Performance of Marine Coatings Applied over Flash Rusted Surfaces
J. Peter Ault, P.E.
Elzly Technology Corporation
Ocean City, NJ
Steve Cogswell
Atlantic Marine and Drydock
Jacksonville, FL
Abstract: This paper investigates a key concern with Ultra High Pressure Waterjetting
(UHPWJ) surface preparation – the impact of “flash rusting” on coating life. Flash rusting
can occur under certain environmental conditions when the steel is left sufficiently wet
following UHPWJ. Reducing or eliminating flash rusting can increase the cost of surface
preparation. However, flash rust’s impact on coating life is debated. This paper presents
an evaluation of coatings applied over flash rust on US Navy ship hulls. The coatings were
inspected after up to six years in service. The results show that the coating performance
over locations which had Moderate levels of flash rust was comparable to those with Light
levels of flash rust.
Introduction
Currently, several shipyards utilize ultra high pressure waterjetting (UHPWJ) to prepare
steel for painting. One drawback of UHPWJ is the occurrence of “flash rust” on the
surface after cleaning. Because coatings have historically been applied over clean “white
metal” it is somewhat of a paradigm shift for the industry to begin painting over flash rust.
NACE No. 5/SSPC-SP 12, Surface Preparation and Cleaning of Metals by Waterjetting
Prior to Recoating defines three parameters of surface cleanliness – four levels of visible
contaminants, four levels of flash rust, and three levels of non-visible contaminants. The
levels of flash rust defined in NACE No. 5/SSPC-SP 12 are:
o No Flash Rust: A steel surface which, when viewed without magnification,
exhibits no visible flash rust.
o Light (L): A surface which, when viewed without magnification, exhibits small
quantities of a yellow-brown rust layer through which the steel substrate may be
observed. The rust or discoloration may be evenly distributed or present in patches,
but it is tightly adherent and not easily removed by lightly wiping with a cloth.
o Moderate (M): A surface which, when viewed without magnification, exhibits a
layer of yellow-brown rust that obscures the original steel surface. The rust layer
may be evenly distributed or present in patches, but it is reasonably well adherent
and leaves light marks on a cloth that is lightly wiped over the surface.
o Heavy (H): A surface which, when viewed without magnification, exhibits a layer
of heavy red-brown rust that hides the initial surface condition completely. The rust
may be evenly distributed or present in patches, but the rust is loosely adherent,
easily comes off, and leaves significant marks on a cloth that is lightly wiped over
the surface.
Some commercial owners allow coatings to be applied over “Moderate” levels of flash
rust. The U.S. Navy only allows painting over “Light” levels of flash rust due to their
lower level of risk tolerance and longer expected service life. Closed-loop UHPWJ
systems routinely provide surfaces which results in light to no flash rust. However, areas
on ships which have complex shapes do not accommodate the closed loop equipment.
Hand lance equipment is commonly used to clean these areas. Because hand lances do not
incorporate water removal devices, surfaces tend to remain wet longer and develop
Moderate or Heavy levels of flash rust. Costly secondary surface preparation processes are
required to reduce the flash rust to the “Light” condition. It has been estimated that this
added process increases the cost of surface preparation up to 20% on an underwater hull.
To better understand the impact of flash rusting on coating life, this project compares the
performance of coatings applied to four (4) Navy hulls prepared with closed-loop UHPWJ
systems on broad areas and hand lance UHPWJ with minimal secondary surface
preparation in complex hull areas.
