×10−6 to 60×10−6 ◦ C−1 [4] and for the metal to the substrate deteriorate [13]. Therefore, the lifetime
base: steel 16×10−6 ◦ C−1 , zinc 23×10−6 ◦ C−1 , aluminium of the coating system to a significant degree is connected
23×10−6 ◦ C−1 [5]). Temperature determines both the stress with capability to resist the forming and relaxation of stress
level and the rate of viscoelastic response of the coating to without mechanical damage to the coating and the coating/
stress level. In the initial period of exposure the coating mate- metal interface.
A need exists to predict rapidly and simply the service
∗ Corresponding author. Tel.: +48-58-347-2483; fax: +48-58-347-1092. life of organic coatings on metal [3,14,15]. Successful pre-
E-mail address: [email protected] (A. Miszczyk). diction of long-term performance of organic coatings in

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 1 8 8 - 1
50 A. Miszczyk, K. Darowicki / Progress in Organic Coatings 46 (2003) 49–54

2. Experimental

Three different types of commercially available paints

were studied:
• solvent-borne polyvinyl coating system, four layers,
380(±40) ␮m thick (supplied by Malchem Polish Paint
Factory),
• solvent-borne epoxy coating system, four layers,
280(±25) ␮m thick (supplied by Sigma Coatings Poland),
• water-borne acrylic system, two layers, 60(±10) ␮m thick
(supplied by Nobiles Polish Paint Factory, Akzo Nobel
Group).
All systems were applied to mild steel panels sandblasted
(solvent-borne coatings) or polished with abrasive paper
up to 360 grade (water-borne coating) according to pro-
ducer guidelines. The coated panels were stored at room
temperature for about 2 months before experiments. Then
PVC cylinders were attached on the coating surface using
a two-component epoxy adhesive. During the experiment
Fig. 1. Temperature changes during the year 2000 in Warsaw, Poland [1]. the cylinder was filled with a 3% NaCl aqueous solution
in the case of solvent-born paints and 0.3 g NaCl + 0.7 g
Na2 SO4 /dm3 in the case of water-borne paint.
different weather conditions requires identification of the The exposed area was 81 cm2 in the case of solvent-borne
critical environmental variables that control degradation rate paints and 38 cm2 in the case of water-borne paint. The elec-
and determination of the value distribution in the given place. trochemical measurements were made using a two-electrode
For polymeric materials, temperature and humidity play an system. A platinum mesh with low impedance was used
important role in dynamics of coating degradation [16,17]. as the counter electrode. The impedance measurements
There are some parts of the word where the temperature were performed using a Schlumberger 1255 Frequency Re-
cycles around 0 ◦ C occur frequently. In this case additional sponse Analyser and ‘home made’ high impedance buffer.
phenomena may take place that is freezing/thawing of wa- Measurement and acquisition of data were controlled by
ter within coating. Because ice has a greater volume than a computer. Impedance spectra were determined in the
the same water weight of water, ice crystals create addi- full frequency range from 1 MHz to 1 mHz. Ten points
tional stress, which may cause damage. It has been postu- per decade were measured above 0.1 Hz and five points
lated that three kinds of water, bulk-like water, freezable per decade below 0.1 Hz. In order to reduce measurement
pore (bound) water and non-freezable pore water, can be time narrower frequency range, 100–1 kHz, was used also.
characterised in organic coatings [18]. Coating degradation A perturbation signal was applied of an amplitude in the
connected with the freezing/thawing cycles are attributed to 10–60 mV range, depending on the impedance of the in-
the freezable water. Non-freezable water interacts strongly vestigated system. The measured spectra were analysed
with the polymer lowering Tg . Therefore, the influence of using the equivalent circuit presented in Fig. 2. Usually,
freezing/thawing cycles is important and should be consid- the high-frequency part of the spectrum was analysed,
ered as a detrimental factor. describing the barrier properties of the coating.
An understanding of the development of stress over time A part of the samples of a acrylic coating were subjected to
at an elastic-viscoelastic interface subjected to a cyclic ther- exposure in condition of an urban-sea side (Gdansk, Poland)
mal loading could contribute to more effective coating for-
mulation, testing and failure analysis. This paper compares
different thermal treatments applied to conventional coating
formulations to investigate the role of temperature changes
in degradation. The state of the coating during and after
thermal treatment was evaluated with the use of impedance
spectroscopy. This technique has found wide application in
investigations of coatings, allowing evaluation of the effect
of aggressive interaction of the environment on the coating Fig. 2. Electrical equivalent circuit used for analysis of impedance data.
Re , electrolyte resistance, Rp , electrolyte resistance in coating pores, CPE
system [19–22]. The advantage of impedance spectroscopy (Y0 , n) constant phase element modelling capacitive properties of the
utilised in this work is the possibility of quantitative detec- coating, ZI impedance of the electrolyte/metal interface at the bottom of
tion of small interaction effects and propagation of defects. pores.
 A. Miszczyk, K. Darowicki / Progress in Organic Coatings 46 (2003) 49–54 51

