Effect of Dissolved Oxygen and Immersion Time on the ... - MDPI

materials Article

Effect of Dissolved Oxygen and Immersion Time on the Corrosion Behaviour of Mild Steel in Bicarbonate/Chloride Solution Gaius Debi Eyu *, Geoffrey Will, Willem Dekkers and Jennifer MacLeod Department of Chemistry, Physics and Mechanical Engineering, Queensland University of Technology, Brisbane QLD 4000, Australia; [email protected] (G.W.); [email protected] (W.D.); [email protected] (J.M.) * Correspondence: [email protected]; Tel.: +61-24429877 Academic Editor: Richard Thackray Received: 4 August 2016; Accepted: 22 August 2016; Published: 1 September 2016

Abstract: The electrochemical behavior of mild steel in bicarbonate solution at different dissolved oxygen (DO) concentrations and immersion times has been studied under dynamic conditions using electrochemical techniques. The results show that both DO and immersion times influence the morphology of the corrosion products. In comparative tests, the corrosion rate was systematically found to be lower in solutions with lower DO, lower HCO3 − concentrations and longer immersion time. The SEM analyses reveal that the iron dissolution rate was more severe in solutions containing higher DO. The decrease in corrosion rate can be attributed to the formation of a passive layer containing mainly α-FeO (OH) and (γ-Fe2 O3 /Fe3 O4 ) as confirmed by the X-ray diffractometry (XRD) and X-ray photoelectron spectroscopy (XPS). Passivation of mild steel is evident in electrochemical test at ≈ −600 mVSCE at pH ≥8 in dearated (≤0.8 ppm DO) chloride bicarbonate solution under dynamic conditions. Keywords: mild steel; bicarbonate; polarization; passivation

1. Introduction Bicarbonate plays a major role in the dissolution reactions of steel pipelines [1]. Studies indicate that increasing HCO3 − concentration exacerbates iron dissolution [2–4]. To date, the role of bicarbonate ions in the corrosion process of carbon steel is still debated [5] Moreno et al. [6] confirmed that high bicarbonate increased localized corrosion contributing to stress corrosion cracking [7,8] and pitting corrosion [9] in carbon steel pipelines. Thomas and Davies [10] reported that bicarbonate ions reduce the passive-active (Flade) potential for ferric oxides, while increasing the passive-active transition potential for magnetite with increasing HCO3 − concentration when the concentration is greater than 10−2 M. However, some researchers argued that a stable-film formation (passivity) can be achieved when a sufficient amount of bicarbonate is present [11–13]. The formation and stability of such passive films depends to a large extent on environmental conditions, such as; applied potential, pH, oxygen, solution chemistry, temperature, fluid flow rate, immersion time and metallurgy of the steel [14–16]. Bicarbonate-induced corrosion is potentially relevant in many applications, such as wellbore systems for geological sequestration of CO2 [17]. Intergranular stress corrosion cracking (IGSCC) in underground pipelines at potential range (≈−0.625 VSCE to ≈−0.425 VSCE ) has been linked with the presence HCO3− and CO3 2− ions at pH (8–10.5) [18]. Delanty and O’Beirne [19] cited a Canadian investigation that revealed that stress corrosion cracking (SCC) in carbon steel X65 was more severe in oxygen restricted areas, presumably because oxygen reduction prevents hydrogen evolution which is thought to facilitate cracking. In contrast, Yunovich et al. reported that SCC susceptibility of API X52 carbon steel was higher in an aerated conditions than in deaerated solution Materials 2016, 9, 748; doi:10.3390/ma9090748

www.mdpi.com/journal/materials

Materials 2016, 9, 748

2 of 18

at pH 8 and 10 [18]. Dissolved oxygen has been shown to increase corrosion rates [20–24] as expected however, Beckett et al. and Sarin et al. [25,26] claimed it decreased with higher DO under stagnant conditions. Dissolved oxygen is an electron acceptor in the corrosion of iron as well as oxidation of ferrous iron [27]: Fe + 1/2 O2 + H2 O ↔ Fe2+ + 2OH− (1) Fe2+ + 1/4 O2 + 1/2 H2 O + 2OH− ↔ Fe(OH)3 (s)

(2)

Kuch [28] reported that in the absence of oxygen, it is possible for ferric scale γ-FeOOH (lepidocrosite) previously formed on the metal surface to act as an electron acceptor: Fe + 2FeOOH (scale) + 2H+ ↔ 3Fe2+ + 4OH

(3)

The corrosion process continues even after DO is depleted according to the Kuch mechanism. However, in the absence of lepidocrosite in the scale, the corrosion reaction still proceeds in deaerated water as reported elsewhere [26], indicating that in a near-neutral pH environment, the Kuch mechanism is not the only mechanism of metal loss in deaerated conditions. Baek et al. [29] concluded that DO facilitates the formation of different iron oxide products. For example, FeOOH and γ-Fe2 O3 were formed at high DO concentrations whereas α-Fe2 O3 was formed in deaerated solution at a more positive potential. The influence of dissolved oxygen and immersion time in the corrosion process of mild steel in a near-neutral pH in corrosive media remains controversial. Here, we present an investigation of the corrosion behaviour of mild steel in chloride-containing bicarbonate solutions of varied DO and immersion time in near-neutral pH, under dynamic conditions at moderate temperature (23 ± 1 ◦ C). Potentiodynamic polarization and electrochemical impedance spectroscopy were used to study the corrosion processes for the electrochemical tests. By incorporating a range of surface analysis techniques, including X-ray diffraction (XRD) and scanning electron microscopy (SEM), we were able to characterize the chemistry and morphology of the corrosion products. 2. Results and Discussion 2.1. Bicarbonate Concentration Figure 1 shows the polarization profile for mild steel in 1 g/L and 5 g/L bicarbonate with 2 g/L chloride concentration under dynamic conditions. Bicarbonate has a distinct effect on the polarization characteristics in bicarbonate/chloride solution. The dissolution in the active and prepassive region is accelerated by the presence of bicarbonate ions due to the formation of soluble complex anion Fe(CO3 )2 2− [3]. At ≈−600 mVSCE , Fe(OH)2 is converted to Fe3 O4 [29] in dearated bicarbonate solution, resulting in passivation. At this potential (≈−600 mVSCE ), Fe2 O3 /Fe3 O4 forms an oxide film at pH ≥ 8,. The anodic polarization current was greater with 5 g/L bicarbonate solution at corrosion potential ≈−600 mVSCE . However, above this potential, the dissolution decreased with increasing bicarbonate concentration. The impedance spectra were obtained from a rotating disc electrode (specimen) immersed in solutions containing 1 g/L and 5 g/L bicarbonate in chhloride solution. The data was fit with equivalent circuit shown in Figure 2, and shows a good fit to the data and represents the physical situation expected [30,31], where Rs , Rct , Ra , Qdl and Qf represent the solution resistance, charge transfer resistance (metal/film interface), adsorption resistance (solution/film interface), double layer capacitance and passive film capacitance, respectively. A constant phase element (CPE) defined by the values of n and Q, is commonly used to compensate for inhomogeneity of electrode surface. The impedance Z, of the constance phase element as given in [32]: ZCPE = Q−1 (jω )−n ,

(4)

Materials 2016, 9, 748

3 of 18

√ where and9,9,n748 are constant and exponent, respectively, j = ( −1) and ω = 2π f is the angular frequency Materials 33 of MaterialsQ2016, 2016, 748 of 18 18 in rad/s calculated using f, the frequency in Hz . At low frequency, the impedance and phase angle angle the for charge transfer activity, while at frequency they angle describe describe the kinetic kinetic response response for the thetransfer chargeactivity, transferwhile activity, while at high high they frequency they describe the kinetic response for the charge at high frequency depend on depend on layer [33]. dependlayer on surface surface layer inhomogeneity inhomogeneity [33]. surface inhomogeneity [33]. 1.5 1.5

1.0 1.0

11 g/L g/L NaHCO NaHCO33 ++ NaCl NaCl 55 g/L g/L NaHCO NaHCO33 ++ NaCl NaCl

E, V(SCE)

0.5 0.5

0.0 0.0

-0.5 -0.5

-1.0 -1.0 -5 -5

-4 -4

-3 -3

-2 -2

-1 -1

00

11

22

33

22

log log i(A/cm i(A/cm ))

polarization for mild steel in 55 g/L solutions Figure 1. 1. Potentiodynamic Potentiodynamic polarization forfor mild steelsteel in 11 g/L g/L and g/L bicarbonate/chloride solutions Figure Potentiodynamic polarization mild in 1and g/L andbicarbonate/chloride 5 g/L bicarbonate/chloride at at 2000 2000 rpm. rpm. solutions at 2000 rpm.

Figure Figure 2. 2. Equivalent Equivalent circuit circuit for for impedance impedance measurements. measurements.

