Corrosion mechanisms of Zn(Mg,Al) coated steel: 2. The effect of Mg and Al alloying on the formation and properties of corrosion products in different electrolytes

Corrosion Science 90 (2015) 482–490

Contents lists available at ScienceDirect

Corrosion Science journal homepage: www.elsevier.com/locate/corsci

Corrosion mechanisms of Zn(Mg,Al) coated steel: 2. The effect of Mg and Al alloying on the formation and properties of corrosion products in different electrolytes M. Salgueiro Azevedo a,b, C. Allély b, K. Ogle a, P. Volovitch a,⇑ a b

Institute de Recherche Chimie Paris, CNRS – Chimie ParisTech, 11 rue Pierre et Marie Curie, 75231 Paris Cedex 05, France ArcelorMittal Research SA, 57283 Maizières-lès-Metz, France

a r t i c l e

i n f o

Article history: Received 12 April 2014 Revised 11 July 2014 Accepted 12 July 2014 Available online 30 July 2014 Keywords: A. Zinc A. Magnesium A. Aluminum B. Polarization C. Atmospheric corrosion C. Oxygen reduction

a b s t r a c t Oxygen reduction rate on galvanized steel and ZnAlMg coated steel decreased in presence of corrosion products identified as basic zinc salts (BZS) and layered double hydroxides (LDH). Under cathodic polarization of galvanized steel the transformation BZS ? ZnO occurred, resulting in a loss of the inhibiting effect. It did not happen on ZnMgAl. Acid–base titrations of Zn2+-containing solutions with and without additives demonstrated that (1) Mg2+ ions delay the transformation BZS ? ZnO; (2) Al3+ forms LDH which remains stable even at pH > 12.5. A residual Al ‘‘skeleton’’ was observed in corrosion products formed on ZnMgAl. It may increase the compactness of the patinas. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction ZnMgAl coatings for steel are known to be more resistant against corrosion than conventional hot-dip galvanized steel in many different environments – accelerated and atmospheric corrosion tests [1–17]. Their better anti-corrosion behavior is considered to be due to the combined addition of Mg and Al in the Znbased coating that changes the microstructure and the nature of precipitated corrosion products. The atmospheric corrosion rate of galvanized steel is governed by oxygen reduction which is often limited by diffusion through a layer of corrosion products. Therefore, the nature and stability of the corrosion products may be fundamental for the corrosion mechanisms of zinc-based coatings [7,18]. Mg alloying results in the formation of intermetallic phases with Zn, usually Zn2Mg. Zn–Mg intermetallics are more active in the galvanic series than pure Zn, causing preferential corrosion. Binary and ternary eutectics are also detected in ZnMgAl coatings (Fig. 1) which are known to corrode preferentially compared to Zn dendrites [1–3]. The dissolution of the intermetallic phase could lead to the formation of Mg2+-containing corrosion products as discussed by many authors [6–8]. Some authors observed Mg2+-rich ⇑ Corresponding author. Tel.: +33 1 44 27 20 74; fax: +33 1 46 34 07 53. E-mail address: [email protected] (P. Volovitch). http://dx.doi.org/10.1016/j.corsci.2014.07.042 0010-938X/Ó 2014 Elsevier Ltd. All rights reserved.

zinc compounds (oxides [6,7] and basic salts [8]). One hypothesis proposed is that Mg2+ doping enhances the insulating character of zinc compounds [9]. A second hypothesis is that the charge transfer at grain boundaries can also be limited in presence of Mg [10]. In other works [1,5,6], an important effect was assigned to pH buffering during precipitation of Mg(OH)2. It should however be noted that insoluble corrosion products rich in Mg2+ were frequently not detected [1,5,6]. Therefore a third hypothesis has been proposed: Mg2+ stabilizes protective corrosion products of zinc such as simonkolleite (Zn5(OH)8Cl2H2O) and zinc hydroxysulfate (Zn4(OH)6SO4xH2O) against their transformation into soluble hydroxide, carbonate or sulfate complexes or less compact products [1,5,6]. This stabilization effect may be attributed to pH buffering or to consumption of the excess anions. The importance of Mg content in the alloy for the corrosion resistance has also been noted [7]. From the microstructure of ZnMgAl samples [1], it seems that Al remains at early stages on the coating in an insoluble form, and at later stages, it participates as a layered double hydroxide (LDH). The LDH species are considered to possess good barrier properties [1,8,11,13,19]. However, in recent work [14], it was demonstrated that, in humid atmosphere, the corrosion rate of ZnMgAl, for which LDH was detected, was higher than the corrosion rate of the same coating for which LDH was not detected. The absence of LDH was also noted on ZnMgAl coatings in various NaCl-corrosion tests with neutral to slightly alkaline solution pH (pH from 7 to 10) [12] or

483

M. Salgueiro Azevedo et al. / Corrosion Science 90 (2015) 482–490

a

b

Ternary eutectic

Zn lamellas with Al “rings” Zn

Zn2Mg

Zn dendrite

steel 100 nm

1 µm

-2

c 4

O2 reduction plateau

2

2 0

1

-2 -1.30

-1.20

-1.10

-1.00

-0.90

E / V vs. SCE Fig. 1. Characterisation of the uncorroded ZnMgAl coating. (a) General structure revealed by backscattered electron image in SEM; (b) high resolution image showing the detailed microstructure of the ternary eutectique; (c) polarization curves of fresh surfaces of ZnMgAl (curve 2) and reference hot dip galvanized (curve 1) coatings. The potential chosen for the test of oxygen reduction kinetics is also shown by an arrow.