Table 1
Mean maximum and mean minimum temperatures for each month of
exposure
Month

April May June July August September

tmin (◦ C) 8 10 13 13 14 9
tmax (◦ C) 18 21 23 21 23 16

Fig. 5. Course of temperature changes of samples with an epoxy coating.

Fig. 3. Course of temperature changes of samples with an acrylic coating. Two cases were considered for coated sample exposure:
(a) temperature changes above 0 ◦ C,
climate for a period of 6 months, from the beginning of (b) temperature changes when minimum temperature is be-
April to the end of September, 2000. On the basis of re- low 0 ◦ C and maximum temperature is above 0 ◦ C.
gional meteorological data the mean lowest and mean high-
est temperatures were determined for each month (Table 1). In the first case temperature programme was controlled
Samples taken for laboratory tests were subjected to ther- by computerised ‘home made’ set up using typical propor-
mal treatment in accordance with Fig. 3, performing during tional, integral and derivative (PID) arrangement. The sam-
one day a programme of temperature changes characteristic ple was heated and cooled according to programmed course
for 1 month; hence 30 or 31 shocks of parameters in accor- between 20 and 55 ◦ C. The thick epoxy coating system on
dance with Table 1. During remaining time, samples were steel was immersed in 3% NaCl and exposed to temperature
stored at 20 ◦ C. Simultaneously, two parts of coating samples variations according to Fig. 5. The sample was kept at 20 ◦ C
were submitted to thermal interaction, one immersed in so- in 3% NaCl during 2 weeks before thermal treatment to ob-
lution simulating atmospheric conditions: 0.3 g NaCl +0.7 g tain stable state of investigated system. Impedance measure-
Na2 SO4 /dm3 , and second one without presence of solution ments were carried out in the frequency range 100–1 kHz
(as presented schematically in Fig. 4). Impedance spectrum every 2 min during dynamic temperature changes. Using
measurements were performed before and after natural and a software programme [23], coating electrical parameters
laboratory exposures. Samples exposed with no immersion were determined on the basis of obtained impedance spectra.
were immersed in 0.3 g NaCl + 0.7 g Na2 SO4 /dm3 for 48 h The freezing–thawing cycle was performed between +20 ◦ C
before impedance measurements. (4 h) and −20 ◦ C (20 h).

3. Results and discussion

The first group of experiments was done in order to isolate

the contribution of temperature changes on coating degra-
dation in the presence of aqueous solution. In Fig. 6 rela-
tions have been presented in the Arrhenius coordinates, i.e.,
in the function of 1/T, where T is the temperature in the
Kelvin scale. It results from Fig. 6 that changes of coat-
ing are in reality irreversible. Every cycle causes very small
shift of coating parameters characteristic for coating degra-
dation. A spiral mode of parameter changes is observed.
Probably such behaviour is due to the fact of different lev-
Fig. 4. Experimental cell for simultaneous realisation of thermal treatment
of coated sample in immersion in 0.7 g Na2 SO4 + 0.3 g NaCl/dm3 and els of water uptake in the coating at different temperatures
with no immersion. [24]. Absorbed water at higher temperature cannot be fully
52 A. Miszczyk, K. Darowicki / Progress in Organic Coatings 46 (2003) 49–54

desorbed at the lower temperature during dynamic temper-

ature decrease. The results can be microcracks, microvoids,
and local delamination in order to accommodate excess wa-
ter. Each next temperature cycle repeats this phenomenon
but starting from the higher level of water content (indicated
by CPE parameters Y0 and n) and microcrack damage (in-
dicated by coating resistance Rp ).
The aim of the next part of these studies was to isolate
the contribution of temperature changes to coating degra-
dation in the presence and absence of water solution and
compare results with exposure in natural conditions. Sam-
ples of a water-borne acrylic coating selected on the basis
of approximately identical initial electric parameters were
subjected to natural exposure in conditions of an urban-sea
side (Gdansk, Poland) climate for a period of 6 months,
from the beginning of April to the end of September 2000.
Samples taken for laboratory tests were subjected to thermal
treatment performing a programme of temperature changes
characteristic for the whole natural exposure period (see
details in experimental part) with and without presence of
solution (as presented schematically in Fig. 4). In this way,
two extreme cases of exposure in natural conditions were
modelled, there are constant presence of water and expo-
sure with no precipitation. Impedance measurements were
performed before and after natural and laboratory exposure
for comparison. In Fig. 7 spectra are presented before ex-
posure (a) and after exposure (b). Generally, a decrease in

