Impact of Thermal Power Plant Waste Products on Corrosion Dynamics of Pipeline Carbon Steel Alloy in Soil Aqueous Solutions

Abstract

This research examines the effect of thermal power plant desulfurization by-products on pipeline carbon steel alloy corrosion dynamics in soil aqueous solutions to improve soil properties and decrease carbon steel pipeline corrosion rates. Electrochemical tests such as chronopotentiometry, potentiostatic polarization, and electrochemical impedance spectroscopy (EIS) assessed corrosion behavior. Adding desulfurization by-products to soil solutions (DBS) created a transpassive layer during potentiostatic polarization, with a passive current density of about 10μA.cm^-2. EIS measurements showed a substantial increase in polarization resistance, with the DBS exhibiting nearly 700 times higher resistance than the standard soil solution (SSS) at 0V (OCP). Raman spectroscopy identified lepidocrocite (γ−FeOOH) in the DBS-treated coupons, while maghemite and akaganeite were found in chloride-enriched conditions.AFM analysis indicated heightened surface roughness with DBS addition, especially with NaCl. XRF and FTIR spectroscopy of the waste products identified them as primarily composed of silica, aluminosilicates, and oxides of calcium, magnesium, and aluminum. The results elucidate DBS's impact on soil corrosion and suggest methods to mitigate corrosion in industrial and environmental contexts. This study enhances the understanding of buried pipeline corrosion mechanisms and presents a new use of thermal power plant waste for corrosion protection.

Keywords:
Desulfurization by-products; carbon steel corrosion; soil aqueous solutions; electrochemical impedance spectroscopy; surface characterization; waste products

1. Introduction

Corrosion in buried pipelines, particularly near thermal power plants, is a significant financial burden. The worldwide prevalence of buried ferrous water pipes has resulted in a significant socioeconomic loss¹,². Soil corrosivity is a serious problem caused by the corrosion damage to water pipelines being damaged by corrosion, which affects the performance of pipe manufacturers³.

In 2016, the National Association of Corrosion Engineers (NACE) released a comprehensive report detailing the economic impact of corrosion, estimating that the global cost of corrosion amounted to $2.5 trillion in 2013. This figure represents approximately 3.4% of the global Gross Domestic Product (GDP), highlighting the significant economic burden of corrosion⁴. Oil and gas pipelines are subjected to leaks and ruptures owing to corrosion, which significantly affects the safe operation of pipeline systems⁵.

Specifically, in the oil, gas, and chemical sectors, corrosion management presents a difficult challenge, incurring costs of approximately $170 billion annually. Beyond the financial implications, the health and environmental hazards stemming from potential failures in oil and gas infrastructure underscore the urgent need for the advancement of corrosion-resistant materials and the enhancement of corrosion mitigation strategies on a global scale⁶.

This issue is amplified by industrial operations, notably thermal power generation. Specifically, the Candiota Thermoelectric Complex in Brazil exemplifies this challenge. Coal combustion yields gases that produce by-products devoid of industrial applications as raw material after desulfurization with hydrated lime. A 350 MW power facility generates approximately 60 tons/h of such by-products.

The composition of the electrolytes under the coatings is a critical aspect in the corrosion of steel, which reflects the natural electrolyte profile of the soil and is also shaped by how quickly the steel corrodes and the bustling activity of bacteria living under peeling coatings, among other factors⁷. Understanding the intricacies of how corrosion eats away from buried pipelines is difficult. This complexity stems from how the speed of corrosion intertwines with an array of soil characteristics and the inherent qualities of pipeline materials, making it difficult to accurately predict the corrosion rates⁷,⁸.

Strategies for corrosion management include selecting resistant materials, design optimization, applying anti-corrosive chemicals and coatings, cathodic protection, and technological parameter control coupled with rigorous inspection and management throughout all stages of implementation. Improving corrosion management significantly mitigates the risks within the oil, gas, and refining industries, thereby reducing their environmental and societal impacts⁹.

According to the results obtained by Córdoba et al.¹⁰, the most effective anticorrosive procedure for acidic soils is to add 5% of fly ash or cement to the soil, while El-Shamy et al.¹¹, minor additions of (CaO) up to 0.5%wt were added to the filling material to reduce its corrosivity. Our research indicates that adding small amounts of (CaO) to the soil significantly lowers the corrosion rate of steel pipelines. This reduction is primarily attributed to the capacity of (CaO) to decrease the moisture content and increase the alkalinity of clay in the soil, thereby reducing its corrosive potential.

This study investigated the potential of thermal plant desulfurization by-products (DBS) to enhance soil properties and adjust pH levels, thereby reducing the corrosion rate of carbon steel alloys in pipelines. This study focuses on the corrosion of buried carbon steel owing to the substantial costs associated with pipeline replacement. Desulfurization by-products have been utilized as corrosion inhibitors to decrease the corrosion rate. Innovative corrosion mitigation strategies are discussed, followed by an experimental methodology detailing the repurposing of desulfurization by-products as corrosion inhibitors in a case study.

2. Materials and Methods

2.1. Materials

The specimens for the electrochemical tests were prepared by cutting a sample from an industrial metallic pipeline made of the AISI 1020 carbon steel alloy. The samples yielded coupons with dimensions of approximately 4.5 cm x 2.5 cm x 0.7 cm. The coupons were sanded using silicon carbide sandpaper with grain sizes of #100, #220, #400, and #600, and were then degreased in acetone. A circular area of approximately 3.80 cm² was isolated on each coupon using an epoxy resin and a transparent polypropylene cylinder.

A soil sample was collected during maintenance from a buried anti-fire water pipe at a depth of 1m within the Candiota Thermoelectric Complex. Solution preparation followed the procedures established by Brazil´s Electric Energy Research Center (CEPEL). A standard aqueous soil solution was prepared using sieved soil with a concentration of 200 g/L. Another solution consisted of sieved soil (100 g/L) and desulfurization by-products (100 g/L). Subsequently, both solutions were augmented with 3%wt. (NaCl) to accelerate the corrosion process. Table 1 lists the abbreviations used to denote samples.

Table 1
Substrate and solution nomenclature used in the experiments.

