Surface Characterization and Corrosion Behavior of 90/10 Copper-Nickel Alloy in Marine Environment

Surface characterization and corrosion behavior of 90/10 copper-nickel alloy in seawater from Xiamen bay at 30 °C for 56 days were investigated in this study. The results indicated that the corrosion product layer was mainly a mixture of CuO, Cu2O, and Cu(OH)2, with a transition to CuCl, CuCl2, and Cu2(OH)3Cl during the corrosion process. However, as corrosion proceeds, the resistance of the product film was reduced due to its heterogeneous and fairly porous structures, which led to local corrosion of the alloy. The corrosion potentials (Ecorr) increase while corrosion current densities (Icorr) decrease with time because of the formation of protective oxide film.


Introduction
Copper-nickel alloys are generally known for their good corrosion resistance, excellent machinability, and outstanding thermal and electrical conductivity in marine environments [1][2][3]. The 90/10 copper-nickel alloy has been widely applied in various fields, such as seawater piping, heat exchangers and condensers in ships, desalination plants, power plants, and ship hulls [4][5][6]. The excellent corrosion resistance of Cu-Ni alloy is mainly due to two reasons. First, the ionization of copper is difficult as a result of the positive equilibrium potential and the high thermodynamic stability of Cu. Second, nickel was incorporated into Cu 2 O film and increases the corrosion resistance in two ways: (i) Ni 2+ occupies the vacant position of Cu + and increase the ionic resistance, which makes two Cu + ions disappear and thus increases the ion resistance of the film; (ii) Ni 2+ replaces Cu + directly, while the ionic resistance does not change and each substitution results in the disappearance of one Cu + and increases the electronic resistance [7]. Moreover, it is relevant to the formation of a duplex oxide layer on the surface of copper alloy, which is mainly composed of Cu x O and Cu-hydroxide/oxide layer and protects the copper matrix. Corrosion products often deposit on the film and play a protective role [4,8,9]. Therefore, copper and copper alloys are more corrosion-resistant than other metal alloys [1].
In marine exposure conditions, Cu 2 O rapidly forms on the surface of the matrix in the initial stage and reacts with chlorides, leading to the formation of CuCl, which usually converts to Cu 2 (OH) 3 Cl as end corrosion products [10,11]. The above components have already been confirmed on bare copper at atmospheric exposures by Fuente [12], and also observed after exposure in laboratory conditions with humidified air and pre-deposited NaCl [13,14]. However, the failure mainly happens on the Cu alloys surface in marine environments and severely disrupts the safe operation of ship and marine engineering [15].
As mentioned in various literatures [16][17][18], it is well-known that the corrosion of Cu-Ni alloys in seawater is a typical electrochemical corrosion and that the main characteristics are: (i) Due to the The surfaces of the specimens were prepared by grinding and polishing with SiC papers and diamond paste sequentially, followed by degreasing in acetone and drying in warm air. The corrosive solution is a mixture of 0.5mol/L H 2 O 2 and natural seawater from Xiamen bay (24 • 27 N latitude, 118 • 04 E longitude, and pH 8.2). In order to eliminate the influence of microbiology, the seawater was sterilized by steam at 120 • C for 20 min. The experiment was carried out in a thermostat water bath at 30 • C, while the oxygen was continuously injected into the corrosive solution with an electromagnetic charge pump.

Morphology Observation Analysis by SEM/EDS
After corrosion in the mix solution, a set of specimens were cleaned with deionized water and then dried in air. Another set of samples was immersed in sulfuric acid solution diluted 10 times for 3 min, of which relative density was 1.84, and then removed the corrosion products by hard brush. The prepared specimens were treated by spray gold, SEM with EDS (CS3400, CamScan, Nottingham, UK) were used to study the microstructure.

