DESALINATION

ELSEVIER Desalination 150 (2002) 247-254

www.elsevier.com/locate/desal

Corrosion and mechanical behavior of fusion bonded epoxy

1.
(FBE) in aqueous meala

Anees U. Ma&*, Ismail Andijani, Shahreer Ahmed, Fahd Al-Muaili

Research and Development Center, SWCC, PO Box 8328, Al-Jubail-31951, Kingdom of Saudi Arabia
Tel. +966 (3) 3613577; Fax: +966 (3) 3621615; email: [email protected]

4 July 200 1; accepted 15 April 2002

Abstract
In recent years there have been many reported cases of corrosion failure in cement concrete pipelines. In the majority
of cases, the failures have been attributed to rebar corrosion which is caused by the permeability of chloride from low
resistivity soil and subsequent attack on a passive layer on an iron bar in the structure. As a possible alternative to
cementitious materials, some organic coatings based on olefm, viny1 or epoxy-based polymers have been considered.
However, due to a paucity of data on the behavior of these coatings in aqueous media-particularly product water -
the possibility oftheir application in water transmission systems in the Kingdom has not been Mollyexploited. This paper
deals with the studies carried out on the corrosion and mechanical behavior of fusion bonded epoxy (FBE) coating on
steel in aqueous mediawhich include product water, distilled water and saline water. The mechanical testings on coating
include adhesion, bending and cathodic disbondment testings. The corrosion studies include immersion testing under
static and dynamic conditions, autoclave tests and accelerated (salt-fog) tests. The analysis of results indicates chemical
inertness of FBE coating in either of the aforementioned water used during testing, good adhesion and no damage to the
coating during bending. Cathodic disbondment tests indicate that FBE coating sustains under cathodic protection (CP)
conditions. In general, the results of mechanical and corrosion tests indicate that FBE is a promising material for internal
coating on steel in water transmission systems.

Keywora!s: Fusion bonded epoxy; Mechanical testings; Immersion testings; Cathodic protection; Cathodic disbondment;
Steel; Product, distilled, and saline waters

1. Introduction
of cases the failures have been attributed mainly
In recent years there have been several cases to rebar corrosion which is caused by permea-
of pipe failures in the form of leakages, bursting bility of chloride from low resistivity soil. The
or cracking of the pipelines [l-4]. In a majority problem is acute in areas where the soil, besides
having high chloride contents, has intermittent
*Corresponding author. dry and wet spells. As a possible alternative to

001 l-9164/02/$- See front matter (9 2002 Elsevier Science B.V. All rights reserved
PII: SO0 1 l-9 I64(02)00980-3
248 A. U. Malik et al. /Desalination 150 (2002) 247-254