Abbreviated Nomenclature	Potential Applied	Solution	Nomenclature
CS	-	-	Carbon Steel
SSS – 0 V	0 V (OCP)	Without NaCl	Standard Soil Solution
SSSNaCl – 0 V		With 3%wt NaCl	Standard Soil Solution with 3%wt NaCl
DBS – 0 V		Without NaCl	Desulfurization By-product Solution
DBSNaCl – 0 V		With 3%wt NaCl	Desulfurization By-product Solution with 3%wt NaCl
SSS – -500 mV	-500 mV (SCE)	Without NaCl	Standard Soil Solution
SSSNaCl – -500 mV		With 3%wt NaCl	Standard Soil Solution with 3%wt NaCl
DBS – -500 mV		Without NaCl	Desulfurization By-product Solution
DBSNaCl – -500 mV		With 3%wt NaCl	Desulfurization By-product Solution with 3%wt NaCl

2.2. Electrical conductivity and pH measurements

The electrical conductivities of the aqueous solutions were determined using a Gehaka® CG2000 conductivity meter, and the pH levels were measured using a Del Lab® microprocessed digital pH meter. These tests assessed changes in soil solution conductivity following the addition of desulfurization by-products and NaCl (DBSNaCl). Furthermore, the pH levels were measured and adjusted based on the Pourbaix Diagram to promote layer formation.

2.3. Electrochemical tests

The electrochemical tests were conducted using a GAMRY Reference 3000 potentiostat with a conventional three-electrode cell setup. The reference electrode was a standard calomel electrode, Ag || AgCl || KCl (saturated) (Orion®), with a reference potential of +195 mV(SHE). The counter electrode was platinum and the AISI 1020 carbon steel alloy electrode, prepared earlier, served as the working electrode. Experiments were performed in triplicate, and the data were analyzed using the Gamry ECHEM® software.

Owing to the transient nature of the corrosion process, the open circuit potential was measured after conditioning the sample for 1 min at a potential of -0.7 V(OCP) and scanning the potential over 120 s for all solutions to mitigate the potential formation of an oxide layer prior to the commencement of the experiments.

For potentiostatic polarization, a scanning rate of 10 mV/s was adopted, with the applied potential ranging from -1.2 V (OCP) to +0.7 V (OCP). This range of values was used to increase the initial cathodic effect on the surface and reduce any possible oxide-layer formation.

Electrochemical Impedance Spectroscopy (EIS) measurements were performed using an overlapping AC voltage with an amplitude of 10 mV (RMS) at a constant potential of the open circuit potential (E_OCP) and at –500 mV (SCE) under the most severe layer formation condition across a frequency range of 100 kHz to 100 mHz. These measurements facilitated the determination of the electrical resistance of the oxide layer and the analysis of the immersion time of the corrosive solution.

Following the electrochemical tests, the samples were rinsed with distilled water and air-dried at room temperature. Macroscopic photographs were obtained to examine surface morphology. In addition, the coupons were subjected to Raman spectroscopy. Atomic force microscopy (AFM) images were used to determine the composition and microstructural features of the surface layer.

2.4. Spectroscopic techniques and X-ray fluorescence

After ten days of immersion in solutions with and without 3 wt% (NaCl), the rust on the carbon steel coupons was characterized by Raman Spectroscopy. Measurements were conducted at room temperature using a B&WTek® micropositioning system and Andor Shamrock 303i monochromator. A thermo-electrically cooled, back-illuminated deep-depletion device operating at -80°C was utilized for signal detection, ensuring minimal dark current. The sample was irradiated with a 532 nm excitation laser beam focused by a microscope equipped with an 80x objective lens, which also collected the Raman signal in the backscattered direction. The excitation power was constant at 0.20 mW∙μm^-2 to avoid any thermal effects. Table 1 lists the compositions of the analyzed samples, including the desulfurization by-products and soil. The waste products were analyzed by Fourier transform infrared (FT-IR) spectroscopy. Spectra were recorded on a Shimadzu IR Prestige-21 spectrometer in the transmission mode using KBr discs. Data were acquired at room temperature by collecting 100 scans at a resolution of 4.0 cm⁻¹.

Three coal-fired power plant waste samples were analyzed using a Bruker S1 Turbo SD portable X-ray fluorescence (XRF) analyzer. Measurements were performed in triplicate, with an acquisition time of approximately 120 seconds per analysis, using the GeoChem General FP method in the dual mode.

2.5. Atomic force microscopy

The surface samples detailed in Table 1 were analyzed by atomic force microscopy (AFM) using an Agilent® Technologies 5500 instrument. Before the analysis, rust was removed from the coupon surfaces using tape. AFM images were acquired at room temperature in non-contact mode by employing high-resolution SSS-NCL probes (Nanosensors ®) with a force constant of 48 N/m and a resonance frequency of 154 kHz). Image capture and analysis were performed using PicoView® 1.14.4 software from the Molecular Imaging Corporation and further analyzed using PicoImage® 5.1 software. The AFM analysis procedure was designed to provide high-resolution imaging of the sample surfaces, focusing on the characteristics of the superficial layer formed post-EIS testing.

3. Results

3.1. Soil aqueous solutions pH

The pH and electrical conductivity of each solution were then measured. The solutions SSS and SSSNaCl showed pH values of 7.91 and 7.97, respectively. Meanwhile, DBS and DBSNaCl exhibited pH values of 12.64 and 12.39, indicating that adding desulfurization by-products increases the pH values. The electrical conductivity readings for SSS and SSSNaCl were 359 and 19,790 µS/cm, respectively, whereas those for DBS and DBSNaCl were 5,090 and 19,750 µS/cm, respectively. These measurements revealed that (NaCl) significantly increased the electrical conductivity of the solution, potentially accelerating corrosion.

3.2. Open circuit potential

Figure 1 shows the results of the chronopotentiometric tests for all solutions. The measured pH value was in the range of 8 for SSS and SSSNaCl, whereas for DBS and DBSNaCl, it was in the range of 12.5. The average corrosion potential for SSS stabilized close to -390 mV (SCE) (-225 mV (SHE)) and for SSSNaCl stabilized close to -515 mV (SCE) (-320 mV (SHE)), while for DBS it stabilized close to -320 mV (SCE) (-125 mV (SHE)) and for DBSNaCl stabilized close to -335 mV (SCE) (-140 mV (SHE)).

Figure 1
Chronopotentiometric curves typical of the open-circuit potential as a function of time for carbon steel immersed in aqueous solutions (SSS, SSSNaCl, DBS, DBSNaCl).