Raman Spectroscopy Analysis
The Raman spectroscopy was analyzed to recognize the oxidation state of Cu element on the surface at room temperature by Maya2000 Pro-NIR spectrometer (Ocean Optics, Oxford, UK), and a He-Ne laser with 785 nm line was used to exciting the specimens. The surfaces of the alloy after corrosion for several days in backscattering configuration were examined with a 50-fold objective lens. Spectral measurements were recorded from 100 to 1200 cm −1 at different points on the surface for each specimen.

XPS Analysis
To further detect the composition at different depths of the corrosion surface, the XPS of the prepared specimens was carried out by K-Alpha X-ray photoelectron spectrometer (thermo escalab 250, Thermo Fisher Scientific, Gloucester, UK). Prior to XPS test, the specimens were treated in deionized water and dried using a vacuum oven [33]. The Al Ka X-ray source were used to motivate the high-resolution photoelectron of Cu2p and O1s, while the output data were analyzed by the XPS PEAK 1 software (Version 4.1).

Electrochemical Measurement
The corrosion behavior of the samples immersed for 0 d, 7 d, 14 d, 28 d, and 56 d was characterized by potentiodynamic polarization curves and was determined using a CHI660D workstation (Huachen, Shanghai, China) in the corrosive solution at 20 ± 1 • C with a conventional three-electrode cell. The solutions were in contact to air to get saturated with oxygen, before the electrochemical measurements, while after the measurements, the electrolyte is no longer aerated. The potentiodynamic polarization curves were measured in the range from −0.4 V to 0.0 V (vs. SCE) of a scan rate of 0.01 mV/s the samples.

