Surface analyses of corrosion products formed on API X-70 steel in sour brine after testing in a pipe loop

Flow assisted corrosion (FAC) is a problem of pipeline systems that handle high flow rates and strong direction changes. In the present investigation, FAC was tested on carbon steel exposed to the NACE 1D-196 environment by means an experimental pipe loop. As the exposure time increased, corrosion products formed a mixture of oxides, sulfides, and an apparent sulfate (rhombohedral mikasaite Fe2 (SO4)3), which was found in greater proportion and appeared to have a significant effect on decreasing corrosion rate. Transmission electron microscopy and x ray diffraction patterns seemed to confirm the presence of a sulfate and some oxides as the major chemical species contained in the corrosion products.


Introduction
Flow assisted corrosion (FAC) is one of the main causes of failure in oil facilities [1], it is an important and inevitable challenge in the oil and gas industry [2,3] that usually occurs under aggressive flow regime [4,5]. Hydrodynamics accelerates the mass transfer process and damages the protective corrosion products on the steel surface so the corrosion rate of pipelines is considerably increased [6][7][8][9]. Fluid flow rates associated with FAC decreases concentration gradient and thus increases corrosion rate [10][11][12][13][14][15][16][17][18]. No evidence of oxide film has been found due to mechanical shear of the FAC damaged surfaces of feed water piping [19]. Erosion-corrosion is a form of mechanical degradation that involves corrosion as well as mechanical wear. This occurs on the surface of the material due to the action of numerous individual impacts of solid or liquid particles. Higher flow velocities are associated with this kind of degradation. Definite surface patterns are formed on components undergoing FAC meanwhile for erosion corrosion, grooves, gullies, or rounded holes are observed. This can occur in metals and alloys that are completely resistant to a particular environment at low flow velocities unlike FAC degradation [20,21].
Corrosion of pipelines also depends on the type of well (sour/sweet), temperature, CO 2 and H 2 S content, flow rate, oil or water wetting nature, and the composition and surface condition of the steel [22]. Sour and sweet wells containing H 2 S and CO 2 gases, respectively, which form acidic media in aqueous solution, provide a favorable environment for iron dissolution [23]. Corrosion in sour environments is normally initiated by the formation of iron sulfide films [24] and their formation is mostly temperature dependent, with protective and adherent films being formed at relatively high temperatures (>80°C) [25,26]. Iron sulfide is accompanied by the generation of molecular hydrogen gas, which is the main cause of hydrogen embrittlement [24,27,28]. The Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. corrosion behavior of carbon steel in the presence of CO 2 [29][30][31] and H 2 S [32] has already been studied in detail.
In this work the flow assisted corrosion of steel API 5L X-70 in sour brine is evaluated; a flow of 30 l min −1 conduced in a loop was provided on the metallic coupons and corrosion tests were performed for periods of 3, 6, and 12 h. Changes on metallic surfaces were recorded by scanning electron microscopy coupled with energy dispersive x-ray spectroscopy microanalysis (SEM/EDS). Furthermore, corrosion products were analyzed by x-ray diffraction (XRD) and high-resolution scanning transmission electron microscope (TEM) to investigate the chemical species formed on surface and its main products such as mikasaite, which were selected for further study. The corrosion rates as a function of time were determined by linear polarization resistance (LPR) for each testing time.

Materials and methods
2.1. Design of couplings for specimen support in solid works Figure 1 shows the polyethylene pipe where coupling is spliced without altering the internal surface to avoid changes in the flow direction. The steel coupons are mounted in the three perforations which are 90°separated from each other. These specimens were coupled to perform different analyzes after testing (linear polarization resistance measurements, SEM, XRD and TEM analysis).

Assembly of couplings in the experimental piping circuit (loop)
The experimental piping circuit is shown in figure 2. A preliminary test using water as fluid is performed to corroborate the correct operation of the device. The coupons were made of API 5L X70 steel having a 600-grit sandpaper surface finish with an exposed area of 0.9739 cm 2 .