under wet exposure to NaCl in ambient air [14]. A heterogeneous distribution of Al on heavily corroded cross sections of the coatings was also observed by the authors. In contrast, other authors tested some conditions in which the formation of LDH is observed or claimed from the initial stages of corrosion [2,8,11,13]. The presence of Al also modifies the stability of different Zn patinas. Al promotes simonkolleite precipitation in place of ZnO at pH 9 and modifies the morphology of the Zn-product, increasing its compactness [20]. This work is focused on the role of Mg and Al on the formation and stability of different corrosion products. The effect of ‘‘natural’’ corrosion products formed on galvanized steel and ZnMgAl-coated steel on the oxygen reduction rate is evaluated by means of electrochemistry. Titration is used to study the precipitation of ‘‘synthetic’’ corrosion products in function of presence/absence of Mg2+ and Al3+ together with Zn2+ and different anions. The data are completed with the spatial distribution of corrosion products after corrosion tests using different electrolytes. 2. Experimental 2.1. Materials The coated steel samples were supplied by ArcelorMittal. The reference sample was the conventional hot dip galvanized steel with coating thickness of 20 lm, and composition Zn–Al (0.2 wt.%). The ZnMgAl coated steel was also prepared by hot dip process with thickness of 20 lm and composition Zn–Mg (3.0 wt.%)–Al (3.7 wt.%). The mean coating composition and thickness are summarized in Table 1. ZnMgAl coating contains Zn dendrites surrounded by a ternary phase of Zn, Al and intermetallic Zn2Mg (Fig. 1a and b). The corrosion potential of ZnMgAl coating was close to the corrosion potential of the reference material (Fig. 1c). Prior to experiments the samples were degreased with

Table 1 Thickness and chemical composition of the coatings. Label

Coating

Thickness

Zn (wt.%)

Al (wt.%)

Mg (wt.%)

1

Conventional hot-dip galvanized ZnMgAl

20 lm

99.8

0.2

–

20 lm

93.3

3.7

3.0

2

ethanol, but no surface alkaline cleaning or surface finishing was done. We have chosen to keep the natural surface oxides because this particular work aimed to study the effect of the evolution of the surface oxide on the reactivity and the natural oxides initially present on the surface are also interesting from this point of view. Moreover, each surface finishing for multiphase alloys can result in selective dissolution and in formation of a surface microstructure typical for this treatment. Taking into account that the surface treatment procedure for ZnMgAl coated steel is still under development, this study is limited to the untreated surfaces as a reference state and the untreated surfaces on which we artificially formed corrosion products (see Section 3.1). 2.2. Electrochemical experiments All tests used a BioLogic VMP3 potentiostat for the potential/ current control and measurement. A three electrode cell was used with a saturated calomel electrode (SCE) as the reference electrode, a platinum wire as the counter electrode and the sample as the working electrode. For the oxygen reduction rate measurement, galvanized steel and ZnMgAl-coated steel samples were immersed vertically in two different electrolytes, 0.171 M NaCl solution and synthetic ‘‘rain water’’ developed in [15] with a composition shown in Table 2. Both electrolytes were used in solution with total

484

M. Salgueiro Azevedo et al. / Corrosion Science 90 (2015) 482–490

Table 2 Composition of the synthetic ‘‘rain water’’ electrolyte [15] with total mineralization 10 g l1. Salts

Mineralization (g l1)

Concentration (mmol l1)

CaCl2 (NH4)2SO4 Na2SO4 NaNO3 NaHCO3 Total

3.91 1.82 1.77 1.81 0.69 Salts: 10.00

35.2 13.8 12.5 21.3 8.0 Ions: 243.1

mineralization 10 g l1 and initial pH was adjusted to 9 by NaOH addition. Supplementary dissolution experiments were performed in Na2SO4 and NaCl electrolytes with total mineralization 10 g l1. The surface of the ZnMgAl coating was immersed under applied potential (E = 0.95 V vs. SCE) until the current density decreased less than 0.01 mA cm2 (the current density in the beginning of the dissolution was about 5 mA cm2). This potential for the dissolution experiments was chosen between the corrosion potentials of the coating (approximately 1.05 V vs. SCE, Fig. 1c) and of the substrate steel (about 0.65 V vs. SCE) in order to ensure the dissolution of the coating but not of the substrate steel. After the dissolution experiment, the corrosion products at the cut-edge were removed in an ultrasonic bath with a solution of 200 g l1 glycine in de-ionized water (15 min). 2.3. Titration experiments Titration experiments were performed with a Mettler Toledo One ClickÒ Titration G20 titrometer using a 1.0 M NaOH solution. The pH was continuously monitored with a combined 3 M KCl glass electrode (Mettler Toledo), calibrated with disposable standard buffer solutions (Merck). The pH evolution was recorded using Mettler Toledo LabXÒ light titration software. All solutions were prepared with analytical grade (>99.5%) reagents produced by MERCK and Millipore water (resistance 18 X cm1). The composition of tested solutions is presented in Table 3. The initial volume of the solutions was 50 ml and the added volume of 1.0 M NaOH was 50 ll every 60 s. The temperature of the solutions was stable (25 ± 2 °C). 2.4. Corrosion products and corrosion profile characterization Corrosion products were characterized by X-ray diffraction (XRD) using the Cu Ka1 radiation in a PANalytical X’Pert on Table 3 Experimental conditions of the titration experiments, with label used on the figures and number of the figures where they are shown. Label

Solution 2+

Solvent

Fig.