Figure 3a,b show the Nyquist and Bode-phase plots measured for specimens g/L and g/L Figure 3a,b 3a,b show showthe theNyquist Nyquistand andBode-phase Bode-phaseplots plotsmeasured measuredfor forspecimens specimensinin in111g/L g/Land and555g/L g/L Figure bicarbonate/ chloride solutions. The impedance spectra show similar trends, although the capacitive bicarbonate/ chloride solutions. The impedance spectra show similar trends, although the capacitive bicarbonate/ chloride solutions. The impedance spectra show similar trends, although the capacitive loop was depressed with higher bicarbonate concentration. The decrease in the impedance spectra is loop was was depressed depressed with with higher higher bicarbonate bicarbonate concentration. concentration. The The decrease decrease in in the the impedance impedance spectra spectra is is loop linked linked to to iron iron dissolution dissolution which which contributes contributes to to uneven uneven surface surface roughness roughness and and irregular irregular distribution distribution linked to iron dissolution which contributes to uneven surface roughness and irregular distribution of current density on the specimen surface. The Bode impedance plots in Figure 3b show that of current density on the specimen surface. The Bode impedance plots in Figure 3b show that at at low low of current density on the specimen surface. The Bode impedance plots in Figure 3b show that at frequencies, frequencies, the the impedance impedance is is slightly slightly lower lower with with higher higher bicarbonate bicarbonate concentration. concentration. The The lower lower low frequencies, the impedance is slightly lower with higher bicarbonate concentration. The lower bicarbonate bicarbonate concentration concentration exhibits exhibits higher higher impedance impedance in in the the capacitive capacitive region, region, which which implies implies that that aa bicarbonate concentration exhibits higher impedance in the capacitive region, which implies that more more protective protective oxide oxide layer layer is is achieved achieved in in low low concentration concentration bicarbonate. bicarbonate. For For this this lower lower a more protective oxide layer is achieved in low concentration bicarbonate. For this lower concentration concentration concentration solution, solution, the the phase phase angle angle (( )) is is higher higher at at low low frequencies frequencies compared compared to to solution solution with with solution, the phase angle (θ) is higher at low frequencies compared to solution with higher bicarbonate, higher higher bicarbonate, bicarbonate, indicating indicating aa more more resistive resistive barrier barrier film film layer layer at at lower lower concentration. concentration. The The crosscrossindicating a more resistive barrier film layer at lower concentration. The cross-sections of the corrosion sections sections of of the the corrosion corrosion products products shown shown in in Figure Figure 44 reveal reveal that that aa more more uniform uniform corrosion corrosion product product products shown in Figure 4 reveal that a more uniform corrosion product was formed in 1 g/L was was formed formed in in 11 g/L g/L bicarbonate bicarbonate in in comparison comparison with with 55 g/L g/L bicarbonate bicarbonate solution. solution. bicarbonate in comparison with 5 g/L bicarbonate solution.

Materials 2016, 9, 748

4 of 18

Materials 2016,9,9,748 748 Materials 2016, Materials 2016, 9, 748

4 of 18 4 4ofof1818 50 5050

1g/L NaHCO3+NaCl 1g/L 1g/LNaHCO NaHCO3+NaCl +NaCl 5g/L NaHCO33+NaCl 5g/L 5g/LNaHCO NaHCO3+NaCl +NaCl

700 700700

1g/L NaHCO3 + NaCl 1g/LNaHCO NaHCO+ + NaCl 40 1g/L 3 NaCl 5g/L NaHCO33+ NaCl 4040 5g/LNaHCO NaHCO NaCl 5g/L + +NaCl 3 3

3 33

30 3030

3

600 600600

2 2 log (|Z|/ .cm log (|Z|/  ) ) log (|Z|/ .cm

500 500500

2

Z im( cm ) 2 2 im( cm Z Zim(  cm ) )

2

4 44

Phaseangle angle(deg) (deg) Phase Phase angle (deg)

800 800800

400 400400 300 300300 200 200200 100 100100 0 0 00 100 200 300 400 500 600 100 200 200 300 300 400 400 500 500 600 600 0 0 100 2 Z re (cm 2) 2 ) Z re (cm Z re (cm )

20 2020

2 22

10 1010 1 11

700 700 700

0 00

800 800 800

-1 -1 -1

0 00

1 2 11 2 log (f/Hz) 2 log (f/Hz) log (f/Hz)

(a) (a) (a)

3 33

4 44

(b) (b) (b)

Figure 3. Impedance spectra for mild steel in bicarbonate chloride solution: (a) Nyquist plots and (b)

Epoxy Epoxy Epoxy

Scale Scale Scale

Steel Steel Steel

Epoxy Epoxy Epoxy

Scale Scale Scale

Steel Steel Steel

Figure 3. Impedancespectra spectrafor formild mildsteel steel inin bicarbonate chloride solution: (a) Nyquist plots and (b)and Figure 3. 3.Impedance for mild steelin bicarbonate chloride solution: (a) Nyquist plots Figure Impedance spectra bicarbonate Bode phase angle and impedance magnitude vs. frequencychloride plots. solution: (a) Nyquist plots and (b) Bode phase phase angle and impedance magnitude vs. vs. frequency plots. (b)Bode Bode angle and impedance magnitude frequency plots. phase angle and impedance magnitude vs. frequency plots.

(a) (b) (a) (b) (a) of the scale formed at (a) 1 g/L (b) 5 g/L sodium bicarbonate/chloride (b) Figure 4. Cross-section solution Figure 4. Cross-section of the scale formed at (a) 1 g/L (b) 5 g/L sodium bicarbonate/chloride solution Figure 4. Cross-section of the scale formed at (a) 1 g/L (b) 5 g/L sodium bicarbonate/chloride solution at 2000 rpm. Figure 4. rpm. Cross-section of the scale formed at (a) 1 g/L (b) 5 g/L sodium bicarbonate/chloride solution at 2000 at 2000 rpm. at 2000 rpm. 2.2. 2.2. Dissolved Dissolved Oxygen Oxygen

2.2.2.2. Dissolved Oxygen Dissolved Oxygen Dissolved oxygen concentration decreases with nitrogen purging time. In the present experiments,

Dissolved oxygen concentration decreases with nitrogen purging time. In the present experiments, the nitrogen purging also led decrease in as depicted in 5. Dissolved decreases with nitrogen purgingtime. time. In the present experiments, the nitrogenoxygen purging also led to to aa corresponding corresponding decrease in temperature temperature asIn depicted in Figure Figure 5. The The Dissolved oxygenconcentration concentration decreases with nitrogen purging the present experiments, test solution temperature was ≈24 C at 0 h purging, with DO ≈ 4 ppm. The DO decreased significantly thethe nitrogen purging also led to corresponding decrease temperature as depicted in Figure test solution temperature wasto ≈24 C at 0 h purging, with DO ≈in 4 ppm. The as DOdepicted decreased nitrogen purging also led aacorresponding decrease in temperature in significantly Figure 5. The 5. to ≈0.8 ppm after 33 h, at point decreased to C. in ◦ theattemperature The test solution temperature wasC≈ 0 h purging, DO 4 ppm. The decrease DO decreased to ≈0.8 ppm after h, was at which which point the temperature to≈≈18.6 ≈18.6 C. The The decrease in test solution temperature ≈24 at240 hC purging, with DOdecreased ≈with 4 ppm. The DO decreased significantly temperature with purging time was due to the cooling effect of the cooler nitrogen gas. ◦ temperature with purging time was due to the cooling effect of the cooler nitrogen gas. significantly to ≈ 0.8 ppm after 3 h, at which point the temperature decreased to ≈ 18.6 C. The decrease to ≈0.8 ppm after 3 h, at which point the temperature decreased to ≈18.6 C. The decrease in in temperature temperaturewith withpurging purgingtime timewas was due cooling effect of the cooler nitrogen due to to thethe cooling effect of the cooler nitrogen gas. gas. 24 24

DO concentration / ppm / ppm DO concentration DO concentration / ppm

4

Temperature

21 22 21

3

2

1

22 23 22

o

3 3

23 24 23 o o Temperature ( C) Temperature Temperature ( C) ( C)

Dissolved oxygen Dissolved oxygen Temperature Temperature Dissolved oxygen

4 4

2 2

20 21 20 19 19 20

1 1 0 0

1 1

2 2

3 3

N2 purging time / h N2 purging time / h 0

1

2

18 18 19

18 3

Figure 5. Dissolved oxygen (DO) concentration andtime temperature as a function of N2 purging time. N2 purging /h Figure 5. Dissolved oxygen (DO) concentration and temperature as a function of N2 purging time.

Figure Dissolvedoxygen oxygen(DO) (DO)concentration concentration and 2 purging time. Figure 5. 5.Dissolved and temperature temperatureasasaafunction functionofofNN 2 purging time.

