www.paintsquare.comJ P C L / N o v e m b e r 2 0 0 4 / P C E44

number of parameters

affect the adhesion of coat-

ings to steel substrates.

According to ISO 8503,1

these conditions include

the following substrate properties:

• rust and mill scale;

• contaminants, such as dust, oil, grease, and salts;

and

• profile (roughness).

In turn, the conditions above influence the selection

of a method to prepare the substrate for coating

application. Several studies have evaluated the

effects of surface preparation methods, primarily dry

abrasive blasting (AB) and waterjetting (UHP), on

coating performance. These studies also investigated

the long-term behaviour of coating systems. These

studies, however, focused on the effects of flash rust

and salt contamination.2,3

Combining AB and UHP (UHPAB) into a single

blasting process is an innovative approach to surface

preparation. Although this method has already been

introduced into ship repair practice, there is not

enough information available to evaluate the quality

of surfaces and coating adhesion associated with

UHPAB. This study’s goal is to systematically investi-

gate parameters that are suitable for evaluating the

quality of a steel substrate after the use of three

surface preparation methods: AB, UHP, and

UHPAB. The investigation focuses on repair of ships

and steel structures. The selected coating systems

were mainly suitable for ballast tank applications.

However, one coating, originally designed for steel

structures in water, such as weirs and flood bar-

rages, was also included to check the capabilities of

the investigated methods to prepare such surfaces.

The exposure environment for the coatings to be

studied was selected according to the classifications

Effects of Surface Preparation Methodson Adhesion of Organic Coatingsto Steel Substrates

A

By A.W. Momber, Mühlhan Equipment Services GmbH,

Hamburg, Germany

S. Koller, Germanischer Lloyd AG

Hamburg, Germany

and H.J. Dittmers, Cor roconsult GmbH

Hamburg, Germany

Fig. 1a: AB (dry abrasive blast cleaning)

Fig. 1c: UHPAB (ultra-high pressure abrasive blasting)

A

B

C

surface of each panel was separated

into three parts: one for AB, one for

UHP, and one for UHPAB. All test pan-

els were originally coated with an

organic coating system at an average

DFT between 400 and 800 µm (16 and

32 mils). After coating removal, the

surfaces treated with UHP and

UHPAB were allowed to dry (Fig. 2).

At the dry surfaces, roughness para-

meters were measured as described

below. Thereafter, the new coatings

were applied to the entire panel sur-

face. Thus, each coating was applied

identically over the three types of pre-

pared surfaces. After the coatings

cured, adhesion tests were performed

as described below.

Five commercially available epoxy

coating systems were selected. Their

major properties are listed in Table 3.

The coatings were applied by airless

spray in a paint booth with a con-

trolled climate. The coatings cured for

50 days. Then, specimens for the labo-

ratory tests were manufactured. Each

was 15 cm x 10 cm (6 in. x 4 in.)

according to the requirements of ISO

face preparation and coating applica-

tion were excellent because they were

performed under ideal, controlled con-

ditions. Such conditions cannot be

expected in field work.

Surface Preparationand Coating Procedures

Three methods were used for surface

preparation:

• AB as shown in Fig. 1a;

• UHP as shown in Fig. 1b; and

• UHPAB as shown in Fig. 1c.

The performance parameters of these

methods are listed in Table 2. The abra-

sive used for AB and UHPAB was a

commercial copper slag with the fol-

lowing properties in accordance with

ISO 11126-36:

• hardness > 7 (Mohs);

• density = 3600 kg/m3;

• particle size range = 0.4–1.2 mm; and

• particle shape—irregular.

For the UHP and UHPAB applications,

tap water with an electrical conductiv-

ity of 650 S/cm was used.

Test panel size was 3 m (~10 ft) in

length and 0.5 m (1.6 ft) in width. The

45www.paintsquare.com J P C L / N o v e m b e r 2 0 0 4 / P C E

listed in ISO 12944-2.4 The following

categories were considered.

• Atmospheric-corrosivity: C5-M

(marine). This category covers coastal

and offshore areas with high salinity as

well as areas with almost permanent

condensation.

• Categories for water and soil: Im2.

This category covers sea or brackish

water; structures considered include

sluice gates, locks, offshore structures,

and other completely or partially

immersed constructions.