Approach
As part of this project, we identified a series of ships including Oliver Hazard Perry class
frigates and Ticonderoga class Aegis guided missile cruisers which had hull preservation
work performed in the 2001-2002 timeframe. The work was performed at a shipyard in the
southeast United States. During this preservation work, the contractor requested approval
to deviate from H-B 2½ L (Light Flash Rust) to H-B 2½ M (Moderate Flash Rust) 1 on the
hand lance hydroblasted areas. Approval of these requests was granted with the following
constraints:
• The hand lance cleaned areas were kept to the minimum feasible
• The amount of time between hand-lance blasting and blow-down with dry air was
minimized
• The amount of time between blow-down with dry air and application of the primer
coat was minimized
• Evidence of coating manufacturer concurrence was provided
In 2007, three frigates and one cruiser which were preserved with the above described
deviation were to be dry-docked. During each of the dockings, we collected as much
evidence as possible to determine if there were any differences in the performance of the
1
An H-B 2½ M is nominally equivalent to WJ-2M. Similarly, an H-B 2½ L is nominally equivalent to WJ-
2L.
coating over the close-loop cleaned areas versus the hand lance cleaned areas. The
inspection team consisted of:
• Consulting engineer who performed the testing
• A Navy representative who participated in the inspections when the surface
preparation and coating was originally accomplished
• Shipyard personnel who were present when the surface preparation and coating was
originally accomplished
For the first two ships, the project team performed a visual observation of the hull. If
significant performance differences existed where the hand lance cleaned surfaces were
allowed to be painted over, coating blistering and/or adhesion loss may be more prevalent
in the hand lance cleaned areas versus the broad areas cleaned with the closed-loop
UHPWJ equipment. Visual inspections were made from the drydock floor after the ship
was hauled and cleaned with low pressure water. Areas which could not have been
cleaned with the closed-loop system were inspected for signs of coating failure such as
peeling paint, rusting, blistering, and undercutting at defects. The inspections were
documented with written comments and photographs. The visual inspections did not
differentiate performance over the hand lance versus closed-loop machine cleaned areas.
It was decided that a physical evaluation could potentially be more informative than the
visual inspections. A test procedure was developed which included:
• Document the condition of the test area prior to testing (DFT, photographic,
descriptive).
• Selectively remove the topcoat (AF or Alkyd) to expose the anticorrosive coating.
• Perform various tests on the anticorrosive to ascertain the integrity of the coating
(pull-off adhesion, coating capacitance, film thickness).
• Remove a small area of the epoxy (nominally 3-inch by 3-inch), allowing a visual
examination of the steel substrate and an assessment of the primer/steel bond.
• Repair test areas as required by the ship owner.
The physical test procedures were designed to better characterize the condition between the
primer and the steel surface. This is the interface which one would expect to be affected if
the flash rust associated with the hand lance cleaning impacted the coating performance. If
the flash rust in the hand lance areas is negatively impacting coating performance, the pull-
off adhesion should fail at the primer/substrate interface at loads below 800 psi.
Additionally, if the flash rusting is negatively impacting performance we would expect to
see active corrosion where the coating is chemically removed to expose the substrate.
Results
USS KLAKRING (FFG-42). The USS KLAKRING (FFG-42) had underwater hull
coating removed and reapplied in November, 2002. A deviation request was approved to
allow an H-B 2½M condition on the hand lance hydro-blasted areas in lieu of the required
H-B 2½L. The deviation request was accompanied by a letter from the paint manufacturer
indicating that the primer was suitable for use over an H-B 2½M prepared surface.
On January 30, 2007 the USS KLAKRING was inspected in drydock. There were no
visual differences evident between areas that were machine cleaned and hand-lance
cleaned. The underwater hull was in good shape with negligible blistering. The paint
appeared intact and well adhered. The freeboard coating was also in good condition with
the exception of localized rusting at locations of mechanical damage and near overboard
discharges.
USS HUE CITY (CG-66). The USS HUE CITY (CG-66) had underwater hull coating
removed and reapplied in December, 2002. A deviation request was approved to allow an
H-B 2½M condition on the hand lance hydro-blasted areas in lieu of the required H-B
2½L. The deviation request was accompanied by a letter from the paint manufacturer
indicating that the primer was suitable for use over an H-B 2½M prepared surface.
