Filiform corrosion (FFC) is one of the possible failure mechanisms of organic-coated steel. Beyond cathodic delamination, FFC causes paint detachment and the dramatic loss of the protective properties given by the applied paint. As both failure mechanisms occur during in-field exposure (depending on the environment), when assessing the performance of the protective paint by lab scale tests, we must be aware of the failure mechanism we induce with different accelerated aging cabinets. In this study, we investigate the effect of the prohesion test on the initiation and propagation of FFC. We highlight the concurrent development of FFC and cathodic blistering, which includes the change between the cathodic and anodic delamination front, during an accelerated aging procedure that cycles between saturated humidity and dry stages. The role of the presence of aggressive contaminants (Cl and ) is discussed. According to our findings, cyclic aging tests seem to better stick to the failure mechanism occurring during in-field exposure, particularly due to the wet/dry cycles.

Several corrosion mechanisms could be involved in the degradation of organic-coated steel structures exposed outdoors. The stress factors present in the service environment trigger the initiation and development of phenomena such as cathodic delamination1  (CD), flash rusting,2  or filiform corrosion3  (FFC). The development of electrochemical activity underneath the protective layer can lead to different mechanisms involving different driving forces and delamination effects. Deep knowledge of the possible phenomena is crucial to the design of suitable mitigation strategies to prolong the structure’s service life. An expanding cathodic front characterizes blister formation due to the coating delamination derived from the alkaline environment produced by oxygen reduction.4  On the other hand, when FFC propagates as a thread-like unidirectional path, the leading edge is recognized to be anodic, and iron oxidation takes place. Among the main parameters involved in atmospheric corrosion on painted metals, relative humidity (RH), contaminants presence, and solar radiation are the most critical.5  However, the type of stress is crucial in determining the failure mode, and the level/concentration/intensity of these environmental parameters are demonstrated to be decisive for the mechanism.6-8  The effect is particularly noticeable in the case of the humidity experienced by the protective system. The RH was recognized as the most critical parameter for a long time. Accordingly, several RH thresholds for FFC propagation have been reported in the literature. Van Loo, et al.,9  observed that no FFC can be detected below 65% RH. For higher RH values, the morphology of the growing filaments changes in thicker threads until 95%, where blisters mostly cover the surface. Following the findings of Van Loo, et al., which date back to 1953, FFC initiation and propagation conditions were further refined.10-12  However, despite the RH range of 65% to 95% is probably the most frequently faced in outdoor exposure,5  the evaluation of the durability of organic-coated steel in such conditions is commonly neglected. Due to the need to achieve significant degradation extent quickly, many accelerated laboratory procedures have been designed and standardized. Neutral salt spray test (NSST)13  (ASTM B117) and UV radiation test14  are the most adopted accelerated methods in the last decades by industries and academics to evaluate the behavior of painted steel. However, the neutral salt spray creates very harsh saline fog (100% RH), and in such conditions, the delamination is driven by cathodic blistering and the occurrence of FFC has not been reported.2,15-16  More recently, cyclic tests gained attention from industry and academia, as they are claimed to simulate outdoor exposure better. Among them, the prohesion test is claimed to simulate an industrial environment introducing chlorides and sulfates and creating a slightly acidified fog. The wet stage is cycled with a dry stage, where the temperature is increased from Troom to 35°C.17  In the transient period between the two extreme levels, RH falls into the range in which FFC could occur. In this study, we investigated the effect of the prohesion test on the rust creep of artificially damaged acrylic-coated carbon steel plates. The phenomenology of rust creep corrosion in such cyclic conditions was discussed, and the evolution of the paint delamination mechanism was related to the environment of the accelerated cabinet test. In fact, our tests investigate some potential concurrency between CD and FFC to induce paint delamination. To the best of our knowledge, this is the first reported description of the concurrent delamination mechanism for painted steel exposed to the cyclic accelerated cabinet test. This study aims to support the scientific community by providing a better understanding of the delamination mechanism occurring during exposure of painted steel to cyclic accelerated wet/dry tests.