XPS Analysis of the Corrosi on Product Layer
3.1.1. Cu2p Spectra Figure 1 shows the high-resolution Cu2p spectra (at 925 eV-975 eV) on the corrosion product layer of the prepared samples after exposure in the seawater for 7 days, 14 days, 28 days, and 56 days, respectively. It is noted that the oxide film grows up with the extension time would present a positive shift of binding energy (B.E.) on Cu2p spectra [1]. According to the high resolution of Cu2p spectra in Figure 1, the peaks of Cu2p3/2 and Cu2p1/2 are appearing at 932.5 eV and 952.5 eV, while the core-level of the Cu2p3/2 spectra is presented at B.E. from 928eV to 938 eV as shown in Figure 1a-d, respectively. Besides, the spectrum exhibited the shake-up features around at 943.5 eV and 962 eV for Cu2p3/2 and Cu2p1/2 peaks, respectively, which indicate that CuO is dominant after corrosion due to the presence of the unfilled (d9) valence level of the Cu 2+ ion, and other forms of Cu may exist, such as Cu 2 O and Cu(OH) 2 [34,35].  Figure 1 shows the high-resolution Cu2p spectra (at 925 eV-975 eV) on the corrosion product layer of the prepared samples after exposure in the seawater for 7 days, 14 days, 28 days, and 56 days, respectively. It is noted that the oxide film grows up with the extension time would present a positive shift of binding energy (B.E.) on Cu2p spectra [1]. According to the high resolution of Cu2p spectra in Figure 1, the peaks of Cu2p3/2 and Cu2p1/2 are appearing at 932.5 eV and 952.5 eV, while the core-level of the Cu2p3/2 spectra is presented at B.E. from 928eV to 938 eV as shown in Figure 1 a-d, respectively. Besides, the spectrum exhibited the shake-up features around at 943.5 eV and 962 eV for Cu2p3/2 and Cu2p1/2 peaks, respectively, which indicate that CuO is dominant after corrosion due to the presence of the unfilled (d9) valence level of the Cu 2+ ion, and other forms of Cu may exist, such as Cu2O and Cu(OH)2 [34,35]. Cu XPS spectra of the corrosion product layer on the surface of the 90/10 Cu-Ni alloy in the natural seawater for different days: (a) Core-level Cu2p3/2 spectrum for 7 days; (b) core-level Cu2p3/2 spectrum for 14 days; (c) core-level Cu2p3/2 spectrum for 28 days; (d) core-level Cu2p3/2 spectrum for 56 days.
The relative quantities in Figure 1 are shown in Table 2. It is demonstrated that the primary components of the product film are CuO and Cu2O with the prolonging exposure time from 7 days to 56 days. It is difficult to distinguish the peaks of metallic Cu and Cu2O due to the similar binding energy, so they are expressed as Cu/Cu2O [33]. After exposure to seawater for seven days, as displayed in Figure 1a, two peaks at the B.E. of 933.5eV and 932.6eV are presented in the deconvolution of the core-level Cu2p3/2, and it can be seen that Cu species were composed of 43.2% CuO and 56.8% Cu/Cu2O as listed in Table 2 [36,37]. Figure 1b shows 3 peaks at B.E. of 933.5, 931.9 and 932.5 eV in the deconvolution of core-level Cu2p3/2 spectrum after corrosion for 14 days. According to Table 2, it is believed that the constituent of the corrosion product film is 69.6% CuO, 6.1% CuCl, and 24.3% Cu/Cu2O after corrosion for 14 days [36][37][38]. The results indicated that the primary constituents of the product film are CuO and Cu2O, while a small amount of CuCl was also existed in the layer. It should be pointed out that the formation of a porous CuCl is mainly identified as the first stage in the corrosion process of Cu alloys [2,17,39]. With a prolonged exposure time of 28 days, as shown in Figure 1c and Table 2, there are three deconvoluted peaks of core-level Cu2p3/2 spectrum at B.E. of 934.5, 933.5, and 932.5 eV, which are corresponded to CuCl2, CuO, and Cu/Cu2O [36][37][38][39][40]. It should be noted that CuCl disappeared and a small amount of CuCl2 (12.6%) was detected instead. Furthermore, the relative quantity of Cu2O increased significantly from 24.3% to 44.7%, while the CuO decreased from 69.6% to 42.7% after 28 days of exposure. These indicated that a cuprous oxide layer (Cu2O) and a cupric oxide layer (CuO) formed the duplex corrosion product layer on the alloy surface [1]. In the later stage of corrosion, the core-level Cu2p3/2 Figure 1. Cu XPS spectra of the corrosion product layer on the surface of the 90/10 Cu-Ni alloy in the natural seawater for different days: (a) Core-level Cu2p3/2 spectrum for 7 days; (b) core-level Cu2p3/2 spectrum for 14 days; (c) core-level Cu2p3/2 spectrum for 28 days; (d) core-level Cu2p3/2 spectrum for 56 days.
The relative quantities in Figure 1 are shown in Table 2. It is demonstrated that the primary components of the product film are CuO and Cu 2 O with the prolonging exposure time from 7 days to 56 days. It is difficult to distinguish the peaks of metallic Cu and Cu 2 O due to the similar binding energy, so they are expressed as Cu/Cu 2 O [33]. After exposure to seawater for seven days, as displayed in Figure 1a, two peaks at the B.E. of 933.5eV and 932.6eV are presented in the deconvolution of the core-level Cu2p3/2, and it can be seen that Cu species were composed of 43.2% CuO and 56.8% Cu/Cu 2 O as listed in Table 2 [36,37]. Figure 1b shows 3 peaks at B.E. of 933.5, 931.9 and 932.5 eV in the deconvolution of core-level Cu2p3/2 spectrum after corrosion for 14 days. According to Table 2, it is believed that the constituent of the corrosion product film is 69.6% CuO, 6.1% CuCl, and 24.3% Cu/Cu 2 O after corrosion for 14 days [36][37][38]. The results indicated that the primary constituents of the product film are CuO and Cu 2 O, while a small amount of CuCl was also existed in the layer. It should be pointed out that the formation of a porous CuCl is mainly identified as the first stage in the corrosion process of Cu alloys [2,17,39]. With a prolonged exposure time of 28 days, as shown in Figure 1c and Table 2, there are three deconvoluted peaks of core-level Cu2p3/2 spectrum at B.E. of 934.5, 933.5, and 932.5 eV, which are corresponded to CuCl 2 , CuO, and Cu/Cu 2 O [36][37][38][39][40]. It should be noted that CuCl disappeared and a small amount of CuCl 2 (12.6%) was detected instead. Furthermore, the relative quantity of Cu 2 O increased significantly from 24.3% to 44.7%, while the CuO decreased from 69.6% to 42.7% after 28 days of exposure. These indicated that a cuprous oxide layer (Cu 2 O) and a cupric oxide layer (CuO) formed the duplex corrosion product layer on the alloy surface [1]. In the later stage of corrosion, the core-level Cu2p3/2 spectra for 56 days exhibited three peaks at B.E. of 934.5, 933.5, and 932.4 eV, which is due to CuCl 2 , CuO, and Cu/Cu 2 O, as shown in Figure 1d [36][37][38][39][40]. Besides, according to the relative quantity of Cu2p3/2 spectra, two obvious features can be discerned: (i) The relative amount of CuO (58.4%) and Cu 2 O (32.4%) were major in Cu2p3/2 spectra over the corrosion period, despite the Cu 2 O decreased from 44.7% to 32.4%; (ii) a small amount of CuCl 2 (9.2%) was also detectable in the Cu2p spectra. Table 2. Fitting parameters for Cu 2p3/2 XPS spectra and relative quantity of compounds in the corrosion layers of 90/10 Cu-Ni alloy exposed to seawater for different days.