Simulation of the flow pattern
A numerical simulation was performed in ANSYS Fluent V 11.5 modelling the system network with a mesh of tetrahedral elements which consisted of 278,027 cells and 63,825 nodes. For the analysis, the K-ε turbulence model was used with the same experimental conditions, whose objective was to corroborate that flow rate circulating inside the experimental pipe loop during the experimentation generates a turbulent flow in both ways and studying its effect on the formation of the corrosion products.

Characterization of steel API 5L X-70
The phases characterization of API 5L X70 steel was performed using an optical microscope by means of the optical emission analysis (spark) technique with a BELEC spark and arc spectrometer. The chemical composition of the steel is shown in table 1.
In the composition of this steel, it can be highlighted the presence of chromium (Cr), copper (Cu) and nickel (Ni), which have a great influence on the behavior of the corrosion rate as well as the tendency to stabilize the steel matrix.

Electrochemical tests
Electrochemical tests were conducted to determine the corrosion rate of API 5L X70 steel in the experimental loop. The testing environment consisted of an aqueous sour brine prepared according to  added with 10% kerosene, and hydrogen sulfide (H 2 S) at a pressure of 1 Bar (pH 4.5-5.5). The test fluid was a conductive liquid such as brine/oil mixture to simulate oil field conditions and to maintain the fluid    Figure 3 shows the position of the specimens; the electrochemical cell was composed of 12 metallics coupons, in which the inner and middle rings are made of pipeline steel X70, and serve as the working and reference electrodes, and the counter electrode is made of SS 316L [34]. Electrodes were immersed in the sour brine and supported by the corresponding couple unit. Electric leads were connected to a potentiostat/ galvanostat to perform measurements of linear polarization resistance (LPR), from which current density (I corr = B/R p ) and corrosion rate (CR = 0.00327 × I corr [μA cm −2 ] × EW/ρ [g cm −3 ]) were determined; where ρ stands for steel density and EW for the equivalent weight of carbon steel X70; B is the constant of Stearn-Geary, which is defined as ∼26 mV for an activation-controlled process [35].
In this investigation only one of the three tracks consisted of the experimental piping circuit. The experimentation was carried out for 12 h to analyze the effect of flow rate on the progress of steel corrosion. Each coupling along with its respective specimens was exposed to the corrosive environment for specific periods of time. Figure 4 describes the nomenclature used to identify each of the metallic coupons on testing.

Scanning electron microscopy (SEM)
The characterization of the corrosion products formed on the steel surface after corrosion testing was carried out by scanning electron microscopy (SEM) using a Jeol JSM 6300 unit operated at 20 kV.

X-Ray diffraction (XRD)
The XRD analysis was performed by a Focus D8 Bruker diffractometer, which employs Cu Kα radiation, at 35 kV and 25 A conditions, in a range of 2θ, within 20°to 120°in increments of 0.02°.

Transmission electron microscopy (TEM)
The corrosion products of the samples were detached and collected and thereafter they were diluted in distilled water. Subsequently, they were dispersed by an ultrasonic bath for 5 min, then a drop was taken and deposited onto a grid; once the sample was dried, it was introduced into the equipment for characterization.

HAADF-STEM imaging
A JEM-2100 TEM, equipped with a high-angle-angular-darkfield (HAADF) detector and x-ray dispersive spectrometer (EDS) systems operated at 200 kV for high resolution images of TEM (HRTEM) and high resolution HAADF scanning transmission electron microscope (HR-HAADF-STEM) imaging, was used to analyze the corrosion products formed on the metallic surface of steel.