Carbon steel panels (Q-Panel R-36) were prepared by pickling (2 M HCl for 20 min) and coated with a commercial two-component acrylic-based clearcoat (supplied by Palini Vernici, Lovere BG, Italy). After 1 h of oven curing at 60°C, the dry thickness of the coating was measured to be ca. 75 µm. A longitudinal scratch of 1 mm width and 50 mm length was introduced in the coating to the metal substrate with a standardized tool (Elcometer 1538 DIN Scratching Tool, Manchester, United Kingdom), and the edges of the samples were covered with tape to prevent early cut-edge damage. Consequently, the samples were placed in cyclic condition following the ASTM G85 standard17  for 1,000 h, and the corrosion evolution was monitored periodically by optical imaging. The wet stage of the cabinet test (Ascott-Analytical, Livonia, MI, USA) is based on a solution of 0.05 wt% NaCl and 0.35 wt% ammonium sulfate. Alternatively, 1 h of fog is followed by 1 h of drying conditions at 35°C in which the RH decreases in time down to 40%. Similar samples were loaded with 1 wt% powdery phenolphthalein-based pH indicator (Merck, Darmstadt, Germany) to detect the occurrence of oxygen reduction reaction at the metal/paint interface. In fact, phenolphthalein turns from colorless to purple when exposed to a pH higher than 9.18  Thus, phenolphthalein is an indicator of the cathodic areas (where oxygen reduction occurs: 2H2O + O2 + 4e → 4OH) underneath the organic coating.19-20  The FFC corrosion products deposited at the metal/paint interface were analyzed at the end of the test through an energy dispersive x-ray spectroscopy (EDXS) analysis was performed on the metal substrate after the coating removal. For this purpose, a scanning electron microscopy (SEM) JEOL IT 300 (Tokyo, Japan) apparatus has been used. Moreover, replicas were aged in a neutral salt spray chamber13  (100% RH and 35°C) and a humidostatic chamber (80% RH and 40°C)21  to compare the accelerated procedure outcomes.

After 24 h of the prohesion test, the artificial defect introduced in the coating is fully covered by iron corrosion products derived from the first general corrosion of the metal directly exposed to the aggressive atmosphere (Figure 1[a]). Next to the scribe, the first appearance of corrosion activity underneath the organic layer is demonstrated by round-shaped green to yellow iron-based compounds, acknowledged as green rust.22-23  Thanks to the phenolphthalein-containing varnish, the early stage of blistering can be attributed to the cathodic activity, as the delamination front of the bulge was markedly purple colored (Figures 1[a] and [b]). Therefore, the pH value in correspondence with the edges of the blisters at the initiation stage is in the 9 to 10 range. For prolonged exposure time in the cyclic cabinet, widespread cathodic activity in the proximity of the corroded area was further observed after 50 h (Figure 2[a]). These cathodic areas develop with exposure time: after 150 h (Figure 2[b]), the extension of the purple-colored areas increases. At this time, the initiation of FFC threads is evident. FFC filaments seem to nucleate from the existing round blisters and remain confined in the back part of the corroded area, closer to the scratch. Considering that the color change of phenolphthalein is reversible, the disappearance of the purple halo around the leading front of delamination once FFC propagation has initiated suggests a switch from CD to anodic undermining. A relationship between the filament activity and some outer cathodic spots may be present and is already hypothesized in the literature.3,11  Such a model assumes the formation of a cathodic area just in front of the propagating filament electrochemically coupled with the anodic site placed in the head of the thread. The growth in this way should follow a stepwise path in which the filament periodically embeds the cathodic outpost and rests until a fresh one has formed one length ahead. However, this model has not been experimentally verified yet but was disproved by several authors in the past years.8,12  In fact, observing the spatial distribution of the colored spots with respect to the position and growth direction of the filament in Figure 2(b), the aforementioned mechanism seems not to match the experimental observations. The propagation of threads was verified to be linear and did not follow the cathodic outposts randomly distributed over the surrounding area. Concerning filament nucleation, Figure 3 reports the morphological evolution of the corrosion process from 70 h (Figure 3[a]) to 160 h (Figure 3[b]) of exposure in the prohesion chamber. The initial round-shaped blisters close to the scratch start to experience the oxygen availability gradient between the furthest and the nearest region from the defect. In this way, the FFC propagation acquires a specific directionality which causes the brown corrosion products to deposit in the rear part, as visible in Figure 3(a). At this stage, the border between the green rust and the furtherly oxidized brown compounds is already visible, but its shape is not univocally defined yet. With time, the active head of the filament (delamination front) develops the “V-shaped” morphology in the rear part, as highlighted in Figure 3(c). Both the morphology and the propagation rate are demonstrated to be ruled by the oxygen availability underneath the coating at the active sites.7,24  The darker striation marks in the brown tail visible in Figure 3(c) are likely due to the dry and wet stages in the accelerated weathering cabinet. These alternating conditions are also at the basis of the evolution of the overall degradation of the protective system during the 1,000 h of the test. A co-existence between FFC and blistering was observed with filaments that have nucleated from the early-formed round blisters around the coating defect. Figure 4 shows a sample's time-lapse evolution during the entire prohesion test duration. It should be noted that during prolonged exposure in the test cabinet, blisters manifest at the edge of the delamination front, thereby countering cohesive forces and promoting the delamination of the paint. Both types of delamination fronts progress over time, with some filaments discovered to be embedded within the blisters. On the contrary, fresh fracture fronts initiate from the round-shaped blisters, ultimately leading to a broad detachment of the coating, as depicted in Figure 4(f). In this situation, the delamination interface of a cathodic disbonded coating serves as the site, where the cathodic process occurs, with the anodic reaction taking place at the location of the coating defect. Conversely, at the tip of the filaments, the anodic reaction is juxtaposed with the cathodic reaction of a FFC front.7,25  The potential for an interaction between the active zones of these two simultaneous mechanisms cannot be ruled out.
FIGURE 1.

