The effect of alternating current (AC) on the corrosion rates of cathodically protected buried steel pipeline was investigated through mass loss testing. Results indicate that cathodic protection (CP) at potentials at or below the CP criteria of −780 mVSCE is adequate to prevent increased corrosion due to AC at AC densities up to 350 A/m2. Further, the interfacial capacitance of the steel was theorized to be a controlling factor in the ratio of the capacitive to faradaic current portions of the AC and thus the corrosion rates due to AC. The value of capacitance was measured in various soil types to determine the effect of the bulk environment on this parameter. It was found that interfacial capacitance is much lower in artificial soils than in a soil-simulating solution but that the composition of the soils does not greatly affect the capacitance.
The corrosion of metals under the influence of an alternating current (AC) has a long history of study but it has lately been considered as a distinct problem in the pipeline industry.1-2 Due to a scarcity of data and lack of understanding of the phenomenon, the issue of pipeline susceptibility and corrosion severity is contentious, especially on cathodically protected pipelines. Several national and international standards and publications have been written to deal with the issue of AC corrosion and offer thresholds and guidelines.3-6 Despite these guidelines, no single criterion is generally accepted for dealing with AC corrosion risk.5 The correlation between AC, cathodic protection (CP), and corrosion rate needs further study.
A model of AC corrosion was developed by Ghanbari, et al., which considered the Butler-Volmer kinetics of the metal interface as well as the interfacial capacitance and ohmic resistance of solution.7 This model allowed predictions of AC corrosion rates of steel at any AC or direct current (DC) levels assuming prior knowledge of tafel slopes, oxygen diffusion-limited current density, and interfacial capacitance. From this model, it was recognized that interfacial capacitance has a large controlling effect on the total ratio of faradaic to capacitive AC at the steel surface and is thus of prime importance in determining corrosion rates at any AC density.
The capacitance of a metal is generally assumed to arise from charge movement in the double-layer of the metal/solution interface. Previous research on steel in soils has revealed a large contribution to the capacitance in the form of pseudocapacitance.8-11 Further, it has been suggested that the value of pseudocapacitance may be largely affected by the type of soil in contact with the metal.8,12-13
In this study, the effects of AC density on the corrosion rates of cathodically protected steel samples were determined with mass loss experiments by subjecting steel samples in a soil-simulating solution to various AC and DC potentials. Also, after identifying capacitance as a key parameter in AC corrosion, the effect of soil environment on capacitance was studied through electrochemical impedance spectroscopy (EIS) of steel samples in artificial soils of varying sand silt and clay content.
Cylindrical samples of X65 pipeline steel were manufactured to fit onto the end of a steel rod sealed with a plastic gasket, rubber O-ring, and a glass rod. The samples had a total exposed area of around 4.8 cm2. A three-electrode cell setup was used to control the potential between the working and reference (saturated calomel) electrode. The potentiostats used for voltage control allowed for a function generator input whereby a large-voltage 60 Hz sinusoidal signal could be accurately controlled between the WE and RE along with the simultaneous application of DC potentials.
The solution used for long-term mass loss experiments was a modified NS4 solution14 where the CaCO3 and MgSO4 were removed to prevent scaling issues.
Samples were weighed before testing and then immersed in solution. Once a stable open-circuit potential (OCP) was achieved (around 1 h), the DC and AC signals were applied for 3 to 6 weeks. The samples were removed from solution at the end of the test, cleaned via ASTM G1,15 and then weighed for total mass loss. Sample weighing was performed with a highly sensitive scale accurate to 0.1 mg (0.009 mpy on a 4.8 cm2 sample).
Interfacial Capacitance in Soils
Representative artificial soils were produced using clays and sands sourced from commercial suppliers. The soils were meant to emulate various textural classifications of common soils designated by a certain percentage of clay, silt, and sand particles. The specific soils made and tested are shown in the U.S. Department of Agriculture (USDA) soil texture triangle (Figure 1).16
The portion of clay in these soils was made of a combination of three different clay types: McNamee Kaolin†, EPK Kaolin†, and KY Ball Clay Illite†. Silt was composed of 200 mesh silica and feldspar potash. The sand portion was a coarse 30 mesh quartz.
The soil minerals were thoroughly mixed and 0.1 M NaCl solution was added until the soil had reached its maximum holding capacity for water where the volumetric water content (VWC) was at a maximum. This level was generally around 45% for all soils except sand, as measured by a soil moisture meter. The maximum VWC for sand was around 25%.
Soil experiments used the same three-electrode setup as mass loss experiments but with a large-volume soil box. EIS measurements were taken at the start of the experiment, after 24 h of open-circuit conditions, and then again after 7 d of CP application.
RESULTS AND DISCUSSION
Mass loss results were converted to corrosion rate in mils per year (mpy) and then plotted against the average AC density during the course of each respective experiment. The results are shown in Figure 2 for laboratory tests in soil-simulating solution.