Materials 2016, 9, 748

5 of 18

Figure 6 shows the polarization curves of mild steel in bicarbonate/chloride solution in different Materials 2016, 9, 748 5 of 18 DO concentrations at 2000 rpm. The anodic polarization increased without any retardation at a DO concentration of 64shows ppm,the whereas onecurves anodic peak appeared in solutionssolution containing 1 ppm and Figure polarization of mild steel in bicarbonate/chloride in different 0.8 ppmDO DO. The anodic polarization current increases with corrosion potentials below ≈− concentrations at 2000 rpm. The anodic polarization increased without any retardation at a600 DO mVSCE , concentration of 4 ppm, whereas one anodic peakthan appeared in solutions containing ppm and indicating iron dissolution while at potentials greater ≈ −600 mVSCE , the anodic1current decreases 0.8 ppm DO. The anodic polarization current increases with corrosion potentials below ≈−600 mV SCE indicating some form of passivation. A potential cause of this behaviour would be as ,a result indicating iron dissolution while at potentials greater than ≈ −600 mVSCE, the anodic current of corrosion products. For example, Fe(OH)2 formed on the specimen surface, according to the decreases indicating some form of passivation. A potential cause of this behaviour would be as a following reaction: result of corrosion products. For example, Fe(OH)2 formed on the specimen surface, according to the Fe + 2H2 O → Fe(OH)2 + 2H+ + 2e− . (5) following reaction: Fe + 2Hfilm 2O →barrier Fe(OH)layer 2 + 2H+between + 2e−. (5)solution, Fe(OH)2 forms a defective hydrous the specimen and the and consequently retards the anodic current density [34,35]. The corrosion current density decreased Fe(OH)2 forms a defective hydrous film barrier layer between the specimen and the solution, − 2 − 2 with decreasing DO concentration from 201.5 µA·[34,35]. cm to µA·cm 4 ppm and 0.8 ppm, and consequently retards the anodic current density The22.3 corrosion currentatdensity decreased −2 at 4 ppm and 0.8 ppm, with decreasing from 201.5 µA·cm 22.3 µA·cm respectively, as shownDO in concentration Table 1. This indicates that−2 atomore compact and stable oxide film was as shown in Table 1. This indicates that a more compact and stable oxide film was formedrespectively, at ≈−600 mV SCE at lower DO. The increase in anodic polarisation current with decreasing formed at ≈−600 mVSCE at lower DO. The increase in anodic polarisation current with decreasing DO DO at a more noble potential could be attributed to the solubility of corrosion product as temperature at a more noble potential could be attributed to the solubility of corrosion product as temperature decreases due todue thetocooling effect of of NN insufficient oxidation atsteel the surface steel surface 2 2gas decreases the cooling effect gas and and insufficient oxidation at the to formto form oxide film other through thethe Kuch mechanism. oxide film than other than through Kuch mechanism. 1.5

4ppm 1ppm 0.8ppm

1.0

E, V (SCE)

0.5

0.0

-0.5

-1.0

-1.5 -4

-3

-2

-1

2

0

1

2

log i (A/cm )

Figure 6. Potentiodynamic polarization of mild steel in 5 g/L sodium bicarbonate /chloride solution

Figure 6. Potentiodynamic polarization of mild steel in 5 g/L sodium bicarbonate /chloride solution in different dissolved oxygen concentrations at 2000 rpm. in different dissolved oxygen concentrations at 2000 rpm. Table 1. Polarization parameters for mild steel in 5 g/L bicarbonate/chloride solution in different

Table 1.dissolved Polarization for mild steel in 5 g/L bicarbonate/chloride solution in different oxygenparameters concentrations. dissolved oxygen concentrations. DO (ppm) Icorr (µAcm−2) Ecorr (mV) ba (mV dec−1) −bc (mV dec−1) 0.8

−767.5 163.0 302.9 Ecorr (mV) ba (mV·dec−1 ) −bc (mV·dec−1 ) −780.6 138.5 222.9 0.8 4 22.28 −767.5 201.52 −369.2 243.5163.0 240.3 302.9 1 25.24 −780.6 138.5 222.9 201.52 and Bode-phase −369.2angle plots for 243.5 240.3 Figure 47a,b show the Nyquist different DO concentrations in bicarbonate/chloride solution. The impedance semi-circle increases with decreasing DO, which implies thatshow decreasing DO facilitates formation ofangle a moreplots stablefor oxide layer at DO the steel surface. Figure 7a,b the Nyquist andthe Bode-phase different concentrations in A similar trend can be seen in the Bode impedance magnitude spectra (Figure 7b), where the bicarbonate/chloride solution. The impedance semi-circle increases with decreasing DO, which implies impedance increases with decreasing DO concentrations at low frequencies. The phase angle shows that decreasing DO as facilitates the formation of aAmore oxidevalue layeratat thefrequency steel surface. A similar a higher peak DO concentration decreases. higherstable phase angle low indicates trend can be seen in the Bode impedance magnitude spectra (Figure 7b), where the impedance increases a higher surface resistance at the steel. The increase in Rct and a decrease in Qdl with increasing DO 2 at 0.8 ppm to 104.7 Ω·cm2 at 4 ppm as shown in Table 2 can be attributed to oxide from 1015 Ω·cm with decreasing DO concentrations at low frequencies. The phase angle shows a higher peak as DO film formation. The film resistance (Ra) increases significantly decreasing DO concentration, concentration decreases. A higher phase angle value at low with frequency indicates a higher surface suggesting that the corrosion product at the steel surface is more dense and protective at a relatively resistance at the steel. The increase in Rct and a decrease in Qdl with increasing DO from 1015 Ω·cm2 low DO as shown in Figure 8. 2 DO (ppm) 1

22.28

Icorr (µA ·cm−2 ) 25.24

at 0.8 ppm to 104.7 Ω·cm at 4 ppm as shown in Table 2 can be attributed to oxide film formation. The film resistance (Ra ) increases significantly with decreasing DO concentration, suggesting that the corrosion product at the steel surface is more dense and protective at a relatively low DO as shown in Figure 8.

Materials 2016, 9, 748

6 of 18

Materials 2016, 9, 748

6 of 18

4.5 Table 2. Impedance parameters for mild steel in bicarbonate/chloride solution in different dissolved 50 oxygen concentrations. 4ppm 800 4.0

Materials 2016, 9, 748

46.3 53.8 63.5

Qdl

Ra

0.77 0.66 0.59

40

Qa (F·cm−2 )

440.04 90.49 33.04

5.9 × 10−304 1.39 × 10−3 203 − 3.75 ×610 of 18

3.0

1.76 × 10−3 1.18 2.5 × 10−3 6.88 2.0 × 10−3

1015 874.7 104.7

200

1.5 4.5

0

200

600

-1

0

3.0

400

1

log (f/Hz)

2

3

4

(b)

2

log(|Z|/.cm )

2

(a)

0.5 3.5

4ppm 1ppm 0.8ppm

30

1.5

200

400

600

20

10

Table 2. Impedance parameters for mild steel in bicarbonate/chloride solution in different dissolved 1.0 oxygen concentrations. 0 0

40

2.5

Figure 7. Impedance spectra for mild steel in 5 g/L bicarbonate/chloride solution in different dissolved 2.0 oxygen concentration at 2000 rpm (a) Nyquist impedance plots (b) Bode and phase angle plots.

200

50 0

Phase angle(deg)

4ppm 400 600 1ppm 800 2 0.8ppm Zre (.cm )

0

10

1.0 4.0

800

Zim (.cm )

n

(Ω·cm2 )

Phase angle(deg)

0.8 1 400 4

Rct (Ω·cm )

(F·cm−2 )

2

Rs

1ppm 0.8ppm

3.5

log(|Z|/.cm )

2

Zim (.cm )

DO600 (ppm)

4ppm 1ppm 0.8ppm2

(Ω·cm2 )

0

0.5

800

-1 0 1 2 3 4 DO Zre (.cm ) (f/Hz) Rct (Ω·cm2) Qdl (F·cm−2) Rlog a (Ω·cm2) Qa (F·cm−2) Rs (Ω·cm2) n (ppm) (a) (b) 0.8 46.3 1015 1.76 × 10−3 0.77 440.04 5.9 × 10−4 Figure 7. Impedance spectra for mild steel in 5 g/L bicarbonate/chloride solution in different dissolved Figure solution dissolved 1 7. Impedance 53.8spectra for mild 874.7steel in 5 g/L 1.18bicarbonate/chloride × 10−3 0.66 90.49in different 1.39 × 10−3 oxygen concentration at 2000 rpm (a) Nyquist impedance −3plots (b) Bode and phase angle plots. oxygen (a) Nyquist impedance (b) Bode and phase angle 3.75 plots.× 10−3 4 concentration 63.5 at 2000 rpm 104.7 6.88 × 10 plots 0.59 33.04 2

(a)

0.77 0.66 0.59

440.04 90.49 33.04

5.9 × 10−4 1.39 × 10−3 3.75 × 10−3

(b)

(b)

Epoxy

Scale

Epoxy Steel

Epoxy

Qa (F·cm−2)

Epoxy

1.76 × 10−3 1.18 × 10−3 6.88 × 10−3

Scale

1015 874.7 104.7

Ra (Ω·cm2)

Scale

46.3 53.8 63.5

n

Steel

Qdl (F·cm−2)

Epoxy

Rct (Ω·cm2)

Scale

Rs (Ω·cm2)

Scale

Steel

Steel

DO (ppm) 0.8 1 4

Steel

Table 2. Impedance parameters for mild steel in bicarbonate/chloride solution in different dissolved oxygen concentrations.

