Corrosion behavior of 15CrMo steel for water-wall tubes in thermal power plants in the presence of urea and its byproducts

2.1 Materials and solution

The urea used in this study was supplied by Datang Anyang Power Plant and contained a total nitrogen content of ≥46.4%. All of the simulation solutions were prepared using 18.25 MΩ cm ultrapure water. The 15CrMo coupons utilized in this study were 40 × 13 × 2 mm in size, and the alloy compositions in the components studied are given in Table 1.

2.2 Urea solution decomposition experiment

The urea hydrolysis experiment was carried out with 600 mL of urea solution in an autoclave (F1.4-20/400, Kemao Experimental Equipment Co., Ltd, Dalian, China), simulating the operating condition in the thermal power plants. The experimental device was shown in Figure 2. The pH value of the solution in the pipe where the failure occurred (described in the introduction above) was between 10.0 and 11.0. At a pH of 10.6, the corresponding concentration of urea was 280 mg/L based upon Equations (5)–(7) (K_ammonia = 1.79 × 10⁻⁵, 25 °C); thus, 280 mg/L was selected as the typical urea concentration with a decomposition temperature of 320 °C based upon the plant operating conditions. After the temperature reached the set value, the vapor and liquid phase samples in the autoclave were sampled once per hour for 8 h (the decomposition reached more than 99%). According to the results of typical urea concentration tests and the range of pH (10.0–11.0), five urea concentrations, which were 70, 140, 280, 560 and 840 mg/L, were set, and the corresponding pH was 10.29, 10.45, 10.60, 10.76 and 10.85, respectively. In the above tests, nitrogen gas was introduced into the autoclave for removing oxygen and carbon dioxide, controlling the dissolved oxygen concentration was not more than 10 μg/L. After 8 h of decomposition, the vapor and liquid phase samples were taken out from the autoclave. All samples were sealed and stored in a refrigerator at 4 °C, tested with TOC analyzer, and subjected to IR spectroscopy.

Figure 2:

High-temperature decomposition of urea experimental device diagram.

2.3 Study of decomposition samples

The Fourier transform infrared (FT-IR) spectra of urea solution before and after decomposition were recorded using a Nicolet Avatar 360 FT-IR spectrometer. The UPLC-ESI-MS analyses of samples were performed on a Xevo Triple Quadrupole (TQ) system (Agilent, USA).

2.4 Electrochemical measurements

A conventional three-electrode system was used to evaluate the electrochemical properties of 15CrMo steel in urea and urea decomposition residual solutions with CHI660D electrochemical workstation (Chenhua Instrument Co., Ltd. Shanghai, China). It consisted of 15CrMo steel working electrode (WE) with an exposed area of 1 cm², a platinum counter electrode and a saturated calomel reference electrode. Before the electrochemical tests, the WE was polished with 1200 grit silicon carbide papers, rinsed with deionised water, degreased in alcohol and dried under vacuum. Samples were exposed to the experimental solutions for 45 min, and the open-circuit potentials were monitored for 10 min. No substantial variation (within 10 mV per minute) of the open circuit potential confirmed the attainment of steady state.

Potentiodynamic polarization scans started at −250 mV relative to open circuit and ended at +250 mV relative to open circuit were obtained by sweeping the applied potential on the WE at constant sweep rate of 1 mV/s (Rochdi et al. 2015; Zhang et al. 2011). The corrosion current density was determined using a linear polarization method, and the corrosion rate was calculated by the following equation (Song et al. 2014):

where I_corr is the corrosion current density (A/cm²); K is combination of several conversion terms (3.2680 × 10³, for mm/y); ε is the equivalent weight (27.92 g/equivalent) and ρ is the steel density (7.85 g/cm³).

Electrochemical impedance spectroscopic measurements were performed in the frequency range 100 kHz–10 mHz with AC amplitude of ±10 mV (RMS, root mean square) at the rest potential (Gama-Ferrer et al. 2012; Fernández-Domene et al. 2014). The results of impedance measurements were fitted by ZSimpWin™ software.

