You are on page 1of 9

"PERFORMANCE TESTING OF ANTI-CORROSIVE COATINGS”

David A. Claydon

International Protective Coatings


Akzo Nobel
England

Abstract: The Protective Coatings Industry still lacks a Over the last few years an increasing amount of
globally recognised laboratory test protocol for the performance based specifications have entered the
evaluation of high performance coating systems prior to market place or are currently being developed. These
field use. The advent of new and more stringent VOC have been developed by both recognised standard
legislation has resulted in new product technology institutes, e.g. ISO, NACE, and customer end-users, e.g.
entering the market place without extensive track record Shell, Norsok and Bridge Authorities. The principle aim
and as such, laboratory performance testing has become with these specifications is for coating systems to meet a
increasingly important. The key to successful testing is set of pre-determined laboratory performance based
in attempting to correlate with actual field exposure, i.e. requirements that vary depending on the service end use.
predicting any mode of failure. This paper will review
the various approaches to performance testing under At this present time there is no globally
development in Europe and North America with recognised performance based specification that has full
particular emphasis on in -house work relating to the 100% industry acceptance, this is particularly true for
combination of accelerated cyclic corrosion testing and the Oil and Gas market. There is a differing opinion in
thermal cycling. the methodology of the specification and the number of
tests that need to be carried out. However, the goal is
INTRODUCTION the same, in that it is to ensure high quality coating
systems are approved for use.
As the Protective Coatings Industry moves into
the 21st century, coating manufacturers can no longer rely Laboratory prequalification testing has been
on the extensive track records of their time served particularly pushed to the forefront via the Offshore
product ranges to convince customers of their suitability Industry due to the seriousness of their requirement for
for use. As each day goes by these products are asset protection.
increasingly phased out due to the various strict VOC and
raw material legislations around the globe, and there is The key to successful laboratory performance
nothing more certain than legislation becoming even testing is being able to obtain some kind of correlation
more strict! with field performance. It is important to note that there
is no magical test that will accurately predict field
Coating manufacturers have had to “re-invent” performance and give assurances on expected lifetime.
their product ranges using new and alternative However, by utilising a range of tests the strengths and
technologies in order to meet both legislation weaknesses of a given coating system can be determined
requirements and customer performance expectations, and prevent inadequate product entering the market
without the luxury of proven long term performance in place.
service. This has, of course, shifted the emphasis onto
laboratory testing of the new coating materials and, in It is the scope of this paper to review the
particular, accelerated anti-corrosive and crack resistance methodology behind various tests that can be used to
performance testing. predict the anti-corrosive and crack resistance behaviour
of new coating systems. This paper will review both
industry recognised methods and in-house methods that
can be used and the valuable information they can
provide.
CONDITIONS IN THE FIELD methods documented but all are variations on a theme
which normally includes a salt fog period, dry out or
No sooner has a coating system been applied it freeze period, and a UV/condensation period, all
is under attack from the various external stresses forced utilising standard industry available equipment.
upon it by the surrounding environment. In the case of
the offshore environment, this is a combination of Descriptions of the common methods employed
extreme corrosive conditions, along with possible are detailed below:-
temperature fluctuations, UV exposure and the movement
of moisture in and out of the coating film. If you 72 hours Salt Fog (ASTM
combine this with the coating’s own internal stress then B117, ISO 7253) with artificial