Fig. 6. Changes of coating resistance Rp (a) and CPE parameters Y0 , (b)

and n (c) for epoxy coating on steel in the function of 1/T, where T is
Fig. 7. Impedance spectra of acrylic coatings before exposure (a) and af-
the temperature in the Kelvin scale.
ter exposure (b). (䊊) sample before (a) and after 6 months atmospheric
exposure (b); (䊏) sample before (a) and after thermal treatment in im-
mersion in 0.7 g Na2 SO4 + 0.3 g NaCl/dm3 (b); (䊉) sample before (a)
and after thermal treatment with no immersion (b).
 A. Miszczyk, K. Darowicki / Progress in Organic Coatings 46 (2003) 49–54 53

Fig. 8. Impedance spectra of acrylic coatings exposed at constant tem-

perature, 20 ◦ C, before (䉱) and after exposure ().

protective properties is observed due to natural or labora-

tory exposure. As can be seen in Fig. 7 the spectrum of the
coating exposed in natural conditions takes a similar shape
and magnitude to those of the samples subjected to thermal
treatment in the laboratory. Additionally, it is situated be-
tween spectrum obtained for the immersed and the sample
without immersion. This result suggests that possible be-
haviour of the coating system during outdoor exposure can
Fig. 9. Changes of coating resistance Rp for polyvinyl coating systems
be assessing by delimitation of the range. Limits are results
on steel in the function of exposure time or number of freezing/thawing
of tests in immersion and without immersion. Depending cycles.
on unforeseeable weather condition, obtained results will be
situated within this range. This approach takes into account losses barrier properties. In the case of the sample exposed
a certain lack of reproducibility and repeatability of field in 3% NaCl without freezing/thawing cycles the coating
exposure results due to weather conditions. resistance does not change significantly with exposure time,
In order to verify the effect of temperature variations apart from the first period (A).
the impedance measurements were performed on immersed It is easy to see from the results given above that the effect
samples at constant room temperature during the same pe- of temperature changes on protective properties of coating
riod of performing thermal treatment. In this case relatively systems is significant. The magnitude of this effect depends
small changes were observed as shown in Fig. 8. On the ba- on the presence of water. The Arrhenius approach is com-
sis of obtained data one may conclude that the degradation monly used for description of temperature relations. In sim-
mechanism under natural conditions and under laboratory plest form, the magnitude of the property is proportional to
selected conditions with the use of variable temperatures is exp(−E/RT), where E is the Arrhenius activation energy for
similar. the property under study, R the gas constant and T the ab-
In the third part, the effect of the freezing/thawing cycles solute temperature. When the cyclic change of temperature
on protective properties of polyvinyl coating system on steel with time is greater than the response of the material (relax-
was investigated. Fig. 9 shows the typical dependence of the ation effect), the property would change also cyclically, but
coating resistance on the number of cycles (full points) or with a trend. In the case of tested coating samples, this effect
the time of immersion, as one half of samples was immersed leads to a water absorption increasing in the coating dur-
in 3% NaCl for comparison at constant temperature of 20 ◦ C ing immersion, and increasing in hygrothermal stress within
during the whole testing (open points). The samples which coating or thermal stress during exposure without presence
was intended for freezing/thawing cycles was immersed for of water. Local cohesive or adhesive damages of the coat-
4 days before testing to obtain saturated water uptake state. ing appear when the cohesive or adhesive forces are smaller
During this time coating parameters were monitoring by then stress caused in the coating. In this way it is possible to
impedance measurements (period A in Fig. 9). A resistance explain an induction period for loss of protective properties,
decrease at the immersion beginning can be attributed to such that relatively small changes occur. Such behaviour was
dynamic water uptake and coating swelling. After reach- observed for the polyvinyl coating system.
ing quasi-stable state, freezing/thawing cycle programme
started. Next periods can be distinguished clearly. The re-
sistance of both coatings does not change significantly with 4. Conclusions
time for the next period (B). But after about 20 cycles
(each full point is equivalent to one freezing–thawing cy- The protective properties of organic coating systems on
cle) a continuous decrease of coating resistance is clearly steel were investigated in conditions of temperature changes
observed (period C). This trend indicates that the coating characteristic for outdoor exposure using impedance
54 A. Miszczyk, K. Darowicki / Progress in Organic Coatings 46 (2003) 49–54