The electrochemical experiment was divided into two parts, where the first part is the electrochemical experiment of 15CrMo in the urea solution with different concentrations, and the second part is in the urea decomposition residue with different concentrations at 320 °C. The decomposition residue was taken for the electrochemical experiment. All the above tests were performed under unstirred conditions at room temperature of around 25 °C.

2.5 Weight loss measurements and surface analysis

The operating conditions of the power plant were simulated, and corrosion experiments were conducted in the autoclave. Weight loss measurement was used to assess the corrosion rate as a function of urea concentration. The surfaces of 15CrMo were first ground with 1200 grit silicon carbide papers, rinsed with deionized water, degreased in alcohol and dried under vacuum (Dutta et al. 2017). The concentrations of the urea solutions were 70, 140, 280, 560 and 840 mg/L. After acquiring their initial weights, the metal test pieces were suspended and soaked in 600 mL of urea solution in the autoclave, deoxygenated with nitrogen and heated at 320 °C for 8 h. After cooling to room temperature, metallic coupons were removed, scrubbed with a stiff bristle brush under running water to dislodge loosely adherent corrosion products, and soaked in cleaning solution to remove the remaining corrosion product. The preparation of the cleaning solution was as follows: 500 mL of hydrochloric acid (ρ = 1.19 g/mL) and 3.5 g of hexamethylenetetramine were mixed in distilled water to achieve a 1000 mL solution. The test pieces were cleaned repeatedly at room temperature. After the complete removal of the corrosion products, the samples were cleaned using water and acetone, dried under vacuum and reweighed [flowing Corrosion of metals and alloys–Removal of corrosion products from corrosion test specimens (Standards China 2015)]. Corrosion rate (mm/y), v, was calculated using the following relation (Wu et al. 2018):

where m₀ and m₁ are the weights of the metal coupons before and after weight loss tests in the urea solutions (g); S is the surface area of each metal specimen in cm²; D is the steel density (g/cm³) and T is the exposure time of the specimens in the autoclave (h).

Morphological analysis was performed to obtain an insight into the changes in the surface of the corroded samples caused by the increase in the initial concentration of urea decomposition. The elemental compositions of the samples were determined through scanning electron microscopy–energy-dispersive X-ray spectroscopy (SEM-EDS, FEI QUANTA 200). X-ray diffraction (XRD, D/max-2200/PC) measurements were performed to identify the mineralogical phases. The range of 2θ values was 10°–80° with a 0.05° step size. The scanning speed was 0.03°/s. The XRD patterns were identified using Jade + ^TM (Version 6.5), and a semiquantitative analysis of the compound was performed using reference intensity.

3.1 Determination of the decomposition samples of urea solution

Figure 3 illustrates the total organic carbon (TOC) contents in the vapor and liquid phases as a function of initial urea concentration after decomposition of urea at 320 °C for 8 h. The figure shows that the TOC contents in the vapor and liquid phases after a fixed decomposition time of 8 h increase with an increase in the initial concentration of urea. The TOC content in the vapor phase is considerably larger than the TOC content in the liquid phase, indicating that the partition coefficient of TOC in the vapor phase during urea decomposition is greater than that in the liquid phase.

Figure 3:

Contents of TOC in the vapor/liquid phase after decomposition of 280 mg/L urea at 320 °C.

Figure 4 illustrates the TOC contents of the vapor and liquid phase samples for an initial urea concentration of 280 mg/L as a function of decomposition time at 320 °C. The figure shows that the TOC in the vapor and liquid phases at typical decomposition concentration decreases with decomposition time, suggesting that the organic carbon contained in urea is converted into inorganic carbon (CO₂) during the decomposition of urea. At the same time, the TOC contents in the vapor and liquid phases fluctuate in a downward trend because the NH₃ and CO₂ produced by urea decomposition can be converted into organic derivatives, such as ammonium carbamate, under high temperature and high pressure. As shown in Figure 4, the TOC content in the vapor phase is higher than that of the liquid phase although the TOC contents in the vapor phase and liquid phase decrease during the hydrolysis of urea, revealing that the partition coefficient of TOC in the vapor phase is greater than that in the liquid phase through the course of decomposition.