you have a severe environment in which the coating has sea water (ASTM D1141)
to perform. As time goes by, these stresses can take their
Norsok Cyclic Test 16 hours dry out (23ºC)
toll on the coating system and make it more prone to
(based on
failure, in the form of cracking or blistering failure, and 80 hours UV at 60ºC/
NACE TM0184)
subsequent corrosion onset. Some new coating systems condensation at 50ºC, 4 hours/
will resist this better than others and thus extend the 4 hours cycle (ASTM G53)
lifetime to first maintenance.
Total duration = 25 weeks
Field performance testing in relevant 168 hours Prohesion (ASTM
environments does, of course, reproduce these conditions G85, Annex A5)
and any subsequent mode of failure over a period of years
168 hours UV at 60ºC/
rather than months, and can be very site specific. In ASTM D5894
condensation at 50ºC, 4 hours/
terms of laboratory performance testing, it is these
4 hours cycle (ASTM G53)
conditions that need to be reproduced in an “accelerated”
manner to provide correlation with any mode of failure in Total duration = 24 weeks
weeks or months rather than years. 168 hours Prohesion (ASTM
G85, Annex A5) with artificial
W e shall now look at laboratory tests used to sea water (ASTM D1141)
evaluate the anti-corrosive properties of coating Draft NACE
materials. 168 hours UV at 60ºC/
(Based on Shell)
condensation at 50ºC, 4 hours/
LABORATORY PERFORMANCE TESTS 4 hours cycle (ASTM G53)
Total duration = 12 weeks
Cyclic corrosion tests (CCTs) 72 hours Salt Fog (ASTM
Accelerated cyclic corrosion tests are fast becoming the B117, ISO 7253)
industry norm for the evaluation of the anti-corrosive
behaviour of atmospheric coating systems. They have 24 hours freeze at –20ºC
been shown to more accurately predict performance in Draft ISO 23040 72 hours UV at 60ºC/
the field than the traditional salt fog test (ASTM B117, condensation at 50ºC, 4 hours/
ISO 7253). This has been well documented by various 4 hours cycle (ASTM G53)
independent corrosion institutes and industry associations
over the last 5-10 years, and although the author does not Total duration = 25 weeks
wish to labour the point here, reference will be made to
this throughout this paper where relevant. All of these tests are employed in
prequalification testing of coating systems.
The success of CCTs has revolved around the
cyclic conditions of temperature fluctuation, UV One can consider those tests which use the salt
exposure and degree of wet and dry out which is fog (5% NaCl) or artificial sea water as the electrolyte as
generated during the test, rather than the static set of being more akin to offshore conditions, whilst the
conditions as in the salt fog test. This all adds up to ASTM D5894 test uses a blend of Sodium Chloride and
impart a more realistic stress onto the coating system in Ammonium Sulphate which is more industrial or
an accelerated manner, something the salt fog test does onshore.
not do. There are quite a few slightly differing CCT
Although the tests are slightly different, one Differences in performance may well be
thing that is constant is that a scribed area is placed into observed for a given coating type when comparing
the coating test panel down to the substrate. This is between the ISO 20340 test, which has a thermal
required to represent a defect area and any subsequent temperature gradient of –20ºC to +60ºC, against the
corrosion onset, which is the principal mode of failure remaining tests which have a smaller temperature
measured in this type of testing. This mode of failure gradient of +23ºC to +60ºC and do not include the low
manifests itself during test in the form of “blistering” at temperature freeze. An example, is given in Photograph
the scribed area as the corrosion “creeps” at the 3. As you can see, there is a considerable difference in
coating/substrate interface. Some of the blisters formed performance and highlights the importance of the larger
may fail further by cracking (see Photograph 1). This thermal cycle in some areas.
blistering mode of failure is commonly seen on field
exposed test panels which have been scribed (see
Photograph 2) therefore the mode of failure is very
similar. The extent of blistering and possible cracking
and the associated measured corrosion creep will allow
comparisons of new systems against controls of known
performance.

