GFGSDF
GFGSDF
GFGSDF
net/publication/223804945
CITATIONS READS
18 1,118
2 authors, including:
Banihan Gunay
Universiteti Metropolitan Tirana
61 PUBLICATIONS 506 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Banihan Gunay on 05 March 2018.
Abstract
Hot rolled steel (HRS) is used extensively in the automotive, agricultural and appliance industries. The corrosive response of HRS was investigated
after it had been exposed to various surface treatments, prior to powder coating. The behaviour of three conversion coatings: zinc phosphate (ZnP),
iron phosphate (FeP) and zirconium (Zr)-based nano-scaled, on HRS was studied. HRS is naturally covered with iron oxide scale and this was
removed from the surface by mechanical and chemical processes, prior to the application of surface treatment and organic coatings. The following
tests on differently treated panels were conducted to evaluate corrosion performance: adhesion tests such as, crosshatch, pull-off, and conical
bending, SEM, XPS, salt spray and electro-chemical impedance studies were also performed. Good correlations were recorded showing that zinc
phosphate conversion coating gave the best performance, and zirconium-based nano-structured conversion coating, was superior to that of iron
phosphate conversion coatings on HRS.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Coatings; Conversion coatings; Hot rolled steel; Phosphating; Oxide scale
0300-9440/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.porgcoat.2007.10.004
B. Tepe, B. Gunay / Progress in Organic Coatings 62 (2008) 134–144 135
formed at temperatures above 570 ◦ C have three different scale mance of the pre-treatments of cold rolled steel and hot dipped
layers: wustite (Fe1 O), magnetite (Fe3 O4 ) and hematite (Fe2 O3 ) galvanized steel.
[12,13]. Oxide scale on the surface is not a continuous layer and
does not represent a protective barrier. Loosely adhered oxide 2. Experimental set-up
scale on HRS is prone to chips and will crack under impact and
in environmental conditions such as heat and humidity. The 2.1. Sample preparation
oxide scale on the surface provides cathodic protection and steel
substrate acts as anode. If there are any breaks on the oxide scale Hot rolled steel (HRS) and hot rolled steel, pickled and
layer the electrochemical action causes steel substrate to corrode oiled (HRS-P), were used for the experiment. Test panels were
without affecting the oxide scale [14]. Application of powder prepared according to ASTM 609 [24] and profiled in turret
coating over HRS surfaces that have not had the scale removed punching machines. Fig. 1 shows a test panel. HRS test panels
is very likely to result in adhesion failure. Therefore, oxide scale were cut from 3 mm thick sheets, HRS-P test panels were cut
on HRS should be removed prior to surface treatment or coat- from 2 mm thickness sheet. Some of the surface investigations
ings. This is usually done by chemical or mechanical methods or such as SEM and XPS experiments were conducted on test pan-
a combination of both [15]. Chemical cleaning using acids such els measuring 10 mm × 10 mm. This is illustrated in Fig. 1. After
as HCl, H2 SO4 , and H3 PO4 is a very effective method for the pre-treatment in different conditions, a side cutter was used to
removal of rust and mill scale [16,17]. The general mechanism cut these small test panels. These samples were suitable for SEM
to remove oxide scale is where acid penetrates through cracks and XPS studies.
and the hydrogen gas pressure increases to a point where the Oxide scale on hot rolled steel was removed by shot blasting
scale is blown off the metal surface. This method is known as and acid pickling. Panels were blasted in a monorail conveyor
“pickling” and sulphuric acid is commonly employed in this type of shot blasting machine and afterwards the surface rough-
process. Bhushan and Gupta [5] and Cheng et al. [18] have ness was recorded in Ra (note that Ra is the arithmetic mean of
studied the process of acid pickling, using warm sulphuric acid the absolute departures of the roughness profile from the mean
(H2 SO4 ) that was used for reference chemical cleaning in this line). Table 1 shows shot blasted test panels.
study. Brushing and blasting is an effective mechanical cleaning The acid pickling process was achieved in a spray cabinet.
method for oxide scale cleaning. However, because brushing is Acid solution was prepared by using 10% (v/v) H2 SO4 , and the
labour intensive it is at a disadvantage to the blasting process. solution heated to 60 ◦ C. The test panels were sprayed for 30 min
Blasting is the most popular process to clean oxide scale on then rinsed with deionised water. Table 2 shows the test panel
HRS. In this study blasting has been used for mechanical used in this paper.
cleaning.