1 2 3 4

0.04 M 0.04 M 0.04 M 0.04 M

Zn Zn2+ + 0.02 M Mg2+ Zn2+ + 0.02 M Al3+ Zn2+ + 0.02 M Mg2+ + 0.02 M Al3+

0.171 M NaCl

6a

1 2 3 4

0.04 M 0.04 M 0.04 M 0.04 M

Zn2+ Zn2+ + 0.02 M Mg2+ Zn2+ + 0.02 M Al3+ Zn2+ + 0.02 M Mg2+ + 0.02 M Al3+

0.171 M Na2SO4

6b

1 2 3 4

0.04 M 0.04 M 0.04 M 0.04 M

Zn2+ Zn2+ + 0.02 M Mg2+ Zn2+ + 0.02 M Al3+ Zn2+ + 0.02 M Mg2+ + 0.02 M Al3+

0.008 M NaHCO3

6c

1 2 3 4

0.04 M 0.04 M 0.04 M 0.04 M

Zn2+ Zn2+ + 0.02 M Mg2+ Zn2+ + 0.02 M Al3+ Zn2+ + 0.02 M Mg2+ + 0.02 M Al3+

0.171 M NH4Cl

6d

powder. The XRD were collected with angular resolution of 0.02° over the angular range 5–80° (2h) and with a counting time of 0.3 s/step. The evaluation of the data was done using the HighScore Plus software package, containing the JCPDS (ICDD) database files. Raman spectrometry was performed using a LabRAM Aramis spectrometer, from Horiba Jobin Yvon, with green laser 532 nm. The results were compared with reference spectra taken from the literature [13,21–23]. Table 4 names the corrosion products detected in this work, with their respective formula and chosen abbreviation. The samples were observed by scanning electron microscopy (SEM), using a Gemini 1530 microscope with Field Emission Gun source (Schottky-type) and energy dispersive spectroscopy (EDS) for elemental analysis with Si(Li) detector and QUANTAX evaluation software (Bruker AXS). The cross section observations were done on samples mounted in resin cut in cross section and polished mechanically. Before observation, a layer of 10 nm of C was sputtered on the cross section in order to ensure charge evacuation from the surface that could contain non-conducting corrosion products. Additional SEM observations together with EDS analysis were performed on corroded ZnMgAl coating after 100 h of accelerated corrosion Salt Spray Test (SST) [24], with modified electrolyte called ‘‘rain water’’ in a concentration of 1.0 wt.%, as described in [15]. Confocal Raman spectrometry was performed on the cross section of corroded ZnMgAl coating from accelerated cyclic corrosion test, VDA 233-102 [25], with 0.17 M NaCl electrolyte, after five cycles (which corresponds to five weeks of alternating humid and dry phases). 3. Results 3.1. Effect of precipitated corrosion products on oxygen reduction rate To simulate the effect of ‘‘natural’’ corrosion products on the rate of cathodic reaction we modified the applied potential pulse technique (technique inspired by cyclic voltametry in which cyclic potential scanning in cathodic and anodic direction is replaced by a cyclic application of alternating cathodic and anodic pulses on the material) [26], performing a series of long alternating potentiostatic pulses and open circuit exposures. This was made in order to 1) form patinas on the alloy surface during immersion; and 2) measure the effect of these patinas on the rate of the cathodic reaction at applied potential 1.15 V vs. SCE. The open circuit potential in both solutions was close to 1.05 V vs. SCE. The value of the cathodic potential was chosen only 10 mV lower (1.15 V vs. SCE) in order to place the system in the domain of the oxygen reduction (Fig. 1c) but also to minimize the reduction of zinc oxides which

Table 4 Equivalence between the name of corrosion product, its composition and the chosen abbreviation. Label

Name

Abbreviation

Chemical formula

1

Layered double hydroxide

LDH

2 3 4 5 6 7

Simonkolleite Zinc hydroxysulfate Hydrozincite Zincite Zinc hydroxide Magnesium hydroxide Aluminum hydroxide

ZHC ZHS HZ ZnO Zn(OH)2 Mg(OH)2

M(II)xM(III)y(A)m(OH)nzH2O M(II) = Zn2+, Mg2+, M(III) = Al3+  2 A = CO2 3 , Cl , SO4 Zn5(OH)8Cl2H2O Zn4(OH)6SO4nH2O, n = 3–5 Zn5(OH)6(CO3)2H2O ZnO Zn(OH)2 Mg(OH)2

Al(OH)3

Al(OH)3

8

485

M. Salgueiro Azevedo et al. / Corrosion Science 90 (2015) 482–490

occurs at higher cathodic potentials. The initial pH of the NaCl electrolyte was chosen to be about 9 in order to approach the ‘‘natural’’ pH of ‘‘rain water’’ electrolytes and facilitate corrosion products precipitation. Fig. 2 shows the measured cathodic current under an applied cathodic potential of Ea = 1.15 V vs. SCE on galvanized steel and ZnMgAl coatings in ‘‘standard’’ for corrosion tests 0.171 M (or 10 g l1) NaCl solution [24] and in developed in [15] ‘‘rain water’’ electrolyte with total mineralization 10 g l1 and composition presented in Table 2. One can observe in Fig. 2a) that, in a NaCl electrolyte, the initial cathodic current measured for ZnMgAl (period I in Fig. 2a and b) was smaller than that for galvanized steel, which means a lower oxygen reduction rate on ZnMgAl than on galvanized steel (with natural oxides on the top layer). The current did not stabilize in ‘‘rain water’’ electrolyte on either coating. After the first step of applied cathodic potential (period I in Fig. 2a and b), the system was returned to the open circuit potential (period II in the figures) for 12 h and the samples were allowed to corrode spontaneously. Then the cathodic potential was again applied during 30 h (period III in the figures). At the beginning of the second cathodic polarization, the cathodic current decreased significantly in both electrolytes and for both coatings as compared to the first cycle (compare periods I and III in Fig. 2). However, it is clear from the figure that the cathodic reactivity of galvanized steel increases with time under cathodic polarization while it remains stable for ZnMgAl. This is true for both electrolytes. The decrease of the initial cathodic current after the OC exposure (beginning of the period III) compared with the fresh surface (period I) can be attributed to the barrier effect of the corrosion products formed during the OC exposure (period II). An inhibitive action on galvanized steel is longer in ‘‘rain water’’ than in NaCl (compare the curve shape in Fig. 2a and b), which can be associated with the buffer capacity of this electrolyte. For both coatings and both (fresh or patinated) surface states the cathodic reactivity is higher in the ‘‘rain water’’ electrolyte than in NaCl solution (compare Fig. 2a and b). Table 5 shows the corrosion products detected by X-ray diffraction before and after the cathodic pulses applied to galvanized steel and ZnMgAl coatings in both electrolytes (the sampling is indicated by circled points 1 and 2 in Fig. 2). Directly after the OC immersion in NaCl solution, ZHC and HZ, which can be both described as basic zinc salts (BZS), were formed on galvanized steel and LDH on ZnMgAl. After immersion in the ‘‘rain water’’ electrolyte, BZS were formed on both coatings. At the end of the polarization, ZnO was detected on galvanized steel in both electrolytes,

Table 5 Corrosion products identified on galvanized steel and ZnMgAl coated steel from the and represent the sampling time polarization experiment. The numbers indicated in Fig. 2. Electrolyte

NaCl ‘‘rain water’’