cementitious materials, in recent years, a number the coating shortly thereafter. Niel and
of olefin, vinyl and epoxy-based polymer Whitehurst [ 131used chloride contamination that
coatings have appeared on the market. Epoxy- remained in the micropits after sand blasting of
based paints or coatings have been employed as steel surface for studying FBE coating perfor-
internal linings as well as external coatings in a mance. They found that in the presence of a
con-siderable number of pipelines in the pitted surface, chloride contamination can cause
Kingdom. In recent years, polyethylene coatings serious loss of performance in FBE coatings in
having an FBE primer and a two-layer hot cathodic disbonding and hot water tests. For
polyethylene-based material extruded over it underground coatings and other immersion
have been used extensively throughout the world coatings in critical applications, a maximum
as external coatings for open or buried pipe lines chloride level of 2 ppm was suggested [ 141.
[5,6]. In such coatings, the epoxy provides good A review of the important work carried out on
adhesion to steel and good cathodic disbonding the performance of FBE coating as cited above
characteristics, which are combined with a water reflects that the coating appears to have a charac-
barrier and good mechanical properties of teristic property required for application in a
polyethylene. The combination has better water transmission system. However, perfor-
adhesion, cathodic disbonding resistance, mance data about FBE coating are sketchy, and
hydrolytic stability and impact strength than therefore, systematic study of its corrosion and
either coating used alone. For internal coatings, mechanical behavior is important and the work
fusion bonded epoxy (FBE) is a promising described in this paper was carried out by
material and has been used for transmitting water keeping this point of view.
though on limited scale [7].
For application in a water transmission
system, using organic coatings in general, and 2. Experimental
FBE in particular, the effect of chloride contami- 2.1. Materials
nation on coating performance is an important
objective of the study. Carbon steel AISI 1018 (0.18% C) panels
Soluble salt contamination can cause pre- were used for FBE coating. Fusion bonded epoxy
mature failure on virtually all types of coatings. FBE-Scotchkote 206N (FBE-X) specimens were
Amongst the anions, chloride is the single most acquired from the Al-Qahtani pipe terminal
damaging anion because it migrates under (AQPT), Dammam. The coated samples were of
coating film. A chloride containing solution has the following dimensions and quantity.
a high osmotic pressure contributing to moisture l Thickness: -20-21 Mils (500-525 microns)
penetration, loss of adhesion and blistering. The l Coupon size (mm) and quantity:
source of chloride contamination could be from 100x50 50
the environment around metallic surfaces. Alblas 200x200 50
and London [S] reviewed the literature concem- 200x25 25
ing the effect of chloride contamination on the
corrosion of coated steel surfaces. Appleman [9],
2.2. Equipment
Helvig [lo], Weldon [ 1 l] and Flores [ 121showed
various correlations between level of chloride The following instruments were used for
and premature coating failures. These inves- carrying out the experimental work: salt spray
tigators applied the contaminant (chloride) in cabinet; holiday detector with calibration meter-
known quantities to the steel surface and applied model AP-W, Tinker and Rasor; coating thick-
 A. U. Mu&ket al. /Desalination 150 (2002) 247-254 249

tress meter, Elcometer Model 345 and Posi 3.2. Flexibility test
Tector-2000; adhesion meter from DYNA
A Mandrel bend machine was used to test the
Proceq, Zurich, Swi~erland~ cathodic disbond-
steel specimens coated with FBE at room
ment testing assembly.
temperature. The main objective of the bending
test is to determine the strength of the coating
under bending condition. Prior to bending, the
3. Mechanical testing
specimens were inspected visually for any visible
3. f. Adhesion test defect followed by holiday test. The FBE-coated
samples were tested by a pulse-type detector set
Adhesion tests on FBE-coated steel samples
at 2500*50 V. The thickness of the specimens
were carried out at 25°C using the ASTM D4541
was also measured before the test. The specimens
- 85 (r-e-approved 1989) technique.
were then clamped in the holder and bent flat-
Pull-off adhesion test and/or crosscut methods
wise at 60°C over a thick shoe; bending was
were employed to determine the adhesive
accomplished in approximately 30 s, The speci-
strength of the coatings. The adhesion tester
mens were again inspected by a holiday detector
consists of dollies made of aluminum, which
to confirm any cracking in the coating after
were glued perpendicular onto the coated surface
bending the specimens.
of the samples. After the curing of adhesive
The Mandrel bend machine was pro~~med
(glue), the testing apparatus was attached to the
to carry out the flexibility test by increasing the
loading fixture and aligned to apply the tension
mandrel radius step by step until the coating
normal to the test surface. The force applied to
stops failing. The smallest available mandrel shoe
the loading fixture is then gradually increased
was an 87-mm radius. The present strain was
and monitored until either plug of coating
calculated as given below:
material is detached or a specified value is
reached. The relative stress applied to each
coating can be calculated as follows: %&rain=%

X=2_ where t is the effective thickness of the speci-

Ed2 mens (DFT + metal), r is the radius of the
where X is the pull-off strength achieved at mandrel shoe, and DFT is the dry film thickness.
failure (psi), I; is the highest force applied to the The percent strain is directly proportional to the
test surface (pounds), and d is the equivalent effective thickness t of the specimen.
diameter of the original surface area stressed
(inches).
3.3. ~ut~odic dis~ond~e~t test
In the crosscut adhesion test, a crosscut was
made with the help of a utility knife on the coated The tests were carried out with FBE
surface deep to the metal substrate. At the specimens for 7 days (4OOC)and 4 weeks (25°C)
crosscut the blade of the knife was inserted under using the CAN / CSA-2245.20 M92 method [ 151.
the coating and with a levering action force was This test provides accelerated adhesion assess-
applied to chip off the coating. The chipped off ment and determines resistance of the coating to
area was observed under microscope (magni- cathodic potential and current flow. Coated steel
fication x40) to see the extent of removal of the samples of 200x200 mm were used for the tests.
coating from the substrate. In the middle of the coated specimen a hole of
250 A. U Malik et al. /Desalination 150 (2002) 247-254

