Mitigation of top‐ and bottom‐of‐the‐line CO2 corrosion in the presence of acetic acid (II): Inhibition using azole derivatives

Carbon steel flowlines transporting hydrocarbon fluids are susceptible to internal corrosion in the aqueous phase at the base (6 o'clock position) and in aqueous condensed droplets at the top (12 o'clock position). Respectively, these issues are known as “bottom‐of‐the‐line” corrosion and “top‐of‐the‐line” corrosion and inhibitors that are used to control internal pipeline corrosion need to be effective at both locations. Here, we explore whether 2‐mercaptobenzimidazole (2‐MBI), 2‐phenyl‐2‐imidazoline (2‐PI), 2‐amino‐5‐ethyl‐1,3,4‐thiodiazole (2‐AETD), are able to control CO2 corrosion simultaneously at both the top‐ and the bottom‐of‐the‐line in the presence and absence of acetic acid, a common minor constituent of produced hydrocarbon fluids. The performance of the species varied between highly effective (2‐MBI), moderately effective (2‐AETD) to ineffective (2‐PI). Inhibition was effective at both bottom‐ and top‐of‐the‐line, and with acetic acid present. Given that the vapor pressure of these species is negligible, it is suggested that they are carried from the bulk phase to the top‐of‐the‐line dissolved in aerosol droplets rather than in the vapor phase.


| INTRODUCTION
Carbon and low-alloy steel pipelines are the material of choice for the transport of unprocessed hydrocarbon fluids because of their desirable properties including ease of fabrication, mechanical properties, and cost-effectiveness. However, they generally have a low resistance to corrosion in aggressive environments. In addition to hydrocarbon fractions, upstream and downstream hydrocarbon fluids contain a range of species that may include carbon dioxide, hydrogen sulfide, water, and organic acids, all of which are relatively corrosive to ferrous pipeline materials. Phase separation of the aqueous content from the liquid hydrocarbon fraction leads to a risk of corrosion at the base (6 o'clock) position, giving rise to bottom-of-the-line corrosion (BLC). This is a well-understood process and gives rise to a range of corrosion damage mechanisms. BLC is generally mitigated by the incorporation of corrosion inhibitors into the fluids that partition into the aqueous phase. However, unexpected corrosion damage at the top-of-the-line corrosion (TLC) was additionally identified in the 1990s and commonly arises due to temperature differences between the 12 o'clock and the 6 o'clock positions leading to condensation of water vapor if the headspace is cooler. [1] TLC, which is difficult to identify in the field, is regarded as a significant threat primarily, but not exclusively, in wet gas lines. [2] The phenomenon is exacerbated by higher fluid temperatures, ineffective or non-existent external insulation, the presence of organic acids, and under conditions of flow stratification. Mitigation against TLC is challenging as inhibitor regimes designed for BLC only have poor effectiveness.
Organic, primarily acetic, acids are known risk factors for both BLC and TLC in hydrocarbon lines. For example, acetic acid was found to increase corrosion at both the top and bottom of the line [3] where its adverse influence on TLC was thought to be due to reduced iron carbonate stability in droplets with lowered pH. [4] This leads to the hypothesis that pH control might be an effective means to limit TLC, acting similarly to pH control in steam systems. [5] In hydrocarbon lines, methyl diethanolamine (MDEA) is used for preferential removal of hydrogen sulfide, is economical and readily available, and therefore could also be used for corrosion mitigation. [6] Using MDEA under immersed conditions. Pojtanabuntoeng et al. found that corrosion mitigation was effective between pH 6.5 and 7, [7] while, in an earlier paper, we demonstrated that MDEA has useful, but limited, effectiveness for both TLC and BLC. [8] Organic corrosion inhibitors are commonly used to mitigate damage to carbon steel by restricting the kinetics of the cathodic reaction, or the anodic metal dissolution reaction, or both. As the review by Schmitt [9] shows, the efficiency of an organic inhibitor is generally related to its tendency to dynamically adsorb and desorb onto electrochemically active sites on metallic substrates. The presence of heteroatoms such as nitrogen, sulfur, oxygen, and phosphorous as well as triple bonds or aromatic rings in the molecular structures aid in this inhibition mechanism. Thus, Finšgar and Jackson report that amide, azole, mercapto, and related species are effective inhibitors for corrosion in strong acids. [10] However, there are fewer studies on the application of such compounds under weak acid (e.g., carbonic and acetic acid environments).
In this contribution, we use the experimental cell design reported previously, [8] with capability for electrochemical measurements at the bottom-and the top-of-the-line, to explore corrosion inhibition under a CO 2 /acetic acid environment. For effective corrosion suppression at the top-of-the-line, a species with low vapor pressure might not be expected to be effective. This hypothesis was examined by selecting species that have low or minimal vapor pressure under the operating conditions. We chose 2-mercaptobenzimidazole (2-MBI), which is effective in hydrochloric and sulfuric acid pickling, industrial cleaning, descaling, and oil well acidification [10] as well as in CO 2 corrosion in brine. [11] This was compared with 2-phenyl-2-imidazoline (2-PI), which is effective for atmospheric corrosion of copper [12] and for aluminum under immersion in HCl. [13] Also, 2-amino-5-ethyl-1,3,4-thiodiazole (2-AETD) which is effective for inhibition of copper in brine, [14] and for steel in HCl [15] and in H 2 SO 4 . [16] 2 | EXPERIMENTAL METHOD

