Effect of defect on corrosion behavior of electroless Ni-P coating in CO2-saturated NaCl solution

Accepted Manuscript Title: Effect of defect on corrosion behavior of electroless Ni-P coating in CO2 -saturated NaCl solution Authors: Chong Sun, Jiankuan Li, Shuo Shuang, Hongbo Zeng, Jing-Li Luo PII: DOI: Reference:

S0010-938X(17)31566-4 https://doi.org/10.1016/j.corsci.2018.01.041 CS 7352

To appear in: Received date: Revised date: Accepted date:

24-8-2017 29-12-2017 29-1-2018

Please cite this article as: Chong Sun, Jiankuan Li, Shuo Shuang, Hongbo Zeng, Jing-Li Luo, Effect of defect on corrosion behavior of electroless Ni-P coating in CO2-saturated NaCl solution, Corrosion Science https://doi.org/10.1016/j.corsci.2018.01.041 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Effect of defect on corrosion behavior of electroless Ni-P coating in CO2-saturated NaCl solution Chong Sun, Jiankuan Li, Shuo Shuang, Hongbo Zeng, Jing-Li Luo  Department of Chemical and Materials Engineering, University of Alberta, Edmonton, Alberta T6G 1H9, Canada Corresponding author. E-mail address: [email protected] (J.-L. Luo)

IP T



N

U

SC R

Graphical abstarct

A

Highlights

M

1. Ni-P coating had a good resistance to corrosion disbonding in CO2 environment.

ED

2. The defect was the major path for the mass transport at the coating/steel interface. 3. The defect caused the localized corrosion and corrosion disbonding of Ni-P coating.

PT

4. The presence of the defect retarded the disbonding of coating away from the defect.

CC E

5. Localized corrosion and disbonding model of coating in CO2 environment was proposed.

Abstract: The effects of defect on the localized corrosion and disbonding behavior of electroless Ni-P coating in CO2 environment were investigated using electrochemical methods

A

and surface characterization techniques. The results showed that the Ni-P coating had a good resistance to corrosion disbonding at open circuit potential, even with an artificial defect in the coating. The localized corrosion and the disbonding of the coating occurred after the acceleration by cathodic polarization because the coating defects provided effective pathways

1

for the electrolyte to transport through the coating and along the coating/substrate interface laterally. A corrosion model was proposed to well interpret this phenomenon. Keywords: A. Metal coatings; B. EIS; B. Polarization; B. SEM; C. Acid corrosion 1. Introduction In the oil and gas industry, carbon steel has been widely used as casing or tubing material because of its low cost [1, 2]. However, it is susceptible to corrosion due to the presence of

IP T

corrosive species (e.g., CO2, H2S, O2 and Cl-) in the corrosive environments [3-7]. Among various corrosion protection techniques, coating protection is an effective way to improve the

SC R

physical and chemical properties of carbon steel. An electroless Ni-P coating, recognized by its good resistance to corrosion, wear and abrasion, etc., has been applied to the carbon steel surface in various environments [8-12]. Moreover, the economic and operational costs can be

U

well balanced by coating carbon steels to improve their corrosion resistance and extend their

N

service life [13, 14]. Therefore, the electroless Ni-P coating has received extensive attention.

A

The corrosion performance of electroless Ni-P coating has been extensively investigated in corrosive environments containing brine or acid [8-12, 15-18]; the results suggest that Ni-P

M

coating possesses a better corrosion resistance than the substrate, thus effectively protecting

ED

the substrate through isolating the substrate from the corrosive environments. Limited studies on the corrosion behavior of Ni-P coating in the oilfield environments [19-21] reveal that Ni-P coating has a good resistance to the corrosive medium such as CO2, H2S and high salinity.

PT

However, defects, such as micropores, lattice defects and surface inclusions as well as some defects caused by mechanical effect, may be present in Ni-P coating [11, 17, 20]. These

CC E

defects are likely to pose great risks to the durability and stability of Ni-P coating in corrosive environments.

Related research indicates that defects in organic coating can provide pathways for the

A

ingress of the corrosive medium into the organic coating. The accumulation of corrosive medium at the coating/substrate interface changes the adhesion and bonding between the coating and substrate, which promotes the disbonding of organic coating [22-25]. Unlike organic coating, electroless Ni-P coating has a strong adhesion due to the metallic bond between the coating and substrate. However, as a metallic coating, it can be corroded in the corrosive environments [15-18]. Therefore, the roles of defects in the disbonding process of 2

Ni-P coating may be different from that of organic coating. Zhao et al. [26] investigated the corrosion failure process of NiCrBSi alloy coating in NaCl solution and found that the localized corrosion first occurred, and then the substrate was selectively corroded due to the occurrence of localized corrosion in the coating. Wang et al. [16] reported that corrosion pits formed on the Ni-P coating surface after immersion in NaCl solution for 30 days. Crobu et al. [27] also reported that Ni-P coating suffered from localized corrosion after a prolonged

IP T

polarization at 0.1 V (vs. SCE) in chloride and sulphate solution. Moreover, the related research for carbon steel or stainless steel in environments with the co-existence of Cl- ions and CO2 or H2S showed that localized corrosion was readily initiated due to a synergistic

SC R

effect between CO2 or H2S and Cl- ions [28-30]. However, limited reports are available about the effects of coating defects on the localized corrosion or corrosion disbonding of Ni-P coating in CO2 environment. It can be expected that, apart from providing pathways for

U

corrosive medium to transport through the coating, the presence of coating defects is likely to

N

cause the localized corrosion of Ni-P coating in the environments with co-existence of CO2

A

and brine (Cl- ions); their synergistic effect probably promotes the coating corrosion. When

M

localized corrosion extends to the substrate, the corrosion disbonding of coating may occur due to the corrosion of substrate underneath the coating.

ED

The objective of this study is to understand the effect of defect on the localized corrosion and disbonding behavior of an electroless Ni-P coating in CO2-saturated NaCl solution. The

PT

electrochemical behavior of the Ni-P coating with or without artificial defect was investigated in a double-cylinder corrosion cell and a single-cylinder corrosion cell, respectively. The

CC E

coating and corrosion products were characterized by scanning electron microscope (SEM), energy dispersive spectroscopy (EDS), X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS). Based on the testing results, a corrosion model was proposed to illustrate

A

the localized corrosion and corrosion disbonding mechanism of Ni-P coating in the CO2-saturated NaCl solution. 2. Experimental methods 2.1 Material and solution preparation The specimens of L80 steel, with a size of 15 mm × 15 mm × 5 mm, were used as the

3

substrates for electroless Ni-P coatings. Prior to the electroless process, the specimen surface was ground sequentially up to 1200 grit silicon carbide paper, rinsed with deionized water and degreased with alcohol in ultrasonic bath for 20 min. It was then cleaned in an alkaline solution consisting of 3 wt.% sodium carbonate and 3 wt. % sodium hydroxide for 20 min at 60 °C and rinsed with deionized water. The substrate was then etched in 10 wt.% HCl solution under stirring condition for 50 s to remove metallic oxides and obtain a micro-roughness on

IP T

the steel surface, it was then washed with deionized water and dried in compressed air. Finally, the cleaned substrate was dipped in electroless bath to plate Ni-P coating. Table 1 lists the compositions of electroless plating bath and plating conditions.

SC R

The artificial defect with a diameter of about 1.6 mm was made by drilling through the coating to the substrate at the center of specimen. Figs. 1a and b show the surface and cross-sectional morphologies of the electroless Ni-P coating with an artificial defect using

U

optical stereoscopic microscope (OSM) and SEM, respectively. The SEM cross-sectional

N

morphology clearly shows that the drilling did not cause the coating around the artificial

A

defect to peel off from the substrate (Fig. 1b).

M

The test solution was a 3.5 wt.% NaCl solution. Before the experiments, the solution was first bubbled with N2 to purge oxygen for at least 12 h, and then was saturated with CO2 gas.

ED

The pH value of the CO2-saturated NaCl solution was about 4.1 at ambient temperature (22 ºC) and pressure. After the specimen was placed in the solution, CO2 gas was continuously

PT

bubbled through the solution at a low flow rate of 20 mL/min to ensure the CO2-saturated state during the test.

CC E

2.2 Electrochemical measurements The electrochemical measurements were conducted in a three-electrode cell: a carbon

rod served as the counter electrode (CE), a saturated calomel electrode (SCE) as the reference

A

electrode (RE) and the Ni-P coating with or without an artificial defect as the working electrode (WE). All the potentials mentioned in this study were versus SCE. Two types of corrosion cells were used for the electrochemical measurements. The double-cylinder corrosion cell was employed to determine the indirect impact of the artificial defect on the electrochemical behavior of Ni-P coating and avoid the influence of the exposed substrate at the artificial defect on the electrochemical measurements of Ni-P coating [25]. As 4

exhibited in Fig. 2, a conical cylinder was cemented around the artificial detect to form the inner corrosion cell. One set of reference electrode (RE1) and counter electrode (CE1) were placed in the outer cell while the other set (RE2 and CE2) in the inner cell. The exposed area of WE in the outer cell was 175 mm2, while that in the inner cell including the artificial defect was 5 mm2 (The artificial defect area was 2 mm2). The single-cylinder corrosion cell was made by removing the inner cell in Fig. 2, the exposed area of WE was 225 mm2. The test

IP T

conditions are listed in Table 2. Tests 1 and 2 were performed under the open circuit potential (OCP) condition. For the tests 3-5, a cathodic potential of -1.5 V (vs. SCE) was applied on WE to accelerate the test process. All the tests were performed at ambient temperature (22 ºC)

SC R

and pressure.