Coating

Galvanized ZnMgAl Galvanized ZnMgAl

I

II

After OCP exposure

After cathodic polarization

ZHC, HZ LDH ZHS, ZHC, HZ ZHS, HZ

ZHC, ZnO LDH HZ, ZnO LDH, HZ

while the corrosion products of ZnMgAl did not change in NaCl electrolyte and, in the ‘‘rain water’’, LDH was also detected after cathodic polarization. 3.2. Morphology of corroded samples after different exposures In order to determine common features in the spatial distribution of the corrosion products, SEM observations and Raman spectroscopic analyses were performed on the cross sections of ZnMgAl after 100 h of salt spray test (SST). The tests used the standard procedure (keeping the same temperature, relative humidity, pH and flow rate) as for a standard SST but the electrolyte was either NaCl or the ‘‘rain water’’, both electrolytes with the total mineralization of 10 g 11 (see [15] for the details). An example in Fig. 3 a shows the cut edge of a ZnMgAl sample after 100 h of salt spray test in ‘‘rain water’’ electrolyte. The preferential corrosion of the ternary phase is observed and is in agreement with the results of previous studies obtained on different ZnMgAl compositions and in different test conditions [1–3]. Al and Zn in corrosion product keep their relative positions as inside the lamellas of uncorroded coating, which is illustrated by high resolution electron microscopy image of the corroded ternary phase area in Fig. 3b) and by the concentration profiles of Al, Zn, S and O (Fig. 3c) measured by EDS at points shown in Fig. 3b: as in non-corroded ZnMgAl, lamellas containing Zn (and S) are alternated with lamellas rich in Al. Raman spectroscopic analysis on a cut edge of ZnMgAl after 100 h of SST with NaCl electrolyte is shown in Fig. 4. LDH (spectra A) was only found in the internal layer of corrosion products in place of the former ternary phase in the coating (location ii), not in the external layer of precipitated products (location i). This figure illustrates the general tendency which was observed in different exposure conditions.

in NaCl

a 150

Corrosion products characterization

”

“

b 150 I

III

1

II

III

2

1 1

-2

-2

2 100

100

50

1

1

|j

|j

1 50 2

2 2

2

0 0

0 10

20

30

time / h

40

50

0

10

20

30

40

50

time / h

Fig. 2. Magnitude of the cathodic current (|j|) measured under cathodic polarization of ZnMgAl (curve 2) and reference hot dip galvanized (curve 2) coatings before (period I) and after (period III) open circuit exposure (period II). Applied cathodic potential in periods I and III was 1.15 V vs. SCE. Points and indicate the moments at which the surface was analyzed in order to identify corrosion products.

486

M. Salgueiro Azevedo et al. / Corrosion Science 90 (2015) 482–490

a

b

c concentration / at. %

i

. . 1 2

iii ii

. . . . . 3 4 56 7

200 nm

steel

50

A%

O

40 Zn

30 20

Al

10 S

0. 1

. 2

0

. 3

. 4

. 5

. 6

. 7

060

Fig. 3. Cross section of ZnMgAl coating after 100 h of salt spray test performed in the ‘‘rain water‘‘ electrolyte. (a) General view (backscattered electron imaging in scanning electron microscope) showing the morphology typical for corrosion profile: i. strongly corroded ternary phase; ii. less corroded ternary phase (Zn lamellas are still metallic); iii. uncorroded Zn dendrites inside corroded ternary phase. (b) High resolution image showing lamellar structure of the corroded ternary phase with darker lamellas rich in Al and lighter lamellas reach in Zn and S. The points with numbers show the positions at which the EDS analysis was made. The concentration profile considering these points is shown in Fig. 3. (c) It is evident that at this magnification it is impossible to avoid the effect of the surrounding material and only the evolution in the concentration of each element in function of the position which was analyzed is valuable, the absolute values of concentrations should not be considered.

a

a

Residual coating (Al)

i ii iii Scratched zone Steel

b 149

Intensity / a.u.

551 488

3550

1057 1108

3491

A

b

Resin

2

390 256

1 1060 (hydrozincite)

136

B

1070 3550

376

Steel

C 0

1000

2000

3000

Atomic % C

O

Al

S

Zn

1

-

28.0

65.9

1.4

4.8

2

94.8

4.9

*

*

*

4000

Shift Raman / cm-1 Fig. 4. (a) Optical micrograph of ZnMgAl coating after five weeks of cyclic VDA corrosion test [25] using 0.17 M NaCl electrolyte: (i) external layer of precipitated corrosion products; (ii) corroded coating; (iii) metallic coating. (b) Raman spectra ‘A’ was obtained at (ii) and spectra ‘B’ and ‘C’ at (i). ‘A’ is identified as LDH; ‘B’ shows presence of ZHC and HZ; ‘C’ is identified as HZ.

Fig. 5 shows SEM images of ZnMgAl coating after the anodic dissolution (applied potential Ea = 0.95 V vs. SCE) in different electrolytes as indicated in the legend. Fig. 5(a) and (c) shows the top view of the remaining coating, in which it is possible to identify the steel substrate in a scratched zone and Al fiber forming a mesh; Fig. 5(b) shows the cut edge of the sample presented in Fig 5a after the dissolution of zinc corrosion product in glycine. Some cavities are seen due to the dissolution of Zn dendrites. EDS analysis (Fig. 5b) confirmed that the remaining coating consists mostly of Al (less than 5% of Zn) and that the cavities are filled with resin (mostly carbon is detected). Some oxidation of the Al mesh is detected, but as the atomic fraction of Al is more than twice bigger than that of oxygen, one can suppose that a significant fraction of Al is still present in a metallic form. If some LDH can be formed, it should be localized around the Al skeleton, inside the ternary phase. The residual structure of the

(-) Element not taken into account (*) less than the detection limit

c Residual coating (Al)

Scratched zone Steel

Fig. 5. Typical residual microstructure of ZnMgAl coating after dissolution under applied potential 0.95 V vs. SCE and following removal of zinc patinas. (a) Top view of the remaining Al fiber after dissolution in 0.171 M Na2SO4. (b) Cut edge view of the same sample with EDS analysis showing at point (1) the Al fiber forming a mesh or ‘‘skeleton’’ due to the dissolution of Zn and Zn2Mg lamellas and point (2) the filled with resin cavity from Zn dendrites dissolution. (c) Top view of the remaining coating (Al fiber) after dissolution in 0.171 M NaCl.