| Pipeline simulation
The environmental conditions for the TLC and the BLC were simulated using a tubular acrylic cell 25 cm in length and 10 cm in diameter (2 l in volume) the ends of which were closed off ( Figure 1). An aqueous brine solution (NaCl: 1 wt.%/0.17 M, with or without added acetic acid and/or inhibitor) was maintained in the lower 30% of the cell. A small temperature difference of around 10 degrees Celsius between the top and bottom of the cell was sufficient to drive continuous condensation in its upper portion. The cell was provided with ports for gas inlet and outlet, and for electrochemical connections. BLC was measured using three electrodes that were submerged in the lower part of the cell where voltammetry and electrochemical impedance could be performed. TLC was measured using two (nominally identical) segmented electrodes separated by an insulated spacer and flush mounted into a port at the top of the acrylic cell in the condensation zone; here electrochemical impedance and linear polarization could be carried out.
The three-electrode cell for BLC measurement comprised a carbon steel (0.17% C, 0.78% Mn, 0.20% Si, 0.020% P, and 0.023% S) working electrode of area 0.283 cm 2 , a platinum counter electrode, and a calomel reference electrode. The two-electrode cell, for TLC measurement comprised two nominally identical carbon steel electrodes, each of area 0.072 cm 2 , embedded in an epoxy mount and separated by 1 mm using an insulating rubber spacer. The working electrodes were mechanically polished using silicon carbide papers of grades to a 4000-grit finish then rinsed with ethanol and air-dried before use.
During experiments, the lower part of the acrylic cell was immersed in a water bath at the controlled temperature while the upper part was exposed above the water line, thus creating conditions for condensation from the 10-12-14 o'clock positions of the cell. Before the experiments, the cell was filled with the appropriate solution and then purged with CO 2 gas for at least 1 h to reduce the oxygen concentration below those that might significantly influence the corrosion process (around 50 ppb). Purging of carbon dioxide continued slowly and continuously throughout the experiment to maintain a slight positive internal pressure and hence prevent ingress of oxygen to the closed system.

| Electrochemistry
Samples attained relatively steady open circuit potential after about 2 h' immersion after which electrochemical experiments were carried out. Electrochemical impedance spectroscopy (EIS) used a Solartron 1250 frequency response analyzer with 1286 potentiostat. Samples were polarized 10 mV rms from the open circuit potential over a frequency range of 0.01 Hz to 10 kHz. The simple equivalent circuit shown in Figure 2 was adopted to analyze the corrosion process. Linear polarization resistance (LPR) measurements were conducted by polarizing the working electrode at ±15 mV from the OCP using a scan rate of 0.167 mV s −1 . Potentiodynamic polarization was performed by polarizing the working electrode from cathodic to anodic at a scan rate of 0.167 mV s −1 . Experimental data is quoted to two significant figures, sample-to-sample repeatability was around 10%.

| Analytical methods
The corrosion products formed in different environments were examined using analytical scanning electron microscopy (SEM) and X-ray photoelectron spectroscopy (XPS). A Zeiss EVO50 was used to examine the corrosion   products on the metal surface using energy dispersive X-ray analysis (EDS).

| RESULTS AND DISCUSSION
3.1 | BLC: CO 2 saturated brine plus acetic acid Figure 3 shows reference potentiodynamic polarization curves for carbon steel in CO 2 saturated 0.17 M NaCl solution with additions of acetic acid. The anodic branches all show a Tafel response with slopes in the range 30-35 mV/decade. In the absence of acetic acid, the cathodic branch showed the expected rate-limiting kinetics due to diffusion of carbonic acid. Additions of 500 and 1000 ppm (8.3 and 16.7 mM) acetic acid increase the cathodic kinetics by 3-5 times and correspondingly the corrosion potential becomes more anodic. The results agree with our previous work on CO 2 + acetic acid corrosion. [8] 3.2 | BLC: Inhibition with 2-PI, 2-MBI, and 2-AETD Potentiodynamic polarization of carbon steel immersed in CO 2 saturated 1% (0.17 M) NaCl at 70°C with varying concentrations of 2-PI, 2-MBI, and 2-EATD are shown in Figures 4-6, respectively. Evidently, 2-PI is not effective in this application as it accelerates the overall cathodic kinetics and, hence, corrosion rate. It can therefore be assumed that the active groups and/or molecular conformation do not adsorb sufficiently on the metal surface in this environment. Given that 2-PI is unsuccessful we did not consider it further in this work.
Conversely, 2-MBI inhibits under these conditions effectively suppressing both the cathodic and anodic reactions with an extended region (~200 mV) of passivity from the OCP until around -620 mV. Although 2-AETD was less effective in suppressing the cathodic reaction it also showed anodic inhibition but with a reduced  passive potential range (~100 mV) compared with 2-MBI. Addition of 500 ppm (8.3 mM) acetic acid reduces inhibitor effectiveness (Figure 7). While 2-EATD was found only to weakly inhibit the cathodic reaction at 1 mM concentration, 2-MBI remained effective at concentrations of 1 and 3 mM albeit with a reduced passive region (50-100 mV). Electrochemical impedance confirmed the efficacy of 2-MBI, Figure 8 and the limited effectiveness of 2-AETD ( Figure 9). The result is in agreement with other reports that 2-MBI has good inhibition performance in acidic brines. [16] Electrochemical corrosion data under immersed conditions are collected together in Table 1. There is generally excellent agreement between inhibition efficiencies (determined from polarization resistance) whether by electrochemical impedance or by linear polarization. This can be attributed to the single time constant process evident from the impedance results. Inhibition efficiencies determined from potentiodynamic polarization follow similar trends compared with alternative methods but with less close agreement.