A Gamry Reference 3000 electrochemical workstation was used for all electrochemical measurements. For the test in the double-cylinder corrosion cell, the cathodic potential was

U

applied in both inner and outer cells and electrochemical impedance spectroscopy (EIS)

N

measurements were only conducted in the outer cell at OCP to avoid direct measurement of

A

defect-exposed substrate. After the cathodic polarization or immersion under open circuit

M

condition for the designed time (The cathodic polarization was periodically suspended to allow for EIS measurements), OCP measurement was first conducted for 1 h to obtain a stable

ED

state. After that, EIS was measured at the corresponding OCP with an AC amplitude of 10 mV in the frequency range of 100 kHz to 10 mHz. To ensure the validity and stability of the EIS

PT

data, all the EIS data were verified by the Kramers-Kronig transforms (KKT) function from Gamry Echem Analyst software. The EIS data were fitted with an appropriate equivalent

CC E

circuit using Zsimpwin software. All the electrochemical tests were repeated at least 2 times to ensure the reproducibility and only the representative results were reported.

A

2.3 Characterization of the coating and corrosion scale After the electrochemical tests, the surface morphologies of the corroded specimens were

observed using SEM. The chemical compositions of the corrosion products were analyzed using EDS and XPS. Subsequently, the corroded specimens were cut along their cross-sections and sealed using epoxy, the corresponding cross-sectional morphologies were observed using SEM. Additionally, the morphology, chemical composition, phase structure and chemical valence of electroless Ni-P coating were characterized by means of SEM, EDS, 5

XRD and XPS. 3. Results and discussion 3.1 Morphology and structure of the electroless Ni-P coating Fig. 3a shows the SEM surface morphology of the as-deposited electroless Ni-P coating. The coating surface was smooth with a spherical nodular-like microstructure. Some

IP T

nanopores were observed around the spherical nodular boundary. The coating with a thickness of about 25 μm was homogeneous, compact and closely combined with the substrate (Fig. 3b). The EDS line scanning analysis reveals that Ni and P elements, with a composition of 88.28

SC R

wt. % Ni and 11.72 wt. % P, uniformly distributed in the coating (Fig. 3c). This means that the Ni-P coating was a high phosphorus-containing coating [31]. The XRD pattern (Fig. 3d) shows that the as-deposited Ni-P coating was amorphous. It might be attributed to the

U

high-phosphorus content in the coating, which caused its structure to become amorphous due

N

to the formation of a high degree of lattice disorder [17, 32]. Related research indicates that

A

the high phosphorus-containing coating has a high corrosion resistance in acid environment

M

due to its amorphous structure [10, 32-34].

ED

3.2 Effects of an artificial defect and the cathodic polarization on the OCP of Ni-P coating Fig. 4 shows the OCPs of the Ni-P coatings with or without an artificial defect after the cathodic polarization or immersion under open circuit condition for different times. It can be

PT

seen that the OCP remarkably shifted to the negative direction when an artificial defect was present in the coating (Figs. 4a and b), indicating the higher tendency for corrosion to occur.

CC E

As reported previously [17], the OCP of L80 steel was about -600 mV vs. SCE) in 3.5 wt.% NaCl solution. As seen in Fig. 4b, the OCP of the Ni-P coating with an artificial defect was similar to that of L80 steel. It means that the OCP mainly revealed the corrosion tendency of

A

the exposed substrate at the artificial defect when the coating and artificial defect were simultaneously exposed to the solution. After the cathodic polarization for different times, the OCPs of the Ni-P coatings with or without an artificial defect decreased with different levels (Figs. 4c-e) as compared to those under open circuit condition (Figs. 4a and b), implying that the cathodic polarization significantly increased the corrosion tendency. Fig. 4e shows the OCP of the coating in the outer cell when the coating with an artificial defect was exposed to 6

the double-cylinder corrosion cell. The isolation of the artificial defect from the outer cell shifted OCP towards a more positive direction when comparing with that of the coating in the single-cylinder corrosion cell (Fig. 4c). It is, thus, concluded that the presence of the artificial defect could reduce the corrosion tendency of the coating away from the defect. 3.3 Corrosion behavior of the Ni-P coating at OCP

IP T

3.3.1 Effect of the nanopore defect Fig. 5 shows the EIS of the Ni-P coating after immersion in the single-cylinder corrosion cell at OCP for different times. The shapes of Nyquist plots did not change with time

SC R

obviously under the test conditions (Fig. 5a), suggesting that the corrosion mechanism did not change in the test duration. This probably means that the corrosive medium did not penetrate to the coating/substrate interface. However, the diameters of the semicircles in the Nyquist

U

plots, the total impedance values and the maximum phase angles in the Bode plots (Fig. 5b)

N

decreased, indicating that the corrosion resistance decreased with the immersion time.

A

To further clarify the corrosion process of the coating, a well-accepted equivalent circuit

M

(Fig. 5c) [6, 11, 35-37] was used to describe two sub-electrochemical interfaces. Rs is the solution resistance; Qc and Rc are the capacitance and resistance of the coating, respectively,

ED

which were employed to describe the properties of the coating; Qdl and Rt are the double layer capacitance and the charge transfer resistance, respectively, which were used to describe the

PT

electrochemical reaction process at the coating/solution interface at the nanopores. In order to accurately fit the EIS data, the constant phase element designated as Qc or Qdl in the

CC E

equivalent circuit was used instead of the pure capacitance, and its admittance [Y(Q)] is defined as [35]:

𝑌(𝑄) = 𝑌0 (𝑄)(𝑗𝜔)𝑛

(1)

A

where Y0(Q) and n are the admittance constant and empirical exponent, respectively; j is the imaginary number; ω is the angular frequency. The fitted values of electrochemical parameters are listed in Table 3. It can be seen that both Rc and Rt were quite high, whereas Y0(Qc) and Y0(Qdl) were quite low during the first 6 h of immersion, indicating that the coating, as a barrier layer between the solution and the substrate, provided high resistance to the mass transport and isolated the contact of the corrosive medium with the substrate. However, upon

7

prolonging the corrosion time, Rc decreased rapidly corresponding to the increase of Y0(Qc). This is because of the existence of nanopores in the coating (Fig. 3a), the corrosive medium could penetrate into the coating through these nanopores. Likewise, Rt dramatically decreased while Y0(Qdl) increased with time, an indication that the penetration of corrosive medium promoted the corrosion of the coating, which consequently led to an increasing dissolution rate of the coating at the nanopores. After 66 h of the corrosion test, Rt increased slightly with

IP T

time corresponding to the rapid decline of Y0(Qdl). This is presumably ascribed to the formation of corrosion products in the nanopores, which retarded the mass transport and slowed down the corrosion of the coating.

SC R

The above results illustrate that both Rc and Rt were the main resistances of the corrosion process. Therefore, a parameter, Rp (Rp = Rc + Rt), was named polarization resistance and employed to reflect the corrosion rate of the coating at OCP. There existed an inverse

U

relationship between Rp and the corrosion rate [38]. Furthermore, a parameter named water

N

absorption of a coating (Wc) was used to evaluate the penetration degree of the corrosive

M

 100

lo g (8 0 )

(2)

ED

WC % 

 Q (t )  lo g  c   Q c (0 ) 

A

medium. It can be calculated by the following empirical equation [39]:

where Qc(0) and Qc(t) are the coating capacitances at 0 h and t h, respectively. See Fig. 6 for the variations of Rp and Wc with time, during the first 6 h of immersion, Wc was lower than

PT

1%, followed by a rapid increase with a higher slope (about 0.65 %/h). After 66 h of immersion, Wc started to increase with a lower slope (about 0.22 %/h). Correspondingly, Rp

CC E

was dramatically reduced before reaching 66 h of immersion, and then slowly decreased with time. This further confirmed that the nanopores in the coating provided pathways for the mass transport. As the immersion time increased, the corrosive medium gradually penetrated into

A

the coating and caused the corrosion of coating in the nanopores, consequently increasing the corrosion rate. After 66 h of immersion, the corrosion and the penetration degree of the corrosive medium almost reached the stable state because of the formation of corrosion products in the nanopores. To further confirm the above analysis, the morphology of the Ni-P coating after immersion in the single-cylinder corrosion cell for 148 h at OCP was examined, as shown in 8

Fig. 7. The amounts of black spots on the coating surface (Fig. 7a) were larger than those on the original coating surface (Fig. 3a). The EDS analysis indicates that the black spots, consisting of Ni, P, C and O elements, were corrosion products. The cross-sectional morphology in Fig. 7b exhibits the corrosion traces which, unlike the polishing traces, extended towards the inside of the coating. This was confirmed by the EDS analysis since C and O elements were detected alongside Ni and P elements in the traces (Fig. 7b). It is

IP T

believed that the corrosive medium penetrated into the coating through the nanopores and further caused the corrosion of the coating. However, no obvious corrosion was found at the coating/substrate interface, which implied that no corrosive medium reached the

SC R

coating/substrate interface after 148 h of immersion. The morphology observation was in good agreement with the EIS analysis.