487

M. Salgueiro Azevedo et al. / Corrosion Science 90 (2015) 482–490

2

1

a

13

3

iii

11

iii

4

1

b

13

iii

iii

11

ii

2

3

iii

i’

iii

iii

ii

i’

ii

pH

pH

ii

9 i

7

9 i

i

i

i

7

i

5

5

3

3 0

1

2

3

4

0

5

1

2

[OH-]/[Zn2+] 1

c

13

2

3 iii

4

3 1

iii

“plateau” due to NH4+/NH4OH

11 ii

pH

9 i

i

7

4

d

13

iii

ii

i

3

5

[OH-]/[Zn2+]

iii

11

pH

4 iii

ii

i

i

7

5

iii ii

9

i

iii 4

2

5

3 0

1

2

3

4

3

5

0

2

[OH-]/[Zn2+]

4

6

8

10

[OH-]/[Zn2+]

Fig. 6. Titration curves of different electrolytes with 1.0 M NaOH at 50 ll/min. The conditions at which the sampling was made for XRD and SEM analysis are indicated as (i), (ii) and (iii). (a) 0.171 M NaCl solution, (b) 0.171 M Na2SO4 solution, (c) 0.008 M NaHCO3 solution and (d) 0.171 M NH4Cl solvent in presence of (1) 0.04 M Zn2+, (2) 0.04 M Zn2+ + 0.02 M Mg2+, (3) 0.04 M Zn2+ + 0.02 M Al3+, (4) 0.04 M Zn2+ + 0.02 M Mg2+ + 0.02 M Al3+.

Al-rich phase observed in all conditions suggests the importance of the initial phase distribution for the barrier effect of the corrosion products.

3.3. Formation and stability of the precipitated artificial corrosion products Titration was used to study the effect of Mg2+ and Al3+ on the precipitation and transformation of different ‘‘synthetic’’ corrosion products with Zn2+ and different anions. Fig. 6 shows the titration curves obtained for the different solutions presented in Table 3. Note that the x-axis is presented as the

ratio of the quantity of the hydroxide-ion added to the quantity of Zn2+ ions initially present. In all solutions containing Al3+ (curves 3 and 4 in Fig. 6a–d) the first plateau from pH 4 to 4.5 is due to the formation of Al2(OH)+5 [27]. In absence of Al3+ (curves 1 and 2) the first plateau around pH 6.5 to 7 corresponds to the formation of basic zinc salts (BZS) – simonkolleite (Fig. 6a and d), zinc hydroxysulfate (Fig. 6b), hydrozincite (Fig. 6c). With addition of Mg2+ (curves 2 and 4) a supplementary plateau at pH around 10 is observed due to precipitation of Mg(OH)2 [27]. The buffer effect is visible in the NH4Cl solution (Fig. 6d) at pH 8–10 (compare with [15]). Table 6 summarizes the composition of precipitates which were sampled, filtered and dried at different moments of the titrations

Table 6 Summary of precipitated products identified at different titration conditions. Symbols (i), (ii) or (iii) correspond to the moment of the sampling which is indicated on the titration curves in Fig. 6(a)–(d) by the same symbol. The concentrations of cations in the electrolyte are (when present): 0.04 M Zn2+, 0.02 M Mg2+, 0.02 M Al3+. Figure: Electrolyte:

6a 0.171 M NaCl

Curve: Present cations: i/i0 pH  7/9 ii

pH  10

1 Zn ZHC ZnO –

iii

pH  12

ZnO

2 Zn + Mg ZHC

3 Zn + Al LDH

4 Zn + Mg + Al LDH

1 Zn ZHS

2 Zn + Mg ZHS

3 Zn + Al ZHS

4 Zn + Mg + Al ZHS

ZHC Mg(OH)2 ZHC Mg(OH)2

–

LDH Mg(OH)2 LDH Mg(OH)2

–

ZHS

–

ZnO

ZHS Mg(OH)2

LDH

ZHS Mg(OH)2 LDH Mg(OH)2

Figure: Electrolyte:

6c 0.008 M NaHCO3

Curve: Present cations: i pH  7 ii

pH  10

1 Zn HZ ZnO –

iii

pH  12

ZnO

6b 0.171 M Na2SO4

2 Zn + Mg HZ ZnO HZ Mg(OH)2 HZ Mg(OH)2 ZnO

LDH

6d 0.171 M NH4Cl 3 Zn + Al LDH

4 Zn + Mg + Al LDH

1 Zn ZHC

2 Zn + Mg ZHC

3 Zn + Al LDH

4 Zn + Mg + Al LDH

–

LDH Mg(OH)2 LDH Mg(OH)2 ZnO

ZHC

ZHC Mg(OH)2 ZHC Mg(OH)2 Zn(OH)2

LDH

LDH Mg(OH)2 LDH Mg(OH)2 Zn(OH)2

LDH

Zn(OH)2

LDH Al(OH)3

488

M. Salgueiro Azevedo et al. / Corrosion Science 90 (2015) 482–490

a

c

b

Atomic %

Atomic %

Atomic %

O

Mg

Al

Cl

Zn

O

Mg

Al

Cl

Zn

O

Mg

Al

S

Zn

47.3

*

10.8

10.8

31.0

51.4

5.9

7.7

5.8

29.2

68.5

*

7.8

8.3

15.3

(*) less than the detection limit

(*) less than the detection limit 2+

Fig. 7. SEM images and EDS analysis of recovered precipitates at some steps of the titration experiment: (a) Zn + Al3+ in 0.171 M NaCl solvent after the second plateau (Fig. 6a, curve 3 i); (b) Zn2+ + Mg2+ + Al3+ in 0.171 M NaCl solvent after the third plateau (Fig. 6a, curve 4 ii); and (c) Zn2+ + Al3+ in 0.171 M Na2SO4 solvent after the second plateau (Fig. 6b, curve 3 i). Carbon is not taken into account for the calculation of atomic fractions.

which where presented in Section 3.2, suggested the absence (in NaCl and Na2SO4) or the delay (in NaHCO3 and NH4Cl) of the transformation of BZS into ZnO or Zn(OH)2. In order to verify this result, additional experiments were made. After the titration in NaCl or in Na2SO4 was finished, the solution was not filtered immediately but was stirred for 48 or 72 h. Than the precipitates were filtered, dried and analyzed by XRD. Fig. 8 shows the X-ray diffraction pattern of the precipitates formed and evolved in the solution containing 0.04 M Zn2+ + 0.02 M Mg2+ in 0.171 M NaCl. Immediately after the end of the titration experiment (pH 12.5), ZHC and Mg(OH)2 are detected (Fig. 8a). Fig. 8b demonstrates however the formation of zinc hydroxide after 48 h of the solution stirring once the titration experiment was stopped. It can be concluded from this result that Mg2+ just slowed down the transformation ZHC ? Zn(OH)2 and after 48 h of stirring the transformation was completed. Similar experiments

a Intensity / a.u.