The potential values for SSS and SSSNaCl are within the corrosion domain, according to the Pourbaix electrochemical equilibrium diagram for the Fe/H₂O system at 25°C; therefore, these solutions can exhibit active dissolution, which is thermodynamically spontaneous. The potential values for DBS and DBSNaCl were within the passivity domain according to the Pourbaix electrochemical equilibrium diagram for the Fe/H₂O system at 25°C. Therefore, this resulted in the formation of a passivation layer¹².

This passivation process occurred because the addition of by-products increased the pH to approximately 12.5. In this case, an oxide protective film was formed around the steel that protected it from corrosion, and a similar process occurred with the addition of cement when the pH increased to approximately 13¹².

3.3. Potentiostatic polarization

Figure 2 shows the typical potentiostatic polarization curves for all solutions. This observation highlights the unique behavior of DBS and DBSNaCl compared with standard solutions. The standard solutions did not exhibit the formation of a pseudo-passivating layer in DBS with a current density of approximately 10 μA/cm² and DBSNaCl with a current density of approximately 20 μA/cm². In DBS, the pseudo-passivating layer does not maintain a constant current density within the formation potential range. In NaCl, there is a stable layer over a larger range, even though the current densities are similar.

Figure 2
Typical potentiostatic polarization curves for carbon steel immersed in aqueous solutions: (A) SSS and DBS and (B) SSSNaCl and DBSNaCl.

After the potentiostatic polarization tests, the coupons were washed with distilled water, dried, and photographed, as shown in Figure 3. To further investigate the effects of different solutions on the metal surface, a visual inspection was conducted after the potentiostatic polarization tests. The visual aspect of DBS shown in (C) was better than that of SSS, as shown in (A). This appeared to form a passivation layer in the DBS, as shown in (C).

Figure 3
Coupons immersed in aqueous solution samples: (A) SSS, (B) SSSNaCl, (C) DBS, and (D) DBSNaCl.

To understand the surface layer formation process from the potentiostatic polarization curves, two potentials were chosen (0 V(OCP) and -500mV(SCE)) to evaluate the formation of the superficial layer versus immersion time using EIS measurements.

3.4. Electrochemical impedance spectroscopy

The Bode and Nyquist diagrams for the coupons immersed in all solutions are presented in Figures 4 and 5 after 10 days of exposure to consider the minimal time required to form a measurable layer. It was measured potentiostatically at 0 V (OCP) to evaluate the system equilibrium potential in two solutions: a Standard Soil Solution (SSS) without NaCl (Figure 4A for Bode, Figure 4B and Figure 4C for Nyquist plots), and Standard Soil Solution (SSS) with 3%wt. NaCl (SSSNaCl) (Figures 4D, 4E and 4F.

Figure 4
Normalized Bode and Nyquist plots measured potentiostatically on the potential 0V(OCP), after 10 days of exposure: (A) Bode and (B) Nyquist plots of SSS and DBS at 0 V ; (C ) Zoomed-in Nyquist plots of (B); (D) Bode and (E) Nyquist plots of SSSNaCl (SSS+3%wt.NaCl) and DBSNaCl (DBS+3%wt.NaCl) at 0 V; (F) Zoomed-in Nyquist plots of (E).

Figure 5
Bode and Nyquist plots measured potentiostatically on the potential -500 mV(SCE), after 10 days of exposure: (A) Bode and (B) Nyquist plots for the SSS and DBS at -500 mV(SCE) (C) Zoomed-in plot of (B); (D ) Bode and (E) Nyquist plots for the SSSNaCl (SSS+3%wt.NaCl) and DBSNaCl (DBS+3%wt.NaCl) at -500mV(SCE); (F) Zoomed-in plot of (E).

This study used a potential of -500 mV (SCE) to determine the worst conditions for layer formation. Thus, in Figure 5, the Bode and Nyquist plots, measured potentiostatically at -500 mV (SCE) after 10 days of exposure, are presented in Figure 5A Bode and Figure 5B Nyquist plots. Figure 5C shows a magnified plot of (B). In addition, SSS and DBS with 3wt. of NaCl are presented in Figure 5D as a Bode plot, and Figure 5E shows a Nyquist plot of SSSNaCl (SSS+3%wt NaCl) and DBSNaCl (DBS+3%wt.NaCl) at -500mV(SCE). (F) Zoomed-plot of (E).

To simulate the electrochemical effects of the immersed coupons, two equivalent circuits were used to interpret the electrochemical behavior observed in the Bode and Nyquist plots. While equivalent circuit model 1 (ECM 1) adequately represents the corrosion process in standard soil solutions, a different approach is necessary for desulfurization by-product solutions. ECM 1 for both standard soil solutions (SSS and SSSNaCl) is presented in Figure 6A and ECM 2 for both desulfurization by-product solutions (DBS and DBSNaCl) is presented in Figure 6B. Notably, a 45° inclination at low frequencies in the Nyquist plots in Figures 4 and 5 is not observed; thus, the Warburg model is not considered in the models shown in Figure 6.

Figure 6
EIS equivalent circuit models (ECM) used to simulate the electrochemical effects on the coupons immersed in ECM 1 for both standard soil solutions (SSS and SSSNaCl) and ECM 2 for both desulfurization by-product solutions (DBS and DBSNaCl).

The Warburg element, which represents the diffusion process, was not included in the equivalent electrical circuits because the contribution of diffusion is less significant than that of the charge transfer process. Figure 7D illustrates the discrepancy between the fitted model incorporating the Warburg element and the observed values and compares these plot values with the ECM2 fit.

Figure 7
Nyquist plots of SSS at -500mV (SCE). (C) Full plot fitting by ECM 2 (D) Zoomed plot of (C) in the specific inflection zone. A 2nd fitting option considering an ECM with a diffusion component (W) (FIT 2) was included to present an unsuccessful fit. Obs. Fit 2, presented in Figure 7C, was inserted only to demonstrate divergence with the possible occurrence of the diffusion process.

The shapes of the Nyquist diagrams suggest that the charge-transfer process was the dominant factor controlling the corrosion of carbon steel in the tested solutions.

Table 2 presents the fitting values of Figures 5 and 7 as examples of typical fitting values obtained using the simplex method of fit. The goodness-of-fit represents the statistical parameter of the fitting quality. High-quality fitting was defined as goodness-of-fit values below 1E^-3.