U

3.3.2 Effect of an artificial defect

N

Fig. 8 shows the EIS of the Ni-P coating with an artificial defect after immersion in the

A

single-cylinder corrosion cell at OCP for different times. Compared with the Nyquist plots of the coating without an artificial defect (Fig. 5a), the presence of the artificial defect led to the

M

occurrence of an inductive semicircle at low frequency (Fig. 8a). The shape of EIS did not

ED

change in the test duration. The total impedance at 0.01 Hz in the Bode plots (Fig. 8b) was much smaller than that of the coating without an artificial defect at the same corrosion time (Fig. 5b) and only fluctuated within the range of 2018 to 1548 Ω∙cm2 in the test duration. It

PT

means that the corrosion of the exposed substrate at the artificial defect was a main factor governing the corrosion process. The EIS results mainly revealed the corrosion process of the

CC E

exposed substrate at the artificial defect. Fig. 9a shows the SEM surface morphology of the artificial defect after immersion in the

single-cylinder corrosion cell at OCP for 148 h. A large amount of corrosion products was

A

observed at the artificial defect. The EDS analysis of point A in Fig. 9a indicates that these products contained Fe, Cr, Mn, O, C and Cl elements (Fig. 9b), but the Ni and P elements derived from the coating were not detected. It is obvious that the corrosion of the exposed substrate mainly occurred at the defect. Moreover, the enrichment of Cl element was detected in the defect (Fig. 9b). It is well known that excess cations derived from the dissolution of the substrate attract anions, such as Cl-, to maintain the charge neutrality of the solution and the 9

accumulation of Cl- ions in pits plays a catalytic role in the corrosion process, and further increases the localized corrosion rate [28, 29, 40]. See Fig. 9c for the surface morphology of the coating away from the artificial defect, only small amounts of corrosion products, consisting of Ni, P, C and O elements (Fig. 9d), were found on the coating surface. The coating away from the artificial defect was corroded slightly (Fig. 9c), compared with that without an artificial defect (Fig. 7a). This confirmed that the presence of the artificial defect

IP T

slowed down the corrosion of the coating. Fig. 9e exhibits the cross-sectional morphology of the artificial defect. Apparently, the substrate was preferentially corroded, whereas the corrosion of the coating around the artificial defect was inhibited. The corrosion of the

SC R

substrate at the defect not only expanded along the depth direction, but also expanded along the coating/substrate interface laterally from the defect, which promoted the corrosion disbonding of the coating around the defect. However, no corrosion occurred at the

U

coating/substrate interface away from the defect (Fig. 9f).

N

The above results indicate that the microscopic coating defects acted as the effective

A

pathways facilitating the mass transport and as the local active sites on the coating surface for

M

corrosion. However, the coating disbonding did not occur at OCP during the test period, even with an artificial defect in the coating. Therefore, to accelerate the corrosion process, the

ED

disbonding behavior of the Ni-P coating with or without an artificial defect was further investigated by using a cathodic polarization method.

PT

3.4 Corrosion behavior of the Ni-P coating after the cathodic polarization

CC E

3.4.1 Effect of the cathodic polarization on the corrosion of the Ni-P coating Fig. 10 shows the EIS of the coating at OCP after the cathodic polarization of -1.5 V (vs.

SCE) in the single-cylinder corrosion cell for different times. The shapes of Nyquist plots

A

changed with time after cathodic polarization (Fig. 10a), corresponding to the change of the corrosion process. The total impedance value at 0.01 Hz (Fig. 10b) was obviously lower than that at OCP (Fig. 5b) at the same corrosion time, corresponding to the decline of the reaction resistance. Especially, the impedance was rapidly reduced after the cathodic polarization for 43 h. This probably means that the corrosive medium had penetrated to the coating/substrate interface.

10

The morphology observation in Fig. 11 shows the localized corrosion pits, which were observed on the coating surface after the cathodic polarization for 92 h. Figs. 11a and c show two typical localized corrosion morphologies. See Fig. 11b for the cross-sectional morphology of the pit in Fig. 11a, the pit was in the developmental stage and did not run through the coating. No obvious corrosion occurred at the coating/substrate interface underneath the pit. In contrast, the pit in Fig. 11c penetrated the entire coating and the

IP T

substrate was corroded (Fig. 11d). The corrosion of the substrate expanded along the depth direction and the coating/substrate interface laterally, similar to that at the artificial defect (Fig. 9e). Furthermore, the EDS analysis of the corrosion products in the pit reveals that besides Ni,

SC R

P, Fe, Cr, C and O elements, the enrichment of Cl element was also detected in the pit (Fig. 11f), whereas no Cl element was found in the corrosion products on the coating surface (Figs. 11e and f). Apparently, the localized corrosion of the coating first occurred after the

U

acceleration by the cathodic polarization and subsequently, the corrosion of the substrate and

N

the local disbonding of the coating around the pits were able to proceed once the pits extended

A

to the coating/substrate interface.

M

When an artificial defect was present in the coating, the EIS behavior of the coating with an artificial defect after the cathodic polarization of -1.5 V (vs. SCE) in the single-cylinder

ED

corrosion cell obviously changed (Fig. 12), in comparison with that under open circuit condition (Fig. 8). As seen in Fig. 12a, the inductive semicircle at lower frequency

PT

disappeared after the cathodic polarization for 6 h. This was probably because that the cathodic polarization promoted the corrosion of the coating and the substrate, leading to the

CC E

rapid formation of the corrosion products at the artificial defect and the coating surface. It can be confirmed by the decline of the total impedance (Fig. 12 b) as well as the variation of the OCP (Fig. 4d). Similar to the case at OCP (Fig. 8b), the total impedance at 0.01 Hz

A

maintained in the range of 748 to 441 Ω∙cm2 from the beginning to the end of the tests (Fig. 12b). It should be noted that hydrogen bubbles mainly appeared at the artificial defect during the cathodic polarization. It confirmed that the results of the EIS mainly revealed the corrosion process of the exposed substrate at the artificial defect in this cell set up. However, the variations of the current density with time for the coating with an artificial defect at the potential of -1.5 V suggested that the absolute values of the current densities increased slowly

11

during the first 66 h of the cathodic polarization, and subsequently increased with time dramatically. It can be inferred that the corrosion disbonding of the coating might occur after the cathodic polarization for 66 h. After the cathodic polarization in the single-cylinder corrosion cell for 92 h, the morphology features similar to those at OCP (Figs. 9a, b and e) were found at the artificial defect, as shown in Figs. 14a, b and e. However, the lateral expansion depth (114 μm) of the

IP T

substrate corrosion at the coating/substrate interface was larger than that (79 μm) at OCP (Fig. 9e). Additionally, the cross-sectional morphology of the coating away from the defect clearly reveals that the disbonding of the coating occurred at the coating/substrate interface (Fig. 14f).

SC R

It can be determined that the cathodic polarization promoted the corrosion disbonding of both the coating around the artificial defect and the coating away from the defect. Moreover, see Fig. 14c for the surface morphology of the coating away from the artificial defect, a large

U

amount of corrosion products formed on the coating surface. The EDS analysis shows that,

N

besides Ni, C and O elements, Fe was also detected in the corrosion products (Fig. 14d). It is

A

expected that the mass transport of corrosive species was also achieved through the coating,

M

causing the corrosion of the substrate underneath the coating. Meanwhile, the produced Fe2+ ions diffused to the coating surface and subsequently formed the corrosion products of iron.

ED

The above results make it clear that it is not possible to obtain the related information of the coating degradation and the mass transport process through the coating using the

PT

single-cylinder corrosion cell, because the information of EIS for the coating with an artificial defect is mainly derived from the corrosion behavior of the exposed substrate by using this set

CC E

up. Therefore, the double-cylinder corrosion cell was used to monitor the corrosion process of the coating with an artificial defect to determine the corrosion behavior of the coating away from the artificial defect.

A

3.4.2 Corrosion behavior of the Ni-P coating away from the artificial defect Fig. 15 shows the EIS of the outer cell at OCP after cathodic polarization for various

time durations. The EIS behavior was obviously distinct from that of the coating with an artificial defect in the single-cylinder corrosion cell (Fig. 8 and Fig. 12) due to the isolation of the artificial defect from the outer cell (Fig. 2), but similar to that of the coating without an artificial defect exposed to the single-cylinder corrosion cell (Fig. 10). As indicated in Fig. 15, 12

the changes of the Nyquist plots and Bode plots with time mainly included three stages. Therefore, three different equivalent circuits [6, 11, 25, 35-37, 41-44] were used to describe the corresponding corrosion process of the coating, as depicted in Fig. 5c and Fig. 16. During the first 6 h of corrosion, the EIS behavior, similar to that at OCP (Figs. 5a and b), mainly showed the electrochemical reaction process at the coating/solution interface at the nanopores. Thus, the equivalent circuit with the same parameters was used, as shown in Fig. 5c. However,

IP T

the nanopores in the coating, similar to the pits, were small and shallow and thus, the diffusion process of the corrosive medium was not obvious. During the test period of 21 - 66 h, the pit depth increased with time, which might significantly confine the diffusion inward to

SC R

the corrosion reaction interface in the pits. Therefore, the diffusion process of corrosive medium through the pits had a significant influence on the impedance of the corrosion system, which exhibited a finite length diffusion behavior (Fig. 15a). In this case, the equivalent

U

circuit in Fig. 5c was modified and the cotangent diffusion element (O) was adopted in the

N

intermediate frequency region to describe this behavior (Fig. 16a) [26], and its admittance

1

A

[Y(O)] can be expressed as [35]:

M

𝑌(𝑂) = 𝑌0 (𝑂)(𝑗𝜔)1/2 coth [𝐵(𝑗𝜔)2 ] , 𝐵 =

𝑙

√𝐷

(3)

where l is the thickness of diffusion layer (here, it is the depth of the pit); D is the ion