These categories determined the type

and intensity of the laboratory tests to

be performed according to ISO 12944-

6.5 These tests are listed in Table 1. The

results of the neutral salt spray test

will be reported at a future date. The

present article reports about the

results of the pull-off strength measure-

ments. For the temporal and technical

conditions of our study, condensation

tests and a water immersion test did

not allow an evaluation of the coating

systems’ performance. This was mainly

due to two reasons. First, the duration

of testing was too short. Second, sur-

Corrosivity

Categor y

C5-M

Im2

Table 1: Test Procedures for Paint Systems Applied to Steel(ISO 12944-65)

Durability

Ranges

medium

medium

ISO 2812-2

(Cold Water Immersion)

—

2,000 h

ISO 6270

(Cold Water Condensation)

480 h

—

ISO 7253

(Neutral Salt Spray)

720 h

720 h

Parameter

Operating pressure in MPA

Nozzle diameter in mm

Water consumption in l/min

Abrasive consumption in kg/min

Table 2: Surface Preparation Method ParametersMethod

AB

8.5 (air)

10

0

12

UHP

200 (water)

6 x 0.3

10

0

UHPAB

150 (water), 8 (air)

19

10

12

Fig. 1a–c: clockwise from p. 44, top left, Surfacepreparation methods (see Table 2 for operationalparameters)

Fig. 1b: UHP (waterjetting)

12944-6.4 The specimens were cut

with plate shears. Commercially avail -

able coatings were manually applied to

each sample for edge protection. One

sample of the plates prepared with AB

and UHPAB was not coated, but was

saved for scanning electron microscopy

(SEM) inspections. From these samples,

small specimens with the dimensions

20 mm x 20 mm (~1 in. x 1 in.) were

cut.

Roughness was measured according

to ISO 8503-47 using a portable stylus

instrument. Adhesion between the

coating and the substrate was estimat-

ed according to ISO 4624,8 by means

of a commercial pull-off testing device

J P C L / N o v e m b e r 2 0 0 4 / P C E46 www.paintsquare.com

1a

EP

Zn

2

2.8

67

70

—

Table 3: Properties of Applied Coating SystemsProperty

Binder

Primer

Number of coatings

Density in kg/l

Solid content in vol.-%

DFT in µm

VOC in g/l

Coating System

1b

EP

—

—

1.5

100

490

—

2

EP

—

1

—

96

346

free

3

EP

—

1

—

83

253

—

4

EP

—

1

1.3

60

273

385

5

EP

—

1

—

80

374

170

Fig. 3: Effects of surface preparation methods on roughness parameters

Fig. 2: Samples after surface preparation and prior to coating

what different trend. Here,

AB forms the deepest pro-

file, followed by UHPA B

and UHP. But the differ-

ences in the maximum

roughness between

UHPAB and AB are only

3%. It seems from these

results that UHP does not

have the capability to form

a new profile under the

conditions of this study.

The results of the pull-off

tests are summarised in

Figs. 4 and 5. In Fig. 4, two

groups of columns can be

distinguished. The first

group covers coating sys-

tems 1, 3, and 4, whereas

the second group covers

coating systems 2 and 5.

The first group is charac-

terised by two features.

First, the pull-off fa i l u r e

was always cohesive (within the coat-

ing), not adhesive (between the substrate

and the coating). Second, the individual

coating systems showed a more or less

constant level of pull-off strength val-

J P C L / N o v e m b e r 2 0 0 4 / P C E 47www.paintsquare.com

with a maximum load capacity of 18

MPa. The dollies were stainless steel,

20 mm (~1in.) in diameter. The drying

time for the glue was four hours. The

samples were pre-cut before the pull-

off tests were performed. Six measure-

ments were taken for each sample.

Because three samples were available

for testing, 18 measurements were

taken for each condition. The portion

of adhesive failure and cohesive failure

was evaluated from the appearance of

the failed sections.

Results and DiscussionThe results of the roughness measure-

ments are summarized in Fig. 3. For

substrate evaluation, the roughness

parameter “average maximum rough-

ness,” RY 5, in accordance with ISO

5803, is very often specified. This para-

meter is equal to the roughness para-

meter RZ measured with the stylus.

This parameter has high values for the

UHPAB surfaces, followed by the AB

s u r faces and the UHP surfaces. The

same trend is true for the average

roughness Ra. The maximum rough-

ness, Rmax, however, shows a some

Fig. 5: Failure types after pull-off testing (coating system 5)a – adhesive failure (AB, dry abrasive blast cleaning)b – mixed adhesive-cohesive failure (UHP, waterjetting)c – cohesive failure (UHPAB, ultra-high pressure abrasive blasting)

Fig. 4: Effects of coating systems and preparation methods on pull-off strength

A

B

C

the coating system. As shown in Fig.

5b, the failure at the UHP blasted sam-

ples is partly adhesive. The failure on

the substrate prepared with UHPAB is

shown in Fig. 5c, where a totally cohe-

sive failure of the bulk coating can be

seen. A very interesting result is that

J P C L / N o v e m b e r 2 0 0 4 / P C E48 www.paintsquare.com

ues. The maximum deviation estimated

for the results of this group is 4.3% fo r

coating system 3. Thus, the surfa c e

preparation method as measured by

these tests does not have

a relevant effect on pull-

off strength. The average

pull-off values measured

on the corresponding

samples do not depend on

the substrate surfa c e

q u a l i t y, but rather charac-

terise the tensile strength

of the coating system.