Mg(OH)2

Mg(OH)2

ZHC

ZHC

ZHC

ZHC ZHC

ZHC

ZHC

ZHC

5

10

15

20

25

30

35

40

2theta / °

b

Zn(OH)2 Zn(OH)2

Intensity / a.u.

as indicated on the curves presented in Fig. 6 by points i, ii and iii. The identification was made using XRD and Raman spectroscopy. Supplementary EDS analysis (Fig. 7) verified the high fraction of the expected anions in the formed LDH in order to ensure that the LDH was formed by precipitation with the expected anion and not only due to the carbonate accumulation during drying. This was necessary because the XRD peaks were large and the peak positions of basic salts and layered double hydroxides with different anions are very close [28,29]. In NaCl solution with only Zn2+ cations simonkolleite (ZHC) was formed together with ZnO at neutral pH and ZHC was completely transformed into ZnO at alkaline pH. In the presence of Mg2+ and/ or Al3+ ions, only simonkolleite (for mixtures of Zn2+ and Mg2+) or LDH (when Al3+ was present) were detected even at pH 12. In Na2SO4 electrolyte with only Zn2+ cations, zinc hydroxysulfates (ZHS) were formed at neutral pH, and were completely transformed into ZnO at alkaline pH. In contrast to NaCl, ZHS was formed at neutral pH even in the presence of Al3+ and LDH was detected only at high pH. In NaHCO3 solutions containing only Zn2+ cations, hydrozincite (zinc hydroxycarbonate, HZ) and ZnO were detected at neutral pH. HZ was transformed into ZnO with pH increase. In the presence of Zn2+ and Mg2+, ZnO and HZ were also initially formed at neutral pH, however ZnO disappeared during the plateau of Mg(OH)2 precipitation (pH around 10) and re-appeared after the end of the experiment (pH > 12). In the presence of Zn2+ and Al3+, LDH was the only product detected in both, neutral and alkaline conditions. In the solution containing all three cations, LDH was detected at all pH. After the end of the third plateau (the plateau of the precipitation of Mg(OH)2), ZnO was also formed in the tri-cation solution. Chemical analysis of the precipitates formed in NH4Cl electrolyte at pH below 12 revealed ZHC when Zn2+ or both, Zn2+ and Mg2+, were present in the solution. ZHC was transformed into Zn(OH)2 at pH 12 with or without Mg2+. With the addition of Al3+, LDH was the only product at neutral pH and small quantities of Al(OH)3 appeared at the end of the experiment in the absence of Mg2+ (pH 12.5). In the tri-cation mixtures, aluminum hydroxides were not detected even at high pH but a partial transformation of the LDH into Zn(OH)2 occurred at very alkaline pH (pH > 12).

Mg(OH)2 Zn(OH)2

Mg(OH)2

Zn(OH)2

0

5

10

15

20

25

30

35

40

2theta / ° 3.4. Kinetic effect of Mg2+ The tendencies in the formation of specific products during acid – base titration of the mixtures containing both Zn2+ and Mg2+,

Fig. 8. XRD analysis of the precipitates formed after titration of (0.04 M Zn2+ + 0.02 M Mg2+ in 0.171 M NaCl) solution (Fig. 6a, curve 2 iii) (a) immediately at the end of titration experiment and (b) after 48 h of stirring. ZHC is zinc hydroxychloride, Zn5(OH)8Cl2.

489

M. Salgueiro Azevedo et al. / Corrosion Science 90 (2015) 482–490

in Na2SO4 solution demonstrated even stronger kinetics effect of Mg2+ on the transformation ZHS ? Zn(OH)2. Similar experiments were made for the solutions when the titration was stopped at pH 10 (Fig. 6a, point ii on curve 2). In this case ZHC was stable even after 48 h of stirring in the presence of Mg(OH)2. The results demonstrated strong stabilization of BZS against the transformation BZS ? ZnO/Zn(OH)2 at alkaline pH by ions of Mg. 4. Discussion 4.1. Influence of Mg and Al on the stability of Zn-containing corrosion products According to the titration results, Mg2+ ions are able to prevent or delay the transformation of basic zinc salts at alkaline pH. This effect can be caused by different factors. First, the precipitation of Mg(OH)2 can buffer the pH at values which are below the values necessary to transform BZS (like ZHC, ZHS, HZ) into ZnO or Zn(OH)2, as suggested in [1,5,6]. This means, in case of ZnMgAl corrosion, that as much as there is Mg2+ in solution, the pH is buffered and the transformation of the BZS is avoided. This mechanism however contradicts the equilibrium calculations [30], demonstrating that at the pH at which Mg(OH)2 precipitates, ZnO is more stable than BZS. The effect is therefore expected to be kinetic in nature. A possible reason is that the rapidly formed Mg(OH)2 physically blocks the access of hydroxide ions to the initially formed BZS, delaying the transformation at alkaline pH (pH > 12). The delay of the transformation BZS ? ZnO and Zn(OH)2 can be an important factor for corrosion resistance because (i) every transformation can disturb the structure and increase porosity, and (ii) zinc oxides are known to be more porous than the other Zn-patinas [1,5,6,19,20]. Al3+ in solution enables the precipitation of LDH, which is considered by many authors as a good barrier for oxygen diffusion [1,2,8,11–13,19]. We observed that this compound was stable at least up to pH = 12. The peak positions in the XRD spectra were very large and a complete characterization of the LDH was not achieved in this work. However, according to the literature [28,29], the stability constant of ZnAl-LDH is higher than that for MgAl-LDH. The participation of Mg2+ in the LDH formation cannot be completely excluded because even if the Mg(OH)2 precipitation plateau at pH  10 is still present in the presence of Al3+, it is shorter than in the absence of Al3+, which can be interpreted as some Mg2+ has been already precipitated at lower pH (during the precipitation of LDH at pH  6).

The results also demonstrated that the ZHS is stabilized and the formation of LDH is delayed in SO2 4 containing solutions. The presence of sulfate anions can hence significantly modify the relative stability of different corrosion products in the presence of Al3+ and result in altered corrosion mechanisms. In our previous publication [15] we discussed the primordial role of ammonium and bicarbonate ions, in this work the specific role of the sulfate anion was revealed. The effect of Mg2+ and Al3+ on the formation and transformation of zinc patinas in different electrolytes is schematically summarized in Fig. 9. 4.2. Effect of precipitated corrosion products on the oxygen reduction rate The polarization experiment (Section 3.1) demonstrates that a layer of corrosion products is able to decrease the cathodic reaction on the surface of both galvanized steel and ZnMgAl coatings. This is coherent with previous observations [31–33]. However, during cathodic polarization on corroded galvanized steel, the barrier effect disappears, as the current magnitude increases with time (Fig. 2). We interpret this current increase in terms of the transformation BZS ? ZnO (detected by XRD) due to the pH increase by cathodically generated hydroxide (reaction given by Eq. (1)).