ED

diffusion coefficient in the pit. Rt1 and Qdl1 in Fig. 16a are the double layer capacitance and the charge transfer resistance at the coating/solution interface in the pits, respectively. During

PT

the corrosion of 88 - 105 h, the pits penetrated the entire coating; a direct pathway for diffusion was achieved between the exposed substrate and the solution. Nyquist plots could be

CC E

characterized with a high-frequency capacitive semicircle and the superposition of low-frequency capacitive semicircle and Warburg (W) impedance (Fig. 15a). Therefore, a semi-infinite length diffusion element (W) was used to simulate this behavior (Fig. 16b), and

A

its admittance can be given by [35]: 𝑌(𝑊) = 𝑌0 (𝑊)(𝑗𝜔)1/2

(4)

Moreover, Rt2 and Qdl2 in Fig. 16b represent the charge transfer resistance and the double layer capacitance at the exposed substrate/solution interface in the pits, respectively. The fitted values of the main electrochemical parameters are listed in Table 4. The calculated polarization resistance and the water absorption of coating are shown in Fig. 17. 13

As exhibited in Fig. 17, the polarization resistance of the coating was notably lower than that at OCP (Fig. 6) at the same corrosion time, corresponding to a higher corrosion rate. At 6 h, Wc was only 0.92 % at OCP while that after the cathodic polarization was up to 5.79 %, implying that the cathodic polarization improved the penetration rate of corrosive medium and thus, accelerated the corrosion of the coating. As seen in Table 4, compared with the higher value of Rt in the early corrosion stage, Rt significantly decreased but mildly changed

IP T

during the test period of 21 - 66 h, indicating that the pits expanded with a high and stable corrosion rate in this stage. It can be seen that the growth of the pits further led to the significant decline of Rc after corrosion for 21 h. Meanwhile, it also favored the further

SC R

penetration of the corrosive medium in the coating. As exhibited in Fig. 17, Wc dramatically increased and entered a relatively stable stage only after 21 h of test. Likewise, the Y0(O) value increased with time, suggesting that the resistance of mass diffusion through the coating

U

decreased because of the growth of the pit. After 88 h of corrosion, Rc was reduced to a quite

N

low value which was an order of magnitude lower than those before reaching 66 h of

A

corrosion, indicating that the coating had negligible resistance to the penetration of the

M

corrosive medium under the cathodic polarization conditions. The corrosive medium reached the coating/substrate interface in this stage. Furthermore, compared with the high and stable

ED

value of Rt induced by the coating dissolution at pit/solution interface during the corrosion of 21 - 66 h, Rt began to decrease at 88 h, followed by a more rapid decrease with time. Thus, the

PT

Rt value should represent the dissolution resistance of the substrate at the coating/substrate interface rather than that of the coating in the pits, an indication that the corrosion of the

CC E

substrate underneath the coating away from the artificial defect occurred in this stage, which was consistent with the morphology observation in Fig. 14f.

A

3.4.3 Effect of the coating defect on the corrosion of the Ni-P coating away from the defect In this study, for the coating free of artificial defect, the hydrogen bubbles appeared on

the overall coating surface in the early corrosion stage. The cathodic potential was completely applied on the coating and resulted in the rapid penetration of corrosive medium along the nanopore defects, which promoted the corrosion of the coating at the nanopore defects by forming the corrosion pits. After that, the hydrogen evolution reactions occurred at the local sites of the coating where the pits formed. However, the hydrogen evolution reaction mainly 14

took place at the artificial defect during the cathodic polarization when the coating and the artificial defect were simultaneously exposed to the electrolyte. It can be deduced that the presence of the artificial defect or the formation of the localized corrosion pits had an indirect influence on the corrosion of the coating away from the defect or the pits. Compared with a relatively long diffusion process (21 - 66 h) of the coating away from an artificial defect (Fig. 15a), the coating without the artificial defect showed a shorter

IP T

corrosion diffusion process (21 - 43 h), as seen in Fig. 10a, and the total impedance at 0.01 Hz decreased to a lower value of 3086 Ω∙cm2 at 43 h (Fig. 10b) while that of the coating with an artificial defect was 12097 Ω∙cm2 at 43 h (Fig. 15b). It indicated that for the coating free of

SC R

artificial defect, the cathodic polarization can greatly accelerate the mass transport process of corrosive species in the pits. The pits played the same role as the artificial defect during the coating disbonding when the pits penetrated the entire coating. Therefore, as prolonging the

U

corrosion time, hydrogen evolution reaction was transferred to these sites and aggravated the

N

corrosion of the substrate in the pits and at the coating/substrate interface around the pits. It is,

A

thus, concluded that although the presence of the artificial defect or the formation of the

M

corrosion pits in the coating could accelerate the corrosion of the substrate at these sites and promote the corrosion disbonding of the coating around the defect or the pits, it could slow

ED

down the mass transport process of corrosive medium in the coating away from them, thereby retarding the corrosion and the disbonding of coating away from the defects or the pits.

PT

3.5 Corrosion model for the localized corrosion and disbonding of the Ni-P coating In this study, the nanopores in the Ni-P coating not only provided effective pathways for

CC E

the mass transport of corrosive medium through the coating, but also acted as the local active sites for corrosion on the coating surface. At OCP, the penetration of corrosive medium mainly depended on the osmotic pressure across the coating [25], which was a very slow

A

process. However, under an applied cathodic potential, the cathodic polarization provided a force of electric field for the mass transport of corrosive medium [25], which played a significant role in accelerating the penetration of corrosive medium and thereby, promoting the localized corrosion of the coating and the corrosion of the substrate at the coating/substrate interface. Prior to discussing the degradation process of electroless Ni-P coating, it is necessary to determine the possible corrosion reactions of the coating in CO2 15

environment. 3.5.1 Corrosion reactions of the Ni-P coating in CO2 environment In CO2 environment, CO2 dissolves in the electrolyte to form H2CO3, and H2CO3 partially dissociates into H+, HCO3- and CO32- ions [45]: (5)

H2CO3 ↔ H+ + HCO3-

(6)

HCO3- ↔ H+ + CO32-

(7)

IP T

CO2 + H2O ↔ H2CO3

The process provides the essential H+ ions for the cathodic hydrogen evolution reaction. In

SC R

this case, the corrosion process is mainly dominated by the three cathodic reactions [45]: 2H+ + 2e- → H2 2H2CO3 + 2e- → H2 + 2HCO3-

(9) (10)

U

2HCO3-+ 2e- → H2 + 2CO32-

(8)

N

To determine the roles of Ni and P in the corrosion reactions, XPS was employed to

A

identify the chemical states of Ni and P in the corrosion products and compared with those in the coating. The binding energies were calibrated using residual carbon with C1s peak value

M

of 284.6 eV. Fig. 18 shows the XPS spectra of the as-deposited Ni-P coating and corrosion products. As seen in Fig. 18a, the peaks at 852.3 eV and 869.6 eV indicated the existence of

ED

Ni0 as elemental state, whereas the peak at 855.7 eV was ascribed to Ni2+ [46-48]. Likewise, the spectrum of P 2p displayed two main peaks at 129.2 eV and 130.0 eV (Fig. 18c), both

PT

corresponding to the phosphorus of elemental state in the coating, while the peak at 132.6 eV was assigned to the oxidized P [47-50]. Thus, it can be summarized that the as-deposited Ni-P

CC E

coating was mainly composed of Ni0, P0 and small amounts of oxidized Ni and P compounds. The presence of considerable amounts of Ni0 renders the coating its electrical conductivity. Therefore, the hydrogen evolution reactions of Eqs. (8) - (10) mainly occur at the

A

coating/electrolyte interface or the defect/electrolyte interface rather than the coating/substrate interface. As a result, the localized corrosion of the substrate occurred around the artificial defect (Fig. 9e and Fig. 14e) or the localized corrosion pits (Fig. 11d), as well as underneath the coating (Fig. 14f), but it did not cause the coating to peel off from the substrate. After corrosion in CO2 environment at OCP for 148 h, the main peaks of Ni 2p spectrum at 855.7 eV and 873.5 eV were assigned to Ni 2p3/2 and Ni 2p1/2, respectively (Fig. 18b), indicative of 16

Ni2+ in the oxidation state [49, 51]. The corresponding satellite peaks (Sat.) were also identified at 861.5 eV for Ni 2p3/2 and 879.2 eV for Ni 2p1/2 (Fig. 18b), respectively. However, no obvious peaks appeared in P 2p spectrum (Fig. 18d). It means that the dissolutions of both Ni and P in the coating occurred in CO2 environment. Concurrently, the products of Ni deposited on the coating surface, while those of P presumably dissolved in the solution under test conditions.