The purpose of the present paper is to evaluate zinc phosphate, 2.2. Pre-treatment
iron phosphate and zirconium-based nano-scaled conversion
coating on HRS. Other authors have studied pre-treatment pro- Prior to the powder coating phase, a pre-treatment phase was
cesses on steel and its alloys [10,19–22]. However, there has been administered by five stages commonly used in phosphating in
nothing specific undertaken on HRS. The corrosion and adhesion industry: cleaning, rinsing, conversion coating, rinsing, and pas-
resistance of the conversion coatings on HRS was correlated by sivation. During this study the average size of a pre-treatment
using different test methods. Many authors have used salt spray plant was taken as a reference. Table 3 shows pre-treatment
and electrochemical methods for corrosion study [10,22]. Tang stages and zone lengths. It has been considered that it was a con-
et al. [23] have used the EIS method to investigate the perfor- veyor line and line speed as 2400 mm/min. According to this line
Table 1
Shot blasted test panels
Test panel Measurement reading 1 Measurement reading 2 Measurement reading 3 Measurement reading 4 Average Standard deviation
Table 2 Table 4
Test panels used in the study Coating weights
Type of test panel Process prior to Legends Coating type Coating weight
pre-treatment
Zinc phosphate 2.54 g/m2
HRS (hot rolled steel) Shot blasted S0–S9 Iron phosphate 0.65 g/m2
HRS (hot rolled steel) Acid pickling E0–E9 Nano-scale Unknown
HRS-P (hot rolled steel pickled and oiled) P0–P9
ASTM 5402 [25] for curing; the MEK solvent rub test was used.
speed, contact time of each stage was calculated and it has been Average powder coating thickness was kept between 60 m and
simulated in a small spray cabinet. After alkaline cleaning and 70 m. For electrochemical impedance, test panels were coated
rinsing (first stage), a water break test was conducted for clean- with 40 m thickness powder paint.
liness of surface. The coating weight for zinc-phosphate and
iron-phosphate was measured according to Eq. (1). The weights 2.4. Adhesion test
of phosphated test panels were recorded in gram (W1 ) and the
panels were cleaned by using 6% by volume hydrochloric acid Adhesion testing was conducted according to ASTM D4541
(HCl) then their weights were recorded in gram (W2 ). From the [26] pull-off and ASTM D3359 [27] tape test and ASTM D522
change of panels’ weights, the weight of the coating per unit [28] conical mandrel test. A Zwick tensile testing instrument is
area (g/m2 ) was calculated. Table 4 shows coating weight for used for the pull-off method to measure the lift force required to
test panels used in this study. pull a small area of coating away from the base metal. A dolly
W2 − W 1 was attached by adhesive (epoxy araldite) to the coating. After
Coating weight (g/m2 ) = (1) curing, the coating was cut around the base of dolly. The dolly
coated area (m2 )
and test panels were attached by jig to the upper and lower jaw
of the tensile test machine. 0.5 mm/min speed was applied and
2.3. Paint system maximum force was recorded on breakage point. The percent-
age of paint lifting and maximum force was recorded. For tape
After pre-treatment and dry-off process, all test panels were testing, cross-cut technique was used as stated on ASTM 3359
painted in an electrostatic powder coating booth. White indus- [27] method B. ElcometerTM cross-cut blades were used in this
trial TGIC free polyester powder coating was used. Coated test test. Classifications of adhesion tape test result were recorded
panels were stowed for 20 min in a curing oven at 200 ◦ C. After according to test method. A 1 mm spacing crosshatch cutter
the coating process some of the panels were tested according to with six cutting edges was used for this test. ASTM D522 [28]
Table 3
Pre-treatment stages
Process type Stage 1 Stage 2 Stage 3 Stage 4 Stage 5 Dry off oven
conical mandrel testing was conducted. Test panels for a con- coated test samples have a hydrophobic surface. A thin layer
ical mandrel test were ground down to 0.8 mm thickness prior of coating was lifted by using very fine emery paper, which
to pre-treatment and powder coating. This is stated in ASTM allowed the surface to absorb NaCl solution on to the painted
D522 [28] standard. Then a powder coated HRS sheet was bent surface.