Scale

(c)

Epoxy

(b)

Steel

(a)

(b)

Figure 8. Cross-section of the scale formed at in 5 g/L bicarbonate/chloride solution for (a) 4 ppm (b) Figure 8. Cross-section of the scale formed at in 5 g/L bicarbonate/chloride solution for (a) 4 ppm 1 ppm (c) 0.8 ppm dissolved oxygen at 2000 rpm. (b) 1 ppm (c) 0.8 ppm dissolved oxygen at 2000 rpm.

(c)

Materials 2016, 9, 748

7 of 18

Materials 2016, 9, 748

7 of 18

2.3. Immersion Time

2.3. Immersion Time polarization curves showing the effect of immersion time on mild steel in Figure 9 contains Figure 9 contains polarization curves the effect conditions of immersion timerpm) on mild steel speeds. in bicarbonate/chloride solution at 0.8 ppm DOshowing under dynamic (2000 rotation bicarbonate/chloride solution at 0.8 ppm DO under dynamic conditions (2000 rpm) rotation speeds. The polarisation curves show that immersion time has a marked effect on anodic and cathodic The polarisation curves show immersion has a at marked effect on anodic and cathodic polarization current. However, the that effect was moretime apparent corrosion potentials between (−250 and polarization current. However, the effect was more apparent at corrosion potentials between + 250 mVSCE ), where anodic potential current decreased with immersion time. The curve(−250 for each and + 250 mVSCE), where anodic potential current decreased with immersion time. The curve for each immersion time contains one anodic peak and, as expected, corrosion products on the steel surface immersion time contains one anodic peak and, as expected, corrosion products on the steel surface become thicker and more compact with immersion time. Siderite can form in relatively low flow become thicker and more compact with immersion time. Siderite can form in relatively low flow velocities at pH 5,>and thethe film thickness time[16]. [16].The The passivation potential velocities at > pH 5, and film thicknessincreases increases with with time passivation potential rangerange becomes broader withwith increasing immersion thepitting pittingbreakdown breakdown potential Eb increasing becomes broader increasing immersiontime, time, with with the potential Eb increasing from from ≈−355 mV at 2 h to ≈− 229 mV after 8 h immersion time as shown in Table 3. ≈−355 mV SCE at 2 h to ≈−229 mVSCESCE after 8 h immersion time as shown in Table 3. SCE 1.5

2h 4h 8h

1.0

E,V (SCE)

0.5

0.0

-0.5

-1.0

-1.5 -5

-4

-3

-2

-1

0

1

2

2

log i (A/cm )

Figure 9. Potentiodynamic polarization for mild steel in 5 g/L bicarbonate/chloride solution at

Figure 9. Potentiodynamic polarization for mild steel in 5 g/L bicarbonate/chloride solution at different immersion times at 2000 rpm. different immersion times at 2000 rpm. Table 3. Polarization parameters for mild steel in 5 g/L bicarbonate/chloride solution for different

Tableimmersion 3. Polarization times atparameters 2000 rpm. for mild steel in 5 g/L bicarbonate/chloride solution for different immersion times at 2000 rpm. Immersion Ecorr Icorr Eb ba (mV dec−1) -bc (mV dec−1) (µA·cm−2− (mV) Time (h) ) 2 (mV) Immersion Time (h) Eb (mV) Ecorr (mV) Icorr (µA·cm ) ba (mV·dec−1 ) −bc (mV·dec−1 ) 2 22.28 −355.0 −767.5 163.0 302.9 2 22.28 −355.0 −767.5 163.0 302.9 4 8.98 −292.3 −820.4 126.6 138.0 4 8.98 −292.3 −820.4 126.6 138.0 8 6.04 −229.1 134.8 8 6.04 −229.1−833.9 −833.9 145.6 145.6 134.8

Figure 10 shows the Nyquist plots and Bode phase angle plots of mild steel in

Figure 10 shows the Nyquist plots and Bode angle plots of times mild steel in rpm. bicarbonate/chloride bicarbonate/chloride solution at 0.8 ppm DO atphase different immersion at 2000 The Nyquist plots that theimmersion diameter oftimes the semi-circle increases with immersion time, which solution atin 0.8Figure ppm 10a DO show at different at 2000 rpm. The Nyquist plots in Figure 10a show implies that the steel surface resistance improves with immersion time. The result is consistent with that the diameter of the semi-circle increases with immersion time, which implies that the steel surface more corrosion product being deposited on the the steel surface, and forming a relatively inert surface resistance improves with immersion time. The result is consistent with more corrosion product being barrier at the steel–solution interface. The charge transfer resistance (Rct) increased from deposited on the the steel surface, and forming a relatively inert surface barrier at the steel–solution 1015 Ω·cm2 to 2700 Ω·cm2 in going from 2 h to 8 h immersion time as shown in Table 4, presumably interface. The charge transfer resistance (Rct ) increased from 1015 Ω·cm2 to 2700 Ω·cm2 in going from as a result of deterioration of the oxide film on the steel surface. A similar result was reported in our 2 h toprevious 8 h immersion time as shown in Table 4, presumably as a result of deterioration of the oxide film work [31]. The Bode impedance magnitude plots show similar trends. The impedance on the steel surface. A similar resultincrease was reported in our previous [31]. The Bode magnitude and the phase angle with immersion time as work evident in Bode plots impedance at low magnitude plotsasshow similar trends. The impedance magnitude and the phase angle increase frequencies shown in Figure 10b, implying that the corrosion product at the steel surface becomes with immersion as evident in Bode at lowtime, frequencies as shown in Figure 10b, implying densertime and more protective withplots immersion as is evident in Figure 11. It would thereforethat be the expected that metal dissolution would be retarded, in agreement with the data obtained from corrosion product at the steel surface becomes denser and more protective with immersion time, as polarization curves. is evident in Figure 11. It would therefore be expected that metal dissolution would be retarded, in agreement with the data obtained from polarization curves.

Materials 9, 748 748 Materials 2016, 2016, 9,

88 of of 18 18

Materials 9, 748 8 of 18 Table2016, 4. Impedance parameters for mild steel in bicarbonate/chloride solution for different immersion

Table times. 4. Impedance parameters for mild steel in bicarbonate/chloride solution for different Table 4. Impedance parameters for mild steel in bicarbonate/chloride solution for different immersion immersion times. Immersion Rct Ra Rs times. Qdl (F·cm−2) Qa (F·cm−2) n 2 2 2 2 2 − 2 (Ω·cm Time (h) ) ) (Ω·cm Immersion Time (h) n(Ω·cm Rs (Ω·cm Rct (Ω)·cm ) Qdl (F·cm ) Ra) (Ω·cm2 ) Qa (F·cm−2 )

4

1600 1800

2h 4h 8h2h 4h 8h

4

Zim (.cm )

8001000

2

600 800 400 600

3

60 50 50 40

3

40 30

2

30

2

20

1

200 400

10

1

0 200 0 0

200 0

400 200

600 400

800

1000

1200

2

600 Zre800 (.cm1000 )

1400

1200

1400

1600 1600

1800

-1

1800

-1

0 0

2

1 1

2

log (f/Hz) 2

3 3

4

20 10 0

0

4

log (f/Hz)

(a)Zre (.cm ) (a)

(b)

(b)

(c) (c)

Epoxy

Epoxy

Scale

Epoxy Epoxy

Scale Scale

(b) (b)

Steel Steel

(a) (a)

Scale

Steel Steel

Epoxy Epoxy

Scale Scale

Steel

Figure 10. Impedance mild steel steel in in 55 g/L g/L bicarbonate/chloride bicarbonate/chloride solution different Figure 10. Impedance spectra spectra for for mild solution at at different Figure 10. Impedance spectra for mild steel in 5 g/L bicarbonate/chloride solution at different immersion times at 2000 rpm (a) Nyquist impedance plots (b) Bode and phase angle plots. immersion times at 2000 rpm (a) Nyquist impedance plots (b) Bode and phase angle plots. immersion times at 2000 rpm (a) Nyquist impedance plots (b) Bode and phase angle plots.