Optical microscope images were collected on an acrylic-coated steel sample at increasing magnifications after 24 h of prohesion test with cyclic RH. (a)  The early stage of FFC, where green-colored blisters are forming along the longitudinal scratch, and (b) and (c) highlight the cathodic front around the initiation bulges thanks to the phenolphthalein as a pH indicator loaded in the varnish (purple at pH > 10).

FIGURE 1.

Optical microscope images were collected on an acrylic-coated steel sample at increasing magnifications after 24 h of prohesion test with cyclic RH. (a)  The early stage of FFC, where green-colored blisters are forming along the longitudinal scratch, and (b) and (c) highlight the cathodic front around the initiation bulges thanks to the phenolphthalein as a pH indicator loaded in the varnish (purple at pH > 10).

Close modal
FIGURE 2.

(a) Optical images of the initiation stage after 50 h of the prohesion test, where the cathodic front is highlighted by the purple coloration of the phenolphthalein pH indicator loaded in the varnish (purple at pH > 10). (b) After the filament shape development at 150 h of prohesion aging the alkaline front is no longer observed.

FIGURE 2.

(a) Optical images of the initiation stage after 50 h of the prohesion test, where the cathodic front is highlighted by the purple coloration of the phenolphthalein pH indicator loaded in the varnish (purple at pH > 10). (b) After the filament shape development at 150 h of prohesion aging the alkaline front is no longer observed.

Close modal
FIGURE 3.

Filament nucleation along the coating longitudinal defect after (a) 70 h and (b) 160 h of prohesion aging test. In (b), the competition between CD and FFC is highlighted. The inset (c) shows the darker “flow lines” in the early formation of the V-shaped leading part of the filament.

FIGURE 3.

Filament nucleation along the coating longitudinal defect after (a) 70 h and (b) 160 h of prohesion aging test. In (b), the competition between CD and FFC is highlighted. The inset (c) shows the darker “flow lines” in the early formation of the V-shaped leading part of the filament.

Close modal
FIGURE 4.

Time lapse of corrosion evolution of painted carbon samples during prohesion test. Increasing aging time from (a) 24 h to (f) 1,000 h cycling between dry and wet stages, both having 1 h duration.

FIGURE 4.

Time lapse of corrosion evolution of painted carbon samples during prohesion test. Increasing aging time from (a) 24 h to (f) 1,000 h cycling between dry and wet stages, both having 1 h duration.

Close modal
At the end of the test, the corrosion products deposited on the steel substrate by the filament growth were analyzed by SEM-EDXS (after the coating removal) to determine their chemical composition. In Figure 5, the EDXS maps collected over the head of the filaments are reported. The tip appearance under the coating is characterized by a sort of crater, where the metal’s anodic dissolution occurs, followed by the deposition of corrosion products. Furthermore, Figures 5(b) and (f) testify to the continuous transport of contaminants anions (Cl and ) from the defect in direct contact with the testing solution through the tail up to the edge of the head. On the other hand, contaminants cations, such as Na2+ are not present along the filament (Figure 5[e]). The role of contaminants in the mechanism of FFC was investigated by Williams and McMurray8  and they reported their crucial contribution in the initiation stage. The transport of chloride and sulfate anions toward the filament head during the propagation is driven by the tendency to maintain electroneutrality while iron cations are released from the surface in the anodic sites. FFC propagation is dependent on the contaminant anions, while cations are not relevant because they play a role only in the first stage of delamination because they are not transported by the filament.8,26  Accordingly, the environment directly influences the redox reaction at the interface and also determines the composition of the liquid and jelly electrolytes at the interface. This fact is reflected in (i) the filament thickness,26  (ii) the corrosion products composition, and (iii) the anodic dissolution severity. Indeed, the strong presence of chlorides combined with the acidic pH in the leading head10,23  promotes a harsh environment at the metal/paint interface in which the coating adhesion is undermined, and the iron dissolution abounds.
FIGURE 5.