They are, therefore, a

property of the bulk coat-

ing.

The situation with

the second group is quite different.

This group is characterised by two fea-

tures. First, pull-off strengths show a

dependence on the surface preparation

method and, therefore, on the proper-

ties of the prepared steel substrate.

Second, there is a notable trend in the

strength values: AB shows the lowest

values, followed by UHP, whereas the

highest values are measured after

UHPAB. An interesting observation

was made regarding coating system 5

(Fig. 5). Figure 5a shows that the AB

substrate is the only one with a plain

adhesive failure between the steel and

Sample Number

1 (AB)

2 (AB)

3 (AB)

4 (AB)

5 (AB)

Table 4: Grit Contaminationof Substrates after AB

Contamination in %

27.7

60.9

49.4

54.1

68.9

Fig. 7: Effects of surface preparation methods ontype-ii-grit contamination a, b – after AB (dryabrasive blast cleaning); c, d – after UHPAB

(ultra-high pressure abrasive blasting)

Fig. 6: Type i grit contamination after AB (dry abrasive blast cleaning)

A

B

C

D

J P C L / N o v e m b e r 2 0 0 4 / P C E 49www.paintsquare.com

the trends in roughness and pull-off

strength were different. Samples with

high roughness values (such as samples

prepared with AB) do not show the

highest values for pull-off strength.

Surface ContaminationTo understand the rather unexpectedly

low adhesion capacities of the AB sub-

strates, we performed SEM inspections

of the blasted surfaces. Embedded grit

particles are common on grit-blasted

surfaces. Similar to dust, the particles

may act as separators between the sub-

strate and the coating system. In stud-

ies of particle embedment during grit-

blasting,9-11 experimental results have

shown that grit embedment depended

mainly on impact angle and abrasive

type.10 Investigations by Fairfull and

Weldon9 showed a notable effect of grit

type and morphology on contamina-

tion. Slag material (except nickel slag)

was very sensitive to grit embedment.

Experiments with copper slag showed

that the breakdown behaviour of indi-

vidual particles during the impinge-

ment at the steel surface seemed to

play an important role. It was apparent

that the embedment was not simply

due to discrete particles lodging in the

substrate, but rather to extreme break-

down of the slag abrasive into minute

particles, or a physical smearing of the

grit over the surface.9 Embedded grit

reduces the adhesion of the coating to

the substrate. Measurements of the

adhesion strength as a function of the

amount of embedded grit were per-

formed by Griffith et al.12 These

authors showed that adhesion strength

diminished significantly if grit particles

were embedded in the substrate sur-

face. However, we observed the follow-

ing two types of grit contamination.

• Type i: mechanical embedment of

rather large abrasive particles (Fig. 6)

• Type ii: adhesive smearing of very

finely grained grit particles over the

surface (Fig. 7)

Note the high amount of contamina-

tion in Figs. 7a and 7b. Imaging soft-

ware was used to estimate the percent-

age of grit contamination. The results of

the measurements performed on the

substrates prepared with AB are listed

in Table 4. Contamination levels are

between 27.7 and 68.9%. Also of inter-

est is the type of contamination. The

images clearly show a layer of finely

ground particles that cover the sub-

strate surface similar to a separation

layer. This layer prevents direct con-

tact between the coating and the sub-

strate. The degree of contamination is

notably lower for UHPAB (Figs. 7c and

7d). This is probably due to the clean-

ing capability of the high-pressure

water flow that is part of the UHPAB

system. Figure 8 shows an SEM image

taken from the polished cross section

of a coated AB substrate. Considerable

accumulations of debris can be noted

between the substrate and coating at

places. The debris were identified by

EDX analysis to be copper slag.

We concluded from these observa-

tions that smearing of crushed abrasive

particles over the substrate surface is a

probable reason for the poor perfor-

mance of the organic coating systems 2

and 5 on substrates prepared by AB.

This behaviour has some counterpart if

flash rust intensity is evaluated.

Coating manufacturer specifications do

not allow the coating of substrates

showing heavy flash rust (designation

H), whereas light (L) and even moderate

(M) flash rust can be tolerated some-

times. The reason for this restriction is

mainly that heavy flash rust does not

adhere strongly enough to the sub-

strate.

However, because some coatings in

our study are not affected by grit cont-

amination, one could assume there are

Fig. 8: SEM image of polished cross-section of a coated AB substrate, with accumulations of debris between substrate and coating

J P C L / N o v e m b e r 2 0 0 4 / P C E50 www.paintsquare.com

other causes of the difference in adhe-

sion. The question arises of whether

some types of coatings are compatible

with grit-contaminated surfaces. Such

coatings could be analogous to those

defined as flash rust-tolerant coatings.