3.2 mm in diameter was drilled through the loss of adhesion at the scribed and unscribed
coating to expose the substrate. A 200-mm-long surfaces. In one set of samples, scratch lines
plastic pipe of loo-mm diameter was glued onto (scribes) were made through one corner of the
the specimen with the holiday at the center of the samples to the diagonally opposite corner of the
tubing. A cathodic disbondment test cell was sample, i.e., “X” shaped. One side ofthe coupons
assembled with a DC power supply, platinum was scribed while the other side was left
wire as anode, high resistance volt/amp meter unscribed. The specimens without the scribe
and a calomel reference electrode. The DC power mark were weighed before starting the salt spray
supply was designed and manufactured at the test.
research and development center by the instru- In the salt spray chamber the specimens were
mentation section. The advantage of using this placed meeting the following conditions:
power supply was that it keeps the applied 1. All the specimens were supported parallel
potential constant irrespective of the current to the principal direction of horizontal flow of
flowing through the cell. Test specimens glued fog.
with the plastic pipe were kept on a hot plate and 2. A specimen holder was made of plastic and,
a 900 ml solution of 3% NaCl was poured in each therefore, specimens were not in contact with
plastic pipe. each other or with any metallic material.
The temperature of the hot plate was raised to 3. A 5% solution of sodium chloride was
maintain the temperature of the NaCl solution at atomized by compressed air in the chamber.
40°C. A potential of - 1SV vs. saturated calomel 4. The temperature of the chamber was kept at
(SCE) was applied. 38OC.
The negative lead of the power supply was Specimens were exposed under the above-
connected to the coated plate and positive lead to mentioned conditions for 25, 50, 75 and
a platinum anode. After 7 days the electrolyte 100 days, respectively. After the required expo-
was drained out and the test cell was immediately sure period, the samples were examined as per
dismantled. The coated plate was cooled to room ASTM D1645-71a (re-approved 1984). This
temperature. The blade of a hard utility knife was method provides a means of evaluating and
inserted under the coating near the holiday edge, comparing basic corrosion performance of the
and using a levering action, coating was chipped substrate, pretreatment, or coating system, or a
off. This action was continued till it became combination thereof, after exposure to a corro-
impossible to flake off the coating. The radius of sive environment. The specimens were carefully
the disbonded area from the holiday edge was removed from the holder and gently washed in
measured along seven different directions and an clean running water to remove salt deposits from
average was taken. their surfaces, and then immediately dried. The
exposed surface at the scribes was cleaned with
a brush to remove all the rust. Mean creepage
4. Corrosion test from the scribe and failed area was measured and
rated as per ASTM Dl654-71a. Similarly,
4.1. Salt spray test measurements were also carried out for the
Salt spray tests were carried out in a salt spray blisters appearing on the scribed and unscribed
fog chamber following ASTM Bll7-90. Speci- sides.
mens with and without scribes exposed to the salt
fog were evaluated with respect to a change in 4.2. Closed-circuit loop test
mass, blistering associated with corrosion, and FBE specimens were fixed in coupon holders
and installed in an indigenously designed and Table 1
fabricated closed-circuit loop. The experiments Pull-off adhesion test results for coating type, FBE-X
were carried out under the following conditions:
(1) temperature: 4O”C, (2) medium: distilled Adhesive strength, psi
water, (3) duration: 4 weeks, (4) flow rate:
No. 1 150
16 m3/h.
No. 2 210
No. 3 345