O2 þ 2H2 O þ 4e ! 4OH

ð1Þ

The cathodic current on ZnMgAl is constant due to stabilization of corrosion products discussed in previous section. We suggest that the barrier effect on both coatings is due to BZS, as the cathodic current after OCP exposure is (i) initially present on galvanized steel in both electrolytes and (ii) is present and stable on ZnMgAl in ‘‘rain water’’ electrolyte even if LDH is not formed. The fact that ZHS and ZHC disappear from galvanized steel and only HZ is detected at the end of cathodic polarization in ‘‘rain water’’ electrolyte is consistent with the instability of ZHS [32,33] and with the fact that, among BZS, the carbonates species are the most stable [21,31,32]. The instability of BZS and the absence of the stabilizing Mg effect result in the lost of barrier protection on galvanized steel. On ZnMgAl, the stabilizing effect of Mg2+ and Al3+ ions protect the system against the loss of barrier protection. Slower increase of cathodic current on galvanized steel covered by patina in ‘‘rain water’’ electrolyte than in NaCl can be assigned to the buffer capacity of this electrolyte due to the presence of NH+3 and HCO 3 [15]. The delayed formation of LDH in ‘‘rain water’’ electrolyte in cyclic experiment correlates with the results of

1. all electrolytes

Zn2+

blocked by Mg2+

Basic Zinc Salts

Mg(OH)2

rapid

ZnO

Zn-Al Layered Double Hydroxides Zn(OH)2 Fig. 9. Schematic illustration of the formation and transformation of different corrosion products in the presence of different species in the electrolyte (as indicated).

490

M. Salgueiro Azevedo et al. / Corrosion Science 90 (2015) 482–490

accelerated corrosion tests from our previous work [15], where the effect of the ‘‘rain water’’ electrolyte was featured as being responsible for the retardation of Al dissolution. The formation of a residual Al skeleton observed in Section 3.2 can reinforce the barrier effect by retaining the patinas in a compact structure. Our results suggest that the Al distribution inside the corroded coating reflects the distribution in the initial alloy. This result is coherent with Keppert et al. [12], who observed a heterogeneous distribution of Al inside the corroded coating; the low mobility of Al3+ is also consistent with the absence of Al in soluble corrosion products after accelerated corrosion tests [15].

5. Conclusion 1. During open circuit exposure, corrosion products were formed on both galvanized steel and ZnMgAl coatings. Cathodic polarization before and after this open circuit exposure in NaCl and in synthetic ‘‘rain water’’ electrolytes demonstrated good initial barrier effect of BZS and BZS/LDH on the oxygen reduction in both environments. 2. On galvanized steel, but not on ZnMgAl, under cathodic polarization, BZS transforms into ZnO with loss of barrier properties. This transformation is delayed in ‘‘rain water’’. 3. The stabilization of BZS on ZnMgAl is explained by a kinetic effect of Mg2+ confirmed by acid–base titrations of Zn2+/Mg2+/ Al3+ mixtures. The effect can not be explained by thermodynamic considerations because BZS were formed during titrations in the pH range 10–12.5 when the formation of ZnO is expected from thermodynamic calculations. 4. The effect of Mg2+ varies in function of the electrolyte composition. Mg2+ suppress the direct ZnO precipitation and retards the transformation of BZS into ZnO/Zn(OH)2 at pH 7–13 in NaCl and Na2SO4 and at pH 10 in NaHCO3. The presence of Mg2+ ions does not affect the nature of precipitated zinc-containing products at neutral or strongly alkaline pH in NaHCO3 and in NaH4Cl electrolytes. 5. In the presence of Zn2+/Mg2+/Al3+ mixture, ZHS is preferentially formed in Na2SO4 solution and its transformation into LDH occurs at alkaline pH (12). 6. An additional effect is achieved through a skeleton of non-oxidized Al which keeps the corrosion product compact.

References [1] P. Volovitch, T.N. Vu, C. Allély, A. Abdel Aal, K. Ogle, Understanding corrosion via corrosion products characterization: II. Role of alloying elements in improving the corrosion resistance of Zn–Al–Mg coatings on steel, Corros. Sci. 53 (2011) 2437–2445. [2] S. Schuerz, M. Fleischanderl, G.H. Luckeneder, K. Preis, T. Haunschmied, G. Mori, A.C. Kneissl, Corrosion behaviour of Zn–Al–Mg coated steel sheet in sodium chloride-containing environment, Corros. Sci. 51 (2009) 2355–2365. [3] M. Dutta, A.K. Halder, S.B. Singh, Morphology and properties of hot dip Zn–Mg and Zn–Mg–Al alloy coatings on steel sheet, Surf. Coat. Technol. 205 (2010) 2578–2584. [4] J. Elvins, J.A. Spittle, J.H. Sullivan, D.A. Worsley, The effect of magnesium additions on the microstructure and cut edge resistance of zinc aluminium alloy galvanized steel, Corros. Sci. 50 (2008) 1650–1658. [5] P. Volovitch, C. Allély, K. Ogle, Understanding corrosion via corrosion products characterisation: I. Case study of the role of Mg alloying in Zn–Mg coating on steel, Corros. Sci. 51 (2009) 1251–1262. [6] N.C. Hosking, M.A. Strom, P.H. Shipway, C.D. Rudd, Corrosion resistance of zinc–magnesium coated steel, Corros. Sci. 49 (2007) 3669–3695. [7] T. Prosek, A. Nazarov, U. Bexell, D. Thierry, J. Serak, Corrosion mechanism of model zinc–magnesium alloys in atmospheric conditions, Corros. Sci. 50 (2008) 2216–2231.

