Study on the failure analysis of four metals for anaerobic fermentation reactor based on numerical simulation and electrochemical method

The core problem of low energy consumption anaerobic fermentation reactor is that the reactor can make reasonable and effective use of energy from two aspects of quantity and quality, so as to ensure the anaerobic fermentation performance of the reactor and make it run efficiently and energy-saving. However, serious corrosion of metal for reactors was found during operation, and the selection of reactor materials became the key to restrict biogas production. In this paper, the corrosion characteristics of the four metals including Q235A steel, Q345A steel, 45# steel and 3Cr13 steel were determined by corrosion morphology, mechanical and electrochemical experiments. The results showed that the corrosion product particles of Q235A steel were polygonal, showing a good cross-linking feature, which was better than that of Q345A and 45# steels. However, there was no obvious boundary between corrosion product particles of 3Cr13 steel. The presence of Fe3C in the corrosion products of 3Cr13 steel and the observation of micro-cracks on the surface at nanometer scale indicated that the intergranular corrosion of 3Cr13 steel was dominated. For another aspect, with the increase of CO2 and CH4 content, the corrosion rate of the four metals was generally accelerated, in which the Q235A steel showed the best performance under different conditions. In addition, Q345A and 45# steels showed relatively good corrosion resistance, under which the total mole of mixed gas (CO2/H2O/CH4) was lowest.


Introduction
Due to the advantages of low cost of crop straw raw materials, fast growth and reproduction of anaerobic bacteria, strong ability of anaerobic bacteria to produce gas, low pollution of anaerobic fermentation, simple structure of anaerobic reactor, easy operation of reactor and low operating cost, crop straw anaerobic fermentation technology has been attracted the attention of researchers all over the world [1,2]. In order to increase the concentration of the anaerobic bacteria per unit volume, strengthen organic matter and contact probability of anaerobic bacteria, keep the reactor temperature constant, improve the life of the reactor and the simplicity of operation, a series of highly efficient anaerobic reactor was developed, such as anaerobic filter (AF), continual stir tank reactor (CSTR), anaerobic fluidized bed (AFB), up-flow anaerobic sludge bed (UASB), anaerobic baffled reactor and Integration bioreactor (IBR), among which UASB is the most popular [3].
However, relevant studies had shown that serious corrosion problems occurred during the operation of UASB reactor [4][5][6][7]. This was mainly because a water film with the thickness of 20-30 molecular will be absorbed on the surface of the metal, to form the electrochemical corrosion environment. Under this circumstance, the hydrogen evolution reaction was mainly occurred. At the same time, there were microbial corrosion, gas-liquid alternating corrosion and waterline corrosion. At present, carbon steel and stainless steel were mainly used to make UASB reactor. However, corrosion of carbon steel was inevitable due to the above Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. harsh corrosion environment. Although stainless steel showed good corrosion resistance in CO 2 and microbial environment, micro cracks were still found during the nondestructive testing.
Nowadays, relevant studies focus on corrosion behavior of different metals in the CO 2 and microbial environment, and even how to increase CH 4 production. However, the applicability of different metals in complex CO 2 /H 2 O/CH 4 environment had not been reported. In this paper, the corrosion characteristics of the four metals including Q235A Steel, Q345A steel, 45# steel and 3Cr13 steel were determined by immersion experiment, tensile experiment and electrochemical experiments, which can provide reference for material selection in reactor design of different sizes.