The capability of the liquid coating to

penetrate the grit layer and to encapsu-

late individual grit particles may play a

notable role. A similar process was

reported by Kaiser and Schütz13 with

regard to rust. A programme is

planned to investigate this issue. For

coating systems 2 and 5, the UHP sam-

ples show slightly lower adhesion val-

ues than the UHPAB samples. These

discrepancies could be explained by the

lower roughness of the substrates pre-

pared with UHP. Although UHP sam-

ples are completely clean in terms of

grit contamination, their profile is too

low to contribute to better adhesion.

Another conclusion from our study

is that the assessment of dust on steel

surfaces according to ISO 8502-314 is

not sufficient because it considers only

dust particles visible under 10x magni-

fication. The microscopically small

abrasive debris found in our study can-

not be detected by this method.

Therefore, even substrates that are cat-

egorised as dust size class “0” in ISO

8502-314 may not be completely free

of dust contamination.

SummaryThe investigation shows that a sensi-

tive balance exists between surface

preparation method and the coating

system to be applied. A proper combi-

nation of both can notably improve

adhesion and corrosion resistance in

general because good adhesion is a pre-

condition for good performance of pro-

tective coating systems. A layer of

microscopic, finely crushed abrasive

particles embedded in the surface

seems to affect bonding between the

substrate and coating for some coating

systems, while for others it does not.

With UHP and UHPAB, this effect can

be eliminated, and an optimum surface

morphology can be achieved with

UHPAB. Salt spray test results for the

performance of the coating systems

and substrates discussed in this article

will be reported later.

References1. ISO 8503: Preparation of Steel

Substrates before Application of

Paints and Related Products.

2. Allen, B., “Evaluating UHP

Waterjetting for Ballast Tank

Coating Systems,” PCE (October

1997), pp. 38–43.

3. Morris, M., “Update: Evaluating

UHP Waterjetting as Preparation

for Ballast Tank Coating,” PCE

(September 2000), pp. 54–59.

4. ISO 12944-2: Paints and

Varnishes—Corrosion Protection

of Steel Structures by Protective

Paint Systems—Part 2:

Classification of Environments,

1998.

5. ISO 12944-6: Paints and

Varnishes—Corrosion Protection

of Steel Structures by Protective

Paint Systems—Part 6: Laboratory

Performance Test Methods, 1998.

6. ISO 11126-3: Preparation of Steel

Substrates before Application of

Paints and Related Products—

Specifications for Non-metallic

Blast-cleaning Abrasives—Part 3:

Copper Refinery Slag, 1993.

7. ISO 8503-4: Preparation of Steel

Substrates before Application of

Paints and Related Products—

Surface Roughness Characteristics

of Blast-cleaned Steel Substrates—

Part 4: Method for the Calibration

of ISO Surface Profile Comparators

and for the Determination of

Surface Profile, Stylus Instrument

Procedure, 1988.

8. ISO 4624: Paints and Varnishes—

Pull-off Test for Adhesion, 2002.

9. Fairfull, C.L., Weldon, D.G., “Salt

Contaminated Abrasives and their

Effect on Coating Performance,”

Proceedings of the SSPC

Conference, Atlanta, 2001

(Pittsburgh, PA: SSPC: The Society

for Protective Coatings, 2001).

10. Amada, S., Hirose, T., Senda, T.,

“Quantitative Evaluation of

Residual Grits under Angled

Blasting,” Surface and Coatings

Technol., Vol. 111, 1–9, 1999.

11. Momber, A.W., Wong, Y.,

Budidharma, E., Tjo, R., “Profiling

of Low-carbon Steel with

Supersonic Water Jets,” Wear, Vol.

249, 2002, pp. 853–859.

12. Griffith, B.J., Gawne, D.T., Dong, G.,

“A Definition of the Topography of

Grit-blasted Surfaces for Plasma

Sprayed Alumina Coatings,” Trans.

ASME, J. of Manuf. Sci. and Engng.,

Vol. 121, 1999, pp. 49–53.

13. Kaiser, W.D., Schütz, A., 2001,

Hochdruckwasserstrahlen und

Restrostverträglicher

Beschichtungsstoff—eine Mögliche

Instandsetzungsstrategie, WEKA,

Bd. 2, Teil 9, Kapitel 8.5, pp. 1–26.

(See also: Momber, A.W.,

Hydroblasting and Coating of Steel

Structures, Elsevier Advanced

Technology, Oxford, 2003,

pp.124–125.)

14. ISO 8502-3: Preparation of Steel

Surfaces before Application of

Paint and Related Products—Tests

for the Assessment of Surface

Cleanliness—Part 3: Assessment

of Dust on Steel Surfaces Prepared

for Painting (Pressure-sensitive

Tape Method), 1992.