Microstructural characterization of steel API 5L X70
The microstructure of API 5L X70 steel at 200X magnification is shown in figure 5. Ferrite (light contrast) and perlite (dark contrast) phases can be observed; grains tend to be equiaxed. Ferrite is normally dissolved, and cementite (Fe 3 C) contained in perlite is increased. Iron of steel is slightly oxidized in hematite (Fe 2 O 3 ) before immersion and later it is oxidized to another chemical species such as magnetite (Fe 3 O 4 ). The anodic reaction of steel in the sour environment generates iron sulfide (FeS) and the cathodic reaction molecular H 2 . As corrosion  progress, mackinawite FeS (1-x) is formed on steel surface. Other sulfides and/or sulfates might precipitate depending on pH, temperature, flow rate and oxygen content. It is thought that marcasite (FeS 2 ) might led to the presence of minor content of mikasaite, Fe 2 (SO 4 ) 3 , in corrosion products, after H 2 SO 4 is formed when H 2 S is dissolved in water so general corrosion and localized corrosion might appear; the former because of the continuous formation of FeS and the other due to film porosity that under certain conditions might lead to hydrogen embrittlement [36].

Corrosion rates as a function of exposure time
The corrosion rate of steel X70 steel versus time is shown in figure 6. The highest corrosion rates were recorded for the three positions after 3 h, meanwhile at 6-12 h a drastic decrease in the corrosion rate was observed. In the former, it can be assumed the likely formation of non protective iron films which could be easily removed by the shear stress generated by the fluid in motion and related with an active oxidation of the steel [37]. Conversely, as the corrosion progress, it can be considered the formation of more stable corrosion products, between 6 and 12 h of experimentation for the positions '3', '6' and '9', which reached similar corrosion rates. Figure 7 shows the SEM images recorded through the immersion of the API 5L X70 steel at position '9'. At lower magnifications, a greater removal of the corrosion products is observed so an active oxidation of the steel seems to be favored. Furthermore, there is an increase in the formation of corrosion products in the coupons exposed at 6 and 12 h; being evident their agglomeration on the steel surface after testing in the sour brine. Similar results were obtained for the API 5L X70 steel at positions '3' and '6'. Figure 8 show the mapped images after the formation of corrosion products on the steel surface after 3 h. These elements might indicate the formation of FeO, Fe 2 O 3 , Fe 3 O 4 , besides some sulfides such as mackinawite (FeS 1-x ), as it has been reported in the literature [38][39][40]. Carbon, calcium, chloride, and sodium are distributed in certain regions and iron and sulfur are distributed all over the surface. Figure 9 shows that point 1 contains S and Fe, so the formation of sulfides and oxides in the protective film of the corrosion products is inferred. Point 2 contains high levels of Na (17.70 wt%) and Cl (41.09 wt%) typical of the brine, so it can be said that this morphology corresponds to undissolved salt crystals. Point 3 shows apparently high levels of O, Fe and S, so the formation of sulfides and oxides as a mixture in the corrosion products is likely. Table 2 shows the results obtained by EDX microanalysis for the corrosion scale formed on steel surface after 3 h of exposure time at a flow rate of 30 l min −1 . Figure 10 shows the mapping images after the formation of corrosion products on the steel surface after 6 h, in which oxygen is uniformly distributed practically over the entire surface and sulfur is not located on this test. Sodium and chloride are observed uniformly distributed with some areas of higher concentration but    apparently in a lower concentration than those shown in figure 8 that corresponds to the coupon mapped after 3 h of exposure. Figure 11 shows the presence of Na and Cl (point 1), as mentioned before these are elements contained in the brine and they are precipitated after 6 h of exposure. Point 2 contains Fe (14.71 wt%) and O (7.75 wt%) so the formation of mainly oxides distributed on the surface is assumed ( figure 10). Table 3 shows the results obtained by EDX microanalysis for the corrosion scale formed on steel surface after 6 h of exposure time at a flow rate of 30 l min −1 . Figure 12 shows the mapping of elements after the formation of corrosion products on the steel following 12 h of the dynamic tests. O, S and Fe are identified over the analyzed surface. These elements might indicate the formation of FeO, Fe 2 O 3 , Fe 3 O 4 , besides some sulfides such as mackinawite (FeS 1-x ), as it has been reported in the literature [38][39][40]. Figure 13 shows that carbon, calcium, iron, and chloride are distributed in certain regions in a higher proportion than in others; it is highlighted the almost uniform distribution of oxygen and sulfide through corrosion products.