Table 2
Fitting values of plots presented in Figure 5E for ECM 1 and Figure 7 for ECM2.

ECM 1 fitting of SSSNaCl; 3%wt.NaCl at 0V after 10 days of immersion
Parameters (ECM1)	Value	Error	Units
Rcor	748.1	15.1	ohms*cm^2
Rsol	7.826	3.95E-02	ohms*cm^2
CPEcor	1.37E-03	1.59E-05	S*s^a/cm^2
nCPEcor	8.45E-01	3.27E-03
Goodness of Fit	5.74E-04
ECM 2 fitting SSSNaCl; 3%wtNaCl at -500mV after 10 days of immersion
Parameters (ECM2)	Value	Error	Units
Rsol	9.28E-01	1.32E+03	ohms*cm^2
Rcor	1.83E+03	128	ohms*cm^2
Rpore	373.2	1.32E+03	ohms*cm^2
Ccor	8.76E-04	2.56E-05	S*s^a/cm^2
n	7.76E-01	1.61E-02
Cc	4.39E-10	3.10E-09	S*s^a/cm^2
m	9.90E-01	2.67E-04
Goodness of Fit	4.87E-05

The charge transfer process controls the corrosion process of carbon steel alloy immersed in the SSS, as evidenced by the semicircle in the low-frequency region of the Nyquist plots in Figures 4B and 5B. In contrast, the Nyquist plots of the desulfurization by-product solution (DBS), shown in Figures 4D and 5D, exhibit a straight-line behavior at low frequencies, indicating an overlap of semicircles related to the charge transfer process and the diffusion process, which occurs through the passive layer, owing to the solution thermodynamically stable conditions for the creation of the oxide film. A similar result was reported by Giarola et al. , who used an NS4 synthetic solution and cement to modify its properties.

Electrochemical circuits were used to adjust the electrochemical impedance spectroscopy data. The low-frequency arc of the Nyquist diagram (below 100 Hz) was associated with the corrosion process at the steel/soil interface. ECM 1, in Figure 6A, represents the corrosion process at electrolyte resistance (Rsol) and the parallel pair composed of corrosion resistance (Rcor) and the constant phase element (CPEdl) in the low-frequency arc, corresponding to the charge transfer resistance between the carbon steel alloy and the standard soil solution¹². For DBS, the impedance assumed higher values, which occurred in the passivation process, and the corrosion process happened at the carbon steel/oxide film interface. The equivalent circuit model that represents the process is illustrated by ECM 2 in Figure 6B. The equivalent circuit is composed of electrolyte resistance (Rsoln), and constant phase element (Cc) of the oxide, which represents the capacitance equivalent of carbon steel alloy, Rpore is the oxide charge transfer resistance between carbon steel and the solution, and Rcor and Ccor represent the polarization resistance of charge transfer from the metal and the oxide film, and the capacitance at this interface. Comparing the Bode essays in SSS and DBS without and with NaCl (Figure 4 and Figure 5), one more time constant was observed, as shown in Figure 5 than in Figure 4. Based on this evidence, two models (ECM 1 and ECM 2) were used. ECM 2 presents an additional pair of resistance and capacitive effects (CPE); therefore, the polarization resistance was used to compare the evidence of layer formation on the surface. Evaluation of the values obtained from the Bode plots indicated that the steel alloy coupons immersed in DBS presented a high impedance normalized modulus (Zmod). Comparing SSS with DBS at 0 V, it can be observed that the Zmod of DBS was 11 times higher than that of SSS (Figure 4A). In addition, comparing SSSNaCl and DBSNaCl at 0 V, the Zmod of DBSNaCl is 2 times higher than that of SSSNaCl (Figure 4C). However, at -500 mV, a greater difference in Zmod between the solutions (SSS and DBS) was not observed (Figure 5).

The polarization electrical resistances obtained using both models of the equivalent circuits are shown in Figure 8. The DBS exhibited the highest electrical resistance over time. The formed passive layer presented 470 kΩ a polarization resistance instead of 0.7 kΩ in the standard soil solution after 10 days of potentiostatic measurement at a potential of 0 V (OCP).

Figure 8
Mean of the behavior of polarization electrical resistance overtime on the potential: (A) at 0 V (OCP): SSS – 0 V, SSSNaCl – 0 V, DBS – 0 V, and DBSNaCl – 0 V; (B) at -500 mV (SCE): SSS – -500 mV, SSSNaCl – -500 mV, DBS – -500 mV and DBSNaCl – -500 mV.

After 10 days of exposure to the EIS tests, the coupons were washed with distilled water, dried, and photographed, as shown in Figure 9.

Figure 9
Coupons immersed in aqueous solutions samples measured potentiostatically on the potential 0 V (E_OC) and -500 mV (SCE), after 10 days of exposure: (A) SSS – 0 V; (B) SSSNaCl – 0 V; (C) DBS – 0 V; (D) DBSNaCl – 0 V; (E) SSS – -500 mV; (F) SSSNaCl – -500 mV; (G) DBS – -500 mV; (H) DBSNaCl – -500 mV.

Following the EIS tests, macrophotography was performed to visually corroborate the electrochemical findings and examine the physical state of the coupons after the exposure.

3.5. Raman spectroscopy

Micro-Raman spectroscopy is well suited for characterizing the oxides and (oxy)hydroxides involved in the corrosion of iron alloys, offering insights into the active phases of corrosion processes¹³.

Various iron oxides and hydroxides can form on steel surfaces during atmospheric corrosion, including lepidocrocite (γ-FeOOH) with an orthorhombic structure, goethite (α-FeOOH) also with an orthorhombic structure, magnetite (Fe₃O₄) with a cubic structure, maghemite (χ-Fe₂O₃) which can have either a cubic or trigonal structure, hematite (α-Fe₂O₃) with a trigonal structure, and akaganeite (β-FeOOH) with a monoclinic structure. Experimental observations have frequently identified magnetite and that maghemite coexists, leading to focused studies on this binary mixture. Furthermore, the ternary compounds of goethite–magnetite–maghemite and akaganeite–magnetite–maghemite have been extensively studied because of their common presence in corroded areas¹⁴.