Steel

2

10001200

70 60

2h 4h 2h8h 4h 8h

2

12001400

2 log(|Z|/ cm  ) cm ) log (|Z|/

1400 1600

Zim (.cm )

Ra −2)−4 1.76 × 10×−2−3)10−30.77 × 10 dl (F·cm (F·cm Q Qa5.9 n 0.77440.04 1.76 5.9 × 10−4 2) 440.04 (Ω·cm −4 1.651.65 × 10×−3−310−3 1 804.0 5.04 × 10 1 804.0 5.04 × 10−4 1.76 × 10 −4 −40.77 440.04 5.9 × 10−4 −4 0.82 618.6 7.28 × 10 5.83 7.28 × 10−3 0.82 618.6 5.83 × 10 × 10−4 1.65 × 10 1 804.0 5.04 × 10−4 7.28 × 10−4 0.82 618.6 5.83 × 10−4 70

Figure 11. Cross-section of the scale formed in 5 g/L sodium bicarbonate/chloride solution at 2000 rpm Figure 11. Cross-section of the scale formed in 5 g/L sodium bicarbonate/chloride solution at 2000 rpm Figure formedtimes. in 5 g/L sodium bicarbonate/chloride solution at 2000 rpm after 11. (a) Cross-section 2 h (b) 4 h andof (c)the 8 hscale immersion after (a) 2 h (b) 4 h and (c) 8 h immersion times. after (a) 2 h (b) 4 h and (c) 8 h immersion times.

Phase angle (deg)

4 8

1800

Rct 1015 1015 2) (Ω·cm 1370 1370 1015 2700 2700 1370 2700

Rs 46.3 46.3 2) (Ω·cm 30.5 30.5 46.3 41.6 41.6 30.5 41.6

Phase angle (deg)

Immersion

2 2 Time (h) 4 4 2 8 8

Materials 2016, 9, 748

9 of 18

MaterialsAnalysis 2016, 9, 748 2.4. Surface

9 of 18

Materials 2016, 9, 748

9 of 18

Figure 12 shows scanning electron micrographs obtained from the surfaces of samples immersed 2.4. Surface Analysis in 1 g/L and 5 g/L bicarbonate chloride solutions at 2000 rpm. The SEM results show that pits with 2.4. Surface Analysis Figure 12 shows scanning electron micrographs obtained from the surfaces of samples immersed relatively small diameters of a few microns were predominant with specimen in low bicarbonate in 1 Figure g/L and g/L bicarbonate chloridemicrographs solutions atobtained 2000 rpm. Thethe SEM results thatimmersed pits with 125shows scanning electron from surfaces of show samples concentration. The density of pits decreased inwere theatsample exposedSEM to specimen 5results g/L bicarbonate solution, but relatively small of achloride few microns predominant in low bicarbonate in 1 g/L and 5 g/Ldiameters bicarbonate solutions 2000 rpm. Thewith show that pits with the pits became much larger both in depth and width. Figure 13a–c show the surface morphologies concentration. density of of apits decreased the sample exposed to 5specimen g/L bicarbonate but relatively smallThe diameters few microns in were predominant with in low solution, bicarbonate of specimens afterThe corrosion in solution at bicarbonate different oxygen the pits became much larger bothdecreased in bicarbonate/chloride depth and width. Figure 13a–c the surfacedissolved morphologies concentration. density oftests pits in the sample exposed to show 5 g/L solution, but (DO)the concentrations. The micrographs revealed theFigure corrosion with decreasing of specimens after corrosion tests solution atdamage different dissolved oxygen (DO) pits became much larger both in in bicarbonate/chloride depth and that width. 13a–c show thedecreases surface morphologies concentrations. The micrographs that the damage decreases with decreasing of specimens after corrosion tests in revealed bicarbonate/chloride solution at different dissolved oxygen dissolved oxygen due to more tenacious oxide filmcorrosion on the steel–solution interface, as (DO) shown in dissolved oxygen due to more tenacious oxide film on the steel–solution interface, as shown in Figure concentrations. The micrographs revealed that the corrosion damage decreases with decreasing Figure 13c for solution containing 0.8 ppm DO. Our cross-sectional images (Figure 11) show that the 13c forthickness solution 0.8 ppm DO.oxide Our film cross-sectional images (Figure 11) show that theFigure oxide oxygencontaining due to more tenacious onthe thefilm steel–solution interface, as shown in oxidedissolved film increases with immersion time, morphology also varies with immersion filmfor thickness with time,cross-sectional the film morphology also varies time 13c solutionincreases containing 0.8immersion ppm DO. Our images (Figure 11) with showimmersion that the oxide time as shown in Figure 14 providing a barrier layer at the metal–solution interface, and as a result as shown in Figure 14 providing a barrier layer themorphology metal–solution andimmersion as a resulttime the film thickness increases with immersion time, theatfilm alsointerface, varies with the surface damage also decreases, ranging from pitting corrosion to general corrosion, as shownin surface also corrosion to general corrosion, shownin as showndamage in Figure 14 decreases, providing aranging barrier from layer pitting at the metal–solution interface, and as aasresult the Figure 15. 15. Figure surface damage also decreases, ranging from pitting corrosion to general corrosion, as shownin Figure 15.

Figure 12. Scanning Electron Microscopy micrographs of mild steel in 5 g/L bicarbonate/chloride

Figure 12. Scanning Electron Microscopy micrographs of mild steel in 5 g/L bicarbonate/chloride solution 2000 rpm:Electron (a) 1 g/LMicroscopy and (b) 5 g/Lmicrographs bicarbonate of solution. Figure 12.atScanning mild steel in 5 g/L bicarbonate/chloride solution at 2000 rpm: (a) 1 g/L and (b) 5 g/L bicarbonate solution. solution at 2000 rpm: (a) 1 g/L and (b) 5 g/L bicarbonate solution.

(a) (a)

(b) (b)

(c) (c)

Figure 13. Scanning Electron Microscopy micrograph in 5 g/L bicarbonate/chloride solution at 2000 rpm: (a) 4 ppm (b) 1 ppmElectron (c) 0.8 ppm dissolve oxygen. in 5 g/L bicarbonate/chloride solution at 2000 rpm: Figure 13. Scanning Microscopy micrograph Figure 13. Scanning Electron Microscopy micrograph in 5 g/L bicarbonate/chloride solution at (a) 4 ppm (b) 1 ppm (c) 0.8 ppm dissolve oxygen.

2000 rpm: (a) 4 ppm (b) 1 ppm (c) 0.8 ppm dissolve oxygen.

Materials 2016, 9, 748

10 of 18

Materials 2016, 9, 748

10 of 18

Materials 2016, 9, 748

10 of 18

(a) (a)

(b) (b)

(c) (c) Figure 14. Scanning Electron Microscopy micrographs of mild steel in 5 g/L bicarbonate/chloride Figure 14. Scanning Electron Microscopy micrographs of mild steel in 5 g/L bicarbonate/chloride solution14. at 2000 rpm before cleaning: (a) 2 hmicrographs (b) 4 h (c) 8 hofimmersion Figure Scanning Electron Microscopy mild steeltime. in 5 g/L bicarbonate/chloride solution at 2000 rpm before cleaning: (a) 2 h (b) 4 h (c) 8 h immersion time. solution at 2000 rpm before cleaning: (a) 2 h (b) 4 h (c) 8 h immersion time.

(a) (a)

(b) (b)

(c) (c)

Figure 15. Scanning Electron Microscopy micrographs of mild steel in 5 g/L bicarbonate/chloride Figure Scanning Electron Microscopy micrographs of mild steel in 5 g/L bicarbonate/chloride solution15. at 2000 rpm after cleaning: (a) 2 h (b) 4 h (c) 8 h immersion time. Figure 15. Scanning Electron Microscopy micrographs of mild steel in 5 g/L bicarbonate/chloride solution at 2000 rpm after cleaning: (a) 2 h (b) 4 h (c) 8 h immersion time.

solution at 2000 rpm after cleaning: (a) 2 h (b) 4 h (c) 8 h immersion time.

decreases while the depth increases with bicarbonate concentration. The pit depth increases from ≈−8 µm to ≈−60 µm for 1 g/L and 5 g/L, respctively, as shown in Figure 16. The corrosion pit density and pit depth in specimens immersed in 1 g/L and 5 g/L bicarbonate were further studied using Leica optical microscopy. The results show that the corrosion pit density decreases while Materials 2016, 9, 748the depth increases with bicarbonate concentration [36]. The pit depth increases 11from of 18 ≈8 µm to ≈60 µm for 1 g/L and 5 g/L, respctively, as shown in Figure 16. The dissolution of iron is accelerated in the presence of bicarbonate owing to the formation of the stable soluble complex anion 2− as reported The3)2corrosion pit density pit depth in specimens g/Lmetal and 5 surface, g/L bicarbonate Fe (CO in [3]. and As more corrosion productimmersed is formedinat1 the galvanic were further studied using Leica optical microscopy. The results show that the corrosion pit density couples could occur between steel surfaces covered with corrosion products (cathode) and uncovered decreases theThis depth increases with bicarbonate concentration. Theinitiation pit depthand increases from active site while (anode). unstable thermodynamic condition accelerates the propagation ≈− 8 µm to ≈− 60 µm for 1 g/L and 5 g/L, respctively, as shown in Figure 16. of localized corrosion as shown in Figure 17.

(a)

(b)

8 µm (c)

(d)

(e)

60 µm (f) Figure 16. Topography profile and 3D image of pit affected area: (a–c) 1 g/L and (d–f) 5 g/L bicarbonate/chloride solution.