Experimental conditions determined by numerical simulation
Generally speaking, the reactor model did not consider the difference of heat and mass transfer between gas and solid phase, in which the quasi-homogeneous model was commonly used for fixed-bed reactor. Furthermore, there was almost no axial flow in the bed, and the flow was uniform. Therefore, in this paper, a one-dimensional quasi homogeneous model was adopted without considering the axial mass and heat transfer of the reactor.
The model was set as the following descriptions. The central tube entrance of the reactor was provided with an adiabatic reaction zone, and the heat exchange tubes were arranged in layers to maintain the reaction temperature. In the sludge layer, the heat exchange tube layer was the heat exchange section, and the tube layer was the adiabatic section. The pressure drop in the reactor was calculated by Ergun equation. The viscosity of mixed gas can be calculated by using thermodynamic similarity and by comparing the equation of state, and the relative viscosity can be obtained by searching the generalized viscosity diagram or substituting the relative empirical formula. The basic parameters of reactor design were shown in table 1. Figure 1 showed the radial temperature and pressure distribution curve of the reactor. Table 2 showed the comparison of simulated and tested results. As can be seen from the data in the table, the maximum error between the temperature and pressure obtained by the test and the simulation results is less than 1% in the six radial test points, indicating that the simulation results can well reflect the temperature and pressure distribution in the reactor.  In the reactor, the thermal insulation layer was 0.664 m. After entering the latter half of the reactor, due to the appearance of the heat exchanger layer and the thermal insulation layer, the temperature distribution in the reactor continued to rise rapidly and then reached a maximum of 322.5 K at R = 1.5 m, and following decreased. Meanwhile, According to reactor radial pressure distribution can be seen, the pressure dropped faster in R = 0.05-0.6 m. This was because special radial flow in the reactor, in which the bigger the diameter of fluid flowing through the sludge layer was, the wider of flow section was. Figure 2 showed the radial distribution of the sum of CH 4 /CO 2 /H 2 O and gas mole fraction in the reactor. In the reactor, CH 4 was produced mainly through the following two reaction processes, namely CH 3 COOH → CH 4 + CO 2 and CH 3 COONH 4 + H 2 O → CH 4 + NH 4 HCO 3 . H 2 O was used as raw material and CO 2 was the associated product. As can be seen from the figure, with the increase of the reactor radial diameter, the mole fraction of CH 4 and CO 2 increased rapidly, while the mole fraction of H 2 O decreased rapidly at first and then slowly. It can be inferred that in the range of R = 0.05-0.5 m, CH 4 generation was controlled by two reaction processes, while in the range of R = 0.5-2.0 m, CH 4 was mainly generated by CH 3 COOH → CH 4 + CO 2 . At this point, the slope of the mole fraction curves of CO 2 and CH 4 was the same. However, it should be noted that in the process of CH 4 generation, other gases were included, such as N 2 [7]. With the increase of radial radius, the total mole fraction of CH 4 /H 2 O/CO 2 decreased first and then increased, reaching the minimum of 53.8% at R = 0.6 m, indicating that the adverse reaction of N 2 generation increased and the productivity of CH 4 decreased. Subsequently, with the increase of radial radius, the total mole fraction of CH 4 /H 2 O/CO 2 gradually tended to 1, indicating that the main reaction process of CH 4 generation was dominant at this time.
3. Failure analysis of different metals based on immersion experiment and electrochemical experiment 3.1. Experimental setting According to the above results based on numerical simulation, the experimental environment was constructed through the high temperature -pressure reactor, as shown in figure 3, and the different conditions were  presented in table 3. The pressure in the reactor was controlled by N 2 with the purity of 99.99%. During the experiment, mixed gas including H 2 O/CO 2 /CH 4 /N 2 was continuously pumped into the reactor in a certain proportion, and then the gas was discharged through the outlet to ensure the pressure in the reactor stable [8].
Four metals (Q235A/Q345/45#/3Cr13) were chosen to carry out the immersion experiment, mechanical experiment and electrochemical experiment, to determine the performance under the conditions as shown in table 3. It should be noted that the CH 4 generation rate was the highest when R = 2 m, so the condition E in table 3 was significantly paid attention to. The main components of four metals were shown in table 4, and the original strength was shown in table 5.
(1) Immersion experiment. The size of the four samples used for immersion experiment was 25 × 25 × 2 mm 3 . The samples were polished to be mirror-like before the experiment. The samples were taken out after 30 days of the experiment under the condition E as shown in table 3, and the corrosion products on the surface of  the samples were analyzed by SEM and XRD. Furthermore, the metal matrix was observed by AFM after removing the corrosion products.
(2) Tensile experiment. The tested samples were prepared according to the GB/T 228-2002 'Metallic materials-Tensile testing at ambient temperature'. The cylindrical samples were adopted with the working section diameter of 5 mm, the standard distance of 30 mm. Before the tensile experiment, the sample was kept under the condition E for 7 days. The tensile experiment was carried with the strain rate of 1 × 10 −6 s −1 , and then the fracture morphology was observed.
(3) Electrochemical experiment. After the immersion experiments were carried for 30 days, the different samples (working electrode, WE) were taken out to conduct the electrochemical experiment, as shown in figure 4. The experimental solution was NaHCO 3 solution with a mass fraction of 1%, and a certain proportion of mixed H 2 O/CO 2 /CH 4 /N 2 gas was continuously injected into it. The whole experimental device was placed in a constant-temperature incubator. In this paper, the saturated calomel electrode (SCE) was chosen as the reference electrode (RE) and the auxiliary electrode (AE) was Pt electrode. The frequency range of electrochemical impedance spectroscopy (EIS) was 10 mHz-100 kHz [9]. Figure 5 showed the microstructure of corrosion products of different metals under the condition E for 30 days. Compared with the corrosion products of 3Cr13 steel, the corrosion products of Q235A steel, Q345A and 45# steel gathered uniformly on the surface and showed relatively obvious limits. Under the magnification of 5000 times (40 μm), the corrosion product particles of Q235A steel were polygonal with large particle size, and layered and stacked together, showing a good cross-linking feature [10]. The corrosion products of Q345A steel (4.838-13.725 μm) and 45# steel (6.407-13.988 μm) presented smaller particle size and were stacked, but the compactness was less than that of Q235A steel (20.064-25.723 μm). However, focused on 3Cr13 steel, there was no obvious boundary between corrosion product particles, showing layered accumulation, while there was an obvious boundary between gathered corrosion products, which was independent of each other [11].

