Assessing the susceptibility of a substrate to filiform corrosion without applying a paint layer

(1)

https://doi.org/10.4224/23002656

Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la

première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca.

Questions? Contact the NRC Publications Archive team at

PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca. If you wish to email the authors directly, please see the first page of the publication for their contact information.

https://publications-cnrc.canada.ca/fra/droits

L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.

READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE.

https://nrc-publications.canada.ca/eng/copyright

NRC Publications Archive Record / Notice des Archives des publications du CNRC :

https://nrc-publications.canada.ca/eng/view/object/?id=6253cca5-15b1-4799-9f2d-ec1d8f5e8dd8 https://publications-cnrc.canada.ca/fra/voir/objet/?id=6253cca5-15b1-4799-9f2d-ec1d8f5e8dd8

NRC Publications Archive

Archives des publications du CNRC

For the publisher’s version, please access the DOI link below./ Pour consulter la version de l’éditeur, utilisez le lien DOI ci-dessous.

Access and use of this website and the material on it are subject to the Terms and Conditions set forth at

Assessing the susceptibility of a substrate to filiform corrosion without applying a paint layer

(2)

Assessing the susceptibility of a

substrate to filiform corrosion

without applying a paint layer

Alban Morel, Ph.D.

Research Associate

– Corrosion and Durability

National Research Council Canada

(3)

2

(4)

ALTec industrial R&D group

– From 2015 to now

20

PROJECTS

INVESTED

6M$ +

TO THE BENEFIT

OF THE MEMBERS

0

Forming:

Hot sta pi g, Casti g, E trudi g…

Assembling:

Weldi g GMAW, laser, F“W… , Adhesi es…

Corrosion and durability:

SCC, galvanic corrosion, filiform corrosion, in-ser i e e posure o ditio o itori g…

(5)

ALTec

partners

(6)

ALTec Members

(20 industrial members as of August 2017)

(7)

What is filiform corrosion?

6 -

_{Filifor orrosio is a t pe of lo alized orrosio hi h affe ts pai ted etals that gi es rise to orrosio}

produ ts of a fila e tous appeara e u der oati g.

- Occurs in humid but not saturated atmospheres, with relative humidity (RH) between 70 and 95%

A. Bautista / Progress in Organic Coatings 28 (1996) 49-58

J.L. Delplancke / Progress in Organic Coatings 43 (2001) 64-74

(8)

7 • Due to lightweighting efforts, there is a growing

use of high yield strength alloys with noble elements

 Noble intermetallic: active for oxygen reduction reaction causing local galvanic coupling

• It is possible to develop alloys that are less sensitive to filiform corrosion by closely controlling there

composition. However, the worldwide supply in raw materials (bauxite) and recycling can make it difficult and

extremely expensive to control impurities that may cause filiform corrosion.

• In order to mitigate corrosion, one will use conversion coatings which are expensive:

 Following the Cr(VI) ban, which new product is suitable for your alloy?

 To which extent can we minimized this step?

 Can a fast test be used in production for quality control?

(9)

8 Can we, with a fast electrochemical approach, predict the susceptibility of an alloy and/or an alloy/conversion

coating couple to filiform conversion coating without the need for paint application?

Reduction of direct and indirect costs related to filiform corrosion problems by:

• diminishing prototyping cost

• reducing time to market

• supplying a complementary tool for quality control

• decreasing risks associated with high production volume.

TARGETED BENEFITS FOR ALTec MEMBERS

(10)

Phase 1: Sample preparation and establishment of a

baseline

A wide span test matrix:

-

4 different substrates expected to corrode at different rates:

-

5 different surface chemical preparation (conversion coatings (CC))

-

Full automotive paint system applied by ACT test panel

9

AAn°1 grinded AAn°1 as laminated AAn°2 as laminated AAn°3 as laminated

no-conversion coating - (alkaline clean, ecoat, prime, base, clear) x x x x

Zinc Phosphate with sealer plus ecoat, prime, base, clear x x x x

Zinc Phosphae without sealer plus ecoat, prime, base, clear x x x x

Zirconium based plus ecoat, prime, base, clear

x x x x

(11)

Phase 1: Sample preparation and establishment of a

baseline

10

4 ”

2 ”

Rolling direction

Standard laboratory test that is known to correlate road exposure

results

(12)

Phase 1: Sample preparation and establishment of a

baseline

Optical characterization of the corrosion extent:

- Two parameters were extracted:

- The corroded area

- The length of the filiform

Example of AA6xx2 grinded sample with zirconium based conversion coating after 3 weeks in ASTM G85-A2

11

Left Scribe Right Scribe Aera (mm2₎

65,23 73,40

Mean length (mm)

(13)

Phase 1: Sample preparation and establishment of a

baseline

Optical characterization of the corrosion extent:

Very large area of paint lifting: looks more like blisters or large interconnected filaments.