| TLC: CO 2 saturated brine plus acetic acid
The background corrosion rate of carbon steel in the vapor phase above CO 2 saturated 0.17 M (1%) NaCl at 60°C (Figure 10), was found to be around eight times less than under immersion at 70°C (Figure 9). This is probably due to the greater ease of formation of iron carbonate in the limited volume of condensate present as observed in Figure 11a which would provide a degree of physical protection. Addition of 1.67 mM (100 ppm) acetic acid into the bulk phase increased the corrosion rate at the top-ofthe-line by a factor of around 3 with Figure 11b showing that the carbonate film had been removed.

| TLC: Inhibition with 2-MBI and 2-AETD
Inhibition by 2-MBI and 2-AETD was investigated using EIS under condensing conditions at 60°C above a solution in the bulk phase held at 70°C and containing 0.17 M NaCl, saturated in CO 2 , plus varying amounts of acetic acid. Linear polarization was not used in this circumstance due to errors arising from solution resistance in the condensate. Figure 12 shows top-of-the-line impedance data for CO 2 saturated 0.17 M (1%) NaCl. The  addition of 1.67 mM acetic acid accelerates corrosion by a factor of about 3 while the addition of 1 mM 2-MBI gives in excess of 90% inhibition for the saturated CO 2 bulk environment and >95% for the CO 2 + 8.33 mM (500 ppm) acetic acid environment. This also confirms that a single time constant was present and, hence, a standard Randles equivalent circuit (Figure 2) can be used for impedance data analysis. Table 2 shows impedance data for a wider selection of environments as a function of time. In each case, there is a progressive increase in Rp values with time indicating a reduction in corrosion rate. Despite neither 2-MBI nor 2-AETD having a significant vapor pressure, they are both clearly acting to inhibit corrosion in the top-of-the-line condensate.
The variation in corrosion rates plotted as 1/Rp, over 30 h as a function of inhibitor concentration in the absence of acetic acid are shown in Figures 13 and 14. In the acetic acid-free environment, 2-MBI is confirmed as an extremely effective inhibitor with efficiencies exceeding 90% however, 2-AETD is less effective with inhibition of around 60% efficiency. For uninhibited environments, the corrosion rates decrease monotonically with time up to 30 h exposure; however, in the inhibited systems, the corrosion rate roughly halves between 1 and 6 h of exposure, then flattens out. Increasing the concentration of inhibitor from 1 to 5 mM generally reduced the corrosion rate further although there is evidence that 2-AETD concentrations above 3 mM have little additional benefit.
The influence of various concentrations of 2-MBI on the corrosion rate at the top-of-the-line with acetic acid present in the bulk solution are shown in Figure 15 where it is confirmed to provide excellent inhibition at sufficiently high concentration. Thus, while the effectiveness of 1 mM 2-MBI faded after 6-10 h with 500 ppm acetic acid present, F I G U R E 12 Electrochemical impedance spectroscopy (EIS) Nyquist plot of carbon steel in the vapor phase (top-of-the-line corrosion (TLC) at 60°C) over CO 2 saturated 1% NaCl with and without 500 ppm (8.33 mM) acetic acid (HAC) and 2-mercaptobenzimidazole (2-MBI) inhibitor.

| CONCLUSIONS
1. 2-PI, reported as a good inhibitor for copper and aluminum, was found to be completely ineffective for carbon steel in CO 2 -saturated sodium chloride solutions. 2. 2-AETD, reported to be an excellent inhibitor for aluminum and carbon steel in strong acids, was found to provide significant (60%-80%) corrosion inhibition under immersed and condensed conditions in CO 2 -saturated sodium chloride solutions but was less effective in the presence of acetic acid.
3. Well-known as an efficient inhibitor for carbon steel in strong acids, 2-MBI was also confirmed to be an extremely effective inhibitor for immersed (bottom-of-the-line) and condensate (top-of-theline) conditions in CO 2 -saturated sodium chloride even when acetic acid was present with efficiencies of around 95% at sufficiently high concentrations. 4. Since neither 2-AETD nor 2-MBI have a significant vapor pressure under ambient conditions, it is suggested that the species are more likely to transport to the top-of-the-line dissolved within aerosol droplets rather than in the vapor phase.