Microscopic morphology
In order to reveal the reasons leading to the difference of corrosion products of different metals in figure 5, XRD analysis was carried out on the corrosion products of four metals, as shown in figure 6, and the quantitative phase determination of XRD results was shown in table 6. It can be seen that FeCO 3 was the main corrosion product. The corrosion products of Q235A steel, Q345A and 45# steels were composed by FeCO 3 and Fe [12], while the corrosion products of 3Cr13 steel contained additional Fe 3 C [13]. Compared with Q235A steel, the content of Fe in corrosion products of Q345A and 45# steels was relative lower, while the content of FeCO 3 was    higher. Moreover, the content of Fe in corrosion products of 3Cr13 steel was the highest. Compared with Q235A steel, treatment process of Q345A and 45# steels was similar with the mostly equal yield strength of 345 MPa and 355 MPa, respectively [14]. Therefore, less Fe precipitated to be free. For 3Cr13 stainless steel, there was obvious grain boundary between different phases. Under the long-term influence of CO 2 /CH 4 or even free C [15], carburization was easy to occur on the grain boundary. Therefore, there was a large amount of Fe 3 C in the corrosion products of 3Cr13 steel, and the corrosion products accumulated in dispersed areas with a clear boundary between different areas. It can be seen by combining figure 5 and figure 6, the main components of corrosion products of Q235A, Q345A and 45# steel were FeCO 3 , but the corrosion products showed different morphologies. The morphology of the corrosion products of Q345A and 45# steel were similar, showing fine particle stacking of corrosion products, and the particle sizes were 4.838-13.725 μm (Q345A) and 6.407-13.988 μm (45#), respectively. However, the particle size of Q235A steel was significantly larger than that of Q345A and 45# steel, ranging from 20.064 μm to 25.723 μm, and the stack was relatively dense. This was because in terms of steel processing technology, the surface of Q235A steel had been treated with Mo-Cr co-infiltration and supersaturated carburizing, resulting in corrosion products of Q235A steel with better particle size and crosslinking than Q345A and 45# steel. For 3Cr13 steel, due to carburizing during the experiment, a corrosion product structure with Fe3C as the skeleton was formed, so an obvious local stack area of corrosion products occurred, and the interface between corrosion product particles was fuzzy.
In order to further clarify the intergranular corrosion behavior of 3Cr13 steel caused by carburizing process, the corrosion products were removed by standard rust removal solution and then the surface morphology was observed by atomic force microscopy. It can be seen from the figure that under different experimental time conditions, obvious micro-cracks appeared on the surface of 3Cr13 steel. With the increase of time, the shape of the micro-cracks changed from strip to irregular shape, and the depth increased. This showed that the intergranular corrosion of 3Cr13 steel was dominated due to the influence of carburizing under the condition of mixed gas [16]. Figure 8 showed the strain-stress curves of different metals under the condition E in table 3 for 7 days, figure 9 presented the changes of tensile strength and yield strength of different metals before and after the immersion experiment, and figure 10 illustrated the fracture pictures.