On February 20, 2007 the USS HUE CITY was inspected in drydock. There were no
visual differences evident between areas that were machine cleaned and hand-lance
cleaned. The underwater hull was in good shape with negligible blistering. The paint
appeared intact and well adhered. The freeboard coating was also in good condition with
the exception of localized rusting at locations of mechanical damage and near overboard
discharges. The rudders had 100% failure of the antifouling to the EC2216 system. The
top layer of ablative antifouling coating was lost as the marine growth dried. In-tact
ablative antifouling coating remained on the hull.
USS BOONE (FFG-28). The USS BOONE (FFG-28) had underwater hull coating
removed and reapplied in April, 2001. A deviation request was approved to allow an H-B
2½M condition on the hand lance hydro-blasted areas in lieu of the required H-B 2½L.
The deviation request was accompanied by a letter from the paint manufacturer indicating
that the primer was suitable for use over an H-B 2½M prepared surface.
On April 2, 2007 (after 6 years in service) a physical evaluation of the USS BOONE was
conducted to determine if a measurable performance difference exists among surfaces
prepared by waterjetting with a hand lance versus a closed-loop system. The hull was
inspected while in drydock. Eight areas, each nominally one square foot were selected for
testing. Six of the areas were on the underwater hull and two areas were on the freeboard.
Of the six underwater hull areas, four were cleaned with a hand lance while two were
likely cleaned with closed loop HPWJ. On the freeboard, one area was cleaned with a
hand lance while the other was likely cleaned with closed loop HPWJ.
A general survey of the hull coating condition showed blistering over approximately 2% of
the underwater hull. Blistering was generally concentrated along the keel. Several blisters
were popped. All contained liquid with a pH of 12-13. About half of the blisters were
between the epoxy primer and steel; the remainder were intracoat blisters. A
representative blistered area (with several popped blisters) is shown in Figure 1.
Detailed data sheets for each of the eight test locations was collected as described in the
“Approach” section. The following observations are made from the data:
• The anticorrosive coating consisted of two coats of epoxy – a red primer and grey
intermediate coat. In the underwater hull locations the anticorrosive was 10-15
mils thick with 2-15 mils of antifouling coating. In the freeboard area, the
anticorrosive was 20-30 mils with an additional 30 mils of alkyd topcoat.
• Nearly all of the adhesion tests failed either within the primer or at the adhesive.
One adhesion pull which was performed over a blister failed at the substrate under
low load (240 psi). Adhesion values as high as 1700 psi were observed.
• Physical removal of the coating showed further evidence that the primer was well
adhered in all locations. Red primer was difficult to remove from the profile
depths. Unburnished peaks had a brownish coloration, but there was no evidence
of “loose” corrosion.
• Only two areas showed indications of compromised adhesion. Figure 2 shows a
close-up of one pull-off test was placed over a latent blister. This test failed to the
substrate at 240 psi. Figure 3 shows two small areas comprising 2% of the coating
removed from patch 8 (0.25 in²) was easily removed to the substrate with
mechanical force. Remaining tests on both of these areas demonstrated excellent
adhesion.
In summary, the physical inspection revealed no evidence of coating failure which would
logically be traced to a systematic or process problem associated with hand lance cleaning.
Of the surfaces inspected, all had tightly adherent epoxy primer. Blistering of the
underwater hull coating was observed along the keel.
Figure 1. Representative blistering along the keel in the vicinity of the stern air masker.
Figure 2. Pull-off adhesion test 1-7, situated near blistering on keel.
Figure 3. Small areas of substrate revealed when excavating at test location 8.
USS SAMUEL B. ROBERTS (FFG-58). The USS SAMUEL B. ROBERTS (FFG-58)
had underwater hull coating removed and reapplied in August, 2002. A deviation request
was approved to allow an H-B 2½M condition on the hand lance hydro-blasted areas in
lieu of the required H-B 2½L. The deviation request was accompanied by a letter from the
paint manufacturer indicating that the primer was suitable for use over an H-B 2½M
prepared surface.