Large corroded area with extremely fine filaments parallel to the rolling direction.

Various corrosion morphologies were observed

(14)

Phase 1: Sample preparation and establishment of a

baseline

Optical characterization of the corrosion extent:

13

no con version Zn Pho sphate with se aler Zn Pho sphate withou t sealer Zirconium based _Chromate 0 200 400 600 800 1000 1200 no conve rsion Zn Pho sphate with sea ler Zn Pho sphate without sealer Zirconiu m base d Chroma te 0 20 40 60 80 100 AAn°1 grinded AAn°1 as laminated AAn°2 as laminated AAn°3 as laminated m e a n co rr o d e d a re a ( m m 2 / 2 in ch e s scr ibe ) no conv ersion Zn Phosp hate with sealer Zn Phosp hate with out sea ler Zirconiu m base d Chroma te 0 1 2 3 4 5 6 7 8 9 10 11 12 13 _{AAn°1 grinded} AAn°1 as laminated AAn°2 as laminated AAn°3 as laminated me a n f ila me n t le n g th (mm)

(15)

Phase 2: Electrochemical testing approach

Filiform corrosion is a process in two steps:

- First step is the corrosion of the substrate at a defect of the paint which causes the paint to lift in

the near by regions.

- Once the paint is lifted, this gives rise to conditions for a differential aeration corrosion (same as

crevice corrosion, pitting corrosion…) that will cause the filiform to grow.

14

H. J . W. LENDERINK, PhD thesis, Delft University of Technology, The Netherlands, 1995.

Initiation through a slight corrosion of the aluminum substrate through the penetration of a small amount of water and salt at a defect of the organic coating.

(16)

Phase 2: Electrochemical testing approach

Defining electrochemical parameters that are susceptible to correlate

the corrosion observed on the painted samples:

As exposed previously, the filiform corrosion can be seen as a two step process:

• The filament propagation process thought to be ruled by the differential oxygenation process can be

modeled by the “i_couple” parameter.

• The initiation process thought to be ruled by corrosion of the substrate at the scratch.

This initiation step can influence the number of filaments starting but also how soon the filaments

start, hence influencing as well the length of the filament after a given time.

-

Two techniques were investigated to model the initiation by corrosion:

- Resistance to polarization measurements

- Electrochemical noise measurements

(17)

Phase 2: Electrochemical testing approach

Resistance to polarization measurements to evaluate the

susceptibility of substrates to corrosion initiation

16

-0.72 -0.70 -0.68 -0.66 -0.64 -0.62 1E-7 1E-6 1E-5 1E-4 1E-3 i (A/ cm 2 ) E vs. SCE (V) AAn°1 as laminated

• Resistance to polarization is inversely proportional to the corrosion rate => Higher is the resistance to polarization, lower is the corrosion current

y = 1E+24x5_{- 2E+19x}4_{+ 6E+13x}3_{- 1E+08x}2_{+ 969.08x - 0.7362}

R² = 0.994 -7.48E-01 -7.46E-01 -7.44E-01 -7.42E-01 -7.40E-01 -7.38E-01 -7.36E-01 -7.34E-01 -7.32E-01 -7.30E-01

-6.00E-06 -4.00E-06 -2.00E-06 0.00E+00 2.00E-06 4.00E-06 6.00E-06 8.00E-06

E vs. S CE i (A/cm2₎ AAn°1 As laminated

R

_p

=969

Ω.cm

2

R_p is the slope of the potential vs. current curve when measured current tends to 0.

(18)

Electrochemical noise measurements to evaluate the

susceptibility of substrates to corrosion initiation

17 Fluctuations due to statistically distributed corrosion process

Al Al3+ O₂ Al Al3+ O₂ 2OH- 2OH - Al Al3+ O₂ Al Al3+ O₂ 2OH- 2OH - Time 1 Time 2 e- e- _e- e- e- _e_- _e_- e- Slightly more anodic Slightly more cathodic Slightly more

anodic Slightly more _cathodic

(19)

Electrochemical noise measurements to evaluate the

susceptibility of substrates to corrosion initiation

18 Fluctuations due to statistically distributed corrosion process

Al Al3+ O₂ Al Al3+ O₂ 2OH- 2OH - Al Al3+ O₂ Al Al3+ O₂ 2OH- 2OH - Time 1 Time 2 e- e- e- _e- e- Slightly more anodic Slightly more cathodic Slightly more

anodic Slightly more _cathodic

Phase 2: Electrochemical testing approach

Negative current

e-

Positive

(20)

19

Statistical analysis of the following parameters • IRms = |Imean|

• σV : Voltage noise standard deviation • σI : Current noise standard deviation

• Rn = σV/ σI : Noise resistance (same physical meaning as Rp)

• PI = σI / IRms : Pitting index (the higher it is the more localized corrosion is likely).