Zhang et al.¹⁵ observed a peak at 987 cm^-1 in the Raman spectra of the corrosion products on the rust layers of a carbon steel alloy sample immersed in an electrolyte for both 0 and 24 h, which nearly vanished after 72 h. The authors identified the corrosion products present at 24 h as lepidocrocite (γ-FeOOH), characterized by bands at 250, 380, 526, 651, 1050, and 1304 cm⁻¹. Furthermore, the study reported that a compact black layer emerged after 168 h of immersion. This layer was determined to be composed of magnetite (Fe₃O₄), as evidenced by bands at 550 and 674 cm^-1, along with mid-strong bands at 250, 380, and 1307 cm^-1, indicative of (γ-FeOOH) on the outer layer. Additionally, the study explored Raman spectra in an air environment, where bands at 298, 399, 481, 554, 675, and 1002 cm^-1 were attributed to goethite.

Raman spectroscopy was employed before and after conducting Electrochemical Impedance Spectroscopy (EIS) analyses over ten days to elucidate the differences in the compounds present in carbon steel and desulfurization by-products. Figure 10 shows the Raman spectra of carbon steel alloy and desulfurization by-products. The soil samples did not exhibit any detectable Raman signals, indicating the absence of Raman-active compounds under the tested conditions.

Figure 10
Raman characterization of the substrate (carbon steel alloy) and desulfurization by-product used in the solutions before EIS tests.

Figure 11 presents the Raman spectra obtained potentiostatically at a potential of 0V (OCP) for coupons submerged in standard soil solutions and desulfurization by-product solutions, with and without the addition of 3 wt% (NaCl). This setup aimed to assess the influence of (NaCl) on the corrosion process by comparing the molecular changes in the presence and absence of this electrolyte.

Figure 11
Raman spectra at 0 V (OCP): SSS – 0 V, DBS – 0 V, SSSNaCl – 0 V and DBSNaCl – 0 V.

Furthermore, Figure 12 displays the Raman spectra acquired potentiostatically at a potential of -500 mV (SCE), targeting coupons exposed to identical treatments as those analyzed at 0 V (OCP). This approach allows for a comparative investigation of the electrochemical behavior and resultant corrosion products under more aggressive conditions after a 10-day exposure period in the specified solutions.

Figure 12
Raman spectra at -500 mV (SCE): SSS – -500 mV, DBS – -500 mV, SSSNaCl – -500 mV and DBSNaCl – -500 mV.

Incorporating cementitious agents into the soil, which results in an elevated pH, significantly alters the formation of rust on the steel alloy. The Pourbaix Diagram for iron illustrates that at higher pH levels, iron oxides and hydroxides possess protective qualities, facilitating the passivation of iron or steel, in the same way as happens in the reinforcing steel embedded in concrete¹⁰. To verify this hypothesis, Raman spectroscopy was employed to characterize the corrosion products on carbon steel alloy samples exposed to SSS and soil amended with DBS.

The analysis revealed distinct peaks within the 265 and 284 cm^-1 range, and the 1070 and 1087 cm^-1 range for DBS at both 0 V and -500 mV and DBSNaCl at 0 V and -500 mV, as depicted in Figures 10 and 11. These peaks correlate with the signature of the original desulfurization by-products, which exhibit peaks at 265 and 1072 cm^-1 in Figure 10. Notably, some samples displayed fluorescence, a phenomenon that was significantly more intense than the Raman signal, likely due to impurities such as organic materials within the samples.

When subjected to potentiostatic treatment at -500 mV (SCE) and immersed in both standard and modified soil solutions (with and without 3%wt (NaCl)), the samples showed a peak at 481 cm^-1 (Figure 11), indicative of goethite (α-FeOOH), corroborating findings by Zhang et al.¹⁵ and Córdoba et al.¹⁰ Similarly, samples immersed in DBS and treated under the same conditions revealed peaks between 1070 and 1081 cm^-1, signifying the presence of lepidocrocite (γ-FeOOH), as identified in previous studies¹⁵. Antunes et al.¹⁶ suggested that the red regions in these spectra denote lepidocrocite, observable in DBSNaCl samples at both potentials, as shown in Figure 10D and Figure 10H.

Furthermore, the samples immersed in DBSNaCl at 500 mV exhibited peaks between 640 and 750 cm^-1 in Figure 11, likely representing maghemite (γ-Fe₂O₃) and akaganeite (β-FeOOH), as discussed by Dubois et al.¹⁴. Akaganeite, typically found in high-chloride environments¹⁶, underscores the role of anions such as Cl^- and SO4^2- as catalysts in steel corrosion, promoting the formation of (γ-FeOOH) as the ultimate corrosion product in solution. The availability of dissolved oxygen critically influences the transformation of corrosion products, with (FeOOH) potentially converting to (Fe₃O₄) under oxygen-limited conditions, leading to the formation of an inner layer at the interface between the outer layer and the substrate¹⁵. Antunes et al.¹⁶ noted that magnetite is common under aggressive atmospheric conditions, highlighting the complex interplay of environmental factors in the corrosion process.

3.6. AFM

Atomic Force Microscopy (AFM) is a powerful tool for assessing the roughness of relatively smooth and flat surfaces. However, AFM analysis of corrosion products poses challenges owing to inherent surface roughness¹⁷.

This technique was applied to examine the surface regions that underwent two distinct potentiostatic treatments: 0 V (OCP) and -500 mV (SCE). The results of these AFM scans, including the mappings and roughness measurements (Root Mean Square, RMS), are shown in Figure 13 and Table 3. These figures and tables include comparisons across various samples: those immersed in a standard soil solution (Figures 13A and E), in a standard soil solution augmented with 3%wt NaCl (Figures 13B and F), in a desulfurization by-product solution (Figures 13C and G), and a DBS enhanced with 3%wt NaCl (Figures 13D and H).

Figure 13
AFM mapping: (A) SSS – 0 V; (B) SSSNaCl – 0 V; (C) DBS – 0 V; (D) DBSNaCl – 0 V; (E) SSS – -500 mV; (F) SSSNaCl – -500 mV; (G) DBS – -500 mV; (H) DBSNaCl – -500 mV.

Table 3
Roughness RMS (area of 50 µm x 50 µm).