Figure 4:

Contents of TOC in the vapor/liquid phase after decomposition of urea at 320 °C for 8 h.

Figure 5 shows the infrared spectrum of the 280 mg/L urea solution before and after decomposition at 320 °C for 8 h. The figure shows that the samples before and after decomposition have an absorption peak of –NH₂ group near the wavenumber of 3440 cm⁻¹, and N–H stretching vibration absorption peaks are observed near the wavenumber of 1600 cm⁻¹. C–N stretching vibration absorption signals are found near 1124 and 1452 cm⁻¹. The sample after urea decomposition exhibits a COO⁻ absorbing vibration peak at the wavenumber of 620 cm⁻¹ (Bacsik et al. 2011; Bacsik and Hedin 2016; Frasco 1964). The ultraperformance liquid chromatography-electrospray ionization tandem mass spectrometry (UPLC-ESI-MS) spectrum of 280 mg/L urea solution after decomposition at 320 °C for 8 h is shown in Figure 6. As shown in Figure 6, a significant peak appears at the mass-to-charge ratio of 79. Thus, the sample contains CH₆N₂O₂. Combined with the results of Fourier transform infrared spectroscopy and UPLC-ESI-MS, the sample after urea decomposition contains ammonium carbamate (H₂NCOONH₄) (Equation (10)).

Figure 5:

Infrared spectrum of 280 mg/L urea solution before and after decomposition at 320 °C for 8 h.

Figure 6:

UPLC-ESI-MS spectrum of 280 mg/L urea solution after decomposition at 320 °C for 8 h.

The following intermediate reaction occurs during the decomposition of urea to produce NH₂COO⁻ (Equation (11)) (Prabhu-Gaunkar and Raman 1998; Qin et al. 2011; Watanabe et al. 2009):

3.2 Electrochemical characteristics of 15CrMo in urea decomposition residues

The polarization curve of 15CrMo in the urea decomposition residue is shown in Figure 7 and summarized in Table 2. Figure 7 shows that the self-corrosion potential (E_corr) of the polarization curve negatively and gradually shifts, and i_corr increases with the increase in the initial concentration of urea at ambient temperature. The corrosion rate of 15CrMo gradually increases simultaneously with the increase in corrosive substances in the decomposition residue. Figure 8 and Table 3 present the polarization curves and corresponding parameters of 15CrMo in undecomposed urea solution, respectively. The graph shows that the self-corrosion potential of 15CrMo positively and gradually shifts, and the corrosion current density decreases with the increase in urea concentration when urea solutions are not decomposed. As the same time, the corrosion rate also decreases with increasing urea concentration. It is obvious that undecomposed urea has a certain effect on corrosion inhibition. This finding is consistent with the reports in related literature studies (Çömlekçi and Ulutan 2018; Hsu et al. 2018; Ramya et al. 2016), and the corrosion of 15CrMo is caused by the decomposition products of urea.

Figure 7:

Potentiodynamic polarization curves for 15CrMo steel in various concentrations of decomposition residue of urea solution at 320 °C.

Figure 8:

Potentiodynamic polarization curves for 15CrMo steel in various concentrations of ambient urea solution.

