Effect of cathodic polarisation on the corrosion behaviour of 316L stainless steel under static and dynamic conditions

2.1 Specimens, chemicals, and surface preparation

A 0.1 m NaCl solution was prepared using analytical grade NaCl powder (Sigma-Aldrich, Fluka, St. Louis, MO, USA).

All experiments in this work were performed on 316L stainless steel coupons with the following chemical composition (in wt%): C 0.024, Si 0.46, Mn 1.22, P 0.038, S 0.0048, Ni 9.69, Cr 16.45, Mo 1.95, Cu 0.35, and Al 0.020. All coupons are 12.5 mm in diameter and 2 mm thick. They were ground with 600-grit silicon carbide paper (Buehler, Düsseldorf, Germany) and then polished with a polishing cloth (MicroCloth, Buehler, Düsseldorf, Germany) using 1.0 μm and then 0.5 μm alumina particles. Finally, the samples were washed with distilled water and acetone and rinsed ultrasonically in ethanol for 5 min. To apply similar experimental conditions on the samples, all measurements were performed within 1 h of polishing.

2.2 Cell design and electrochemical measurements

In this study, we constructed two different dynamic condition systems. One contains a rotating disc electrode (RDE) and the other one is a homemade flow cell setup. The RDE setup (CTV101-Radiometer Copenhagen, France), with a maximum of 5000 rotations/min (rpm), was immersed in a fiberglass cell with 100 ml electrolyte, and for every measurement, a proper rotation rate was applied through the digital speed controller. The homemade flow cell (Figure 2) had an almost 1 l electrolyte container and was equipped with a water pump (ProFlow u 500, 6 W, max. 440 l/h, D-67141, Neuhofen, Germany). In this setup, the electrodes were immersed in a fiberglass cell with an inner diameter of 3.5 cm and the electrolyte was pumped through a plastic tube with an inner diameter of 6 mm. This way, a maximum flow rate of 8.6 cm³/s was applied over the sample.

Figure 2:

Schematic diagram of the homemade flow cell.

WE, working electrode (steel coupon); RE, reference electrode (Ag/AgCl/KCl sat.); CE, counter electrode (graphite rod).

Both static and dynamic setups used a graphite rod with a diameter of 2 mm as a counter electrode and a saturated Ag/AgCl/KCl reference electrode attached to an Autolab Eco Chemie potentiostat (PGSTAT 10, Utrecht, The Netherlands). All measurements were performed in a 0.1 m NaCl solution.

To apply cathodic polarisation to the samples, a potentiostatic polarisation was carried out at three different potentials (-1.3, -1.5, and -1.7 V), each time for a duration of 600 s. In this work, after each potentiostatic polarisation measurement, the generated hydrogen gas bubbles were removed from the surface of the steel electrode by gently stirring the solution after the measurement under static conditions. Under dynamic conditions, because the produced fluid flow removed the gas bubbles from the surface during the cathodic polarisation process, the removal of hydrogen gas bubbles after potentiostatic polarisation was not necessary.

The test specimens were immersed into the NaCl solution for approximately 16 h until a stable value of open circuit potential (OCP) was attained. After this step, the potentiodynamic polarisation curves of the samples were recorded versus the reference electrode with and without cathodic polarisation with a scan rate of 1 mV/s. In addition, electrochemical impedance spectroscopy measurements of the steel samples were carried out in the frequency range of 10 kHz to 0.01 Hz, with an amplitude of 10 mV at a stable OCP. Nyquist plots were recorded with and without cathodic polarisation of each sample at different cathodic potentials and under static and flow conditions.

2.3 Scanning electron microscopy (SEM) and atomic force microscopy (AFM)

The surface topography of the samples was examined by AFM with and without cathodic polarisation by three different cathodic potentials under static and flow conditions. For this, we used a Dimension Edge microscope (Bruker, Billerica, MA, USA) with an antimony-doped silicon tip in a tapping mode. The values of the roughness parameters (in this work, R_a and R_q) were provided by the software of the instrument.

SEM energy-dispersive X-ray spectroscopy (EDS) analyses were performed by a JSM-7600F SEM (JEOL, Tokyo, Japan) and an X-MAX^N-50 EDS detector (Oxford Instruments, Abingdon, Oxfordshire, UK).