SEM-EDX microanalysis of corrosion products on API X70 after 3, 6 and 12 h of exposure
At higher magnifications (600X), the steel surface exposed for 12 h shows agglomerates precipitated with a porous morphology. A magnification of this image is shown in figure 13, in which two zones are evident: (1) a porous and uniform appearance on the steel surface and (2) particles with a geometrical shape. Additionally microanalysis was performed to determine the composition of the particle precipitated. Point 1 indicates a high content of Cu, as well as S, Fe and O so it can be inferred that Cu is present in the corrosion products as reported in the literature [10].
For point 1, table 4 shows the results obtained by EDX microanalysis for the corrosion scale formed on steel surface; where the elements with major percentages are (wt%): Cu-68.34 and S-14.71, this suggests that the combination of these elements generated the presence of copper sulfide in the external layer of corrosion products, which seems to be stable and might contribute to corrosion protection of steel [41] though it seemed to be less effective in comparison to sulphides and sulfates that appeared in a higher proportion, as indicated by the DRX analyses and mapped images as observed in figure 14. 3.3.2. Morphology of corrosion products on the steel X70 after 12 h of exposure at higher magnifications As mentioned before a porous morphology is evident; however, at higher magnifications of 1500X and 4000X ( figure 15) a series of rounded holes containing particles of irregular form are observed. This morphology is different from that observed in the literature and is probably attributed to the presence of mikasaite as a corrosion product, which has been recently reported [42]. This morphology has a series of holes that probably allow the corrosive fluid to pass through them, mitigating fluid turbulence and a greater detachment of the corrosion products consequently there is a decrease in the corrosion rate in combination with the presence of sulfides and oxides [43,44].

Characterization of corrosion products by XRD
The XRD patterns recorded for the steel samples positioned at '6 o´clock' is shown in figure 16. It is important to note that only a few compounds are identified after 3h testing, i.e., a rhombohedral hematite (oxide), rhombohedral mikasaite (sulfate) and orthorhombic marcasite (sulfide). It seems that typical oxides are absent, Figure 11. SEM micrograph and EDX microanalysis obtained for API 5L X70 steel at 600X after 6 h of exposure time at a flow rate of 30 l min −1 .
or they were easily detached; unlike these oxides the remained iron shown better adherence and they were partially removed by the fluid apparently generating an increase in the corrosion rate of the steel at 3 h.
After 6 h, the formation of new crystalline species such as cubic beta iron oxide, cubic magemite and orthorhombic magnetite are likely to form. Additionally, the signal intensity of sulfide and sulfate (mikasaite) are enhanced over other corrosion products, contributing thereby to the partial protection of the steel.
At 12 h of steel exposure to corrosive environment, the intensity of the corrosion products relative to those formed at 6 h decreased except for the sulfate (mikasaite, plane 024) which tended to predominate. It is assumed that the formation of this sulfate can be related with the partial protection of the steel due to its oxidation provided by both dissolved oxygen and the turbulent flow present in the system. In section 3.3.1, Cu and S appear to form CuS but in the XRD patterns is not present probably because corrosion products are formed in quite low amounts after 3 h. However, according to Sun Jin Kim et al [41], CuS compound should appear at the 2-theta angles of 31.5°and 45°. Carneiro et al [45] also have stated that the presence of Cu in steel exposed to the acid sour environment promoted the formation of a protective surface film of the type (Fe·Cu)·S on the steel. In this case, the greatest protection seems to be provided by a sulfate type mikasaite generated in a measurable quantity together with the presence of other sulfides of the type (FeS x ).
It is thought that the presence of mikasaite, Fe 2 (SO 4 ) 3 , is derived from FeS 2 as the pH of sour brine was kept within 4.5 to 5.5, so FeS and FeS 2 can be formed ( figure 17); the latter appears to be a precursor of mikasaite. The overall process was reported by the following reaction: Where Fe(OH) 3 is generally regarded to be a surrogate of ferrihydrite, a Fe 3+ oxyhydroxide; other oxyhydroxides, such as goethite, α-FeOOH, and rarely lepidocrocite, γ-FeOOH also are formed as insoluble precipitates. For H 2 S concentrations of about 670 ppm, the presence of two different types of sulfides is detected: FeS 2 on surface, which presents a high tendency to oxidation; oxygen was removed by bombardment with argon ions, though more internally a layer of FeS is formed. The constant presence of oxygen in the structure is an indicator of the manifest tendency of these sulphides to form oxides as Craig previously suggested [45]. Reaction (1) forms sulfuric acid, and in acidic conditions ferrous iron can be generated: Soluble ferric iron and pyrite cannot coexist for any significant length of time because pyrite relatively rapidly reduces to Fe 3+ (aq) and SO 4 2− as described below (equations (3)- (6)). The two dominant sources of extreme acidity in sour brine are H 2 SO 4 (from oxidation of H 2 S and SO 2 ) and the oxidation of pyrite, which produces sulfuric acid (3)(4)(5); the most acidic pH values reported for the sour environment, as shown in figure 17.
Sulfuric acid is produced by the oxidation of pyrite according to the reaction [46]:

Transmission electron microscopy (HRTEM) analysis
Transmission electron microscopy (TEM) can provide high spatial and energetic resolution, which has a tremendous potential to identify the origin and formation of corrosion products. In particular, the image mode of high-angle annular dark-field imaging scanning transmission electron microscope (HAADF-STEM) of highangle scattered electrons produces a strong contrast of atomic number (Z) [47], in response to the local Figure 13. SEM micrograph and EDX microanalysis obtained for API 5L X70 steel at 600X after 12 h of exposure time at a flow rate of 30 l min −1 (600X). Table 4. Chemical composition (wt%) of corrosion scale on API 5L X70.    figure 18(c)). This is an important discovery since the Fe 2 (SO4) 3 can be protective as it apparently reduced the corrosion rate of steel as showed in figure 6 [42]. Figure 19 shows the x-ray powder diffraction pattern, which was used to identify the mikasaite as the principal corrosion product. In this research, it is observed four intensity peaks for this compound, in agreement with table 5, which was highlighted for planes (024), (303), (134) and (318).
The diffraction pattern obtained in figure 20 was developed by using the software simulation powder cell 2.0 to display the crystal structure of the chemical specie under analysis. According to the parameters provided by the diffraction pattern, chemical species might correspond to mikasaite. Figure 20 shows the location of atoms of  iron, oxygen, and sulfur, which according to Miller´s indexes that conformed symmetry planes suggested the presence of a sulfate such as mikasaite.
3.3.5. Effect of the fluid flow rate profile on corrosion Figure 21 shows the behavior of the flow pattern on the pipe walls, where it is observed that the minimum flow rate required to obtain a turbulent flow is 0.986 m s −1 (12 l min) −1 for the experimental conditions of this investigation.
An analysis of the fluid behavior on the metal surface of the specimens was carried out to study the effect that the turbulent flow has on the formation of the corrosion products and, consequently, on the corrosion rate of the steel. This analysis allows to observe the surfaces of the samples in two ways and compare the results with each other. The first is to physically observe the specimens after testing in an optical microscope, and the second    is to consider the surfaces states derived from the numerical simulation by ANSYS Fluent at the same experimental conditions. Figure 21 shows the flow rate profile in colors, from 0 m s −1 (laminar flow in blue) to 1.3 m s −1 (turbulent flow in red) along pipe walls, while in the middle section of pipe predominated an intermediate flow rate of 0.9 m s −1 . Profile might indicate that the highest corrosion rate occurred on the pipe walls where the highest flow rate and turbulence occurred. Figure 22 shows the accumulation of corrosion products after 3, 6 and 12 h, respectively, by numerical simulation and after experimental tests at the same times. When images are compared, it is observed a great similarity with the experimental results shown in macrographs. According to the position of the specimens (3, 6, and 9 o´clock) and to the exposure time, it was observed a greater detachment of the corrosion products and therefore a larger surface damage area was generated (in green color). It was observed that the steel samples show regions of darker contrast (dark brown) as the exposure time was increased, which might be an indication of the presence of sulfides (commonly black in color after testing in sour brine) in combination with iron oxides (Fe 2+ black and Fe 3+ dark brown). Sulfates (SO 2 -4 ) normally generates a yellow color and sulfides a black