The Nyquist diagrams of 15CrMo steel in the urea decomposition residues are shown in Figure 9. And Figure 9 (a) and (b) represents Nyquist curves and Bode plots, respectively. According to the characteristics of the impedance spectrum, the equivalent circuit shown in Figure 10 was selected (Roy et al. 2014a,b), where in R_s represents the solution resistance; R_c and C_c represent the resistance and capacitance of the corrosion product film, respectively; R_ct is the charge transfer resistance; CPE is a nonideal capacitance which represents the interface electric double-layer capacitor; A CPE’s impedance is given by:

where Q is the pre-exponential factor of the CPE; ω is the angular frequency; and n is the exponent, a measure of the surface irregularity of the electrode. When n = 1, 0, −1, the CPE corresponds to the capacitance (C), resistance (R) and inductance (L), respectively.

Figure 9:

Electrochemical impedance curves for 15CrMo steel decomposition residue of urea solution exemplified as: (a) Nyquist, (b) Bode plot at 320 °C.

Figure 10:

Equivalent circuit model used to fit the impedance spectra of 15CrMo steel in 320 °C decomposition residue of urea solution.

The parameters of each electrochemical element are listed in Table 4. The circuit has good fitness through Zsimpwin™ software fitting.

Figure 9 shows that the shapes of the impedance spectra obtained by the 15CrMo steel in the residual solutions after the decomposition tests of urea with different concentrations remain basically the same, indicating that the corrosion mechanism remains constant with the increase in urea concentration (Mahalik et al. 2010). The arcs on the real axis in the Nyquist plots are incomplete semicircles because of the roughness and nonuniformity of the 15CrMo steel surface. The impedance spectra consist of a high-frequency region capacitive reactance arc reflecting the corrosion product film information of the surface of the 15CrMo steel electrode and another large capacitive reactance arc reflecting the low-frequency region of the metal/solution interface electrochemical reaction information. The size of the semicircular diameter of the capacitive reactance arc reflects the magnitude of charge transfer resistance. The smaller the diameter of the semicircle is, the smaller the charge transfer resistance will be, indicating that the dissolution of 15CrMo steel is likely to occur, and the dissolution rate of the passivation film on the surface is accelerated. The Nyquist plots and the fitting parameters show that the corresponding capacitive antiarc radius, R_ct, decreases with the increase in the initial concentration of urea, indicating that the urea decomposition products accelerate the corrosion of 15CrMo steel. The greater the initial concentration of urea is, the higher the degree of metal corrosion will be. These phenomena reveal that elevated concentrations of corrosive decomposition products promote the charge transfer reaction at the metal/solution interface (Otani et al. 2017).

3.3 Weight loss measurements and analysis

Figure 11 shows a graph of the corrosion rate trend of 15CrMo steel with initial urea concentration calculated from the weight loss data. The surface images of 15CrMo samples show that the corrosion form is localised corrosion, and the calculation result of the corrosion rate is the average value. With the increase in urea concentration, the corrosion rate of 15CrMo steels increases, which is consistent with the conclusion obtained in the electrochemical measurements. The corrosion rate of 15CrMo steel in the high-temperature corrosion test reaches 0.686 mm/y when the urea decomposition concentration is 840 mg/L.

Figure 11:

Variation of corrosion rate of 15CrMo in urea solution at 320 °C by weight loss measurement.

Figure 12 shows the metallographic diagrams of the surfaces of 15CrMo test specimens before and after the high-temperature corrosion tests. The surface of the test piece before corrosion is only the slight scratches left after the grinding of metallographic sandpaper. After soaking and corroding in the urea solutions, the corrosion products adhere to the surface of 15CrMo steel, and the corrosion of metal test pieces becomes increasingly serious with the increase in urea concentration.

Figure 12:

Surface images of 15CrMo samples before (a) and after (b–f) corrosion in urea solution at 320 °C.