3.1 Flow motion and produced flow regime

It is known that the rotation of the electrode produces a fluid flow in a vertical direction from the solution towards the electrode surface (Baune, Breunig-Lyriti, & Plath, 2002; Real-Ramirez, Tello, Reyes, & Trejo, 2010). The fluid flow constructed by the flow cell, however, has a different regime. Here, the electrolyte is pumped in a straight direction over the surface and exits from the cell through the outlet tube. However, we know that the electrolyte stream produced by the flow cell makes a complex vortex flow regime when it reaches the immersed electrode, which acts here as a semiwall inside the cell.

In this work, these two different flow regimes were applied on the steel samples at various velocities, and the corrosion behaviour of the steel with and without cathodic polarisation at three different cathodic potentials was investigated.

3.2 Potentiodynamic polarisation under static and flow conditions

Figure 3 shows the potentiodynamic polarisation curves of freshly polished 316L stainless steel coupons in a 0.1 m NaCl solution under static and flow conditions. These graphs were obtained using the homemade flow cell and RDE. The applied flow rates were 1.3, 5.5, and 8.6 cm³/s for the flow cell and 50 up to 3000 rpm for the RDE. These curves show that the potentiodynamic polarisation of steel in the presence of a fluid flow results in higher cathodic current densities, in which cathodic current densities increase with increasing flow rate for both flow regimes. This can be attributed to the higher rate of the cathodic reduction reactions, especially that of the dissolved oxygen in the flow system (Kim et al., 2006). In Figure 3A (flow cell), the corrosion potential (E_corr) of steel changes from -0.39 to -0.21 V when a flow velocity of 1.3 cm³/s is applied, increasing to -0.13 V for a flow rate of 8.6 cm³/s. However, the anodic current densities obtained for different flow velocities show no significant difference.

Figure 3:

Potentiodynamic polarisation curves of steel without cathodic polarisation under static and flow conditions obtained using (A) the flow cell and (B) an RDE.

Similar results were obtained using the RDE (Figure 3B). Here, the E_corr of steel shifts to -0.24 V under 50 rpm rotation rate, and parallel to the increase of the rotation rate, this potential increases. Finally, when applying a high rotation rate of 3000 rpm, the E_corr appears to be 0.15 V.

Based on these measurements and considering the anodic and cathodic reactions that occurred on the steel surface during the anodic dissolution, the observed current and potential shifts can be explained as follows.

During the corrosion of steel in solutions, such as NaCl, the oxidation of iron and the reduction of oxygen are the main anodic and cathodic electrochemical reactions, respectively.

The anodic dissolution of iron is known as a pure charge transfer-controlled process (Bockris & Drazic, 1962). However, the oxygen reduction reaction has been described in the literature as a mixed charge transfer and mass transfer-controlled process (Caceres, Vargas, & Herrera, 2007; Davydov et al., 2005; Miyata & Asakura, 2002). In the presence of a fluid flow, the mass transport-limited cathodic reduction of dissolved oxygen is dominant. Therefore, under flow conditions, the mass transfer of oxygen from the electrolyte to the metal surface, and therefore the cathodic current density, increases. In our opinion, the hydroxyl ions produced during the oxygen reduction reaction increase the pH of an electrolyte. This supports the surface passivation process and shifts the corrosion potential towards nobler potentials.

3.3 Effect of the flow velocity on the pit nucleation

According to our measurements on stainless steel, the maximum flow velocity at which pits can form was found at approximately 1.3 cm³/s using the flow cell described in Figure 2. Figure 4 shows the optical micrographs of four steel coupons after potentiodynamic polarisation in 0.1 m NaCl with a scan rate of 1 mV/s under static conditions (a) and under a fluid flow (b–d) constructed by the homemade flow cell. Figure 4A clearly shows pits that cover the entire surface; after the polarisation of the sample under a 1.3 cm³/s velocity, the number of pits was reduced, and under higher velocities (5.5 and 8.6 cm³/s), no pitting was observed.

Figure 4:

Optical micrographs of pits observed on the steel coupons after potentiodynamic polarisation in a 0.1 m NaCl solution under different flow velocities produced by the flow cell: (A) static, (B) 1.3 cm³/s, (C) 5.5 cm³/s, and (D) 8.6 cm³/s (scan rate of 1 mV/s).