IP T

Related research on the corrosion of the Ni-P coating reported that the preferential dissolution of Ni occurs in acid environment at OCP, leading to the enrichment of P on the coating surface [12, 21, 48, 52]. The enriched P reacts with H2O to form a layer of adsorbed

SC R

H2PO2- [Eq. (11)] [48], which plays an important role in hindering the anodic dissolution of Ni. The reason credited for the negligible amount of P-containing products in the corrosion products is presumably attributed to the dissolution of H3PO3 derived from the further

U

oxidation of H2PO2- [Eq. (12)], as confirmed by the results reported by Diegle and Sorensen

N

[48]. Therefore, the main anodic reaction in the early corrosion stage is the dissolution of Ni

A

[Eq. (13)]. However, according to the results in Fig. 11, the dissolution of the exposed

M

substrate at the localized corrosion pits [Eq. (14)] becomes the major anodic reaction rather than Ni once the corrosion pits extends to the substrate surface. P + 2H2O → H2PO2- + 2H+ + e-

ED

(11) (12)

Ni → Ni2+ + 2e-

(13)

PT

H2PO2- + H2O → H3PO3 + H+ + 2e-

Fe → Fe2+ + 2e-

(14)

CC E

3.5.2 The degradation process of the Ni-P coating Based on the above results, a corrosion model was proposed to describe the degradation

process of the electroless Ni-P coating in CO2-saturated NaCl solution, as shown in Fig. 19. In

A

the early corrosion stage (Fig. 19a), the corrosive species, such as H+, Cl-, HCO3- and H2CO3 in corrosive medium preferentially adsorb onto the nanopore defects under the driving action of osmotic pressure or polarization, where hydrogen evolution reactions [Eqs. (8) - (10)] are easier to proceed, resulting in a rapid anodic dissolution of the coating. Concurrently, the enrichment of Cl- ions in these sites further catalyzes the anodic dissolution. Similar to the catalytic mechanism of Cl- ions on the corrosion of Fe in CO2 environment, Cl- ions are likely 17

to accelerate the corrosion of Ni via Eqs. (15) - (17) [28, 53-55]. It promotes the initiation of the pits at the defects. Ni + Cl- + H2O → [NiCl(OH)]ad- + H+ + e-

(15)

[NiCl(OH)]ad- → NiClOH + e-

(16)

NiClOH + H+ → Ni2+ + Cl- + H2O

(17)

As the corrosion proceeds, the pits continually develop towards the inside of the coating, as

IP T

exhibited in Fig. 19b. In the developmental stage of pits, the mass transport process of corrosive medium between the electrolyte and the reaction surface in pits becomes difficult due to the increase of the pit depth and the accumulation of corrosion products. Therefore, the

SC R

diffusion mainly controls the corrosion process in this stage (Fig. 15). Furthermore, the rapid corrosion of Ni in the pits provides excess cations (e.g., Ni2+) which further attract anions (e.g., Cl-) to maintain the charge neutrality of the electrolyte. The concentrated metallic

U

chloride can hydrolyze to form H+ and metal hydroxide or oxide [54]. The produced H+

N

initiates the internal acidification of the pits, forming a self-catalysis process similar to the one

A

in an occluded corrosion cell [26, 40, 56] and thereby, accelerating the coating corrosion in

M

pits. The pits eventually penetrate the entire coating and the electrolyte can accumulate at the exposed substrate surface through the pits.

ED

Subsequently, the corrosion process enters into the stage of the substrate corrosion, which is governed by the anodic dissolution of the substrate and the diffusion of corrosive

PT

medium to the substrate surface (Fig. 15). In this stage, the pits formed in the coating play the same role as the artificial defect. Therefore, the corrosion process at the coating/substrate

CC E

interface of the coating without an artificial defect is similar to that of the coating with an artificial defect (Fig. 9e and Fig. 14e). As shown in Fig. 19c, given the fact that the electrode potential of the substrate (L80 steel) is more negative than that of the Ni-P coating [17], the

A

substrate at the coating/substrate interface is preferentially corroded while the corrosion of the coating around the pits is inhibited. As the corrosion proceeds, the corrosion of the substrate underneath the pits gradually expanded along the coating/substrate interface laterally from the pits (defects), as well as along the depth direction, causing the local disbonding of the coating around the pits. Although the above process can slow down the mass transport process of corrosive medium through the coating away from the pits, the electrolyte still penetrates the

18

coating to reach the coating/substrate interface through the nanopore defects, thereby resulting in the local disbonding of the coating away from the pits after the cathodic polarization, as shown in Fig. 14f. 4. Conclusions The disbonding of the Ni-P coating did not occur after corrosion in CO2-saturated NaCl solution for 148 h at OCP, even with an artificial defect in the coating. However, the corrosion

IP T

occurred at the nanopore defects and extended towards the inside of coating. Therefore, the electroless Ni-P coating demonstrated a good resistance to corrosion disbonding.

SC R

The presence of the artificial defect or the formation of localized corrosion pits in the coating could slow down the mass transport process through the coating, retarding the corrosion and disbonding of the coating away from the defects or the pits.

U

Under the cathodic polarization condition, the defects in the Ni-P coating provided

N

effective pathways for the mass transport of corrosive medium through the coating and along

A

the coating/substrate interface laterally from the defects, thereby causing the localized corrosion of coating.

M

A corrosion model has been proposed for corrosion of Ni-P coated L80 steel in CO2

ED

environment. The electrolyte penetrated into the nanopores and caused the corrosion of coating at the nanopores, promoting the initiation of the pits in CO2 environment. As the corrosion proceeded, the accumulations of corrosive species in the pits increased the localized

PT

corrosion rate. After the pits penetrated the entire coating, the corrosion process was governed by the substrate dissolution and the mass diffusion between the substrate interface and the

CC E

electrolyte. The corrosion of the substrate expanded along the coating/substrate interface laterally and the depth direction due to the accumulation of the electrolyte at the exposed

A

substrate surface, causing a local corrosion disbonding of the coating. Acknowledgements This work was supported by Natural Sciences and Engineering Research Council of Canada/RGL (Reservoir Management Inc.) Collaborative Research and Development Grants.

19

References [1] S.Q. Guo, L.N. Xu, L. Zhang, W. Chang, M.X. Lu, Characterization of corrosion scale formed on 3Cr steel in CO2-saturated formation water, Corros. Sci. 110 (2016) 123-133.

IP T

[2] M.W. Khalil, T.A.S. Eldin, H.B. Hassan, K. El-Sayed, Z.A. Hamid, Electrodeposition of Ni-GNS-TiO2 nanocomposite coatings as anticorrosion film for mild steel in neutral environment, Surf. Coat. Technol. 275 (2015) 98-111.

SC R

[3] Y.S. Choi, S. Nesic, S. Ling, Effect of H2S on the CO2 corrosion of carbon steel in acidic solutions, Electrochim. Acta 56 (2011) 1752-1760. [4] A.Q. Fu, Y.R. Feng, R. Cai, J.T. Yuan, C.X. Yin, D.M. Yang, Y. Long, Z.Q. Bai, Downhole corrosion behavior of Ni-W coated carbon steel in spent acid & formation water and its application in full-scale tubing, Eng. Fail. Anal. 66 (2016) 566-576.

A

N

U

[5] Q.Y. Wang, X.Z. Wang, H. Luo, J.L. Luo, A study on corrosion behavior of Ni-Cr-Mo laser coating, 316 stainless steel and X70 steel in simulated solutions with H2S and CO2, Surf. Coat. Technol. 291 (2016) 250-257.

M

[6] V. Garcia-Arriaga, J. Alvarez-Ramirez, M. Amaya, E. Sosa, H2S and O2 influence on the corrosion of carbon steel immersed in a solution containing 3 M diethanolamine, Corros. Sci. 52 (2010) 2268-2279.

ED

[7] P.P. Bai, H. Zhao, S.Q. Zheng, C.F. Chen, Initiation and developmental stages of steel corrosion in wet H2S environments, Corros. Sci. 93 (2015) 109-119.

CC E

PT

[8] H. Ashassi-Sorkhabi, S.H. Rafizadeh, Effect of coating time and heat treatment on structures and corrosion characteristics of electroless Ni-P alloy deposits, Surf. Coat. Technol. 176 (2004) 318-326. [9] E. Georgiza, J. Novakovic, P. Vassiliou, Characterization and corrosion resistance of duplex electroless Ni-P composite coatings on magnesium alloy, Surf. Coat. Technol. 232 (2013) 432-439.

A

[10] T.S.N.S. Narayanan, I. Baskaran, K. Krishnaveni, S. Parthiban, Deposition of electroless Ni-P graded coatings and evaluation of their corrosion resistance, Surf. Coat. Technol. 200 (2006) 3438-3445. [11] H. Ashassi-Sorkhabi, M. Es`haghi, Corrosion resistance enhancement of electroless Ni-P coating by incorporation of ultrasonically dispersed diamond nanoparticles, Corros. Sci. 77 (2013) 185-193.

20

[12] J.N. Balaraju, T.S.N.S. Narayanan, S.K. Seshadri, Evaluation of the corrosion resistance of electroless Ni-P and Ni-P composite coatings by electrochemical impedance spectroscopy, J. Solid State Electrochem. 5 (2001) 334-338. [13] M. Fenker, M. Balzer, H. Kapp, Corrosion protection with hard coatings on steel: past approaches and current research efforts, Surf. Coat. Technol. 257 (2014) 182–205.

IP T

[14] A.L.M. Oliveira, J.D. Costa, M.B. de Sousa, J.J.N. Alves, A.R.N. Campos, R.A.C. Santana, S. Prasad, Studies on electrodeposition and characterization of the Ni-W-Fe alloys coatings, J. Alloys Comp. 619 (2015) 697-703.

SC R

[15] S.Q. Guo, L.F. Hou, C.L. Guo, Y.H. Wei, Characteristics and corrosion behavior of nickel-phosphorus coating deposited by a simplified bath, Mater. Corros. 68 (2017) 468-475.