The corrosion pit density and pit depth in specimens immersed in 1 g/L and 5 g/L bicarbonate were further studied using Leica optical microscopy. The results show that the corrosion pit density decreases while the depth increases with bicarbonate concentration [36]. The pit depth increases from ≈8 µm to ≈60 µm for 1 g/L and 5 g/L, respctively, as shown in Figure 16. The dissolution of iron is

Materials 2016, 9, 748

12 of 18

accelerated in the presence of bicarbonate owing to the formation of the stable soluble complex anion 2− as reported in [3]. As more corrosion product is formed at the metal surface, galvanic Fe (CO 3 )2 2016, Materials 9, 748 12 of 18 couples could occur between steel surfaces covered with corrosion products (cathode) and uncovered 16. Topography profile and 3D image of pit affected area: (a–c) the 1 g/L and (d–f) 5 g/L active siteFigure (anode). This unstable thermodynamic condition accelerates initiation and propagation bicarbonate/chloride solution. of localized corrosion as shown in Figure 17.

Cathode

Pit

Anode Rust layer Steel

(b)

(a) Pit

Cathode

Anode

Rust layer

(c)

(d)

Figure 17. Schematic mechanismofofbicarbonate bicarbonate concentration concentration effect: 5 g/L NaHCO 3 (c,d) 1 g/L Figure 17. Schematic mechanism effect:(a,b) (a,b) 5 g/L NaHCO 3 (c,d) 1 g/L NaHCO3 concentration. NaHCO3 concentration.

2.5. XRD Analyses

2.5. XRD Analyses

Figure 18 shows the XRD patterns for specimens immersed in bicarbonate chloride containing Figure 18 shows the XRD18a–c patterns immersed in bicarbonate chloride containing solution at 2000 rpm. Figure showfor thespecimens peak patterns for unimmersed and immersed specimens solution at 2000bicarbonate rpm. Figure 18a–c show the patterns for unimmersed and immersed specimens in different concentrations. The peak elemental composition of the corrosion products tested under different bicarbonate concentrations show similarities in their in different bicarbonate concentrations. The elemental composition of thechemical corrosioncomposition. products tested However, XRD patterns in Figure 18b indicate diffraction in peaks mainly Fecomposition. (iron) (44.3%) However, and under differentthe bicarbonate concentrations show similarities theirofchemical a single peak for -FeO(OH) (akaganeite) (16.32%) and -FeO(OH) (lepidocrocite) (39.34%), in peak the XRD patterns in Figure 18b indicate diffraction peaks of mainly Fe (iron) (44.3%) and a single contrast to the peak patterns in Figure 18c, with multiple peaks of both ( -FeO(OH) goethite (58.16%) for β-FeO(OH) (akaganeite) (16.32%) and γ-FeO(OH) (lepidocrocite) (39.34%), in contrast to the peak and -FeO(OH) (akaganeite) (6.20%). The multiple peaks could probably be due to the density of the patterns in Figure 18c, with multiple peaks of both (α-FeO(OH) goethite (58.16%) and β-FeO(OH) corrosion products on the steel surface. The formation of akaganeite confirms the presence of chloride (akaganeite) (6.20%). multiple peaks probably be dueoftoakaganeite the density of specimen the corrosion ions in the corrosionThe products [31,37]. Thecould relatively high amount in the products on the The formation of akaganeite confirms the presence of chloride ions in immersed in steel 1 g·L−1surface. bicarbonate/chloride solution implies that the corrosion product contains more the corrosion products [31,37]. relatively amount of akaganeite in the specimen in chloride ions. Figure 18d,e The show the XRDhigh peak patterns for specimens immersed in immersed 5 g·L−1 −1 bicarbonate/chloride 1 g·Lbicarbonate/chloride solution implies that the corrosion product contains chloride ions. containing different dissolved oxygen concentrations. Goethitemore ( -FeO(OH)), siderite 3 and -Fe2O3) for were formed at immersed each DO concentration. However, more Figure 18d,eFeCO show the maghemite XRD peak (patterns specimens in 5 g·L−1 bicarbonate/chloride siderite different FeCO3 anddissolved maghemiteoxygen ( -Fe2O3concentrations. ) peaks are evidentGoethite as DO decreases, as shown in Figure 18e. 3 and containing (α-FeO(OH)), siderite FeCO

Materials 2016, 9, 748

13 of 18

maghemite (γ-Fe2 O3 ) were formed at each DO concentration. However, more siderite FeCO3 and maghemite (γ-Fe Materials 2016, 9, 748 13 of 18 2 O3 ) peaks are evident as DO decreases, as shown in Figure 18e. 1000





3500 800



3000

Fe

  Fe

  FeO(OH) 600

Intensity (a.u)

Intensity (a.u)

2500

400

 200

  FeO(OH)   Fe2O3

2000 1500 1000





500

 





0

0 0

20

40

60

80

100

120

0

20

40

60

(a)

  Fe

3

750



Intensity (a.u)

  FeO(OH)   FeO(OH)   Fe O 2 3

1000







250



 Fe  FeO(OH)  FeCO

500

 FeO(OH)

500

120



600 1250



100

(b)

1500



80

2 Theta ()

2 Theta ()

Intensity (a.u)



 Fe2O3

400 300 200 100



0











0 0

20

40

60

80

100

120

0

20

40

2 Theta (°)

60

80

100

120

2 Theta ()

(c)

(d)

700

 600

 Fe  FeO(OH)  FeCO

Intensity (a.u)

500

3

400

 Fe2O3

300 200 100





 



0 0

20

40

60

80

100

120

2 Theta ()

(e) Figure 18. XRD analysis of corrosion products of mild steel in bicarbonate solutions containing

Figure 18. XRD analysis of corrosion products of mild steel in bicarbonate solutions containing chloride chloride at 2000 rpm: (a) unimmersed (b) 1 g/L and (c) 5 g/L bicarbonate (d) 4 ppm (e) 1 ppm. at 2000 rpm: (a) unimmersed (b) 1 g/L and (c) 5 g/L bicarbonate (d) 4 ppm (e) 1 ppm.

2.6. XPS Analyses

2.6. XPS Analyses

Prior to XPS measurements, samples were sputtered for 60 min using 2 keV Ar+ ions to remove + ions to remove Prior to XPS measurements, samples were the sputtered for 60chemistry min using keV Ar adsorbed surface contamination and reveal underlying in2more detail. Argon sputtering is known to convertand FeCO 3 to FeO and, correspondingly, diddetail. not observe FeCO 3, adsorbed surface contamination reveal the[38], underlying chemistry in we more Argon sputtering whichto can be identified its high-binding-energy contribution to Cnot 1s, in our spectra. spectra is known convert FeCO3from to FeO [38], and, correspondingly, we did observe FeCOAll , which can be 3 were calibrated to the C-C adventitious carbon peak at 284.8 eV. In each spectrum, the dominant C identified from its high-binding-energy contribution to C 1s, in our spectra. All spectra were calibrated 1sC-C component was identified adventitious C-C.InIneach each spectrum, case, this peak the expected to the adventitious carbonas peak at 284.8 eV. thehad dominant C 1s companion component was C-O/C=O peaks at higher binding energy. The O 1s spectra of the two samples (Figure 19) were identified as adventitious C-C. In each case, this peak had the expected companion C-O/C=O peaks at similar, and could be deconvolved into two peaks: a relatively sharp peak near

higher binding energy. The O 1s spectra of the two samples (Figure 19) were similar, and could be

Materials 2016, 9, 748 Materials 2016, 9, 748

14 of 18 14 of 18

deconvolved peaks: a relatively sharp peak peak centred near 530near eV (dark gray component) and a broader 530 eV (dark into graytwo component) and a broader 531 eV (light gray component). The 2− peak centred near 531 eV (light gray component). The lower binding energy peak has the expected lower binding energy peak has the expected shape and location for O in iron oxides (Fe3O4, α− in iron oxides (Fe O , α-FeO(OH) and FeO). The broader, higher shape andand location O2broader, binding FeO(OH) FeO).for The higher binding in 3 4 energy peak is consistent with the OH− oxygens − energy peak as is consistent with OH oxygens infrom α-FeO(OH), as well as with C-O contribution α-FeO(OH), well as with thethe C-O contribution adventitious carbon. Thethe analysis of Fe 2p is 2+ from adventitious The analysis of Feof2p complicated due to thecompounds multiplet structure of FeFe2p complicated due tocarbon. the multiplet structure Feis2p peaks for high-spin comprising 2+ 3+ 3+ peaks high-spin compounds comprising Fein aand Fe asymmetric [39]. The multiplet structure manifests in and Fefor[39]. The multiplet structure manifests broad, peak shape, particularly when a broad, asymmetric peak shape, particularly with unmonochromated radiation as measured with unmonochromated radiation aswhen in themeasured present measurements, making deconvolution in contributions the present measurements, contributions fromonly multiple chemical states of from multiplemaking chemicaldeconvolution states difficult,ofand hence we offer a qualitative analysis difficult, hence we offer only a qualitative analysis of the 2p spectra shown in Figure 19. As with of the Feand 2p spectra shown in Figure 19. As with O 1s, the Fe samples show similar peak shapes. The O 1s, the samples show similar peak shapes. The spectral weight in each sample is centred near 710 eV, spectral weight in each sample is centred near 710 eV, consistent with contributions from Fe3O4 and consistent with contributions from Fe3 O4 and α-FeO(OH). α-FeO(OH).