Sample	Roughness RMS (nm)	Sample	Roughness RMS (nm)
CS	17.24
SSS – 0 V	68.57	SSS - -500 mV	82.67
SSSNaCl – 0 V	131.0	SSSNaCl - -500 mV	190.2
DBS – 0 V	16.05	DBS - -500 mV	29.23
DBSNaCl – 0 V	652.4	DBSNaCl - -500 mV	352.3

Analysis of the AFM imagery revealed that samples immersed in the DBS containing 3%wt NaCl (DBSNaCl) and subjected to both potentiostatic conditions exhibited significantly higher surface roughness than those immersed in the SSS. This observation, supported by quantitative roughness measurements, suggests the possible formation of a denser and thicker layer on the surface. Such a layer could indicate enhanced corrosion protection or altered material interactions resulting from desulfurization by-products and (NaCl) treatment. This finding underscores the potential of AFM to provide critical insights into the microstructural changes induced by corrosion and treatment processes, offering a pathway for understanding and mitigating corrosion.

3.7. Waste products analyses

The XRF results (Table 4) indicate the predominance of calcium oxide (CaO), magnesium oxide (MgO), silicon dioxide (SiO₂), aluminum oxide (Al₂O₃), iron oxide (Fe₂O₃), and sulfur (S) in the analyzed samples. The (CaO) content ranged from 42.1 to 43.9%, (MgO) from 17.6 to 18.9%, (SiO₂) around 13.5%, (Al₂O₃) from 9.2 to 9.7%, (Fe₂O₃) from 1.05 to 1.17% and (S) from 11.6 to 11.7%. Minor concentrations of potassium oxide (K₂O), titanium dioxide (TiO₂), and phosphorus pentoxide (P₂O₅) were also detected along with traces of elements such as chlorine, vanadium, chromium, manganese, cobalt, zinc, molybdenum, and rhodium. The typical chemical composition of coal ash and waste is calcium, silicon, aluminum, and iron compounds. The high sulfur content is due to the presence of sulfides and sulfates, which are common in coal waste. The XRF technique proved to be suitable for rapid and accurate analysis of the elemental composition of the samples, with detection limits in the range of tens to hundreds of ppm for most elements, using the GeoChem General method in dual mode.

Table 4
Triplicate results of X-ray fluorescence analyses of coal waste samples (% m/m).

Sample	MgO	Al2O3	SiO2	P2O5	S	Cl	K2O	CaO	TiO2	Fe2O3
1	18.40	9.72	13.50	2.34	11.60	0.16	0.81	43.90	0.06	1.17
2	18.90	9.16	13.50	2.42	11.60	0.15	0.77	42.20	0.04	1.05
3	17.60	9.64	13.90	2.33	11.70	0.16	0.80	42.10	0.05	1.08

Fourier-transform infrared (FTIR) spectra were collected to corroborate the X-ray fluorescence results. The FTIR spectroscopic analysis of the coal ash samples revealed a complex spectrum with multiple absorption bands corresponding to various mineral and organic components. Table 5 summarizes the key absorption bands identified and their assignments based on literature and experimental data (Figure 14). The broad band centered at 3200-3400 cm⁻¹ is attributed to the O-H stretching vibrations of the adsorbed water molecules and hydroxyl groups. The weak band at 2513 cm⁻¹ was assigned to the stretching vibrations of the S-H bonds, possibly originating from sulfur-containing compounds in the waste sample ²². The sharp absorption band at 1796 cm⁻¹ is characteristic of the C=O stretching vibrations of carbonyl groups, which are associated with carboxylic acids or esters²². The 1630-1600 cm⁻¹ band corresponds to the H-O-H bending mode of the adsorbed water molecules²². The strong absorption band at 1430-1400 cm⁻¹ was assigned to the asymmetric stretching vibrations of the Ca-O bonds in calcium oxide¹⁸. The intense band at 1080-1140 cm⁻¹ was attributed to the stretching vibrations of the S-O bonds in sulfate²⁰. The band at 987 cm⁻¹ is assigned to the asymmetric stretching of Si-O-T bonds (where T = Si or Al) in more aluminum-rich aluminosilicate structures²². The absorption band at 950 cm^-1 is characteristic of Si-O-Si or Al-O-Si vibrations in aluminosilicates²². The absorption bands at 875 cm⁻¹ and 715 cm⁻¹ are attributed to the in-plane and out-of-plane bending vibrations of the Ca-O bonds in (CaO), respectively¹⁸. The band at 650 cm⁻¹ corresponds to the stretching vibrations of Mg-O bonds in (MgO)¹⁹. The absorption band at 550-450 cm⁻¹ is assigned to the vibrations of Al-O bonds in aluminum oxide (Al₂O₃)²¹. The band at 450 cm⁻¹ was attributed to the presence of (SiO₂)²⁰. The absorption band in the 400-450 cm⁻¹ range is assigned to the bending vibrations of Mg-O-Mg bonds in magnesium oxide (MgO)¹⁹. Finally, the weak band at 362 cm⁻¹ was attributed to the vibrations of the Ca-O bonds¹⁸.

Table 5
FTIR vibrational assignments of waste power plant.

Wavenumber (cm−1)	Assignment	Reference
362	Ca-O	Hussein et al.18
400-450	Mg-O-Mg bending (MgO)	Yusoff et al.19
450	SiO2	Kirk20
550-450	Al-O vibrations (Al2O3) - Al–O stretching mode in octahedral structure	Djebaili et al.21
650	Mg-O stretching (MgO)	Yusoff et al.19
715	Ca-O out-of-plane bending (CaO)	Hussein et al.18
875	Ca-O in-plane bending (CaO)	Hussein et al.18
950	Al2O3 - O–H deformation vibrations	Djebaili et al.21
987	Si-O-T asymmetric stretching (where T = Si or Al) in more aluminum-rich aluminosilicate structures	Yin et al.22
1080-1140	S-O stretching vibration	Smidt et al.23
1430-1400	Ca-O asymmetric stretching (CaO)	Hussein et al.18
1796	C=O stretching of carbonyl groups	Yin et al.22
1630-1600	The bands near 1606 cm−1 were assigned to the aromatic C=C ring stretching vibrations. The band near 1632 cm−1 is related to the OH- bending vibration of adsorbed H2O.	Yin et al.22
2513	S-H stretching vibrations	Yin et al.22
3200-3400	O-H stretching (adsorbed water, hydroxyl groups)	Yin et al.22

Figure 14
FTIR vibrational bands of waste products.