U

[16] L.L. Wang, H.L. Chen, L. Hao, A. Lin, F.X. Gan, Electrochemical corrosion behavior of electroless Ni-P coating in NaCl and H2SO4 solutions, Mater. Corros. 62 (2011) 1003-1007.

A

N

[17] H. Luo, M. Leitch, Y. Behnamian, Y.S. Ma, H.B. Zeng, J.L. Luo, Development of electroless Ni-P/nano-WC composite coatings and investigation on its properties, Surf. Coat. Technol. 277 (2015) 99-106.

M

[18] M. Crobu, A. Scorciapino, B. Elesener, A. Rossi, The corrosion resistance of electroless deposited nano-crystalline Ni-P alloys, Electrochim. Acta 53 (2008) 3364-3370.

PT

ED

[19] F.B. Mainier, M.M. Araújo, On the effect of the electroless nickel-phosphorus coating defects on the performance of this type of coating in oilfield environments, SPE Advanced Technology Series 2 (1994) 63-67.

CC E

[20] F.B. Mainier, M.P. C. Fonseca, S.S.M. Tavares, J.M. Pardal, Quality of electroless Ni-P (nickel-phosphorus) coatings applied in oil production equipment with salinity, J. Mater. Sci. Chem. Eng. 1 (2013) 1-8.

A

[21] S.R. Allahkaram, M.H. Nazari, S. Mamaghani, A. Zarebidaki, Characterization and corrosion behavior of electroless Ni-P/nano-SiC coating inside the CO2 containing media in presence of acetic acid, Mater. Des. 32 (2011) 750-755. [22] S. Martinez, L.V. Zuij, F. Kapor, Disbonding of underwater-cured epoxy coating caused by cathodic protection current, Corros. Sci. 51 (2009) 2253-2258. [23] S. Ranade, M. Forsyth, M.Y.J. Tan, The initiation and propagation of coating morphological and structural defects under mechanical strain and their effects on the electrochemical behavior of pipeline coatings, Prog. Org. Coat. 110 (2017) 62-67.

21

[24] C.F. Dong, A.Q. Fu, X.G. Li, Y.F. Cheng, Localized EIS characterization of corrosion of steel at coating defect under cathodic protection, Electrochim. Acta 54 (2008) 628-633. [25] J.L. Luo, C.J. Lin, Q. Yang, S.W. Guan, Cathodic disbonding of a thick polyurethane coating from steel in sodium chloride solution, Prog. Org. Coat. 31 (1997) 289-295. [26] W.M. Zhao, Y. Wang, J. Xue, K.Y. Wu, EIS study of the corrosion failure process of steel coated by nickel base alloy, Acta Metall. Sin. 41 (2005) 178-184.

IP T

[27] M. Crobu, A. Scorciapino, B. Elsener, A. Rossi, The corrosion resistance of electroless deposited nano-cystalline Ni-P alloys, Electrochim. Acta 53 (2008) 3364-3370.

SC R

[28] Q.Y. Liu, L.J. Mao, S.W. Zhou, Effects of chloride content on CO2 corrosion of carbon steel in simulated oil and gas well environments, Corros. Sci. 84 (2014) 165-171. [29] W. He, O.Ø. Knudsen, S. Diplas, Corrosion of stainless 316 L in simulated formation water environment with CO2-H2S-Cl-, Corros. Sci. 51 (2009) 2811-2819.

N

U

[30] G.Ø. Lauvstad, R. Johnsen, Ø. Borck, E.F. da Silva, J. Walmsley, Breakdown in Passivity of Austenitic Stainless Steels in Cl- and H2S – Modeling and Characterization of the Pit Initiation Process, Corrosion/2007, Paper No. 07660, NACE, Houston, TX, 2007.

M

A

[31] I.N.A. Oguocha, R. Taheri, S. Yannacopoulos, W.A. Uju, R. Sammynaiken, S. Wettig, Y.-F. Hu, Temperature effects on the chemical composition of nickel-phosphorus alloy thin films, Thin Solid Films 518 (2010) 2045-2049.

PT

ED

[32] K.G. Keong, W. Sha, Crystallisation and phase transformation behavior of electroless nickel-phosphorous deposits and their engineering properties, Surf. Eng. 18 (2002) 329-343.

CC E

[33] Y.X. Wang, X. Shu, S.H. Wei, C.M. Liu, W. Gao, R.A. Shakoor, R. Kahraman, Duplex Ni-P-ZrO2/Ni-P electroless coating on stainless steel, J. Alloys Comp. 630 (2015) 189-194. [34] C.D. Gu, J.S. Lian, G.Y. Li, L.Y. Niu, Z.H. Jiang, High corrosion-resistant Ni-P/Ni/Ni-P multilayer coatings on steel, Surf. Coat. Technol. 197 (2005) 61-67.

A

[35] C.N. Cao, J.Q. Zhang, An introduction to electrochemical impedance spectroscopy, First ed., Science Press, Beijing, 2002. [36] H. Marchebois, S. Joiret, C. Savall, J. Bernard, S. Touzain, Characterization of zinc-rich power coatings by EIS and Raman spectroscopy, Surf. Coat. Technol. 157 (2002) 151-161. [37] Y. Chen, Y.L. Mao, W. Huang, Y. Ji, W.Z. Yang, X.S. Yin, Y. Liu, X. Ling, Corrosion 22

behavior of Ni-P-nano-Al2O3 composite coating in the presence of anionic and cationic surfactants, Surf. Coat. Technol. 310 (2017) 122-128. [38] G.A. Zhang, Y. Zeng, X.P. Guo, F. Jiang, D.Y. Shi, Z.Y. Chen, Electrochemical corrosion behavior of carbon steel under dynamic high pressure H2S/CO2 environment, Corros. Sci. 65 (2012) 37-47.

IP T

[39] D.M. Brasher, A.H. Kingsbury, Electrical measurements in the study of immersed paint coatings on metal. I. Comparison between capacitance and gravimetric methods of estimating water-uptake, J. Appl. Chem. 4 (1954) 62–72.

SC R

[40] C.F. Chen, M.X. Lu, W. Chang, G.X. Zhao, Z.Q. Bai, The Ion Passing Selectivity of CO2 Corrosion Scale on N80 Tube Steel, Corrosion/2003, Paper No. 03342, NACE, Houston, TX, 2003. [41] Z.F. Yin, W.Z. Zhao, W.Y. Lai, X.H. Zhao, Electrochemical behavior of Ni-base alloys exposed under oil/gas field environments, Corros. Sci. 51 (2009) 1702-1706.

N

U

[42] A. Meroufel, S. Touzain, EIS characterization of new zinc-rich powder coatings, Prog. Org. Coat. 59 (2007) 197-205.

M

A

[43] Z.F. Yin, W.Z. Zhao, Z.Q. Bai, Y.R. Feng, W.J. Zhou, Corrosion behavior of SM 80SS tube steel in stimulant solution containing H2S and CO2, Electrochim. Acta 53 (2008) 3690-3700.

ED

[44] C. Liu, Q. Bi, A. Matthews, EIS comparison on corrosion performance of PVD TiN and CrN coated mild steel in 0.5 N NaCl aqueous solution, Corros. Sci. 43 (2001) 1953-1961.

CC E

PT

[45] J.B. Sun, C. Sun, G.A. Zhang, X.D. Li, W.M. Zhao, T. Jiang, H.F. Liu, X.K. Cheng, Y. Wang, Effect of O2 and H2S impurities on the corrosion behavior of X65 steel in water-saturated supercritical CO2 system, Corros. Sci. 107 (2016) 31-40. [46] S.L. Mu, N. Li, D.Y. Li, L.Y. Xu, Corrosion behavior and composition analysis of chromate passive film on electroless Ni-P coating, Appl. Surf. Sci. 256 (2010) 4089-4094.

A

[47] H. Li, H.X. Li, W.-L. Dai, W.J. Wang, Z.G. Fang, J.-F. Deng, XPS studies on surface electronic characteristics of Ni-B and Ni-P amorphous alloy and its correlation to their catalytic properties, Appl. Surf. Sci. 152 (1999) 25-34. [48] R.B. Diegle, N.R. Sorensen, C.R. Clayton, M.A. Helfand, Y.C. Yu, An XPS investigation into the passivity of an amorphous Ni-20P alloy, J. Electrochem. Soc. 135 (1988) 1085-1092.

23

[49] M.F. Hao, M.S. Xiao, Y.H. Yan, Y.Q. Miao, Synthesizing amorphous Ni-P micro-/nano-composites with perfect roundness or embryo-like structures, Adv. Powder Technol. 28 (2017) 3095-3103. [50] B. Elsener, M. Crobu, M.A. Scorciapino, A. Rossi, Electroless deposted Ni-P alloys: corrosion resistance mechanism, J. Appl. Electrochem. 38 (2008) 1053-1060.

IP T

[51] M.S. Xiao, R. Cheng, M.F. Hao, M. Zhou, Y.Q. Miao, Onsite substitution synthesis of ultrathin Ni nanofilms loading ultrafine Pt nanoparticles for hydrogen evolution, ACS Appl. Mater. Interf. 7 (2015) 26101-16107.

SC R

[52] G.T. Burstein, Examination of passive films on iron-nickel alloys by auger electron spectroscopy, Corrosion 37 (1981) 549-556. [53] G.T. Burstein, D.H. Davies, The effects of anions on the behavior of scratched iron electrodes in aqueous solutions, Corros. Sci. 20 (1980) 1143-1155.

U

[54] S. Asakura, K. Nobe, Electrodissolution kinetics of iron in chloride solutions Part I. neutral solutions, J. Electrochem. Soc. 118 (1971) 13-18.