Figure 19. 19.XPS XPScollected collected from Fe (left) 2p (left) 1s (right) regions of Samples: 1 (b) g/L5 and Figure from thethe Fe 2p and and O 1s O (right) regions of Samples: (a) 1 g/L(a) and g/L. (b) g/L. 2p presented data are presented background subtraction, and without deconvolution The5 Fe 2pThe dataFeare without without background subtraction, and without deconvolution of the of the peaks. O 1shave datahad havea had a Shirley background subtracted and have deconvolved into peaks. The OThe 1s data Shirley background subtracted and have been been deconvolved into two two contributions. indicated positions for different chemical states approximate, based contributions. The The indicated positions for different chemical states are are approximate, andand are are based on on values taken from [38,39]. values taken from [38,39].

3. Experimental ExperimentalSection Section 3.1. Materials Materials and and Solutions Solutions 3.1. carbon steel steel rod ( (∅ 16 mm) of grade AISI 1020 was utilized as the rotating disc Cylindrical carbon electrode (RDE). composition of the rod was using a Field electrode (RDE).The Thechemical chemical composition of as-received the as-received rodanalyzed was analyzed usingEmission a Field Electron Probe Microanalyzer (JXA-8530F) with results shown in Table 5. The test samples were Emission Electron Probe Microanalyzer (JXA-8530F) with results shown in Table 5. The test samples polished usingusing emery papers of grit ranging fromfrom 220 220 to 800. TwoTwo solutions were prepared by were polished emery papers ofsizes grit sizes ranging to 800. solutions were prepared dissolving analytical grade sodium bicarbonate (1 g, 5 g) and 2 g of sodium chloride in 1 L of distilled by dissolving analytical grade sodium bicarbonate (1 g, 5 g) and 2 g of sodium chloride in water. The solutions deaeratedwere by bubbling nitrogen gas three hoursgas before 1deionized L of distilled deionized water.were The solutions deaerated by bubbling nitrogen threesample hours immersion to study the effect of bicarbonate immersion time. To study the effects before sample immersion to study the effectconcentrations of bicarbonateand concentrations and immersion time. To of dissolved oxygen, nitrogen gas was bubbled for 0 h, 2 h and 3 h which resulted in 4 ppm, 1 ppm and study the effects of dissolved oxygen, nitrogen gas was bubbled for 0 h, 2 h and 3 h which resulted 0.8 4ppm DO1 concentrations, respectively, before sample immersion.before The electrochemical tests were in ppm, ppm and 0.8 ppm DO concentrations, respectively, sample immersion. The performed at ambient temperature (22at±ambient 3 ◦ C) at temperature pH (8.2 ± 0.1). electrochemical tests were performed (22 ± 3 °C) at pH (8.2 ± 0.1).

Materials 2016, 9, 748

15 of 18

Table 5. Chemical composition of mild steel (wt %). C

Si

Mn

Cr

Cu

Ni

Fe

0.20

0.32

0.79

0.01

0.01

0.01

Bal

3.2. Electrochemical Test The corrosion tests were carried out in a 1 L glass cell containing the rotating disk electrode (working electrode) fitted into rotating shaft of an analytical rotator AFASRE 747 (Pine Instrument). The reference electrode was a saturated calomel electrode (SCE) connected to the cell by a bridge and a Lugging capillary Platinum was used as counter electrode. The angular speed (ω) of the rotating disc electrode was controlled by the ASR RDE speed controller. The rotation speed of the RDE performed in the tests was 2000 rev/min, respectively. Corrosion evaluation was monitored by potentiodynamic polarization and electrochemical impedance spectroscopy (EIS). The tests were performed at a scan rate of (0.167 mV/s) using a Bio-Logic instrument Model VSP 0508 potentiostat. The potentiodynamic polarisation was obtained from a starting potential of −0.30 V vs. SCE to a final potential of +1.20 V vs. SCE. Electrochemical impedance spectroscopy measurements were conducted at open-circuit. Measuring frequencies ranged from 10 kHz to 0.1 Hz with a perturbing alternating current (AC) amplitude of 10 mV and a sampling rate of 10 points per decade. 3.3. Surface Analysis Scanning electron microscopy (SEM) imaging was performed after corrosion tests using a Zeiss Sigma VP Field Emission Scanning Electron Microscope (Oxford XMax 50 silicon drift energy dispersive spectroscopy EDS detector) in secondary electron (SE) image mode with an accelerating beam voltage of 15 kV. Pit volume and pit depth were examined using Leica Optical microscopy Model DFC 490. 3.4. X-ray Diffraction The corrosion products were characterized in-situ using a PANalytical X’pert PRO MPD powder X-ray Diffractometer (XRD) equipped with a 40 keV, 40 mA, Co Kα . A parabolic mirror optic on the incident side, 0.04 rad Soller slits on the source and detector side, and 0.09◦ collimator before the detectors were used to give a parallel beam. The incident angle (ω) was fixed to 3◦ and data collected from 6◦ to 90◦ (2θ). The other slits were optimized to ensure the beam footprint did not exceed the sample dimensions or the detector window opening. Quantitative phase analysis was performed in TOPAS (V5, Bruker). An instrument function collected from SRM 660a was used to accurately model peak shape and width. Fixed incident parallel beam intensity and peak width corrections according to Rowles and Madsen [40], Toraya et al. [41] and Haggerty et al. [42] were employed. The values of these corrections were fixed to those for the instrument function when refining the sample data. Other parameters refined included background, scale factors for each phase, specimen displacement, unit cell parameters for each phase, and a Lorentzian crystallite size term for each phase. 3.5. X-ray Photoelectron Spectroscopy XPS measurements were performed using a non-monochromatized Al Kα (1486.7 eV) source (DAR 400, ScientaOmicron GmbH) with a 125 mm hemispherical electron energy analyser (Sphera II, 7 channel detector, ScientaOmicron GmbH). The XPS was housed in an ultrahigh vacuum chamber with a base pressure of 10−11 mbar. XPS data were collected over Fe, C and O core levels at a pass energy of 20 eV using 0.1 eV steps and a dwell time of 0.1 s. The O 1s core level was fit using symmetric components after subtracting a Shirley background.

Materials 2016, 9, 748

16 of 18

4. Conclusions The electrochemical behavior of mild steel in different HCO3 − , DO concentrations and immersion time was studied, using potentiodynamic polarization and electrochemical impedance spectroscopy. Pitting corrosion depth increases with increasing bicarbonate concentrations. Lowering the amount of dissolved oxygen facilitates passivation, whereas an increase in dissolved oxygen and HCO3 − in bicarbonate/chloride results in a significant increase in the cathodic oxygen reduction process and anodic metal dissolution due to oxidation. The corrosion current density decreases significantly with decreasing DO concentration, dropping from 201.5 µA·cm−2 at 4 ppm to 22.3 µA·cm−2 at 0.8 ppm. A relatively stable oxide films were formed at ~600 mVSCE at lowered DO (1–0.8 ppm) concentrations. Immersion time has an important effect: the oxide film thickness increases and the corrosion rate decreases with increasing immersion time. Internal corrosion of mild steel pipelines in high bicarbonate/chloride environment can be minimized when DO concentration is relatively low at pH ≥ 8 and potential ≈ −600 mVSCE . Oxygen is necessary for passive film formation, but can be detrimental to mild steel above the critical level in high bicarbonate/chloride solution under dynamic conditions. Acknowledgments: This work was funded by the Australian Government and Queensland University of Technology. The authors express their gratitude for the financial support and some characterizations performed at the Central Analytical Research Facility (CARF) operated by the Institute for Future Environments (QUT). Author Contributions: Geoffrey Will and Willem Dekkers are the supervisory team, both helped in designing the experiment, Jennifer MacLeod analysed the XPS data, Gaius Debi Eyu performed the experiments and the data analyses. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

13.