Photograph 3
+23ºC to +60ºC -20ºC to +60ºC
thermal cycle thermal cycle

The measurement of corrosion creep associated


with CCTs can provide valuable information such as:-

(i) Ranking of similar generic coating types.


Graph 1 shows the measured corrosion creep values for
three different zinc epoxy/epoxy intermediate/topcoat
systems in the CCTs mentioned above. As you will see,
all the tests give a similar ranking of the systems and
Photograph 1 allows us to make a considered conclusion on
Failure at scribe in CCT performance. This is particularly useful when
”Blisters” comparing new coating systems against similar generic
systems of known performance. Compare this with
Graph 2, which shows the measured corrosion creep in
the salt fog test. As you will see here, all the systems
perform similarly and it is not possible to give a clear
ranking.

In-house testing is currently concerned with


obtaining correlation in ranking between the various
CCTs and field exposure of some 20 coating systems of
various generic types, and known performance, ranging
from zinc epoxy primed, pure epoxies, modified
epoxies, glass flake epoxies and solvent free epoxies.

Field exposure is taking place at a highly


corrosive coastal exposure site in the North East of
Photograph 2 England (4 metres from North Sea tide).
Failure at scribe on field exposure
”Blisters” So far, between 8 and 18 months exposure time
has elapsed.
Table 1 shows some approximate results so far performance has also been highlighted in the exposure
with regard to corrosion creep from the scribed area. site work. It is envisaged that between 3 and 5 years
exposure will be required in order to apply some
Some correlation is already being seen, High accurate correlation (if any!).
Solids Epoxy Systems 1 & 2 both show excessive
corrosion creep in the cyclic corrosion tests. This “poor”
Table 1

Corrosion Creep Values (mm)


D.F.T. Exposure
Coating Type Norsok CCT ISO 20340 ASTM D5894 Draft NACE
(µm) Site
4200 Hrs 4200 Hrs 4032 Hrs 2688 Hrs
10mm Creep
High Solids Epoxy 1 1 x 500 14 13 8 6
19 Months
8mm Creep
High Solids Epoxy 1 3 x 150 8 8 10 4
12 Months
6mm Creep
High Solids Epoxy 2 2 x 200 7 6 7 6
12 Months
~1mm Creep
Pure Epoxy 1 2 x 150 5 5 3 3
15 Months
~1mm Creep
Modified Epoxy 1 2 x 150 6 7 5 3.5
15 Months
~1mm Creep
HS Modified Epoxy 1 1 x 450 6 8 6 4.5
15 Months
HS Zinc Epoxy 1 1 x 75
~0.5mm Creep
Epoxy MIO 1 1 x 200 4 3 1.5 2
13 Months
Urethane 1 1 x 50
Zinc Epoxy 1 1 x 75
0mm Creep
Epoxy MIO 1 1 x 200 1 1 1 0.5
13 Months
Urethane 1 1 x 50

Graph 1:- Graph 2:-

Corrosion Creep Comparison of Zinc Corrosion Creep Comparison of Zinc


Primed Systems using CCTs
Primed Systems using Salt Fog Test
7 (4032 Hrs)
6
Corrosion Creep (mm)

5 7
4 6
Corrosion Creep (mm)

3 5
2 4
1 3
0 2
Norsok ASTM Draft ISO
Cyclic D5894 NACE 20340 1
Test Type 0
Salt Fog Test
(ii) Differences in surface preparation. (iii) Differences in dry film thickness. In the
Application of coating systems over shot or grit blasted real world, coating systems are rarely applied at the
steel can lead to differences in performance at the scribed recommended specified thickness. Application at
areas (see Photograph 4). The three coat zinc epoxy double or triple thickness is not uncommon, particularly
primed system here shows worse performance over shot at complex areas. Photograph 6 shows the difference in
blasted steel. This sort of information can well lead to performance of a three coat zinc epoxy primed system in
the recommendation that for optimum performance grit a CCT when applied at standard and double thickness.
blasting should be employed. This difference in Again, the salt fog test does not highlight this (see
performance is not seen in the salt fog test (see Photograph 7). It is interesting to note that in many
Photograph 5). prequalification approval protocols only the standard
specified thickness is tested.

New high solids coating systems may be more


prone to increased corrosion creep at the scribe area due
to increase internal stress.

Photograph 4 – CCT Result


Top: Grit blasted steel
Bottom: Shot blasted steel

Photograph 6 – CCT Result


Top: Standard thickness
Bottom: Double thickness

Photograph 5 – Salt Fog Results


Top: Grit blasted steel
Bottom: Shot blasted steel
Photograph 7 – Salt Fog Result
Top: Standard thickness
Bottom: Double thickness
Crack resistance testing Typical conditions required are:-
The issue of crack resistance of coatings is a very
important one. Cracking is a common mode of failure in 1. Thermal cycling (hot/cold as per field
many different service end use areas, and a problem the temperatures).
customer does not like. Once a coating system has
cracked, the ability to provide adequate corrosion Coatings respond to decreasing temperatures by
protection has been seriously compromised. volumetrically contracting which produces a tensile
stress. On increasing temperatures, coatings expand
The natural ageing process of a coating film via producing a compressive stress. The low
the various stresses it experiences will make the coating temperatures are particularly important as they
more prone to cracking. Some coatings will be more increase internal stress within the coating and
prone to cracking than others. therefore can be more destructive with regards to
crack initiation and propagation.
Reproducing cracking failure in the laboratory is
a key area of test method development within the 2. Movement of moisture in and out of film (wet and
Protective Coatings Industry. dry).