Similar experiments were performed using an RDE to indicate the maximum rotation speed where pitting occurs on the steel surface. The results show that, with rotation rates below 500 rpm, pitting is easily detected after potentiodynamic polarisation. However, higher rotation rates prevent pitting over the steel surface.

3.4 Effect of the flow velocity on the cathodically polarised steel

Figure 5A shows the potentiodynamic polarisation curves of steel with and without cathodic polarisation of the samples at three different cathodic potentials under static conditions.

Figure 5:

Potentiodynamic polarisation curves of steel with and without 600 s cathodic polarisation by three different cathodic potentials: (A) static, (B) fluid flow applied by the RDE (500 rpm), and (C) 8.6 cm³/s fluid flow applied by the flow cell.

In this figure, the cathodic parts of the potentiodynamic polarisation curves after cathodic polarisation of the samples exhibit a greater current density than the potentiodynamic polarisation curve of the sample without cathodic polarisation, which can be attributed to the higher rate of the cathodic reactions on the electrode-electrolyte interface. The corrosion potential (E_corr) of the nonpolarised sample is approximately -0.39 V, which increases to -0.09 V after polarisation at -1.3 V for 600 s. By increasing the cathodic polarisation potential to -1.5 V, the E_corr shifts to -0.05 V, and finally, after polarisation at -1.7 V for 600 s, the E_corr reaches 0.01 V. Unlike in Figure 3, the cathodically polarised samples show higher anodic current densities than the nonpolarised samples, increasing corrosion current densities for all cathodically polarised samples.

The potentiodynamic graphs in Figure 5A show that more oxygen gets reduced on the cathodically polarised steel. Therefore, it can be concluded that the rate of the oxygen reduction reaction increases at cathodically polarised surfaces under static conditions and raises the cathodic current density during potentiodynamic polarisation.

Furthermore, when the potential reaches to the anodic part of the potentiodynamic polarisation curves, the hydrogen-containing micropits are assumed to act as preferential sites for chloride ions and increase the anodic current density and favour pitting.

Figure 5B shows the potentiodynamic polarisation curves of the same series of the experiments measured under the RDE generating flow conditions with a rotation rate of 500 rpm. This figure reveals that the cathodic polarisation of steel under the flow regime of the RDE increases the cathodic current density produced by the reduction of oxygen during the potentiodynamic polarisation of steel. Similar to Figure 5A, here also on the cathodically polarised surface, the rate of the oxygen reduction reaction increases and results in higher cathodic current density in the measured potentiodynamic polarisation curves. Moreover, the anodic current density of steel increases, and the pitting potential, which is known to be independent of flow (Harb & Alkire, 1989), shifts towards less noble potentials after the application of greater cathodic potentials. We obtained similar results with only higher current densities when higher rotation rates (up to 5000 rpm) were applied.

Figure 5C shows the potentiodynamic polarisation curves of steel with and without 600 s cathodic polarisation by three cathodic potentials under flow conditions constructed by the flow cell with a flow velocity of 8.6 cm³/s. Unlike the static conditions and the curves obtained under the flow regime of the RDE, here after the application of cathodic polarisation, the cathodic current densities decrease. In addition, the E_corr of the steel shifts towards less noble potentials. Under a flow of 8.6 cm³/s, the E_corr is found to be -0.13 V, which decreases to -0.31, -0.38, and -0.46 V after cathodically polarising the steel samples with cathodic potentials of -1.3, -1.5, and -1.7 V for 600 s, respectively. When considering anodic current densities, on the contrary, similarly in all polarisation curves represented in Figure 5A–C, the cathodically polarised samples show higher anodic current densities and less noble pitting potentials.

The potentiodynamic polarisation curves of Figure 5C reveal that, under the flow regime of the flow cell, the rate of the oxygen reduction reaction decreases with the application of a larger cathodic potential during the cathodic polarisation process. This is attributed to the decrease of the mass transfer rate of the dissolved oxygen, leading to the reduction of the cathodic current density.