FTIR analysis revealed that the thermoelectric power plant waste sample was primarily composed of inorganic compounds, including silica (SiO₂), aluminosilicates, calcium oxide (CaO), magnesium oxide (MgO), and aluminum oxide (Al₂O₃). The spectrum also shows the presence of adsorbed water molecules and hydroxyl groups. In addition, the sample contained traces of organic compounds, as indicated by the absorption bands associated with carbonyl groups and S-H bonds.

4. Discussions

Chronopotentiometric analysis revealed a reduction in the open-circuit potential upon the addition of (NaCl), indicating its influence on electrochemical behavior. Potentiostatic analysis further elucidated the formation of a transpassive electrochemical layer on the material surface of the material at 0 V(OCP) and -500 mV(SCE) potentials, facilitated by the presence of DBS and its (NaCl) variant.

Insights from colorized macrophotography, presenting gray and orange hues, suggest that the heterogeneity of the surface layer potentially encompasses diverse substances, thicknesses, or morphologies. This hypothesis is corroborated by Raman spectroscopy and AFM data, which provide evidence of layered complexity.

Electrochemical Impedance Spectroscopy (EIS) conducted in a DBS solution at 0 V (OCP) demonstrated a polarization electrical resistance of 470 kΩ after a 10-day immersion, indicating the protective nature of the formed layer. Raman spectroscopy suggested that this layer bears compositional similarities to lepidocrocite, whereas AFM analysis revealed a surface roughness of 16.05 nm, attributing a tangible texture to the layer.

Conversely, at a potential of -500 mV(SCE), the polarization resistance in DBS dropped to 1.3 kΩ after ten days, suggesting a diminished protective effect. Resistance to DBSNaCl was notably lower, underscoring the deleterious impact of chloride ions. Raman spectroscopy of the DBSNaCl-immersed coupons revealed a composition akin to maghemite and akaganeite, minerals typically formed in chloride-rich environments, hinting at the aggressive corrosion mechanisms at play.

Further, AFM analyses post a ten-day immersion period showcased DBSNaCl samples exhibiting the most pronounced surface roughness, with measurements of 652.4 nm and 352.3 nm at 0 V(OCP) and -500 mV(SCE), respectively. These findings highlight the significant role of (NaCl) in altering the electrochemical properties and surface morphology of materials, offering critical insights into their corrosion behavior and structural integrity in chloride-enriched settings.

Considering that the findings of this case study are specific to a particular region, they may not be directly applicable to areas with different soil composition and environmental conditions. Further research is required to validate the effectiveness of desulfurization by-products as corrosion inhibitors in diverse soil types and climates. In addition, this study did not investigate the long-term stability and effectiveness of these by-products as corrosion inhibitors.

5. Conclusions

In soil from Candiota, southern Brazil, incorporating a desulfurization by-product (DBS) from a thermal power plant led to the formation of a transpassive layer in potentiostatic polarization studies. This layer exhibited a passive current density of approximately 10 μA/cm², in contrast to the behavior observed in standard soil solutions, where no such passive current was observed. Macroscopic photography revealed evidence of pitting corrosion on specimens subjected to potentiostatic polarization, highlighting the aggressive nature of this corrosion form under certain conditions. Remarkably, DBS greatly enhanced the polarization resistance of carbon steel alloys. Electrochemical Impedance Spectroscopy at 0 V (OCP) showed that the resistance in DBS was nearly 700 times higher than in standard soil solution . At -500 mV (SCE), the polarization resistance in DBS remained 12 times greater than in SSS, underscoring the protective potential of the by-product. Raman spectroscopy confirmed the presence of lepidocrocite (γ-FeOOH) in samples treated with DBS at both 0 V (OCP) and -500 mV (SCE), while the detection of maghemite and akaganeite in DBSNaCl at -500 mV (SCE) reflected the complex corrosion processes in chloride-enriched environments. FTIR analysis showed that the thermoelectric power plant waste was mainly composed of silica (SiO₂), aluminosilicates, calcium oxide (CaO), magnesium oxide (MgO), and aluminum oxide (Al₂O₃), along with adsorbed water and hydroxyl groups. Traces of organic compounds were also identified through carbonyl and S-H absorption bands. AFM images clarified the morphology of the oxide layer in DBSNaCl under potentiostatic conditions at 0 V (OCP) and -500 mV (SCE), revealing a pronounced increase in surface roughness compared to SSSNaCl at both potentials. The addition of 3 wt% (NaCl) desulfurization by-products resulted in a notably rougher surface layer. These results highlight the complex interplay between desulfurization by-products and soil corrosion mechanisms, providing practical insights into advanced protection strategies for industrial and environmental applications.

6. Acknowledgments

The authors are grateful to Eletrobras CGT Eletrosul for supplying samples for this study and for the opportunity to conduct this research. This work was partially funded by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior–Brazil (CAPES), Fund Code 001, LLPTM and Manna team, INCT-NAMITEC, FAPERGS TO21/2551-0000518-1, CNPq Project 406193/2022-3, and CNPq Process 406311/2023-4.