over a conical mandrel and the coating observed for cracks and
discolouration. 3. Results and discussion
Neutral salt spray testing was conducted according to ASTM The results of the cleaning process of oxide scale and con-
B117 [29] in Q-FOGTM CCT 600 neutral salt fog cyclic corro- version coatings used in this study was observed by using
sion tester. Salt spray testing was conducted for all test samples scanning electron microscopy (SEM). Fig. 3A illustrates the
until failure or 1000 h if no failure occurred. An “X” incision chemical cleaning of oxide scale with 10% (v/v) hot sulphuric
was scribed through the coating by using (ElcometerTM scriber) acid. Acid pickling is an effective chemical cleaning process for
making sure that the coating was scribed all the way to the sub- the removal of oxide scale however, it was recorded that some
strate. The test panels were reported for surface corrosion and oxide scale remained on the surface as shown in Fig. 3A. This
creepage from the scribe at 250 h, 500 h, 750 h and 1000 h. If residue could be avoided by increasing the spray time in the
creepage on any of the panels was over 3 mm from the scribe, spray cabinet.
testing for those panels was ended and reported as a failure before Fig. 3B and C shows samples after zinc phosphate pre-
the end of the test period. Ratings based on creepage from the treatment; the samples were used as delivered from the supplier
scribe were detailed on a report sheet. and did not have the oxide scale removed. It is necessary to
remove oxide scale prior to coating as it was found that it
2.6. Electrochemical impedance spectroscopy (EIS) prevents zinc phosphate crystallization and in the dry-off after
the pre-treatment process, oxide scale cracks; this is shown in
On each differently pre-treated test panel one cylindrical tube Fig. 3B. The method of shot blasting to remove oxide scale on
of a transparent acrylic tube was fixed by using an epoxy (epoxy HRS was successful but 20–25% of oxide scale still remained on
araldite) adhesive. The geometrical area for impedance measure- the surface. Fig. 3E shows that shot blasted test samples still have
ments was 10 cm2 . The electrochemical tests were conducted some oxide scale on the surface. Fig. 3D and F shows typical
with 0.5% wt. NaCl solution. A three electrodes electrochem- zinc and iron phosphate coating. It is clearly shown that the zinc
ical cell was formed with a platinum mesh counter electrode phosphated test panel had some pores, voids on the surface, due
(CE) fixed on a rubber bund, which was fixed on an acrylic to the remaining oxide scale. Contrary to the zinc phosphated
tube. The working electrode was attached (WE) on a differently panels, iron phosphated panels displayed good surface coverage
pre-treated test panel (pre-treated and powder coated HRS). The (Fig. 3F).
potentials were measured and referred (RE) to a (SCE), saturated X-ray photoelectron spectroscopy (XPS) was carried out
calomel reference electrode (4 M KCl). Schematic electrochem- using a Kratos Axis Ultra DLD spectrometer for qualitative
ical experiment set-up is shown in Fig. 2. analysis of the nano-scaled conversion coating and oxide scale.
All EIS studies were performed using a Solartron 1260 fre- In the SEM study for nano-scaled conversion coating, it was
quency response analyser, Solartron 1287 potentiostat and a not possible to quantify deposition on the substrate because of
PC. Impedance data were collected at frequencies ranging from its nano-scaled coating characteristic. Therefore, an XPS study
104 Hz to 10−1 Hz. The experimental impedance was inter- was initiated. Fig. 4 shows XPS study for nano-scaled conver-
preted in ZPlot electrochemical impedance software. All the tests sion coating were it can be seen that zirconium and fluoride was
were conducted at room temperature (20 ± 2 ◦ C). The powder deposited on the surface because fluorozirconic acid (H2 ZrF6 )
was used in the nano-scaled conversion coating composition.
Zirconium and fluoride content is shown in Table 5. Zr content
for HRS-P test panels was found to be the highest. This could
result from the fact that the surface of the HRS-P test panels was
free from oxide scale (Fig. 5). Compared to salt spray results
HRS-P test panels performed better with Zr-based coating. This
could be attributed to the Zr content on HRS-P test panels being
Table 5
XPS quantification data (at.% concentration) for nano-scaled conversion
coatings
HRS shot blasted (%) HRS-P (%) HRS acid pickled (%)
more than on acid pickled and shot blasted HRS test panels. Binding energy values of Fe 2p3/2 and Fe 2p1/2 indicated
More Zr deposition took place on HRS-P panels were the con- existence of oxide scale magnetite (Fe3 O4 ) and hematite
tent of Zr (9.04%) and salt spray results were very good. Most (Fe2 O3 ) [30,31]. Fig. 5 shows XPS spectra of oxide scale
of the test panels exceeded 500 salt spray hours; some of the test layer. The XPS survey scan confirmed that oxide scale
panels reached 750–1000 salt spray hours. remained on HRS shot blasted and acid pickled test pan-
The existence of oxide scale for mechanically and chem- els. However HRS-P test panels were free from oxide
ically cleaned HRS test panels was analysed by using XPS. scale.