O1s spectra
The O1s spectra are considered together with the spectra of Cu2p to predict the possible corrosion products [41]. Figure 2 shows the O1s spectra (at 525 eV-538 eV) of the corrosion product film of the samples surface immersed in the seawater for 7 days to 56 days. The core-level spectrum presents three peaks at B.E. of 530.5, 531.6, and 532.3 eV in the high-resolution O1s spectra. It should be noted that, due to the existence of Ni element in Cu-Ni alloy and for the same type of oxides, their binding energy was probably the same. Thus, the peaks at the B.E. of 530.5, 531.6, and 532.3 eV, as listed in Table 3, could be identified as CuO/NiO, Cu 2 O, and Cu(OH) 2 /Ni(OH) 2 [2,41].
As given in Table 3, at the early stage of corrosion (seven days), the main component in the corrosion product layer is presented as CuO (43%). With the immersion time extending to 14 days, Cu(OH) 2 /Ni(OH) 2 (45%) became more dominant than CuO (31%), while at longer exposure times (56 dasy), the relative quantity of CuO increased significantly from 31.0% to 50.4%, while the Cu(OH) 2 /Ni(OH) 2 decreased from 45.2% to 36.3% after 56 days of exposure. These results are generally consistent with Cu2p spectra. However, the difference is that the Cu(OH) 2 did not appear in the Cu2p spectra, and this phenomenon is in accordance with the literature by Xiao [33]. They considered that the intensity of the Cu(OH) 2 in Cu2p spectra is pretty weak [33], while the combination of Cu(OH) 2 and Ni(OH) 2 lead to the result in O1s spectra.
corrosion products [41]. Figure 2 shows the O1s spectra (at 525 eV-538 eV) of the corrosion product film of the samples surface immersed in the seawater for 7 days to 56 days. The core-level spectrum presents three peaks at B.E. of 530.5, 531.6, and 532.3 eV in the high-resolution O1s spectra. It should be noted that, due to the existence of Ni element in Cu-Ni alloy and for the same type of oxides, their binding energy was probably the same. Thus, the peaks at the B.E. of 530.5, 531.6, and 532.3 eV, as listed in Table 3, could be identified as CuO/NiO, Cu2O, and Cu(OH)2/Ni(OH)2 [2,41]. As given in Table 3, at the early stage of corrosion (seven days), the main component in the corrosion product layer is presented as CuO (43%). With the immersion time extending to 14 days,  The current XPS study demonstrates that the corrosion of 90/10 Cu-Ni alloy exposed in marine environment contains the cathodic (reduction of O 2 ) and anodic reaction (oxidation of Cu) as follows [2,41]: The CuCl always generates on the alloy surface and forms the loose CuCl film due to the low solubility [44]. Along with the corrosion time prolonging, the CuCl adsorption film dissolves gradually and forms the copper complex of CuCl 2 − , as follows: Besides, the CuCl − 2 would react with OH − to form the cuprous Cu 2 O film, which would reduce the corrosion rate of the Cu alloy in seawater, as follows [1]: In the marine environment, there is a large amount of dissolved oxygen as described earlier in the "Introduction" section. Carley et al [45] pointed out that the formation of a Cu(II) state on the surface of Cu alloy was due to the existence of adsorbed oxygen. With long exposure time, Cu 2 O hydrolysis generated Cu(OH) 2 as follows [1,46]:

Raman Analysis of the Corrosion Product Layer
The Raman spectrum was investigated of the alloy surface with several corrosion days for better identified the structures of the product film, as illustrated in Figure 3. Each sample was tested three times by Raman spectroscopy to confirm the accuracy. The presence of Cu 2 O (peaks at 151 cm −1 and 411 cm −1 [47]), CuO (peak at 511 cm −1 [48]), CuCl (peak at 1079 cm −1 [48]), and Cu(OH) 2 (peak at 292 cm −1 [47]) are detected in the corrosion product, which is in accordance with the XPS results. However, the Cu 2 (OH) 3 Cl (peak at 217 cm −1 [48]) is also found in Raman measurement, while it was not detected in the XPS spectra, demonstrating that just a small amount of Cu 2 (OH) 3 Cl exists on the corrosion surface [3].  [47]) are detected in the corrosion product, which is in accordance with the XPS results. However, the Cu2(OH)3Cl (peak at 217 cm −1 [48]) is also found in Raman measurement, while it was not detected in the XPS spectra, demonstrating that just a small amount of Cu2(OH)3Cl exists on the corrosion surface [3].

Macrostructure Analysis
The macroscopic morphologies of the specimens corroded for 7 d, 14 d, 28 d, and 56 d are shown in Figure 4. The red corrosive products were presented obviously on the surface of the alloy after immersed for 7 d, as carried out in Figure 4a, which were considered as Cu oxidations (Cu2O), as mentioned by Qin et al. [43,49]. With the corrosion time prolonging, the massive red corrosion products were accumulated, some yellow corrosion products were formed on the alloy surface, and the color turned darker, as presented in Figure 4b,c. After a longer corrosion time (immersion for 56 d), the whole surface turned brown and piled up some green corrosion products, as displayed in Figure 4d.

Macrostructure Analysis
The macroscopic morphologies of the specimens corroded for 7 d, 14 d, 28 d, and 56 d are shown in Figure 4. The red corrosive products were presented obviously on the surface of the alloy after immersed for 7 d, as carried out in Figure 4a, which were considered as Cu oxidations (Cu 2 O), as mentioned by Qin et al. [43,49]. With the corrosion time prolonging, the massive red corrosion products were accumulated, some yellow corrosion products were formed on the alloy surface, and the color turned darker, as presented in Figure 4b,c. After a longer corrosion time (immersion for 56 d), the whole surface turned brown and piled up some green corrosion products, as displayed in Figure 4d. after immersed for 7 d, as carried out in Figure 4a, which were considered as Cu oxidations (Cu2O), as mentioned by Qin et al. [43,49]. With the corrosion time prolonging, the massive red corrosion products were accumulated, some yellow corrosion products were formed on the alloy surface, and the color turned darker, as presented in Figure 4b,c. After a longer corrosion time (immersion for 56 d), the whole surface turned brown and piled up some green corrosion products, as displayed in Figure 4d. The surface morphologies of the sample after removing the corrosion products were shown in Figure 5. It is revealed that the surface of the sample showed slight corrosion pits after corrosion for The surface morphologies of the sample after removing the corrosion products were shown in Figure 5. It is revealed that the surface of the sample showed slight corrosion pits after corrosion for seven days, and the corrosion area gradually increased with the extension of time. With the time prolonging to 56 days, the corrosion area increased significantly, and the surface of the sample was exfoliated.  Generally speaking, the corrosion products of B10 Cu-Ni alloy are multilayer [50]. The inner layer film is a kind of uniformly adhered brick red corrosion product, and often turns into a black and membranous cuprous oxide layer. Due to the superposition of red and black oxide layers, the protective film on the corrosion of Cu alloys are often yellow-brown in appearance. In the presence of Cl − in seawater, due to the combined action of Cland dissolved oxygen in seawater, the red/black oxide product film is partially damaged, which leads to the deepening of the local corrosion of the Cu matrix. The green and loose corrosion products are often formed when the serious corrosion occurs and in form of piles or patches.