AAn°2 no CC AAn°2 Zirconium based CC AAn°2 Zn Ph w/o Sealer

AAn°2 Zn Ph w/ Sealer

Phase 2: Electrochemical testing approach

Electrochemical noise measurements to evaluate the

susceptibility of substrates to corrosion initiation

(21)

20 Statistical analysis on mean values: Pearson correlation coefficient (a tool to explore data at a glance)

� =

�� ,

_{σ σ}

Suggested interpretation for r absolute

value (Evans 1996):

• 0.00 - . 9 : er eak orrelatio

• 0.20 - . 9 : eak orrelatio

• 0.40 - .59 : oderate orrelatio

• 0.60 - .79 : stro g orrelatio

• 0.80 - . : er stro g orrelatio

(22)

21

21 When looking at all 20 AA/CC combination only moderate correlation

observed between i_couple and corr_area

i_couple is a parameter designed to firstly correlate mean_length

before corr_area. i_couple should not correlate corr_area better than

mean_length. This means that the moderate correlation observed with

corr_area has no physical sense when looking at all the data.

Statistical analysis on mean values: Pearson correlation coefficient

1 AAn°1 grinded 2 AAn°1 as laminated 3 AAn°2 as laminated 4 AAn°3 as laminated A No CC B Zn Ph w/ sealer C Zn Ph w/o sealer D Zr based E Cr based

(23)

22

22 When looking at subsets of data, very strong correlations of i_couple with

mean_length and corr_area are observed for:

• samples with no conversion coating

• samples with Zn Phosphate + sealer CC

• samples prepared with grinded AAn°1

Statistical analysis on mean values: Pearson correlation coefficient

(24)

23

0 1 2 3 4 -100 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1A,1B,1C,1D1,1E corr_area (mm 2 / 2" scribe)

i_couple (mA/cm2 of anode)

0 1 2 3 4 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1B,2B,3B,4B corr_area (mm 2 / 2" scribe)

-1 0 1 2 3 4 5 6 0 100 200 300 400 500 600 700 1A,2A,3A,4A corr_area (mm 2 / 2" scribe)

0 1 2 3 4 5 6 1 2 3 4 5 6

Mean filiform leng

th (mm)

Ver good orrelatio et ee i_couple a d corr_area suggests that grinded AAn°1 is highly susceptible to the propagation process.

Statistical analysis on mean values: Pearson correlation coefficient

(25)

24

R_p seams to play a role in the susceptibility to filiform corrosion of some subsets only.

AAn°2 samples: 1/Rp shows a strong correlation with the corrosion extent observed on the painted samples. This suggests that the susceptibility to filiform corrosion of these samples is mainly depend on the initiation step.

Statistical analysis on mean values: Pearson correlation coefficient

(26)

25

Combined effect seams to be also observed in the case of samples without conversion coatings.

While R_p seamed to correlate quite well the corrosion extent, taking into account i_couple increases the correlation factor.

A two steps process means that there should be a combined effect of

the two parameters:

• if a sa ple is er sus epti le to the propagatio phe o e o ut er

resistant to the initiation corrosion, this one will show few filaments but these will be long,

• if a sa ple is sus epti le to orrosio lo Rp value), but the propagation phe o e o is ot hat’s dri i g the orrosio , the orrosio ill prese t either no corrosion, or many filaments but small in length,

• If a sa ple z is sus epti le to oth t pe of orrosio , this o e ill sho a long filaments.