On July 15-16, 2007 (after approximately 5 years of service) a physical evaluation of the
hull coating on the USS SAMUEL B. ROBERTS was performed to determine if a
measurable performance difference exists among surfaces prepared by waterjetting with a
hand lance versus closed loop UHPWJ. The hull was inspected while in drydock. Eight
areas, nominally one square foot were selected for testing. All eight areas were on the
underwater hull. Of the eight areas, four were cleaned with a hand lance while four were
likely cleaned with a closed loop UHP robot.
No significant coating anomalies were observed during a visual survey of the exterior hull.
There was minor blistering at various locations (including along the keel and below
overboard discharges). Areas which did not receive a complete antifouling coating due to
the locations of docking blocks were also noted.
Detailed inspections were conducted at eight locations, somewhat randomly selected by the
project team. The locations were in matched pairs where one patch was several feet below
the boot stripe and its companion patch was near an appendage which would have
precluded cleaning with closed-loop HPWJ equipment. The following observations are
made based on the data collected:
• The anticorrosive coating consisted of two coats of epoxy – a red primer and grey
intermediate coat. In the underwater hull locations the anticorrosive was 10-15
mils thick with 2-15 mils of antifouling coating.
• Adhesion tests failed at various locations through the coating system including
within the primer and between the primer and substrate. Adhesion values between
1,119 and 2,453 psi were observed.
Subsequent to the physical testing, the coating was removed from a 10 by 10 inch area of
patches 3 and 4. Coating removal was accomplished with repeated application of a paint
stripper and scrubbing with a brass brush. Figure 4 and Figure 5 show the substrate of
patch 3 and 4, respectively after coating removal.
Figure 4. Test Patch 3 with nominally 100 square inches of coating removed.
Figure 5. Test patch 4 with nominally 100 square inches of coating removed.
Physical removal of the coating showed further evidence that the primer was well adhered
to a sound substrate. In both cases, the primer was difficult to clean from the depths of the
profile. The exposed steel had a definite anchor profile which visually appeared to exceed
1 mil. Figure 6 shows a close-up of the steel surface at patch 4.
Figure 6. Close-up of the substrate revealed at patch 4.
The lower patch (patch 3) had a stripe of stain through the center. Adhesion test 3-1 was
situated such that part of the dollie was on the stained area and part was on the clean area.
While the adhesion test failed to the substrate on the stained area, the pull test failed at
1968 psi – far in excess of most manufacturers’ requirements for a new coating. Figure 7
shows a close-up of the stain in the area of adhesion test as well as the adhesion test after
failure.
Figure 7. Close-up of adhesion test 3-1 and the test location after coating removal.
After further review of the stain, it was observed that the stain lined up with a streak of
blisters that continued up to an overboard discharge. The type of discharge was not
determined, but it is probable that the dark staining is associated with the discharge rather
than an anomaly related to flash rust associated with hand-lance surface preparation.
Figure 8 shows the relationship between the test patch, observed blistering, and the
overboard discharge.
In summary, the physical inspection revealed no evidence of coating failure which would
logically be traced to a systematic or process problem associated with hand lance cleaning.
Of the surfaces inspected, all had tightly adherent epoxy primer.
Figure 8. Location of test patch relative to overboard discharge. Note line of blistering
highlighted by arrows in the photo.
Conclusion
None of the visual or physical inspections revealed differences in performance after five or
six years which could be attributed to the different waterjetting surface preparation
processes (closed-loop UHP vs hand-lance). This suggests that secondary surface
preparation to remove moderate flash rust is probably unnecessary after hand-lance
waterjetting with the procedures used by the shipyard. Eliminating secondary surface
preparation of moderately flash rusted areas is expected to result in significant cost
savings.
Acknowledgements
The authors would like to thank the National Shipbuilding Research Program (NSRP) and
US Navy Southeast Regional Maintenance Center (SERMC) for their support of this
project. It could not have been a success without their participation.