Microstructure Analysis
The surface morphologies and composition of the corrosion product films deposited on the alloy matrix were studied by the use of SEM and EDS energy (dispersive X-ray spectrometry) analysis, as shown in Figure 6 and Table 4. Generally speaking, the corrosion products of B10 Cu-Ni alloy are multilayer [50]. The inner layer film is a kind of uniformly adhered brick red corrosion product, and often turns into a black and membranous cuprous oxide layer. Due to the superposition of red and black oxide layers, the protective film on the corrosion of Cu alloys are often yellow-brown in appearance. In the presence of Cl − in seawater, due to the combined action of Cl − and dissolved oxygen in seawater, the red/black oxide product film is partially damaged, which leads to the deepening of the local corrosion of the Cu matrix. The green and loose corrosion products are often formed when the serious corrosion occurs and in form of piles or patches.

Microstructure Analysis
The surface morphologies and composition of the corrosion product films deposited on the alloy matrix were studied by the use of SEM and EDS energy (dispersive X-ray spectrometry) analysis, as shown in Figure 6 and Table 4.

Microstructure Analysis
The surface morphologies and composition of the corrosion product films deposited on the alloy matrix were studied by the use of SEM and EDS energy (dispersive X-ray spectrometry) analysis, as shown in Figure 6 and Table 4.    Table 4. The corresponding microstructure of the surface after removing corrosion products are shown in Figure 6c1-c4.
After corrosion for seven days, as shown in Figure 6a1,b1, the pitting areas are small and the Cu matrix is still visible. Besides, some white and grey products gradually appeared on the surface. According to the EDS results, the product film was, perhaps, a combination of the cupric oxide film (CuO and Cu(OH) 2 ) and cuprous oxide (Cu 2 O) [1]. The corresponding microstructure of the surface after removing corrosion products are shown in Figure 6c1, it is shown that the pores initiate from the specimen surface with white corrosion products after corrosion for seven days, which was in agreement with the Reference [51]. After being corroded for 14 days, the pitting area increased significantly and lots of corrosion products piled up on the surface, as shown in Figure 6a2,b2. The EDS revealed the presence of copper and oxygen, which were likely present as CuO and Cu 2 O. Besides, it is indicated that the typical intergranular corrosion appeared on the alloy surface after pickling as demonstrated in Figure 6c2, which was in agreement with the Reference [52]. When testing time up to 28 days, as shown in Figure 6a3,b3, much more particle-like corrosion products deposited on the surface and some products fall off from the sample because of the inadhesive structure and stresses [53]. The EDS analysis indicated that the corrosion products mainly consisted of the oxide and chlorine, and the results are also in accordance with literature [3,8]. According to the microstructure after pickling, the exfoliation corrosion presented on the surface as shown in Figure 6c4. With the extension corrosion time to 56 days, plenty of particle-like corrosion products accumulated and almost fully cover the entire surface of the alloy, as observed in Figure 6a4,b4. Furthermore, there was obvious speckle corrosion in the Figure 6c4, and exfoliation also occurred locally.
The results demonstrate that the corrosion product films on the surface are heterogeneous and fairly porous, as well as all present cracks and exfoliation, which are due to the compactness, high fragility, poor cohesion, and adhesion [12]. Generally speaking, the corrosion of the sample first started from the grain boundary. With the prolongation of immersion time, corrosion continues to expand along grain boundaries, and gradually develops to the surface of the whole sample, forming copper oxide film. At the beginning of film formation, the film is not uniform and the adsorption of Cl − is also inconsistent. The nickel corrosion occurs on the surface of the sample, resulting in poor protective performance of the film in some areas. The Cu 2 O film is divided into inner layer and outer layer [54], in which the outer layer is produced by the deposition of dissolved Cu 2+ , while the inner layer is produced by the more compact and protective Cu 2 O film. The grain boundaries, twins, or other defects on the Cu matrix are under the corrosion product film. The growth rate of the corrosion product film in these areas is slower than that in the intergranular due to the difference of structure and composition. As the corrosion time was prolonged, the oxidation products in the outer layer formed on the surface and the content gradually increased. However, the strength of the corrosion product film is lower than its internal stress, so the products are porous and easy to fall off from the sample, which led to cracks in the film [50]. Figure 7 demonstrates the polarization curves of the 90/10 Cu-Ni alloy after being exposed for several days. The corresponding values calculated from the polarization curves, such as anodic and cathodic Tafel slope (β a and β c , respectively), corrosion potentials (E corr ), and corrosion current densities (I corr ), are given in Table 5. The resistance of the polarization R p can be calculated as follows:

Polarization Curve Measurements
Materials 2019, 12, x FOR PEER REVIEW 10 of 14 strength of the corrosion product film is lower than its internal stress, so the products are porous and easy to fall off from the sample, which led to cracks in the film [50]. Figure 7 demonstrates the polarization curves of the 90/10 Cu-Ni alloy after being exposed for several days. The corresponding values calculated from the polarization curves, such as anodic and cathodic Tafel slope (βa and βc, respectively), corrosion potentials (Ecorr), and corrosion current densities (Icorr), are given in Table 5. The resistance of the polarization Rp can be calculated as follows: Figure 7. The polarization curves of Cu-Ni alloy after being exposed in corrosion solution for different days.   It can be observed that the E corr shift about 50 mV to positive direction with the extension corrosion time to 56 days, which was in virtue of the inhibitive effect of protective oxide film and corrosion on the Cu surface as a result of decreasing the rate of the anodic reaction [55]. Besides, the I corr decreased quickly from 3.812 to 1.056 µA/cm 2 after 56 days of corrosion. This were mainly due to the generation of a protective oxide film and the reduction of the Cl − attack on the alloy surface, which resulted in the reduction of corrosion rate and the homogeneous corrosion of Cu alloy. In addition, the increasing β a was concerned with the decreasing corrosion reaction caused by the reduction of anodic currents. This is identical with the experiment results of Zhu Xiao in studying the corrosion behavior of a novel Cu alloy in a salt spray environment [34].