7. References

• ¹ Voloshyn VA, Zvirko OI, Sydor PY. Influence of the compositions of neutral soil media on the corrosion cracking of pipe steel. Mater Sci. 2015;50(5):671-5. http://doi.org/10.1007/s11003-015-9770-7
  » http://doi.org/10.1007/s11003-015-9770-7
• ² Wasim M, Djukic MB. Corrosion induced failure of the ductile iron pipes at micro- and nano-levels. Eng Fail Anal. 2021;121:105169. http://doi.org/10.1016/j.engfailanal.2020.105169
  » http://doi.org/10.1016/j.engfailanal.2020.105169
• ³ Jiru MG, Tolcha MA, Yeshanew DA. Investigation of soil physicochemical effects on the corrosion of buried iron pipes. J Basic Appl Sci. 2021;17:95-106. http://doi.org/10.29169/1927-5129.2021.17.11
  » http://doi.org/10.29169/1927-5129.2021.17.11
• ⁴ Arriba-Rodriguez L, Villanueva-Balsera J, Ortega-Fernandez F, Rodriguez-Perez F. Methods to evaluate corrosion in buried steel structures: a review. Metals. 2018;8(5):334. http://doi.org/10.3390/met8050334
  » http://doi.org/10.3390/met8050334
• ⁵ Zhou Q, Wu W, Liu D, Li K, Qiao Q. Estimation of corrosion failure likelihood of oil and gas pipeline based on fuzzy logic approach. Eng Fail Anal. 2016;70:48-55. http://doi.org/10.1016/j.engfailanal.2016.07.014
  » http://doi.org/10.1016/j.engfailanal.2016.07.014
• ⁶ Dwivedi D, Lepková K, Becker T. Carbon steel corrosion: a review of key surface properties and characterization methods. RSC Advances. 2017;7(8):4580-610. http://doi.org/10.1039/C6RA25094G
  » http://doi.org/10.1039/C6RA25094G
• ⁷ Arabey AB, Bogdanov RI, Ignatenko VE, Nenasheva TA, Marshakov AI. Effect of corrosion medium composition on rate of crack growth in X70 pipeline steel. Prot Met Phys Chem Surf. 2011;47(2):236-45. http://doi.org/10.1134/S2070205111020031
  » http://doi.org/10.1134/S2070205111020031
• ⁸ Rajeev P, Imteaz M, Gad E. Effect of soil corrosion in failures of buried pipelines. Proc Int Struct Eng Constr. 2015;2(1):1-6. http://doi.org/10.14455/ISEC.res.2015.221
  » http://doi.org/10.14455/ISEC.res.2015.221
• ⁹ Groysman A. Corrosion problems and solutions in oil, gas, refining and petrochemical industry. Koroze Ochr Mater. 2017;61(3):100-17.
• ¹⁰ Córdoba VC, Mejía MA, Echeverría F, Morales M, Calderón JA. Corrosion mitigation of buried structures by soils modification. Ingeniare Rev Chil Ing. 2011;19(3):486-97. http://doi.org/10.4067/S0718-33052011000300016
  » http://doi.org/10.4067/S0718-33052011000300016
• ¹¹ El-Shamy AM, Shehata MF, Metwally HIM, Melegy A. Corrosion and corrosion inhibition of steel pipelines in montmorillonitic soil filling material. Silicon. 2018;10(6):2809-15. http://doi.org/10.1007/s12633-018-9821-4
  » http://doi.org/10.1007/s12633-018-9821-4
• ¹² Giarola JM, Santos BAF, Gomes JAP, Martelli PB, Bueno AHS. Estudo e proposição de soluções sintéticas de solos para avaliação da corrosividade do aço X65. Materia. 2017;22(4):e11900. http://doi.org/10.1590/s1517-707620170004.0234
  » http://doi.org/10.1590/s1517-707620170004.0234
• ¹³ Monnier J, Bellot‐Gurlet L, Baron D, Neff D, Guillot I, Dillmann P. A methodology for Raman structural quantification imaging and its application to iron indoor atmospheric corrosion products. J Raman Spectrosc. 2011;42(4):773-81. http://doi.org/10.1002/jrs.2765
  » http://doi.org/10.1002/jrs.2765
• ¹⁴ Dubois F, Mendibide C, Pagnier T, Perrard F, Duret C. Raman mapping of corrosion products formed onto spring steels during salt spray experiments: a correlation between the scale composition and the corrosion resistance. Corros Sci. 2008;50(12):3401-9. http://doi.org/10.1016/j.corsci.2008.09.027
  » http://doi.org/10.1016/j.corsci.2008.09.027
• ¹⁵ Zhang X, Xiao K, Dong C, Wu J, Li X, Huang Y. In situ Raman spectroscopy study of corrosion products on the surface of carbon steel in solution containing Cl− and ${\text{SO}}_4^{2-}$ Eng Fail Anal. 2011;18(8):1981-9. http://doi.org/10.1016/j.engfailanal.2011.03.007
  » http://doi.org/10.1016/j.engfailanal.2011.03.007
• ¹⁶ Antunes RA, Ichikawa RU, Martinez LG, Costa I. Characterization of corrosion products on carbon steel exposed to natural weathering and to accelerated corrosion tests. Int J Corros. 2014;2014:1-9. http://doi.org/10.1155/2014/419570
  » http://doi.org/10.1155/2014/419570
• ¹⁷ Borch T, Camper AK, Biederman JA, Butterfield PW, Gerlach R, Amonette JE. Evaluation of characterization techniques for iron pipe corrosion products and iron oxide thin films. J Environ Eng. 2008;134(10):835-44. http://doi.org/10.1061/(ASCE)0733-9372(2008)134:10(835)
  » http://doi.org/10.1061/(ASCE)0733-9372(2008)134:10(835)
• ¹⁸ Hussein AI, Ab-Ghani Z, Che Mat AN, Ab Ghani NA, Husein A, Ab. Rahman I. Synthesis and characterization of spherical calcium carbonate nanoparticles derived from cockle shells. Appl Sci. 2020;10(20):7170. http://doi.org/10.3390/app10207170
  » http://doi.org/10.3390/app10207170
• ¹⁹ Yusoff HM, Hazwani NU, Hassan N, Izwani F. Comparison of sol gel and dehydration magnesium oxide (MgO) as a catalyst in michael addition reaction. Int J Integr Eng. 2015;7(3):43-50.
• ²⁰ Kirk CT. Quantitative analysis of the effect of disorder-induced mode coupling on infrared absorption in silica. Phys Rev B Condens Matter. 1988;38(2):1255-73. http://doi.org/10.1103/PhysRevB.38.1255
  » http://doi.org/10.1103/PhysRevB.38.1255
• ²¹ Djebaili K, Mekhalif Z, Boumaza A, Djelloul A. XPS, FTIR, EDX, and XRD analysis of Al₂O₃ scales grown on PM2000 alloy. J Spectrosc. 2015;2015(1):868109. http://doi.org/10.1155/2015/868109
  » http://doi.org/10.1155/2015/868109
• ²² Yin Y, Yin H, Wu Z, Qi C, Tian H, Zhang W, et al. Characterization of coals and coal ashes with high si content using combined second-derivative infrared spectroscopy and raman spectroscopy. Crystals. 2019;9(10):513. http://doi.org/10.3390/cryst9100513
  » http://doi.org/10.3390/cryst9100513
• ²³ Smidt E, Bohm K, Schwanninger M. The application of FT-IR spectroscopy in waste management. In: Nikolic G, editor. Fourier transforms: new analytical approaches and FTIR strategies. Croatia: InTech; 2011. p. 405-30. http://doi.org/10.5772/15998.

Publication Dates

• Publication in this collection
  12 May 2025
• Date of issue
  2025

History

• Received
  19 Aug 2024
• Reviewed
  03 Feb 2025
• Accepted
  16 Mar 2025