Figure 6 shows the variation of the E_corr of the cathodically polarised steel versus the applied cathodic potentials under both static and flow conditions with two flow regimes. Each E_corr value represents an average of at least three measurements. When a greater cathodic potential is applied during cathodic polarisation under flow conditions constructed by the flow cell (8.6 cm³/s), the E_corr decreases. However, under static conditions and also under the flow regime of the RDE, the E_corr increases with increasing the applied cathodic potential.

Figure 6:

Variations of the E_corr of steel versus the applied cathodic potential under static conditions and two different flow regimes produced by the flow cell and the RDE.

Figure 7 depicts the hydrogen reduction process on the steel surface during cathodic polarisation and its effect on the cathodic current density produced under static condition and two different flow conditions (RDE and flow cell) for three different cathodic potentials. The dotted arrows show the cathodic current reduction after the accumulation of hydrogen gas bubbles on the steel surface under static conditions. In the case of 320 and 500 s cathodic polarisation at a potential of -1.7 V under static conditions, a rapid increase in the cathodic current density is detected, which is attributed to the exit of hydrogen bubbles. The formed gas bubbles during cathodic polarisation decrease the active surface area and therefore reduce the cathodic current density under static conditions. According to our observations during the measurements, under static conditions, small gas bubbles appear from the beginning of the polarisation process. These small bubbles become larger and cover a larger area of the surface and decrease the current density. Further increase in the size of the bubbles on the surface exits the hydrogen bubble and once again increases the cathodic current density. However, under both flow conditions, the chronoamperograms exhibit almost constant values during 600 s polarisation.

Figure 7:

Chronoamperograms of steel for three different cathodic potentials (-1.3, -1.5, and -1.7 V) under static conditions and two different flow regimes produced by the flow cell and the RDE.

3.5 Surface morphology of the cathodically polarised steel

Figure 8 shows the AFM images of the steel surface with and without cathodic polarisation for 600 s at three different cathodic potentials under both static and flow conditions constructed by the flow cell. The cathodic corrosion of the surface forms metal/alloy particles that are transferred into the electrolyte. After 600 s polarisation of the surface in a static mode, the surface roughness increases as the cathodic potentials become larger. This observation intensifies under flow conditions, with the maximum surface roughness reaching 44.5, 56.9, and 69.9 nm after 600 s cathodic polarisation at -1.3, -1.5, and -1.7 V, respectively.

Figure 8:

AFM images of steel (A) without cathodic polarisation, (B) with 600 s cathodic polarisation by -1.3 V under static conditions, (C) with 600 s cathodic polarisation by -1.5 V under static conditions, (D) with 600 s cathodic polarisation by -1.7 V under static conditions, (E) with 600 s cathodic polarisation by -1.3 V under a flow velocity of 8.6 cm³/s, (F) with 600 s cathodic polarisation by -1.5 V under a flow velocity of 8.6 cm³/s, and (G) with 600 s cathodic polarisation by -1.7 V under a flow velocity of 8.6 cm³/s.

The flow conditions were produced by the flow cell.

Figure 9 shows the correlation between the roughness parameters (R_a and R_q) of the steel and the cathodic potentials applied under both static and flow conditions. Here, R_a and R_q are the arithmetic mean and root mean square, respectively, of the deviations in height from the profile mean value as follows:

Figure 9:

Roughness parameters (R_a and R_q) of the steel surface versus the applied cathodic potential with 600 s cathodic polarisation in a 0.1 m NaCl solution.

where Y_i is a function that describes the surface profile analysed in terms of height (Y) and position (i) of the sample over the evaluation length (n).

Both R_a and R_q parameters increase with the cathodic potential; however, they increase with the slope after the samples are cathodically polarised under a fluid flow that has the greater cathodic polarisation effect on the surface topography of steel under flow conditions.

It is worth noting that sharpening of the surface grooves, and therefore increasing the surface roughness of the cathodically polarised steel, increases the true surface area. Consequently, as the surface roughness increases both the cathodic and anodic current densities of the potentiodynamic polarisation curves at Figure 5 increase, the E_pit values shift towards more negative potentials, and in general, the corrosion resistance of the steel decreases. Moreover, because rougher surfaces have lower electron work function than smooth surfaces (Li & Li, 2006; Ozawa, Kaykham, & Hiraishi, 1999), in rougher surfaces, it is easier for electrons to transfer from the steel surface. This causes the rougher surfaces to be electrochemically more active. Therefore, in this work, as the surface roughness of steel increases by the cathodic polarisation process, its electron work function decreases and the surface becomes more active. Thus, the contribution of electron work function of the steel surface needs to be considered as another significant factor in the electrochemical behaviour of the steel especially after the cathodic polarisation process.