King, F. Overview of a Carbon Steel Container Corrosion Model for a Deep Geological Repository in Sedimentary Rock; Nuclear Waste Management Organization Technical Report: Toronto, ON, Canada, 2007. Castro, E.; Vilche, J.; Arvia, A. Iron dissolution and passivation in K2 CO3 /KHCO3 solutions. Rotating ring disc electrode and XPS studies. Corros. Sci. 1991, 32, 37–50. Davies, D.; Burstein, G. The effects of bicarbonate on the corrosion and passivation of iron. Corrosion 1980, 36, 416–422. [CrossRef] Linter, B.; Burstein, G. Reactions of pipeline steels in carbon dioxide solutions. Corros. Sci. 1999, 41, 117–139. [CrossRef] Tan, Y.T.; Wijesinghe, S.L.; Blackwood, D.J. The inhibitive effect of bicarbonate and carbonate ions on carbon steel in simulated concrete pore solution. Corros. Sci. 2014, 88, 152–160. [CrossRef] Moreno, M.; Morris, W.; Alvarez, M.G.; Duffó, G.S. Corrosion of reinforcing steel in simulated concrete pore solutions: Effect of carbonation and chloride content. Corros. Sci. 2004, 46, 2681–2699. [CrossRef] Might, J.; Duquette, D. Stress corrosion cracking of high-purity carbon steel in carbonate solutions. Corrosion 1996, 52, 428–434. [CrossRef] Pilkey, A.; Lambert, S.; Plumtree, A. Stress corrosion cracking of X-60 line pipe steel in a carbonate-bicarbonate solution. Corrosion 1995, 51, 91–96. [CrossRef] Mao, X.; Liu, X.; Revie, R. Pitting corrosion of pipeline steel in dilute bicarbonate solution with chloride ions. Corrosion 1994, 50, 651–657. [CrossRef] Thomas, J.; Davies, J. Influence of hydrogen carbonate and chloride ions on the stability of oxide films on mild steel in near-neutral solutions. Br. Corros. J. 1977, 12, 108–114. [CrossRef] Abdulrahman, A.; Ismail, M.; Hussain, M.S. Inhibition of corrosion of mild steel in hydrochloric acid by Bambusa arundinacea. Int. Rev. Mech. Eng. 2011, 5, 59–63. Lu, Z.; Huang, C.; Huang, D.; Yang, W. Effects of a magnetic field on the anodic dissolution, passivation and transpassivation behaviour of iron in weakly alkaline solutions with or without halides. Corros. Sci. 2006, 48, 3049–3077. [CrossRef] Xue, H.; Cheng, Y. Passivity and pitting corrosion of X80 pipeline steel in carbonate/bicarbonate solution studied by electrochemical measurements. J. Mater. Eng. Perform. 2010, 19, 1311–1317. [CrossRef]

Materials 2016, 9, 748

14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

26.

27. 28. 29.

30. 31.

32. 33. 34. 35. 36. 37.

17 of 18

Cáceres, L.; Vargas, T.; Herrera, L. Influence of pitting and iron oxide formation during corrosion of carbon steel in unbuffered NaCl solutions. Corros. Sci. 2009, 51, 971–978. [CrossRef] Gadala, I.M.; Alfantazi, A. Electrochemical behavior of API-X100 pipeline steel in NS4, near-neutral, and mildly alkaline pH simulated soil solutions. Corros. Sci. 2014, 82, 45–57. [CrossRef] Kermani, M.; Morshed, A. Carbon dioxide corrosion in oil and gas production-A compendium. Corrosion 2003, 59, 659–683. [CrossRef] Han, J.; Zhang, J.; Carey, J.W. Effect of bicarbonate on corrosion of carbon steel in CO2 saturated brines. Int. J. Greenh. Gas Control 2011, 5, 1680–1683. [CrossRef] Yunovich, M.; Xia, Z.; Szklarska-Smialowska, Z. Factors influencing stress corrosion cracking of carbon steel in diluted bicarbonate environments. Corrosion 1998, 54, 155–161. [CrossRef] Delanty, B.; O’Beirne, J. Major field study compares pipeline SSC with coatings. Oil Gas J. 1992, 90, 39–44. Gedge, G. Corrosion of cast Iron in potable water service. In Proceedings of the Institute of Materials Conference, London, UK, 1993; pp. 18–28. Abdallah, M. Corrosion behaviour of 304 stainless steel in sulphuric acid solutions and its inhibition by some substituted pyrazolones. Mater. Chem. Phys. 2003, 82, 786–792. [CrossRef] Finnegan, T.J.; Corey, R.C. Corrosion of Steel Quantitative Effect of Dissolved Oxygen and carbon Dioxide. Ind. Eng. Chem. 1935, 27, 774–780. [CrossRef] Oldfield, J.; Todd, B. Corrosion considerations in selecting metals for flash chambers. Desalination 1979, 31, 365–383. [CrossRef] Oldfield, J.W.; Todd, B. Technical and economic aspects of stainless steels in MSF desalination plants. Desalination 1999, 124, 75–84. [CrossRef] Beckett, M.; Beckett, M.; Snoeyink, V.; Jim, K.; Sarin, P.; Kriven, W.; Lytle, D.; Clement, J. A pipe loop system for evaluating iron uptake in distribution systems. In Proceedings of the Water Quality Technology Conference, San Diego, CA, USA, 1–4 November 1998. Sarin, P.; Sarin, P.; Bebee, J.; Beckett, M.; Jim, K.; Lytle, D.; Clement, J.; Kriven, W.; Snoeyink, V. Mechanism of release of iron from corroded iron/steel pipes in water distribution systems. In Proceedings of the Annual Technology Confrence, Denver, CO, USA, 11–15 June 2000. Mcneiill, L.S.; Edwards, M. Iron pipe corrosion in distribution systems. Am. Water Works Assoc. 2001, 93, 88–100. Kuch, A. Investigations of the reduction and re-oxidation kinetics of iron (III) oxide scales formed in waters. Corros. Sci. 1988, 28, 221–231. [CrossRef] Baek, W.-C.; Kang, T.; Sohn, H.-J.; Kho, Y.T. In situ surface enhanced Raman spectroscopic study on the effect of dissolved oxygen on the corrosion film on low carbon steel in 0.01 M NaCl solution. Electrochim. Acta 2001, 46, 2321–2325. [CrossRef] Eliyan, F.F.; Mahdi, E.; Alfantazi, A. Investigating the corrosion of API-X100 pipeline steel in aerated carbonate solutions by electrochemical methods. Int. J. Electrochem. Sci. 2013, 8, 578–590. Eyu, G.D.; Will, G.; Dekkers, W.; Macleod, J. Effect of hydrodynamics and surface roughness on the electrochemical behaviour of carbon steel in CSG produced water. Appl. Surf. Sci. 2015, 357, 506–515. [CrossRef] Popova, A.; Sokolova, E.; Raicheva, S.; Christov, M. AC and DC study of the temperature effect on mild steel corrosion in acid media in the presence of benzimidazole derivatives. Corros. Sci. 2003, 45, 33–58. [CrossRef] Khaled, K.; Hackerman, N. Investigation of the inhibitive effect of ortho-substituted anilines on corrosion of iron in 1 M HCl solutions. Electrochim. Acta 2003, 48, 2715–2723. [CrossRef] Eliyan, F.F.; Mahdi, E.-S.; Alfantazi, A. Electrochemical evaluation of the corrosion behaviour of API-X100 pipeline steel in aerated bicarbonate solutions. Corros. Sci. 2012, 58, 181–191. [CrossRef] Zhang, L.; Li, X.G.; Du, C.W.; Cheng, Y.F. Corrosion and stress corrosion cracking behavior of X70 pipeline steel in a CO2-containing solution. J. Mater. Eng. Perform. 2009, 18, 319–323. [CrossRef] Park, J.-J.; Pyun, S.-I.; Lee, W.-J.; Kim, H.-P. Effect of bicarbonate ion additives on pitting corrosion of type 316L stainless steel in aqueous 0.5 M sodium chloride solution. Corrosion 1999, 55, 380–387. Ståhl, K.; Nielsen, K.; Jiang, J.; Lebech, B.; Hanson, J.C.; Norby, P.; Lanschot, J.V. On the akaganéite crystal structure, phase transformations and possible role in post-excavational corrosion of iron artifacts. Corros. Sci. 2003, 45, 2563–2575. [CrossRef]

Materials 2016, 9, 748

38. 39. 40. 41. 42.

18 of 18

Heuer, J.; Stubbins, J. An XPS characterization of FeCO3 films from CO2 corrosion. Corros. Sci. 1999, 41, 1231–1243. [CrossRef] Grosvenor, A.; Kobe, B.A.; Biesinger, M.C.; Mclntyre, N.S. Investigation of multiplet splitting of Fe 2p XPS spectra and bonding in iron compounds. Surf. Interface Anal. 2004, 36, 1564–1574. [CrossRef] Rowles, M.R.; Madsen, I.C. Whole-pattern profile fitting of powder diffraction data collected in parallel-beam flat-plate asymmetric reflection geometry. J. Appl. Crystallogr. 2010, 43, 632–634. [CrossRef] Toraya, H.; Huang, T.; Wu, Y. Intensity enhancement in asymmetric diffraction with parallel-beam synchrotron radiation. J. Appl. Crystallogr. 1993, 26, 774–777. [CrossRef] Haggerty, R.P.; Sarin, P.; Bérar, J.-F.; Apostolov, Z.D.; Kriven, W.M. Powder diffraction by fixed incident angle reflection using a curved position-sensitive detector. J. Appl. Crystallogr. 2010, 43, 560–569. [CrossRef] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).