[8] T. Tsujimura, A. Komatsu, A. Andoh, Influence of Mg content in coating layer and coating structure on corrosion resistance of hot-dip Zn–Al–Mg alloy coated steel sheet, in: Proceedings of the Galvatech ‘01, International Conference on Zinc and Zinc Alloy Coated Steel, June 26–28, Brussels, Belgium, 2001, pp. 145–152. [9] E. Diler, S. Rioual, B. Lescop, D. Thierry, B. Rouvellou, Chemistry of corrosion products of Zn and MgZn pure phases under atmospheric conditions, Corros. Sci. 65 (2012) 178–186. [10] B. Li, A. Dong, G. Zhu, S. Chu, H. Qian, C. Hu, B. Sun, J. Wang, Investigation of the corrosion behaviors of continuously hot-dip gavanizing Zn–Mg coating, Surf. Coat. Technol. 206 (2012) 3989–3999. [11] D. Persson, D. Thierry, N. LeBozec, T. Prosek, In situ reflection spectroscopy studies of the initial atmospheric corrosion of Zn–Al–Mg coated steel, Corros. Sci. 72 (2013) 54–63. [12] T.A. Keppert, G.H. Luckeneder, K-H. Stellnberger, G. Mori, Influence of the pH value on the corrosion of Zn–Al–Mg hot-dip galvanized steel sheets in chloride containing environments, in: NACE International Corrosion Conference, 2012, C2012–0001493. [13] S. Schürz, G.H. Luckeneder, M. Fleischanderl, P. Mack, H. Gsaller, A.C. Kneissl, G. Mori, Chemistry of corrosion products on Zn–Al–Mg alloy coated steel, Corros. Sci. 52 (2010) 3271–3279. [14] N. LeBozec, D. Thierry, M. Rohwerder, D. Persson, G. Luckeneder, L. Luxem, Effect of carbon dioxide on the atmospheric corrosion of Zn–Mg–Al coated steel, Corros. Sci. 74 (2013) 379–386. [15] M. Salgueiro, C. Allély, K. Ogle, P. Volovitch, Corrosion mechanisms of Zn(Mg, Al) coated steel in accelerated tests and natural exposure: 1. The role of electrolyte composition on the nature of corrosion products and relative corrosion rates, Corros. Sci. 90 (2015) 472–481. [16] K. Ueda, A. Takahashi, Y. Kubo, Investigation of corrosion resistance of prepainted Zn–11%Al–3%Mg–0.2%Si alloy coated steel sheet through outdoor exposure test in Okinawa, Metall. Ital. – Corros. (2) (2012) 13–19. [17] N. LeBozec, D. Thierry, A. Peltola, L. Luxem, G. Luckeneder, G. Marchiaro, M. Rohwerder, Corrosion performance of Zn–Mg–Al coated steel in accelerated corrosion tests used in the automotive industry and field exposures, Mater. Corros. 64 (9999) (2013). [18] D. Landolt, Corrosion and Surface Chemistry of Metals, first English edition., EPFL Press, 2007 (Originally published in French as Corrosion et Chimie de Surfaces des Métaux, Copyright 1993, 1997, 2003 Presses Polytechinques et Universitaires Romandes. Translated and updated from the revised second French version.). [19] T. Ishikawa, M. Ueda, K. Kandori, T. Nakayama, Air permeability of the artificially synthesized Zn–Al–Mg alloy rusts, Corros. Sci. 49 (2007) 2547– 2556. [20] T. Ishikawa, K. Matsumoto, A. Yasukawa, K. Kandori, T. Nakayama, T. Tsubota, Influence of metal ions on the formation of artificial zinc rusts, Corros. Sci. 46 (2004) 329–342. [21] C. Merlin, Approches analytique et électrochimique de la dégradation des tôles d’acier revêtues cataphorésées en atmosphères corrosive contenant des ions sulfates, Thèse de doctorat de l’université Henri Poincaré, Nancy I, 1999. [22] M.C. Bernard, A. Hugot-Le Goff, D. Massinon, N. Phillips, Underpaint corrosion of zinc-coated steel sheet studied by in situ Raman spectroscopy, Corros. Sci. 35 (1993) 1339–1349. [23] R. Autengruber, G. Luckeneder, A.W. Hassel, Corrosion of press-hardened galvized steel, Corros. Sci. 63 (2012) 12–19. [24] ISO 9227:2006, Corrosion tests in artificial atmospheres. [25] F. Beier, K-H. Dtellnberger, S. Geisler, A new accelerated cyclic corrosion test for automotive substrates, EUROCORR Congress September 2009, Nice France. [26] L. Jiang, M. Wolpers, P. Volovitch, K. Ogle, An atomic emission spectroelectrochemical study of passive film formation and dissolution on galvanized steel treated with silicate conversion coatings, Surf. Coat. Technol. 206 (2012) 3151–3157. [27] P. Volovitch, M. Serdechnova, K. Ogle, Aqueous corrosion of Mg–Al binary alloys: roles of Al and Mg, Corrosion 68 (6) (2012) 557–570. [28] C. Forano, T. Hibino, F. Leroux, C. Taviot-Guého, Layred double hydroxides, in: Handbook of Clay Science, Elsevier Ltd., 2006, pp. 1021–1095. Chapter 13.1. [29] C.A. Johnson, F.P. Glasser, Hydrotacilte-like minerals (M2Al(OH)6(CO3)0.5xH2O, where M = Mg, Zn Co, Ni) in the environment: synthesis, characterization and thermodynamic stability, Clays Clay Miner. 51 (1) (2003) 1–8. [30] M. Salgueiro Azevedo, C. Allély, K. Ogle, P. Volovitch, Corrosion mechanisms of + Zn(Mg, Al) coated steel: The effect of HCO 3 and NH4 ions on the intrinsic reactivity of the coating, Electrochim. Acta (submitted). [31] J.D. Yoo, P. Volovitch, A.A. Aal, C. Allely, K. Ogle, The effect of an artificially synthesized simonkolleite layer on the corrosion of electrogalvanized steel, Corros. Sci. 70 (2013) 1–10. [32] J.D. Yoo, K. Ogle, P. Volovitch, The effect of synthetic zinc corrosion products on corrosion of electrogalvanized steel: I. Cathodic reactivity under zinc corrosion products, Corros. Sci. 81 (2014) 11–20. [33] J.D. Yoo, K. Ogle, P. Volovitch, The effect of synthetic zinc corrosion products on corrosion of electrogalvanized steel. II. Zinc reactivity and galvanic coupling zinc/steel in presence of zinc corrosion products, Corros. Sci. 83 (2014) 32– 37.