spectroscopy. Basing on these results, one can propose [3] B.P. Alblas, J.J. Kettenis, PCE 5 (2) (2000) 49.
the following mechanism of accelerated degradation, in [4] J. Antoniewicz, Properties of Dielectrics, WNT, Warsaw, 1971 (in
comparison to isothermal exposure. Temperature changes Polish).
[5] Mechanic Hand-book, vol. 1, WNT, Warsaw, 1984 (in Polish).
induce the development of hygrothermal, thermal or me- [6] I.M. Hodge, J. Non-Cryst. Solids 169 (1994) 211.
chanical stresses in the coating in wet, dry or frosty con- [7] K.D. Dorkenoo, P.H. Pfromm, J. Polym. Sci. 37 (1999) 2239.
ditions, respectively. In spite of relaxation processes and [8] W.J. Davis, R.A. Pethrick, Polymer 39 (1998) 255.
depending on cycling speed, the average magnitude of [9] J.M.G. Cowie, R. Ferguson, S. Harris, I.J. McEwen, Polymer 39
stresses increases. As adhesive bond strength diminishes, (1998) 4393.
[10] L. Xiaowei, H.A. Hristov, A.F. Yee, D.W. Gidley, Polymer 36 (1995)
due ageing of coating material, approaches critical value of 759.
the local cohesive stress and adhesive damages occur. This [11] M.E. Nichols, C.A. Darr, C.A. Smith, M.D. Thouless, E.R. Fischer,
contributes to coating degradation by creation of conductive Polym. Degrad. Stab. 60 (1998) 291.
pathways in the coating as cracks fill with electrolyte and/or [12] H. Lee, S. Krishnaswamy, Trans. ASME 67 (2000) 444.
local delamination. Each further temperature changes can [13] T. Nguyen, J.B. Hubbard, J.M. Pommersheim, J. Coat. Technol. 68
(1996) 45.
cause propagation of defects and their accumulation over [14] M. Kendig, Jeanjaquet, R. Brown, F. Thomas, J. Coat. Technol. 68
time. Macroscopic failure occurs when a specific limit of (1996) 39.
accumulated damages is reached. Generally, the effect of [15] G.P. Bierwagen, L. He, J. Li, L. Ellingson, D.E. Tallman, Prog. Org.
environmental temperature variations on coating protective Coat. 39 (2000) 67.
properties is negative. With these observations in mind, [16] D.Y. Perera, D.V. Eynde, J. Coat. Technol. 59 (1987) 55.
[17] E.V. Schmidt, Exterior Durability of Organic Coatings, Redhill,
accelerated methods for coating testing should include the Surrey UK, FMJ, 1988.
effect of temperature variations characteristic for the region [18] C. Decker, K. Zahouily, Polym. Degrad. Stab. 64 (1999) 293.
of coating application. [19] J.H.W. De Wit, in: Proceedings of the 12th International Corrosion
Congress, vol. 1, September 19–24, 1993, Houston, TX, p. 420.
[20] F. Mansfeld, J. Appl. Electrochem. 25 (1995) 187.
References [21] P.L. Bonora, F. Deflorian, L. Fedrizzi, Electrochim. Acta 41 (1996)
1073.
[22] G. Grundmeier, W. Schmidt, M. Stratmann, Electrochim. Acta 45
[1] Meteorological data from Internet Weather Station at Warsaw
(2000) 2515.
University of Technology, Faculty of Electronics and Information
[23] B.A. Boukamp, AC-immittance Data Analysis System—Equivalent
Technology. ftp://nms.ise.pw.edu.pl/meteo/meteo/.
Circuit, Twente University of Technology, Enschede, The
[2] G. Haynes, Cabinet, in: R. Baboian (Ed.), Corrosion Tests and
Netherlands, 1989.
Standards: Application and Interpretation, ASTM Manual Series:
[24] Ch.M. Hansen, Prog. Org. Coat. 26 (1995) 113.
MNL 20, Philadelphia, 1995, p. 91.