A

N

[55] G.W. Ashley, G.T. Burstein, Initial stages of the anodic oxidation of iron in chloride solutions, Corrosion 47 (1991) 908-916.

A

CC E

PT

ED

M

[56] J.L. Crolet, N. Thevenot, S. Nesic, Role of conductive corrosion products in the protectiveness of corrosion layers, Corrosion 54 (1998) 194-203.

24

Figure captions Fig. 1. (a) OSM surface morphology and (b) SEM cross-sectional morphology of the electroless Ni-P coating with an artificial defect. Fig. 2. Schematic diagram of the double-cylinder corrosion cell: (a) L80 steel substrate; (b) Ni-P coating; (c) artificial defect; (d) inner cylinder wall; (e) outer cylinder wall and (f) epoxy.

IP T

Fig. 3. (a) SEM surface morphology, (b) cross-sectional backscattered electron image, (c) elemental distributions denoted by the red arrow line in (b) and (d) XRD pattern of the as-deposited Ni-P coating.

SC R

Fig. 4. Variations of OCP with time for the Ni-P coatings with or without an artificial defect after the cathodic polarization or immersion under open circuit condition in CO2-saturated 3.5 wt.% NaCl solution for the designed times.

N

U

Fig. 5. (a and b) EIS of the Ni-P coating exposed to CO2-saturated 3.5 wt.% NaCl solution in the single-cylinder corrosion cell at OCP for different times: (a) Nyquist plots and (b) Bode plots. (c) Equivalent circuit used for fitting the EIS data in (a) and (b).

M

A

Fig. 6. Variations of polarization resistance (Rp) and water absorption (Wc) with time for the Ni-P coating exposed to CO2-saturated 3.5 wt.% NaCl solution in the single-cylinder corrosion cell at OCP.

PT

ED

Fig. 7. (a) SEM surface morphology and (b) cross-sectional backscattered electron image of the Ni-P coating exposed to CO2-saturated 3.5 wt.% NaCl solution in the single-cylinder corrosion cell at OCP for 148 h and EDS analysis of corrosion products denoted by A and B in (a) and (b), respectively.

CC E

Fig. 8. EIS of the Ni-P coating with an artificial defect exposed to CO2-saturated 3.5 wt.% NaCl solution in the single-cylinder corrosion cell at OCP for different times: (a) Nyquist plots and (b) Bode plots.

A

Fig. 9. SEM images, EDS analyses and backscattered electron images for the Ni-P coating with an artificial defect exposed to CO2-saturated 3.5 wt.% NaCl solution in the single-cylinder corrosion cell for 148 h: (a) surface morphology of artificial defect; (b) EDS analysis of the corrosion product of Point A in (a); (c) surface morphology of the coating; (d) EDS analysis of the corrosion product of Point B in (c); (e) cross-sectional backscattered electron image of the artificial defect along the direction of red arrow in (a) and (f) cross-sectional backscattered electron image of the coating away from the artificial defect. Fig. 10. EIS of the Ni-P coating after cathodic polarization of -1.5 V (vs. SCE) in the single-cylinder corrosion cell with CO2-saturated 3.5 wt.% NaCl solution for different times: (a) Nyquist plots and (b) Bode plots. 25

Fig. 11. (a, c and e) SEM surface morphologies of the Ni-P coating after the cathodic polarization of -1.5 V (vs. SCE) in the single-cylinder corrosion cell with CO2-saturated 3.5 wt.% NaCl solution for 92 h, (b) and (d) cross-sectional backscattered electron images of the pits along the direction of red arrow in (a) and (c), respectively, and (f) EDS analysis of the corrosion products denoted by A and B in (c) and (e), respectively. Fig. 12. EIS of the Ni-P coating with an artificial defect after the cathodic polarization of -1.5 V (vs. SCE) in the single-cylinder corrosion cell with CO2-saturated 3.5 wt.% NaCl solution for different times: (a) Nyquist plots and (b) Bode plots.

SC R

IP T

Fig. 13. Variations of the current density with the immersion time for the Ni-P coating with an artificial defect exposed to CO2-saturated 3.5 wt.% NaCl solution in the single-cylinder corrosion cell at the potential of -1.5 V (vs. SCE). The current peaks were corresponding to the time duration for the EIS measurements at OCP and the test period was about 1.5 h (1 h OCP and 0.5 h EIS measurements).

M

A

N

U

Fig. 14. SEM images, EDS analyses and backscattered electron images for the Ni-P coating with an artificial defect after the cathodic polarization of -1.5 V (vs. SCE) in the single-cylinder corrosion cell with CO2-saturated 3.5 wt.% NaCl solution for 92 h: (a) morphology of the artificial defect surface; (b) EDS analysis of the corrosion product of Point A in (a); (c) morphology of the coating surface; (d) EDS analysis of the corrosion product of Point B in (c); (e) cross-sectional backscattered electron image of the artificial defect along the direction of red arrow in (a) and (f) cross-sectional backscattered electron image of the coating away from the artificial defect.

ED

Fig. 15. EIS of the outer cell after the cathodic polarization of -1.5 V (vs. SCE) for different times when the Ni-P coating with an artificial defect exposed to CO2-saturated 3.5 wt.% NaCl solution in the double-cylinder corrosion cell: (a) Nyquist plots and (b) Bode plots.

PT

Fig. 16. Equivalent circuits used for fitting EIS data in Fig. 15: (a) 21 - 66 h and (b) 88 - 105 h.

CC E

Fig. 17. Variations of polarization resistance (Rp) and water absorption (Wc) with time for the Ni-P coating away from the artificial defect after cathodic polarization of -1.5 V (vs. SCE) in the double-cylinder corrosion cell with CO2-saturated 3.5 wt.% NaCl solution.

A

Fig. 18. XPS spectra and decomposition peaks for (a) Ni 2p and (c) P 2p of the as-deposited Ni-P coating and (b) Ni 2p and (d) P 2p of Ni-P coating exposed to CO2-saturated 3.5 wt.% NaCl solution in the single-cylinder corrosion cell at OCP for 148 h. Fig. 19. Schematic model for the corrosion of the Ni-P coating with microdefects in CO2-saturated NaCl solution: (a) initiation of the localized corrosion; (b) growth of the localized corrosion and (c) corrosion disbonding of the coating.

26

A ED

PT

CC E Figure 1

27

IP T

SC R

U

N

A

M

A ED

PT

CC E

28

IP T

SC R

U

N

A

M Figure 2

IP T SC R U N A M ED

Ni P Fe

PT

(c)

A

Intensity (a.u.)

CC E

Ni-P coating

0

5

10

15

20

25

Line scanning distance (m)

29

30

35

(d)

10

20

30

40

50

60

A

CC E

PT

ED

M

A

N

U

Figure 3

30

80

SC R

degree

70

IP T

Intensity (a.u.)

Amorphous

-200 0h

(a) Test 1 (b) Test 2 (c) Test 3 (d) Test 4 (e) Test 5

-400

1h

-500

-600

0

30

60

90

Time (h)

A

CC E

PT

ED

M

A

N

U

Figure 4

120

31

150

SC R

-700

IP T

OCP (mV vs. SCE)

-300

25000 0h 6h 21 h 27 h 43 h 66 h 88 h 92 h 112 h 125 h 148 h KKT line Fitting line

(a) 3000

20000

2000 1000

0.04 Hz

2

-Zim(cm )

15000

0 0

1000 2000 3000 4000 5000

10000

5000

0

5000

10000

15000

20000 2

Zre (cm )

30000

35000

90

(b)

U

4.5

N

4.0 3.5 2

A

3.0

M

2.5 2.0

0h 6h 21 h 27 h 43 h 66 h 88 h 92 h 112 h 125 h 148 h KKT Line Fitting line

80 70 60 50 40 30 20

1.0

10

ED

1.5

0.5

-2

10

-1

10

0

10

1

10

2

Frequency (Hz)

A

CC E

PT

10

Figure 5 32

10

3

10

4

0

10

5

-Phase angle (Degree)

5.0

Log|Z | (cm )

25000

SC R

0

IP T

0.04 Hz

33

A ED

PT

CC E

IP T

SC R

U

N

A

M

35000

100 Rp

30000

Wc

80 70

50 15000

40

Wc ()

60

20000

30

10000

IP T

Rp (cm

2

25000

90

20

5000

0

30

60

90

Immersion time (h)

A

CC E

PT

ED

M

A

N

U

Figure 6

120

34

0 150

SC R

0

10

Figure 7

35

A ED

PT

CC E

IP T

SC R

U

N

A

M

2000 1h 6h 21 h 43 h 66 h 88 h 92 h 112 h 125 h 148 h KKT line

(a)

1000

500

0.42 Hz

IP T

2

-Zim(cm )

1500

0 500

1000

1500

2000 2

Zre (cm )

2500

3000

SC R

0

90

N A

2.5 2.0

M

2

Log|Z | (cm )

3.0

1.0

ED

1.5

-2

10

-1

10

80 70 60 50 40 30 20 10 0

0

10

1

10

2

Frequency (Hz)

Figure 8

A

CC E

PT

10

1h 6h 21 h 43 h 66 h 88 h 92 h 112 h 125 h 148 h KKT line

36

10

3

10

4

10

5

-Phase angle (Degree)

U

3.5

(b)

37

A ED

PT

CC E

IP T

SC R

U

N

A

M

A ED

PT

CC E Figure 9

38

IP T

SC R

U

N

A

M

(a)

1h 6h 21 h 43 h 66 h 88 h 92 h KKT line

2

-Zim(cm )

10000

5000

0.04 Hz

0

5000

10000

15000

SC R

2

Zre (cm )

5.0

90

(b)

U

4.5

N

3.5

2

Log|Z | (cm )

4.0

A

3.0

1.0 0.5

ED

2.0 1.5

-2

PT

10

70 60 50 40

10

-1

10

30 20 10

0

10

1

10

2

Frequency (Hz)

Figure 10

A

CC E

80

M

2.5

1h 6h 21 h 43 h 66 h 88 h 92 h KKT line

39

10

3

10

4

0

10

5

-Phase angle (Degree)

0

IP T

0.1 Hz

40

A ED

PT

CC E

IP T

SC R

U

N

A

M

IP T SC R U N A M ED

A

PT

(f)

O

Cl

CC E

Intensity (a.u.)