The test was carried out in an autoclave at

glue used to fix the dollies to coatings was
1500 psi, 40°C in distilled water for 48 h. The
around 500 psi. The pull-off adhesion test results
specimens were half immersed in the test solution
obtained from DYNA adhesion test with 50 mm
during the test. The thickness of coating was
cpdolly are given in Table 1.
measured before and after each test using a
electromagnetic thickness gauge Posi-Tector
2000 at six different places (three in aqueous and 5.2. Flexibility test
three in the vapor phase) on the specimen. AAer
completion of the test, the samples were assessed After bending, the samples were examined
visually for swelling and blistering, etc. The pult- visually followed by the holiday test at the bend
off adhesion test was also carried out on each site. No defect was found either visually or by the
phase, i.e., vapor and aqueous phases. The tests holiday tester on any of the samples. The tests
were carried out under the following conditions. showed that with a thickness (DFT + metal) of
Pressure: 1500 psi 4.5 mm, FBE can sustain up to a 2.4 1% strain.
Temperature: 40°C
Atmosphere: Nitrogen gas 5.3. Cathodic disbandment tests (CDT)
Test medium: Distilled water
Duration: 48 h The average radial disbondment (RD) value
for FBE-X as evaluated from CDT was found to
be 2.0 mm at 4O*C. The magnitude of the RD
5. Resultsand discussion value indicates that the disbondment area around
the holiday under an applied voltage condition is
5.1. Adhesion test
small. The CDT studies carried out by the author
In all the tests the dolly was detached at the [ 161 on polyurethane 600A and ISA, and three-
coatin~dolly interface. This confirms that the layer polythene (PE) coatings under similar
bonding between the metal substrate and coating conditions as those in FBE-X show RD values of
was more than the coating and dolly. The 17.0 mm, 14.9 mm and 3.2 mm, respectively. The
adhesion test results were not consistent, i.e., a coating having the lowest RD has the best
large difference among the data was observed. resistance toward cathodic disbondment. The
The maximum adhesion strength between coating results are summarized in Table 2.
and dolly was 345 psi. An adhesion test was also At 25*C, FBE and other coatings showed
carried out on samples used for the autoclave test almost negligible average RD. No deposits on or
at the AQPT facility. Here again the strength of underneath the coated surface were found at this
glue and coating was not enough to pull-off the temperature. This indicates that at ambient temp-
coatings. The maximum adhesion strength of eratures, no disbondment occurs in either of the
252 A. U.Ma&ket al. / ~esu~i~afionI50 (2002)247-254

Table 2
Results of cathodic disbondment tests

s. no. Coating type Average DFT, mils Average RD, mm

1 FBE 25.0 2.0

2 Aqualine 600A 50.3 17.0
3 Irethane 155 37.3 14.9
4 Three-layer PE 4.5 3.2

Table 3
Results of salt spray test

Coating Exposure (days)/creepage of coating, mm Rating Visual examination remarks

25 50 75 100

FBE-X 0.09 1.27 1.57 2.06 6 No blistering

coatings at the vicinity of holidays under an water in the close circuit loop at 40°C for
applied voltage. 1 month under flowing conditions (flow rate:
16 m3fh). All the samples were intact and no
remarkabie change in the physical condition of
5.4. Sali spray tests
the coating was observed.
Specimens with and without scribe exposed to
the salt fog were evaluated with respect to mean
5.6. Autoclave test
creepage (from scribe) and blistering. The salt
spray results for the FBE coating are summarized The autoclave tests were carried out in order
in Table 3. to know the behavior of coatings under high
The FBE coating (green) shows little creepage pressure and temperature. The test duration was
(0.09 mm) after 25 days of exposure, but it is 48 h and the temperature was fixed at 40°C. The
increased considerably (2.06 mm) after 100 days pressure of the test vessel was kept at 1500 psi.
exposure although no blistering in the coating After the test samples were examined for color,
was found. The scribed samples show a number blistering, loss of adhesion strength and thick-
of blisters where as unscribed samples are devoid ness. The data which present the thickness of the
of any blisters. It is interesting to note that a FBE-X coating in the vapor and aqueous phases,
decrease in number of pits was found from 50 to respectively, are summarized in Table 4.
100 days of exposure in the salt spray chamber. A slight increase in thickness can be seen in
The maximum number of blisters were found on both phases. FBE coatings do not show any loss
the scribed side. of color in the autoclave test. While carrying out
the pull-off adhesion test on the panels of FBE
after the autoclave test, there was no blistering or
5.5. loosed-circuit loop fest
swelling, but the glue failed in both the gas and
Coupons of FBE were exposed to distilled aqueous phases (at 500 psi} for all three samples.
 A. U Malik et al. /Desalination 150 (2002) 247-254 253