Some in -house work in which a cyclic corrosion Serious cracking failure can be caused by the
test was modified to simulate an increased thermal stress volumetric expansion of water within the coating
of –30 to +80ºC, along with salt fog, condensation and film upon freezing.
UV, did not produce the des ired results on a known
system of poor crack resistance, even when over-applied. Changes in the moisture content of coating films
No cracking from the scribe defect occurred. create stresses (sometimes called “hygroscopic
stresses”).
Areas of stress concentration such as welds,
weld spatter, complex angles etc., tend to be the areas in 3. Test panel with areas of high stress concentration
which cracking is commonly seen. Taking this fact in (welds, angles).
hand and replacing the scribe area with a weld area, and
including stripe coating, has produced some success in The first signs of cracking in service can usually be
Norsok cycling testing with regard to reproducing seen in complex areas such as welds, corners, sharp
cracking failures (see Photograph 8). Testing is ongoing edges, weld spatter etc., where stress levels are
against controls of known performance. At this stage, not highest.
all test parallels crack and further work is required to
ascertain the overall validity of this type of testing. 4. Over-application of coating system (internal stress,
practical application).

Over-application, particularly at complex areas is a


major cause of cracking problems due to increased
internal stress within the coating.

All the above conditions impart an overall


stress on the coating system which will cause
Photograph 8 accelerated ageing and produce early cracking failure.
Cracking at weld area in Norsok Cyclic test
The issue of crack resistance is becoming ever
In general, specific tests for crack resistance more important, and the development of test methods to
need to be developed. concentrate on this issue may well become part of future
performance specifications.
When developing specific tests for evaluating
the crack resistance of coatings it is important to include
conditions in which the coating experiences an
“accelerated stress” similar to that experienced in the
field and put together in a “cyclic test” method.
Already, some specific tests are finding their The coated panels at standard and double
way into performance specifications. Two such thickness are subjected to the following cycle:-
specifications are the NACE TG260 Standard for
offshore maintenance coatings and the Dutch Water Cycle Time Process
Board (RWS) test specification.
10 minutes 40ºC Black Panel
The NACE TG260 Standard includes a dry 3 minutes UV + Water Spray
thermal cycling test in which the coating system is
applied to a C-channel test piece and subjected to –30ºC 7 minutes Cooling 40º to –15ºC (no UV)
to +60ºC in a two hour cycle. The total test duration is 3 30 minutes -15ºC + UV
weeks, 252 cycles. On completion of the test, the coating
system is inspected for any signs of cracking with the 10 minutes Heating –15ºC to +75ºC + UV
naked eye and with a stereo microscope. This test utilises 57 minutes 75ºC Black Panel + UV
a programmable thermal cycling chamber. The C-
channel test piece (3” x 2”) provides stress concentration 3 minutes Cooling 75ºC to 40ºC + Water Spray
at the corners (see Photograph 9).
One Cycle = 2 hours
Test Duration is up to 1500 hours.

This test has shown that the use of UV can


induce cracking failure in high build epoxy coatings.
Epoxy coatings that had been overcoated with a durable
topcoat (e.g. polyurethane) tended not to show cracking
failure. However, failure was seen where expected and,
as such, correlation with field result was achieved.

The principal drawback of this method is the


expensive equipment (WOM) required and will not be
accessible to all. This opens up the avenue of test
method development using less expensive, more
Photograph 9 accessible equipment (e.g. salt fog, UV chamber,
freezer) and reproducing similar cracking failure.
The RWS method was developed around
reproducing failures seen in the field. The test involves In-house methods have concentrated on the
the use of a weather-o-meter programmable chamber (the combination of thermal cycling, deionised water
test is referred to as the WOM test). The coating system immersion, application at standard and up to x4
is applied to a 150mm x 100mm panel with three evenly recommended thickness, and the use of welded complex
spaces grooves machined into the panel (see Photograph angle test pieces. The idea behind the test was to get all
10). The grooves provide the area of stress concentration the elements of stress within one 24 hour cycle and test
in which crack initiation and propagation occurs. for up to 4 weeks with extended immersion at weekends
but, above all, would be able to produce cracking failure
where expected and none where not, in a relatively short
period of time.