Figure 10 presents the SEM images of the steel samples after 600 s cathodic polarisation under static conditions and flow conditions constructed by the RDE with a rotation rate of 500 rpm. These images clearly show the formed corrosion products on the steel surface. It is assumed that part of the formed corrosion products is transferred into the solution during cathodic polarisation. In Figure 10C, the characterisation of the observed particles was performed by EDS in two different locations of the produced corrosion products. The quantitative EDS results of this analysis for each location are shown in Table 1. The EDS results of the dark part of the corrosion products (location 1) reveal that iron, carbon, chromium, and nickel form the major part of the corrosion products at this area. Similar to location 1, the EDS analysis performed on the white part of the corrosion products (location 2) also shows the major part of the elements to be iron. In this part, after iron, oxygen and aluminium show the highest percentage of the elements, respectively; this is expected to be in the form of Al₂O₃. From Figure 10 and Table 1, it is needed to consider that, although aluminium is an alloying element in the stainless steel samples used in this work, part of the alumina observed on the surface of the cathodically polarised steel are alumina particles, which remained inside the pores of the steel surface after polishing the samples.

Figure 10:

SEM images of steel samples (A) with 600 s cathodic polarisation at -1.7 V under static conditions and (B) with 600 s cathodic polarisation at -1.7 V under flow conditions constructed by the RDE (500 rpm) and (C) a focused view of image (A).

In these analyses, other elements such as silicon and molybdenum were also detected but significantly lower in percentages.

3.6 Electrochemical impedance spectroscopy

The corrosion resistance of steel with and without cathodic polarisation under static and dynamic conditions produced by the flow cell was investigated by electrochemical impedance spectroscopy. The Nyquist plots in Figure 11A show lower imaginary impedance values after cathodic polarisation at -1.5 and -1.7 V, whereas the Nyquist plots of the nonpolarised sample and the cathodically polarised sample at a potential of -1.3 V almost overlap. Under flow conditions, however, the effect of cathodic polarisation is clearer. At a lower velocity (Figure 11B), the Nyquist plots of the cathodically polarised steel at -1.3 and -1.5 V overlap. As expected, the smallest diameter of the impedance graph is observed for the cathodically polarised steel after the application of -1.7 V cathodic potential. At a flow velocity of 8.6 cm³/s (Figure 11C) the diameter of the cathodically polarised plots decreases more significantly, where the maximum imaginary impedance of the polarised steel after applying a cathodic potential of -1.3 V is approximately 8 kΩ cm², which decreases to approximately 3 kΩ cm² after the steel is polarised at -1.5 and -1.7 V. Moreover, the Nyquist plots for both the static and flow modes show that the corrosion resistance of steel decreases with increasing cathodic polarisation potential. The results are more profound for high flow velocities. Similar results were observed when the RDE was used to produce the flow conditions (results not shown).

Figure 11:

Nyquist plots of steel with and without 600 s cathodic polarisation for three different cathodic potentials (-1.3, -1.5, and -1.7 V) under (A) static, (B) low flow velocity (1.3 cm³/s), and (C) high flow velocity (8.6 cm³/s) conditions.

The flow conditions were produced by the flow cell.

According to the measured potentiodynamic polarisation curves (Figure 5), the Nyquist plots (Figure 11), and the AFM images (Figure 8), it can be concluded that, as the hydrogen entry on the surface occurs with greater intensity with the application of a greater cathodic potential, the cathodic corrosion process changes the morphology of the oxide film and increases the true surface area of the steel with higher intensity. Greater amount of the absorbed hydrogen into the oxide film decreases the stability of the passive layer and increases the corrosion current density and therefore decreases the corrosion resistance of the steel.

These results are in good agreement with the impedance data reported by some other authors (Cheng & Niu, 2007; Kliskic, Radosevic, Gudic, & Smith, 1998; Lee & Pyun, 2000; Li & Cheng, 2007), where the application of the cathodic polarisation process with greater intensity decreases the diameter of the impedance graphs.