Statistical analysis on mean values: Pearson correlation coefficient

(27)

26

All Rights Reserved CNRC -50.0µ 0.0 50.0µ 100.0µ 150.0µ 200.0µ 250.0µ 0 100 200 300 400 500 600 700 1A,2A,3A,4A Corr_area (mm 2 / 2" scribe) i_couple/R p (V -1 ) 0.0000 0.0005 0.0010 0.0015 0.0020 -50 0 50 100 150 200 250 300 350 3A,3B,3C,3D,3E Corr_area (mm 2 / 2" scribe ) i_couple/R_p (V-1)

More points are needed to confirm the relationship Looking at i_couple/Rp is detrimental in the

case of AAn°1 grinded

Statistical analysis on mean values: Pearson correlation coefficient

(28)

27

AAn° : Ver good orrelatio et ee IRms a d

corr_area , o fir i g o e ore the i flue e of the initiation process on the samples susceptibility to

filiform corrosion.

Zr ased CC: good orrelatio is fou d et ee IRms a d corr_area , suggesting that the main protection mechanism provided by the CC is through protection against general corrosion.

0.0 50.0µ 100.0µ 150.0µ -50 0 50 100 150 200 250 300 350 3A,3B,3C,3D,3E Corr_area (mm 2 / 2" scribe) IRms (A/cm2)

Statistical analysis on mean values: Pearson correlation coefficient

(29)

28 Other correlation trends up to now not

found with other parameters

(30)

When looking at the all whole test matrix involving different alloys and different conversion coating, no

correlation was found between the selected electrochemical parameters and the corrosion observed.

However, when keeping either the aluminum alloy or the CC as a constant, correlations arise:

-

“IRms” correlates perfectly the mean corroded area observed for the AAn°2 samples.

- It would be interesting to confirm the trend with various alloys of the series.

-

“i_couple” correlates the mean length of filaments observed for samples prepared without

conversion coating, and the

“i_couple/R

_p

” correlates quite well the mean corroded area.

- If confirmed with more alloys, this could potentially be a tool for the alloy developers.

-

“i_couple” correlates the corroded area observed for the grinded AAn°1.

- If this trend was to be confirmed with other grinded alloys, this could be a selection tool to

evaluate the worst case scenario as it is recognized that corrosion filiform is more important in

places where alterations were done (repairs, machining

…).

As a general rule, each correlations observed should be confirmed with more data points

29

(31)

Defining threshold values

For such a wide span test matrix, can we define threshold values in order to rule out the most at risk

Alloy/CC combinations at an early stage of the selecting process and that way reduce the number of

combination to paint and test in environmental chamber?

- Can we find parameters for which the value is systematically high (or low) when corr_area is high?

-

Through “Rattle” package in R software, one can generate different decision trees.

- Considering the small amount of data points (20 average values per parameters) all data was used to built

and train these decision trees.

(32)

31

no con version Zn Pho sphate with se aler Zn Pho sphate withou t sealer Zirconium based _Chromate 0 200 400 600 800 1000 1200 no conve rsion Zn Pho sphate with sea ler Zn Pho sphate without sealer Zirconiu m base d Chroma te 0 20 40 60 80 100 AAn°1 grinded AAn°1 as laminated AAn°2 as laminated AAn°3 as laminated m e a n co rr o d e d a re a ( m m 2 / 2 in ch e s scr ibe )

a

c

b

(33)

no con version Zn Pho sphate with se aler Zn Pho sphate withou t sealer Zirconium based _Chromate 0 200 400 600 800 1000 1200 no conve rsion Zn Pho sphate with sea ler Zn Pho sphate without sealer Zirconiu m base d Chroma te 0 20 40 60 80 100 AAn°1 grinded AAn°1 as laminated AAn°2 as laminated AAn°3 as laminated m e a n co rr o d e d a re a ( m m 2 / 2 in ch e s scr ibe )

a

c

b

32

(34)

Summary

33 When considering subsets, correlations were found between electrochemical

parameters and the corrosion observed after the standard laboratory test

This electrochemical approach showed to be promising when studying test

matrix with limited changes in variables (constant CC or constant alloy).

- Hence it could potentially find its application in:

- developing aluminum alloy composition resistant to filiform corrosion

- selecting conversion coating for a set alloy (seams especially suitable for

(35)

34 Looking at decision trees built to predict the corroded area for all 20

combinations, threshold values for some electrochemical parameters were

determined.

When ruling out this branch, 3 of the worst

combinations are discarded without the need for environmental chamber test.

When ruling out this branch, 2 more combinations presenting important corrosion are discarded

without the need for environmental chamber test.

25% are discarded without the need for environmental chamber test

Considering the small amount of data points (20 average values per parameters) all data was used to built and train the decision tree. More data point would be needed to test and confirm the tree (especially to confirm that the branch 5 can be ruled out without the risk of discarding valuable options).

If I-couple and IRms are high, chances are high that the combination will be prone to corrosion