3.4 Surface microscopic analysis

Figure 13 shows the SEM images and EDS result of 15CrMo steel before immersion in urea solution, and Figures 14 and 15 show the SEM and EDS results of the surfaces of the test specimens after high-temperature corrosion of 15CrMo steel in 560 and 840 mg/L urea solutions for 8 h. Figure 16 presents the XRD analysis of 15CrMo steel after high-temperature corrosion for 8 h in 840 mg/L urea solution. The SEM images in Figures 14(a) and 15(a) show that serious corrosion occurs on the surfaces of 15CrMo steels, and adhesion deposits of corrosion products are present after 8 h of high-temperature corrosion. The EDS analysis results in Figures 14(b) and 15(b) show that the corrosion products on the surfaces of 15CrMo steel test pieces are mainly composed of three elements, namely, C, O and Fe. The mass fraction of the three elements is more than 98%, and the atomic fraction is more than 99%. Combined with the XRD pattern in Figure 16, FeCO₃ is mainly formed on the surface of 15CrMo. The XRD patterns shows that the surface of the test specimen after corrosion contains iron oxides, such as FeOOH, Fe₂O₃ and Fe₃O₄.

Figure 13:

SEM images (a) and EDS result (b) of 15CrMo steel before immersion in urea solution.

Figure 14:

SEM images (a) and EDS result (b) of 15CrMo steel after immersion in 560 mg/L urea solution at 320 °C for 8 h.

Figure 15:

SEM images (a) and EDS result (b) of 15CrMo steel after immersion in 840 mg/L urea solution at 320 °C for 8 h.

Figure 16:

XRD spectra of 15CrMo after immersion in 840 mg/L urea solution at 320 °C for 8 h.

3.5 Corrosion mechanism analysis of the failure

NH₂COO⁻ can be preferentially adsorbed on the surface of the passivating film and replace oxygen in Fe₂O₃. Subsequently, soluble products were generated by the combination of NH₂COO⁻ with the cations in the passivating film, resulting in the formation of small pits on the exposed base metal. The metal surface inside pitting holes was in an active state, and its electric potentials were negative. The metal surface outside the pitting holes was in a passive state, and its potential was positive. Thus, microcells were established between the active surface inside and the passive surface outside. The corrosion rate of 15CrMo remarkably increased because of the high anode current density caused by the large cathode–anode area ratio in these microcells. However, the metal surface outside the pitting holes were stabilized by cathodic protection to continuously maintain a passive state. The presence of soluble products resulting from the combination of active anions in the solution and cations in the passivation film was an essential prerequisite for the formation of pitting corrosion. In the present case, urea was heated to decomposition, and hydrated oxides enriched in Fe formed during the exposure, followed by the production of Fe₂O₃ in the subsequent stage of hydrothermal reaction. The alkaline environment produced by urea hydrolysis provided favorable conditions for hydrothermal precipitation. Ammonium carbamate, formed as an intermediate during urea hydrolysis, reacted with the dissolved Fe to form complexes (Krogul and Litwinienko 2015), followed by decomposition and deposition on the sample surface. As the corrosion progressed, the corrosion forms changed from pitting to overall corrosion, thereby forming a corrosion profile with pitting characteristics.

Under the action of NH₂COO⁻, the corrosion products peeled off and migrated to where local flow was blocked and formed high deposits. The electrochemical stress corrosion cracking, caused by the original defects in the water-wall tubes under these high depositions, was the direct cause of the accident that occurred in the coal-fired power plant (Bagchi et al. 2020; Rao et al. 2016).

3.6 Prevention method of corrosion

The local design defects of the denitration system and operational errors are the main reason where the denitrification urea solution counter currently invades the thermal system. On the one hand, the improvement of the system can avoid the leakage of urea solution. For example, the compressed air pressure of the spray gun is ensured to be approximately 0.38 MPa by replacing and changing the opening angle of the spray medium of the nozzle. On the other hand, the material selection of the water-cooled wall pipe is improved. In the urea industry, ultralow carbon stainless steel has a special limit on the minimum content of Cr, Ni and Mo to enable it to obtain a fully austenitic structure with good resistance. Properly increasing the content of Nb and Mo in ferritic stainless steel can significantly improve the performance of high-temperature urea corrosion resistance (Wang et al. 2020).