Strain-stress analysis
Compared with the original strength of different metals, the yield strength decreased by 0.85% (Q235A steel), 1.45% (Q345A steel), 11.8% (45# steel) and 62.82% (3Cr13 steel), respectively, while the tensile strength decreased by 41.69% (Q235A steel), 44.04% (Q345A steel), 46.33% (45# steel) and 73.02% (3Cr13 steel), respectively. The influence on the yield strength of different metals was far greater than that of the yield strength. At the same time, the yield strength and tensile strength of 3Cr13 steel were reduced much more than that of Q235A, Q345A and 45# steels, especially the yield strength [17][18][19]. On the other hand, under the research conditions in this paper, the greater the yield strength of the metal was, the greater the decrease rate of tensile strength and yield strength was under the condition E in table 3 for 7 days. From the fracture pictures of different metals, it can be seen that Q235A and 3Cr13 steels exhibited brittle fracture, while Q345A and 45# steels presented ductile fracture [20]. This was because among the four metals studied in this paper, the tensile strength and yield strength of Q235A steel were the lowest, indicating that its toughness was lower. Therefore, in the mechanical experiment, as shown in figure 10 in the revised paper, Q235A steel exhibited brittle fracture, while Q345A and 45# steel were ductile behavior. In addition, due to the serious intergranular corrosion of 3Cr13 steel, its tensile strength and yield strength decreased significantly, which was equal to that of Q235A steel, so it also exhibited brittle fracture. Figure 11 showed the EIS curves of different metals under the conditions A-E in table 3 for 30 days. Under different experimental conditions, the Nyquit diagrams of different metals under the conditions A-E showed the capacitive characteristic. Considering that the metal was mainly corroded by CO 2 [21] under the conditions studied in this paper, the equivalent electric circuit was chosen as R s (Q(R p (C dl R ct ))), where R s was the solution     resistance, Q was the constant phase angle element, including capacitor (Q-Y 0 ) and index (n), C dl was the double-layer capacitance, and R ct was the charge transfer resistance. Figure 12 presented the fitting results of EIS curves.

EIS analysis
With the increase of CO 2 and CH 4 content, the surface resistance (R p ) of Q235A and 3Cr13 steels decreased, while the capacitance (Q-Y 0 ) of Q235A steel decreased, indicating that the higher the CO 2 content was, the less protective effect of corrosion product layer was. However, the capacitance (Q-Y 0 ) of 3Cr13 steel increased, which was due to the second phase precipitation and the formation of local corrosion product accumulation area on the surface [22]. Moreover, the peak values of capacitance (Q-Y 0 ) and surface resistance (R p ) of Q345A and 45# steels occurred under the condition C. Compared with other conditions, the corrosion rate of the two metals was low at this time. By comparing the capacitance (Q-Y 0 ) and surface resistance (R p ) of the four metals, it can be seen that under different conditions, the corrosion products of Q235A were the densest and can play a certain protective role [23][24][25]. Focusing on the electrochemical reaction process, the corrosion of Q235A and 3Cr13 steels was aggravated with the increase of CO 2 and CH 4 content, which was manifested by the decrease of charge-transfer resistance (R ct ). The minimum corrosion rate of Q345A and 45# steels appeared under the condition C.
Therefore, in conclusion, with the increase of CO 2 and CH 4 content, the corrosion rate of the four metals was generally accelerated, while the corrosion resistance of 3Cr13 steel was reduced due to carburization, and intergranular corrosion mainly occurred. However, Q345A and 45# steels showed relatively good corrosion resistance under the conditions B and C, under which the total mole of mixed gas (CO 2 /H 2 O/CH 4 ) was lowest [26] corresponding to figure 2. Therefore, in this paper, the Q2354A steel presented the better performance under different conditions as shown in table 3.

Conclusion
In this paper, the internal environmental parameter of the reactor was determined through numerical simulation. Subsequently, the corrosion characteristics of the four metals were determined by corrosion morphology, mechanical and electrochemical experiments. The main conclusions were as following.
(1) According to the actual working conditions, the pressure and temperature at some locations of the reactor were detected. The maximum error between the tested value and the simulated value was less than 1%. Therefore, the model established in this paper can effectively reflect the working conditions of the reactor. Furthermore, the experimental conditions were determined.
The corrosion product particles of Q235A steel were polygonal, showing a good cross-linking feature. The corrosion products of Q345A and 45# steels were stacked, but the compactness was less than that of Q235A steel. However, there was no obvious boundary between corrosion product particles of 3Cr13 steel. The presence of Fe 3 C in the corrosion products of 3Cr13 steel and the observation of micro-cracks on the surface at nanometer scale indicated that the intergranular corrosion of 3Cr13 steel was dominated.
(3) With the increase of CO 2 and CH 4 content, the corrosion rate of the four metals was generally accelerated, while the corrosion resistance of 3Cr13 steel was reduced due to carburization, and intergranular corrosion mainly occurred. However, Q345A and 45# steels showed relatively good corrosion resistance under the conditions B and C, under which the total mole of mixed gas (CO 2 /H 2 O/CH 4 ) was lowest corresponding to figure 2. Therefore, in this paper, the Q2354A steel presented the better performance under different conditions as shown in table 3.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).