SEM-EDS maps of the corrosion product deposited under the coating in the leading part of the filament. The maps were collected after coating removal for an acrylic-coated steel sample aged 300 h through the cyclic prohesion test.

FIGURE 5.

SEM-EDS maps of the corrosion product deposited under the coating in the leading part of the filament. The maps were collected after coating removal for an acrylic-coated steel sample aged 300 h through the cyclic prohesion test.

Close modal
Only cyclic tests are able to highlight the interaction between FFC and CD thanks to the several levels of humidity experienced by the coated steel. Other static aging procedures such as NSST13  and humidostatic chamber21  trigger exclusively the delamination mechanism favored in that maintained level of RH. As a matter of fact, in saline-saturated fog, only CD is developed (Figure 6) because the mechanism of FFC initiation does not take place. The anodic front of FFC could appear after a prior CD phase, thanks to the depletion of the group (I) salts cations coming from the contaminant species needed to balance the negative charge produced by oxygen reduction (2H2O + O2 + 4e → 4OH) at the leading front.8  During NSST the cations supply (Na2+) is guaranteed constant by the saturated fog and the transition from CD to FFC cannot take place. On the other hand, an aging setup that, after a NaCl contamination, maintains the RH at 80% because of a neat FFC formation, as reported in Figure 7. In this case, the FFC initiation mechanism described by Williams and McMurray8  is in the condition of being fulfilled, and the anodic undermining is the main delamination event detected. The unique aspect presented in this study lies in the observation that under cyclic humidity conditions, FFC and CD develop simultaneously. CD not only serves as an initial stage for FFC nucleation but also manifests as an independent delamination event. The prominence of the latter phenomenon on FFC varies in the filament based on the RH level. On the contrary, when dealing with the cathodic activity as the precursor of the filament creation, it spontaneously stops regardless of the surrounding RH and is substituted by the anodic undermining process. Conversely, when considering the cathodic activity as the antecedent to filament formation, it ceases spontaneously without regard for the ambient RH and is supplanted by the anodic corrosion process. Other investigations utilizing the ASTM G85 standard17  for assessing organic-coated steel did not reveal any instances of FFC occurrence.27-28 
FIGURE 6.

Time lapse of corrosion evolution of painted carbon samples during NSST. Increasing aging time: (a) 100 h, (b) 300 h, and (c) 500 h.

FIGURE 6.

Time lapse of corrosion evolution of painted carbon samples during NSST. Increasing aging time: (a) 100 h, (b) 300 h, and (c) 500 h.

Close modal
FIGURE 7.

Time lapse of corrosion evolution of painted carbon samples during humidostatic test (80% RH at 40°C). Increasing aging time: (a) 100 h, (b) 250 h, and (c) 500 h.

FIGURE 7.

Time lapse of corrosion evolution of painted carbon samples during humidostatic test (80% RH at 40°C). Increasing aging time: (a) 100 h, (b) 250 h, and (c) 500 h.

Close modal

The prohesion test has helped us to simulate an alternating humidity that well matches the natural exposure which normally deals with rain, dew, and dry sunny conditions. A concurrent evolution of CD and FFC has been described for acrylic-coated steel in a cyclic accelerated cabinet test. The choice of an aging test humidity levels in a quite broad range seemed essential to induce the mechanism. As the two main failure mechanisms related to corrosion are characterized by different leading electrochemical activities (cathodic vs. anodic), the durability of the painted structures is suggested to be assessed considering both because the mitigation effect might significantly vary. The bipolar nature of the delamination front could be a challenge for the design of suitable mitigation strategies because most of the inhibition effects are efficient mostly in one single environment.

  • Concurrent development of CD and anodic front propagation of the filaments was found in the cyclic accelerated cabinet test (prohesion).

  • The predominance of one mechanism over the other turned out to be determined by the level of RH experienced by the panels in the chamber.

  • Aggressive contaminants ions have been detected to be transported from the defect along the growing filament by their presence in the leading head at the end of the aging test.

Trade name.