color so the presence of both chemical species might change them into an ochre color, which was observed on the macrographs of steel surface ( figure 22). Sulfur is capable of existing in different oxidation states, from sulfate (+6) to sulfide (-2), the standard potential of the sulfate-sulfide couple was determined as −1.177 V versus SCE  [48], that is to say, -0.73 V versus SHE, which seems to be consistent with Pourbaix diagram (figure 17) [46]. This way, according to previous reaction, sulfide (S 2− ) was then present at the surface, it was possible for this to be reduced either to form thiosulfate (S 2 O 3 2− ) [49]. Despite the presence of oxygen and associated oxides; sulfides seems to predominate in sour brine [50], this way FeS 2− (equation (3)) tended to generate sulfates, previous formation of aqueous sulfuric acid, in the form of a sulfate type of mikasaite Fe 2 (SO 4 ) 3 , which apparently was confirmed by the XRD and HRTEM diffraction patterns. An addittional fact is that color changes on steel surface might be associated to oxides, sulfides and sulfates after the testing time (macrographs) as well as the presence of oxygen, which contribute to stabilize sulfur compounds [51]. Literature does not report the formation of sulfates when carbon steel is tested in normal conditions, though it normally confirms the presence of sulfides when carbon steel is tested in sour brine [52]. However, what motivated this research was the presence of a diffraction pattern associated to a sulfate by means the predominant Miller´s indexes and interplanar distance (figure 18, table 5), so authors postulate that corrosion mitigation migh be ascribed to the presence of a sulfate such as mikasaite. On the other hand, mitigation of corrosion rate might be also derived from copper content in solid solution, as some evidence is provided by the presence of copper sulfide as inidicated by EDX analysis on the corrosion products. HRTEM provided local information on the morphology, crystallography, and chemical composition of corrosion products formed on steel surface. In this case, it was possible to identify the presence of a sulfate as corrosion product. Further, corrosion products shown characteristic morphology with the presence of holes, which means that the flow was not able to detach the layers of corrosion products formed on the steel surface and passed through them mitigating fluid turbulence and probably corrosion rate, so a sulfate appeared to be more stable in comparison to other chemical species on steel surface ( figure 16). This seems to be supported by the reduction of corrosion rates, which were abated from roughly 20 to 2 mm y −1 . Nevertheless, this matter requires a systematic and further study to determine which parameters are most relevant to allow the formation of sulfates under the experimental cited conditions.

Conclusion
Carbon steel in sour brine normally formed oxides though sulfides predominate; the latter is ascribed to a thermodynamic condition of species stability in aqueous solution. Apparently, sulfate formation found the way to be present according to the reactions above postulated. H 2 S, oxygen and Cu might contribute to sulfates formation, which seems to be supported by the evidence provided by HRTEM and XRD diffraction patterns, a Pourbaix diagram ad hoc for the tesing conditions, and surface evidence shown by color change in macrographs so the proposed mechanism apparently fitted with the results found in this work. However, further study is required to get an insight of the effect of oxygen and Cu content on the sulfate formation, as well as on the hollowed pattern found in the corrosion products that seem to provide such a favorable condition to mitigate corrosion.