Polarization Curve Measurements
Generally, in the initial stage of corrosion, the surface of Cu matrix is relatively smooth, while the halogen ions, dissolved oxygen, and corrosion products are free to diffuse on the surface of the matrix, so the corrosion of the matrix surface is controlled by activation. Cu is oxidized to CuO in the passivation zone, while Cu 2 O is oxidized in the activation zone, and chloride is formed by reacting with Cl − in marine environment. Cu 2 O has thermodynamic instability and is easy to be oxidized into copper ions. With the increasing immersion time, the corrosion product film gradually forms on the surface of the matrix and the Cu 2 O film became thicker, so the combination of protective oxide film and corrosion product film inhibits the diffusion of halogen ions, dissolved oxygen, and corrosion products. At this point, the anode reaction is dominated by the mass transfer process, while the cathode process remains relatively unchanged, which will lead to an increase in corrosion potential. Studies [55,56] have shown that the electronic properties of the surface film are related to its composition and structure, so that the surface film shows different types of semiconductors in different potential ranges. Generally speaking, Cu 2 O is a typical P-type semiconductor, while high valence oxides, such as Ni and Fe, are N-type semiconductors. This bipolar structure can play a good protective role. For the B10 alloy, in the range below the flat-band potential, the space charge layer of Ni and Fe oxides in the surface film is in the state of enrichment and acts as the conductor, while the space charge layer of the Cu oxide part is depleted and shows a P-type semiconductor. On the contrary, in the range higher than the flat potential, the space charge layer of Cu oxide part in the surface film is in the enrichment state, which is equivalent to the conductor, while the space charge layer of Ni and Fe oxide is depleted and behave N-type semiconductor. According to the point defect model, the corrosion product film contains various high concentration point defects, such as metal ion vacancy and oxygen vacancy. In marine environment, on the one hand, the Cu 2 O on the surface of the film may occur redox transformation of copper oxide (CuO) or copper hydroxide (Cu(OH) 2 ). On the other hand, when the corrosion product film contacts with Cl − , the oxygen vacancies in the N-type semiconductor on the outer surface of the surface film would adsorb Cl − in solution and Mott-SchottkyPair reaction would occur with Cl − to produce oxygen/metal ion vacancy pairs. The oxygen vacancies can also generate more metal ion vacancies with other Cl − reactions. Thus, more and more metal vacancies accumulate at the metal matrix/film interface, isolating the matrix from the surface film and preventing the film from continuing to grow. Furthermore, due to the increased potential as mentioned above, Cl − adsorbed on the surface film and resulted in the thickened Cu 2 O film, which formed a porous green corrosion product, namely, cupric chloridehydroxide (Cu 2 (OH) 3 Cl) [50]. When the product layer grows to the inner film, the dynamic equilibrium of the dissolution of the outer layer is destroyed, causing the local tension of the product film to exceed the theoretical breaking strength, resulting in penetrating rupture of the surface film. Therefore, the greater the concentration of defects (oxygen vacancy and metal ion vacancy) in the surface film is, the more vulnerable the surface coating is to erosive ions. As corrosion proceeds, continuous Cl − penetrates into the defects of the CuO film and reduces the resistance of the protective film, resulting in localized collapse, and pittings are formed on the surface of the substrate and expanded with the growth of the microvoids under the aggressive attack of Cl − [42]. At this time, the dissolution of copper produces new corrosion active sites, which hinders the further formation of protective film and leads to more serious local corrosion on the surface.

Conclusions
Surface characterization and corrosion behavior of 90/10 copper-nickel alloy immersed in seawater from Xiamen bay at 30°C for 56 days were investigated using XPS, SEM analyses, and electrochemical measurements. According to the results, following conclusions can be summarized:

1.
The corrosion product formed on the surface of the 90/10 Cu-Ni alloy mainly contained CuO, Cu 2 O and Cu(OH) 2 . With increasing exposure time, CuCl, CuCl 2 , and Cu 2 (OH) 3 Cl were also formed on the surface, which to some extent protected the alloy from corrosion.

2.
The corrosion product films on the surface are heterogeneous and fairly porous, as well as all present cracks and exfoliation, which are due to the compactness, high fragility, poor cohesion, and adhesion. More serious local corrosion would occur in the 90/10 Cu-Ni alloy during the process of seawater corrosion.

3.
Electrochemical results showed that, with the extension of exposure time, the E corr shifted to positive direction, while the I corr shifted to negative, which were due to the inhibitive effect of protective oxide film and corrosion on the Cu surface as a result of decreasing the rate of the anodic reaction and the Cl − attack on the surface.