1.
Funke
W.
,
Ind. Eng. Chem. Prod. Res. Dev.
24
(
1985
):
p
.
343
347
.
2.
Leidheiser
H.
,
Corrosion
38
(
1982
):
p
.
374
383
.
3.
Bautista
A.
,
Prog. Org. Coat.
28
(
1996
):
p
.
49
58
.
4.
Lyon
S.B.
,
Bingham
R.
,
Mills
D.J.
,
Prog. Org. Coat.
102
(
2017
):
p
.
2
7
.
5.
Cristoforetti
A.
,
Rossi
S.
,
Deflorian
F.
,
Fedel
M.
,
Mater. Corros.
74
(
2023
):
p.
1429
1438
.
6.
LeBozec
N.
,
Persson
D.
,
Thierry
D.
,
Axelsen
S.B.
,
Corrosion
60
(
2004
):
p
.
584
593
.
7.
Cristoforetti
A.
,
Rossi
S.
,
Deflorian
F.
,
Fedel
M.
,
Prog. Org. Coat.
179
(
2023
):
p
.
107525
.
8.
Williams
G.
,
McMurray
H.N.
,
Electrochem. Commun.
5
(
2003
):
p
.
871
877
.
9.
Van Loo
M.
,
Laiderman
D.D.
,
Bruhn
R.R.
,
Corrosion
9
(
1953
):
p
.
277
283
.
10.
Slabaugh
W.H.
,
Grotheer
M.
,
Ind. Eng. Chem.
46
(
1954
):
p
.
1014
1016
.
11.
Funke
W.
,
Prog. Org. Coat.
9
(
1981
):
p
.
29
46
.
12.
Ruggeri
R.T.
,
Beck
T.R.
,
Corrosion
39
(
1983
):
p
.
452
465
.
13.
ASTM B117-19
, “
Standard Practice for Operating Salt Spray (Fog) Apparatus
” (
West Conshohocken, PA
:
ASTM International
,
2019
).
14.
ASTM G154-23
, “
Standard Practice for Operating Fluorescent Ultraviolet (UV) Lamp Apparatus for Exposure of Materials
” (
West Conshohocken, PA
:
ASTM International
,
2023
).
15.
Leidheiser
H.
,
Wang
W.
,
Igetoft
L.
,
Prog. Org. Coat.
11
(
1983
):
p
.
19
40
.
16.
Watts
J.F.
,
J. Adhes.
31
(
1989
):
p
.
73
85
.
17.
ASTM G85-19
, “
Standard Practice for Modified Salt Spray (Fog) Testing
” (
West Conshohocken, PA
:
ASTM International
,
2019
).
18.
Xiao
H.
,
Wang
Y.
,
Gu
L.
,
Feng
Z.
,
Lei
B.
,
Zhu
L.
,
Guo
H.
,
Meng
G.
,
Prog. Org. Coat.
177
(
2023
):
p
.
107418
.
19.
Zhang
J.
,
Frankel
G.S.
,
Corrosion
55
(
1999
):
p
.
957
967
.
20.
Hoch
G.
,
Localized Corrosion, NACE-3
,
eds.
Staehle
R.W.
,
Brown
B.F.
,
Kruger
J.
,
Agrawal
A.
(
Houston, TX
:
NACE
,
1974
).
21.
ASTM D2803-09
, “
Standard Guide for Testing Filiform Corrosion Resistance of Organic Coatings on Metal
” (
West Conshohocken, PA
:
ASTM International
,
2020
), https://compass.astm.org/document/?contentCode=ASTM%7CD2803-09R20%7Cen-US&proxycl=https%3A%2F%2Fsecure.astm.org&fromLogin=true.
22.
Génin
J.M.R.
,
Refait
P.
,
Simon
L.
,
Drissi
S.H.
,
Hyperfine Interact.
111
(
1998
):
p
.
313
318
.
23.
Abdelmoula
M.
,
Refait
P.
,
Drissi
S.H.
,
Mihe
J.P.
,
Génin
J.M.R.
,
Corros. Sci.
38
(
1996
):
p
.
623
633
.
24.
Watson
T.M.
,
Coleman
A.J.
,
Williams
G.
,
McMurray
H.N.
,
Corros. Sci.
89
(
2014
):
p
.
46
58
.
25.
Leblanc
P.P.
,
Frankel
G.S.
,
J. Electrochem. Soc.
151
(
2004
):
p
.
B105
B113
.
26.
Koehler
E.L.
,
Corrosion
33
(
1977
):
p
.
209
217
.
27.
Tayler
M.L.
,
Blanton
M.
,
Konecki
C.
,
Rawlins
J.
,
Scully
J.R.
,
Corrosion
71
(
2014
):
p
.
71
91
.
28.
Tayler
M.L.
,
Blanton
M.
,
Konecki
C.
,
Rawlins
J.
,
Scully
J.R.
,
Corrosion
71
(
2014
):
p
.
326
342
.