A description of the method is as follows:-

• Immerse test piece in aerated deionised water at


23ºC over weekend.

• On Monday through Friday carry out the


Photograph 10 following:-
9.30 a.m. Remove test piece from immersion and The short term goal is to ascertain what level of
freeze samples at –30ºC for 2½ hours. cracking is acceptable in this test.

12.00 p.m. Heat up to +90ºC in 1 hour. It could be that the test is too aggressive.
Further testing of control systems of known performance
1.00 p.m. Dry out at +90ºC for 2½ hours. will help to answer this question.

3.30 p.m. Place back in immersion.

The thermal cycle of –30ºC to +90ºC was


chosen as representative temperatures in which cracking
occurred in the field.

The thermal cycling between –30ºC to +90ºC is


carried out in a programmable thermal cycling chamber
(see Photograph 11). These chambers are useful in the
fact that they can be programmed to carry out any
thermal cycle between –70ºC up to +120ºC (depending
on model).
Photograph 12
3 coat system at x 2 d.f.t.
Cracking expected

Photograph 13
1 coat system at x 4 d.f.t.
Photograph 11 Cracking expected
Thermal cycling chamber
Although there are panel movements every day
Testing is in its early days, but a degree of of the working week, this test does have the benefit of
success has been achieved in reproducing expected being able to produce cracking failure after only 1 week
cracking phenomena after one week of testing, in of testing in some cases.
addition to no cracking observed where expected,
however, more generic type of coating systems do need Early work has shown that over-application
to be tested. The test piece is approx. 75mm x 100mm in does have an important role in whether or not a coating
size. will crack and to what extent.
Photographs 12-16 show typical examples of the
results obtained.

The cracking tends to initiate and propagate at


the weld areas.
long term evaluation. It is hoped that those systems
which have shown themselves to crack quite early in the
thermal cyclic test will be the first to show signs of
cracking in the field in order to add some credence to the
test, however, this could take a few years.

Further investigation will also attempt to find


any correlation with the Internal Stress Test results of
the coating systems themselves.

SUMMARY

Cyclic corrosion tests provide information on


Photograph 14
the anti-corrosive ability of a coating system in terms of
3 coat system at x3 d.f.t.
blister, rusting, chalking and corrosion undercreep from
No cracking expected a defect area or scribe. The tests can also give valuable
information relating to the effect of over-application,
surface preparation and overall ranking of systems with
regard to preventing corrosion onset at a defect. They
are superior to the traditional salt fog tests in all aspects,
particularly that of correlation with known field
performance.

Simulation of cracking phenomena requires


development of specialised tests which principally use
thermal cycling and complex test pieces with areas of
stress concentration. A significant number of control
coating systems of know performance need to be
evaluated in order to obtain some kind of correlation as
Photograph 15 “Thermal Cycling” tests can easily be too aggressive and
2 coat system at x3 d.f.t. give erroneous results. Over-application of coatings is a
No cracking expected key issue in crack resistance testing.

Coating systems which have good anti-


corrosive properties may be prone to cracking and
subsequent delamination failure which may not be
picked up in a CCT test and vice versa. The
combination of CCT and a “Thermal Cycling” test,
similar to those described in this paper, will go a long
way to providing valuable information for new coating
systems and their performance expectations/limitations.
This is particularly important in aggressive corrosive
environments.

Over the coming years, it is predicted that customers


will increasingly demand anti-corrosive data, and data
Photograph 16 relating to crack resistance for new coating systems as
2 coat system at x 2 d.f.t. an effective replacement for field track record. This
Cracking expected could well become a common feature in all performance
based specifications, particularly for those end users
In an attempt to obtain some correlation with who have experienced problems in the field.
field exposure, large girder and pipe sections with weld
areas have been exposed at our coastal exposure site for

You might also like