Fe C Ni

Fe

P

Ni Fe Ni

Cr

B

A

Ni Fe

Ni P

O C 0

Fe Fe 1

2

3

4

5

6

Energy (keV)

Figure 11 41

7

Ni 8

9

10

600 (a)

1h 6h 21 h 27 h 43 h 66 h 88 h 92 h KKT line

500 1.27 Hz 0.25 Hz

2

-Zim(cm )

400

300

200

IP T

0.05 Hz

0.05 Hz

100

100

200

300

400

500

600

2

Zre (cm )

90

(b)

A

2.0

80 70 60 50 40

M ED

1.5

-2

PT

10

10

-1

10

30 20 10 0

0

10

1

10

2

Frequency (Hz)

Figure 12

A

CC E

1h 6h 21 h 27 h 43 h 66 h 88 h 92 h KKT line

N

2

Log|Z | (cm )

2.5

1.0

800

U

3.0

700

42

10

3

10

4

10

5

-Phase angle (Degree)

0

SC R

0

-0.002

-0.006 -0.008 -0.010

-0.002 -0.003

-0.012

-0.004 -0.005

-0.014

-0.006 -0.007 0

1

2

3

4

5

-0.016 10

20

30

40

50

60

Time (h)

A

CC E

PT

ED

M

A

N

U

Figure 13

70

43

80

90

100

SC R

0

IP T

2

Current density (A/cm )

-0.004

44

A ED

PT

CC E

IP T

SC R

U

N

A

M

A ED

PT

CC E Figure 14

45

IP T

SC R

U

N

A

M

25000 0h 6h 21 h 27 h 43 h 66 h 88 h 92 h 105 h KKT line Fitting line

3000 2000

20000

0.03Hz 1000 0 0

1000 2000 3000 4000

10000

0.03Hz

5000

0.03 Hz

0

5000

10000

15000

20000

25000

2

Zre (cm )

5.0

(b)

U

4.5

N

3.5

2

Log|Z | (cm )

4.0

A

3.0

10

-2

0h 6h 21 h 27 h 43 h 66 h 88 h 92 h 105 h KKT line Fitting line

10

-1

10

90 80 70 60 50 40 30 20 10

0

10

1

10

2

Frequency (Hz)

Figure 15

A

CC E

PT

0.5

ED

2.0

1.0

35000

M

2.5

1.5

30000

SC R

0

IP T

2

-Zim(cm )

15000

46

10

3

10

4

0

10

5

-Phase angle (Degree)

(a)

A ED

PT

CC E

IP T

SC R

U

N

A

M Figure 16

47

35000

100 Rp

30000

Wc

80 70

50 15000

40

Wc ()

60

20000

30

10000

IP T

Rp (cm

2

25000

90

20

5000

0

20

40

60

80

Immersion time (h)

A

CC E

PT

ED

M

A

N

U

Figure 17

48

100

0 120

SC R

0

10

Ni 2p 2p3/2 852.3 eV 2p3/2 861.5 eV

(a) Ni

0

2p3/2 855.7 eV 2p1/2 869.6 eV 2p1/2 879.2 eV

2p1/2 873.5 eV

Ni

0

(b) 2+

Ni

Sat.

850

855

860

865

870

885

U

2p3/2 130.0 eV

0

2p3/2 132.6 eV

A

P

M

Oxidized P

(d)

PT

ED

Intensity (a.u.)

880

P 2p 2p3/2 129.2 eV

0

N

P

Sat.

875

Bonding energy (eV) (c)

2+

IP T

Ni

SC R

Intensity (a.u.)

Ni 2+

128

130

132

134

Bonding energy (eV)

Figure 18

A

CC E

126

49

136

138

A ED

PT

CC E Figure 19

50

IP T

SC R

U

N

A

M

Table 1 Compositions of electroless plating bath and plating conditions for Ni-P coating. Concentration

Plating conditions

Nickel sulphate (NiSO4·6H2O) Sodium hypophosphite (NaH2PO2·H2O) Lactic acid (C3H6O3) Citric acid (C6H8O7) Succinic acid (C4H6O4) Saccharin Sodium (C6H4SO2NNaCO·2H2O) Sodium dodecyl sulfate(SDS, C12H25SO4Na) Lead nitrate (Pb(NO3)2)

25 g/L 30 g/L 20 mL/L 19 g/L 14 g/L 0.06 - 0.08 g/L 0.03 g/L 1 mg/L

pH: 5.5 - 6.0 Temperature: 90 ± 1 ºC Plating time: 120 min Stirring speed: 200 rpm Loading capacity*: 1 dm2/L

Loading capacity is a ratio of the area of the specimen to the volume of the plating solution.

A

CC E

PT

ED

M

A

N

U

SC R

*

IP T

Reagent

51

Table 2 Test conditions

1 2 3 4

Ni-P coating without an artificial defect Ni-P coating with an artificial defect Ni-P coating without an artificial defect Ni-P coating with an artificial defect Ni-P coating with an artificial defect

Corrosion cell

Applied potential (V vs. SCE)

Single-cylinder

OCP

Single-cylinder

OCP

Single-cylinder

-1.5

Single-cylinder

-1.5

Double-cylinder

-1.5

A

CC E

PT

ED

M

A

N

U

SC R

5

Working electrode

IP T

Test

52

Table 3 Electrochemical parameters fitted from the measured EIS data in Figs. 5a and b by using the equivalent circuit in Fig. 5c. Rs (Ω cm2)

Rc (Ω cm2)

Y0(Qc) (Ω-1 cm-2 sn)

n-Qc

Rt (Ω cm2)

Y0(Qdl) (Ω-1 cm-2 sn)

n-Qdl

Quality of fit

0 6 21 43 66 88 92 112 125 148

5.007 5.104 5.180 5.166 5.052 5.044 5.016 5.043 5.004 4.976

23200 18200 14100 8218 5040 3640 3360 2710 2130 1880

0.95 × 10-4 0.99 × 10-4 1.20 × 10-4 2.07 × 10-4 5.50 × 10-4 7.80 × 10-4 8.26 × 10-4 1.02 × 10-3 1.24 × 10-3 1.38 × 10-3

0.9354 0.9348 0.9142 0.9068 0.9003 0.9083 0.9094 0.9146 0.9157 0.9162

8227 7255 2159 1860 1590 1793 1840 2441 2534 2529

0.42 × 10-4 0.45 × 10-4 7.25 × 10-3 8.07 × 10-3 8.36 × 10-3 4.41 × 10-3 3.85 × 10-3 1.98 × 10-3 1.61 × 10-3 1.43 × 10-3

0.6686 0.7857 0.8000 0.8973 0.8865 0.7672 0.7382 0.6387 0.6422 0.6044

0.78 × 10-3 0.61 × 10-3 1.53 × 10-3 3.02 × 10-3 2.51 × 10-3 1.39 × 10-3 1.24 × 10-3 0.76 × 10-3 0.49 × 10-3 0.35 × 10-3

SC R

U N A M ED PT CC E A

53

IP T

Time (h)

Table 4 The main electrochemical parameters fitted from the measured EIS data in Fig. 15 by using the equivalent circuits in Fig. 5c and Fig. 16. Rc (Ω cm2)

Y0(Qc) (Ω-1 cm-2 sn)

Rt (Ω cm2)

Y0(Qdl) (Ω-1 cm-2 sn)

Y0(O) (Ω-1 cm-2 s0.5)

Y0(W) (Ω-1 cm-2 s0.5)

Quality of fit

0 6 21 27 43 66 88 92 105

27500 11900 5395 4607 3622 1659 363 294 258

0.53 × 10-4 0.69 × 10-4 2.51 × 10-4 3.24 × 10-4 3.35 × 10-4 3.58 × 10-4 3.72 × 10-4 3.34 × 10-4 3.56 × 10-4

5028 3884 2455 2377 2480 2338 1965 1554 1079

1.10 × 10-3 1.36 × 10-3 3.06 × 10-3 3.24 × 10-3 2.97 × 10-3 3.42 × 10-3 5.23 × 10-3 8.47 × 10-3 13.9 × 10-3

4.04 × 10-4 4.36 × 10-4 5.57 × 10-4 7.73 × 10-4 -

3.69 × 10-3 4.60 × 10-3 8.42 × 10-3

0.53 × 10-3 2.28 × 10-3 0.86 × 10-3 0.61 × 10-3 1.38 × 10-3 0.87 × 10-3 0.96 × 10-3 1.34 × 10-3 1.40 × 10-3

SC R

U N A M ED PT CC E A

54

IP T

Time (h)