B. Tepe, B. Gunay / Progress in Organic Coatings 62 (2008) 134–144 139
Table 6
Pull-off adhesion and crosshatch test results
Test samples Pull-off adhesion test Crosshatch test
Adhesion (MPa) Percentage area removed (%) Classification Percentage area removed
Table 7
Salt spray results
over 500 h in the salt spray test, whereas results for acid pickled scale, for adhesion and creepage failure, which were exposed to
and shot blasted panels were very poor. Many of the panels failed 500 h of salt spray testing.
in a 250 h time span. Zirconium-based nano-scaled conversion
coating performed better than that of iron phosphate, and some 3.4. Electrochemical impedance measurements
of the shot blasted and acid pickled panels reached 750 salt spray
hours. Differently pre-treated test samples (Table 8) were immersed
Adhesion failure was recorded on some of the shot blasted and in 5% wt. NaCl electrolyte, and data was collected from electro-
acid pickled panels during the salt spray test. It was shown that chemical impedance measurements (Fig. 9), for each conversion
untreated test panels were prone to failure of adhesion in high coating and oxide cleaning systems. Bode plots display absolute
humidity. Fig. 6 illustrates test panels not cleaned from oxide value of the impedance |Z| ( cm2 ), which was plotted versus the
frequency (Hz). The total impedance of test panels defined as |Z| frequency over 104 Hz capacitance was shorted, low frequency
( cm2 ) value, extrapolated to 10−1 Hz. In the Bode magnitude up to 10−1 Hz was used to measure capacitive response. It was
plots, which were used as comparison of each coating system. noted that HRS-P test panels (P7), which were coated with ZnP
Fig. 7 shows impedance |Z| ( cm2 ) at 10−1 Hz. At the high did not show any bode magnitude plot in day 1. This was due
Fig. 9. Nyquist (Z cm2 vs. Z cm2 ) and bode (frequency (Hz) vs. |Z| cm2 ) magnitude plots for differently pre-treated samples. (a) HRS-P FeP conversion
coated samples; (b) HRS-P nano-scaled conversion coated samples; (c) HRS-P ZnP conversion coated samples; (d) HRS acid pickled FeP conversion coated samples;
(e) HRS acid pickled nano-scaled conversion coated samples; (f) HRS acid pickled ZnP conversion coated samples; (g) HRS shot blasted FeP conversion coated
samples; (h) HRS shot blasted nano-scaled conversion coated samples; (i) HRS shot blasted ZnP conversion coated samples.
142 B. Tepe, B. Gunay / Progress in Organic Coatings 62 (2008) 134–144
Fig. 9. ( Continued )
to conversion coating resistance to permeability. In day 2 it dis- et al [32] high performance coating systems can be categorised
played readings of |Z| 108 cm2 at 10−1 Hz. Also, impedance by high impedance |Z| > 109 cm2 at low frequency <10−2 Hz.
for acid etched and ZnP test panels (E7) were very high and Zinc phosphating exhibited the best corrosion resistance on
recorded 109 cm2 (Fig. 9(f)) in the first 2 days but began to HRS-P and HRS acid pickled test panels as shown in Fig. 9(c)
degrade over the duration of the test. The two samples were an and (f). Capacitive region for those panels at 10−1 Hz was over
indication of good corrosion resistance. According to Granata 108 cm2 and after 15 days immersion in NaCl solution, resis-
B. Tepe, B. Gunay / Progress in Organic Coatings 62 (2008) 134–144 143
Fig. 9. ( Continued ).
tance to corrosion was degraded to around 106 cm2 ; despite tallization of ZnP conversion coating and may have created some
this finding the panels still measured the highest resistance. Shot holes, voids and pores. SEM study in Fig. 3C and D shows
blasted and ZnP conversion coated samples (Fig. 9(i)) did not that ZnP conversion coating on oxide scale was insufficiently
perform as well as other ZnP coated samples due to pores and deposited on the surface after pre-treatment.
voids on the conversion coating surface. Oxide scale on the shot Iron phosphating had mediocre barrier properties compared
blasted surface was not cleaned properly and this prevented crys- to those of other conversion coatings. On iron phosphated test
144 B. Tepe, B. Gunay / Progress in Organic Coatings 62 (2008) 134–144