Table 4
Thickness of FEB-X coating in vapor and aqueous phases

Phase Before test, mils After test, mils Change in thickness, mils Average change in thickness, mils

Vapor:
No. 1 20.26 20.93 +0.67 +0.53
No. 2 21.26 21.66 +0.39

Aqueous:
No. 3 20.56 21.93 +1.36 +0.93
No. 4 21.40 20.90 +0.50

6. Conclusions 7. FBE has good mechanical properties, low

water permeation, no chemical degradation and
1. The results of adhesion tests carried out on
good corrosion resistance. Moreover, FBE shows
FBE show that the bonding between the metal
a small increase in radial disbondment under
substrate and the coating was more than the
applied potential (- 1.5 V SCE), thus indicating
coating and dolly.
its stability towards cathodic disbondment. This
2. The flexibility test (bending test) carried
combination of properties makes FBE a suitable
out on FBE shows no defect or presence of
choice for internal and external lining material
holidays at the bending site. The coating can
for steel pipes in water transmission systems.
sustain up to 2.4 1% strain.
3. The results from salt fog tests show the
following behavior of the FBE coating: References
l In scribed samples, the creepage increases
PI A.U. Malik, Case history on the failure ofpipelines in
with increasing exposure time: from 0.09 mm desalination plants, IDA World Conference on
(25 days) to 2.06 mm (100 days). Desalination and Water Reuse, Washington, DC,
l No blistering was observed after 100 days of 1991.
exposure. PI A.U. Malik, M. Kutty and IN. Andijani, Reinforced
4. The pull-off adhesion tests carried out on cement concrete pipelines for desalination water
coated samples after autoclave tests show that the transmission - a critical review and some failure
adhesive strength of FBE coatings is greater than analysis, First Gulf Water Conference, Dubai, 1992.
500 psi. [31 V. Novokshchenov, Corrosion of reinforced concrete
5. The results of the autoclave tests indicate in Persian Gulf region, Mat. Performance, 34(l)
(1995) 51.
that FBE coating shows very small variations in
thickness in the vapor phase and a definite 141 H. Saricimen, O.A. Ashiru, N.R. Jarrah, A. Quddus
and M. Shameem, Effect of inhibitors and coatings
increase in the aqueous phase. Moreover, no loss on rebar corrosion, Mat. Performance, 54(5) (1998)
of color was found. 32.
6. Closed-circuit loop test results after [51 J.D. Kellner, A.J. Doheny and B.B. Patil, A-3 layer
1 month exposure in distilled water indicate no polyethylene coating for plant applications, Mat.
marked change in the color and texture of the Performance, 54(6) (1998) 28-32.
coating. There was no perceptible change in 161 M. Alexander, 3 layer epoxy/polyethylene extruded
weight. coatings for high temperature application, Proc.
254 A. U. Malik ef al. /Desalination 150 (2002) 247-254

Int. Pipeline Conference Part-2 (2), ASME, Fairfield, 1123S. Flores and T.L. Starr, J. Protective Coating
NJ, USA, 1998, pp. 645-652. Linings, March, (1994) 76.
f?] M. George, Factors influencing the choice of internal [ 131 D.NeaJ and T. Whitehurst,Chloride contaminationof
and external anti-corrosion coatings for potable and line pipe, Mat. Performance, 34(2) (1995) 47.
raw water service, Middle East Corrosion Con- [14] J. Paul, Inspecting and repairing concrete before
ference, Manama, Bahrain, 1988. lining, Protective Coating Europe, February, (1996)
[8] B.P. Alblas and A.M. Van London, Protective 16.
Coating Europe, 1997. [ 151 External fusion bonded epoxy coating for steel pipe,
[9] B.R. Appleman, J. Protective Coating Linings, ARAMCO Material System specification, Cathodic
October, (1997) 68. Disbondment of the Coating. CAN/CSA-Z245.20-
[lo V.E. Heivig Metal Finishing, July, (1980) 41. M92, September, (1992) 36.
[l I] D.G. Weldon, A. Bochan and M. Schbiden, J. 1163AU. Ma&, unpublished results.
Protective Coating Linings, June, (1987) 46.