These results indicate that cathodic protection, at all potentials at or below the standard CP potential of −0.78 VSCE (−0.85 VCSE), works to mitigate corrosion even at high AC densities up to 300 A/m2. More negative potentials do not increase the risk of AC corrosion. While higher AC densities generally increase the corrosion rates, CP effectively prevents large corrosion rates at large AC densities. Tests run at the OCP of X65 steel indicate a high degree of susceptibility to AC corrosion with an increase of corrosion rate from 15 mpy at 0 A/m2 to nearly 100 mpy at 400 A/m2.
These results are consistent with a model of AC corrosion developed in this lab.7 In this model, the interfacial capacitance is the lowest impedance circuit element at a frequency of 60 Hz. The majority of AC will then consist of capacitive current due to the movement of charged of particles at the metal/solution interface. A small fraction of the total AC is faradaic in nature; composed of oscillations of oxidation and reduction charge-transfer currents. Increased corrosion is only caused by the periodic spikes in oxidation current occurring at the crests of the alternating voltage signal. Therefore, the fraction of AC contributing to corrosion is very small at all DC potentials.
The effect of CP is to change the bias of the alternating voltage perturbation to a more negatively polarized region. At highly negative potentials, the interfacial voltage oscillation is not large enough to produce significant oxidation currents. In fact, at a sufficiently negative CP potential, oxidation may be thermodynamically unfavorable, even with very large alternating voltage perturbations.
Interfacial capacitance in soils was measured by fitting an equivalent circuit (EC) model to EIS results. A frequency range of 10 kHz to 0.001 Hz was required to reveal the full capacitive arc on a Nyquist plot. This allowed the data to be fit to the equivalent circuit shown in Figure 3 using a nonlinear least squares fitting procedure. The constant phase element (CPE) value obtained was then converted to an effective capacitance by the following formula from Hirschorn, et al.:17
A Warburg impedance was often present on the Nyquist plots and usually developed after application of CP or after long immersion times. Because the Warburg impedance becomes larger at low frequencies (ZW = σω−1/2 − jπσω−1/2), the point at which the Warburg element becomes significant can be observed by an inflection point from a semicircular arc to a 45° line on the Nyquist plot. To achieve a better fit for the EC considered here, it was possible to disregard low-frequency data at the point where the Warburg impedance becomes significant and fit the EC only to data pertaining to a region of frequencies where the Warburg element is negligible.
The results for the effective interfacial capacitance measurements in different soils types taken after 24 h of OCP are shown in Figure 4. Steel in soils often exhibits a large pseudocapacitive behavior.11 This phenomenon leads to large values of interfacial capacitance on the order of hundreds or thousands of uF/cm2 for steel in soils.10,18-19
The capacitance values in soils are lower than what is measured in either NS4 or NaCl solution (3 to 5 × 10−3 F/cm2). There are a few possible reasons for this. It is likely that the less than 100% VWC of soils, even at full saturation, leads to a situation where the steel sample does not possess a 100% active contact area for the electrochemical interface. In effect, the soil particles serve to block solution from contacting the entirety of the steel surface. If the interfacial capacitance is an intrinsic property of the steel/solution interphase, then a decrease in the effective contact area would lead to underestimating the true interfacial capacitance.
It is also likely that the soils leached cations, mainly Ca+ and Mg2+, due to the inherent cation exchange capacity of soil particles, which then formed a thin scale on the steel samples. This would decrease the measured capacitance. Such a scale was not visible on the samples when observed with a light microscope, but it is possible that the formed scale is thin or sparse enough to not be visible under light microscopy. Such a scale is likely to form after application of CP and is expected to influence interfacial capacitance. Measurements of interfacial capacitance after 7 d of CP were inconclusive owing to large diffusional impedances that precluded calculation of effective capacitance.
The difference between capacitance values in different soils was less than expected. When considering the variance within tests in the same soils, there is no large statistically significant difference in the capacitance between soils. The measured values are within the 2 to 4 × 10−4 F/cm2 range. This indicates that the soil’s mineral compositions have little to no effect on the measured capacitance between the various soil types.
Mass loss results for steel samples in a simulated soil solution indicate that superimposed AC does not significantly contribute to corrosion rates for cathodically protected steel in the range of 0 A/m2 to 350 A/m2. However, steel without cathodic protection can experience high corrosion rates due to induced AC. Most of the AC is capacitive in nature, owing to the unusually high pseudocapacitive behavior of steel in aqueous solution. Corrosion attributed to AC is due to periodic spikes in oxidation current density. At more negative potentials, the positive portion of the AC potential does not reach high values where oxidation will occur.
It was also shown that the interfacial capacitance of steel does not significantly differ between soil types with differing mineral composition.
The authors would like to thank Kevin Garrity, Dan Wagner, and Kevin Northon from Mears Group, Inc. for advice on the usage of CP test facilities and Ian Stallman from Marathon Pipeline LLC. We would also like to acknowledge contributing laboratory work from Abigail Helbling, Helen Nee, and Jayme Jennings. This study was completed as part of the DOT CAAP #DTPH5615HCAP02.
Recipient of first place in the Harvey Herro Applied Corrosion Technology category in the Student Poster Session at CORROSION 2018, April 2018, Phoenix, Arizona.