Effect of chromium on the corrosion behaviour of low Cr-bearing alloy steel under an extremely high flow rate

The effect of chromium on the corrosion behavior of low Cr-bearing alloy steel in a wet gas pipeline with a high flow rate was studied using a rotating cylinder electrode (RCE) and self-built wet gas flow loop device. The results show that the addition of chromium in the steel can increase the flow accelerated corrosion (FAC) resistance of steel effectively. It was hard for pure FeCO3 to deposit onto the carbon steel surface to form an intact corrosion film when the flow rate or wall shear stress was high. However, a mixture of Cr(OH)3 and FeCO3 can still be deposited onto the 3Cr steel surface and form an intact and protective corrosion film even under conditions with a 212 Pa wall shear stress in the wet gas pipeline.


Introduction
Carbon steel is the most widely used material in the natural gas industry. However, with the development of natural gas extraction, injection, and transportation technology, the natural gas ow rate in some areas has reached a high level. Under the combined action of corrosive substances such as CO 2 , H 2 S, and Cl À , carbon steel is faced with severe ow accelerated corrosion (FAC) problems. [1][2][3][4] FAC is a deterioration action due to the effect of uid ow that leads to the destruction or thinning of pipelines. 5,6 FAC depends on hydrodynamic parameters of the uid ow, such as ow rate and wall shear stress. 7 High ow rate and wall shear stress lead to efficient mixing, which prevents the formation of protective lms on the surface of the pipeline and increases the corrosion rate as a result. 8 FAC has caused a large number of failures in piping and equipment in all types of fossil, industrial steam, and nuclear power plants, and it is a predominant mode of failure of pipelines. 9 As a countermeasure, replacement of the carbon steel with low Cr-bearing alloy steel, which boasts an excellent performance to price ratio and improved CO 2 corrosion resistance, has been suggested because trace amounts of Cr element can effectively prevent FAC. [10][11][12] Jiang et al. 13,14 reported that the addition of Cr generates amorphous products in a rust layer, and Cr is suitable for improving the corrosion resistance of low alloy steel to ow-accelerated corrosion caused by the inhibition of a cathodic reaction in an O 2 -containing environment. A large number of studies have shown that the amorphous product lm will form on the surface of low Cr steel, with fewer pores and better protection. [15][16][17] However, there is still a lack of research on the FAC of Cr-bearing alloy steel in the CO 2 -containing environment. Wang et al. 18 found that pitting corrosion of 1Cr and 3Cr steel may occur under the ow conditions with sand. Stack et al. 19 found that increases in ow velocity resulted in higher current densities for the anodic reaction. Flow rate has a signicant effect on the precipitation kinetics, morphology, and mechanical properties of a FeCO 3 scale on carbon steel. [20][21][22] However, few investigations have been conducted on the effect of ow on the corrosion product performance of low Cr-bearing alloy steel in the CO 2 -containing environment.
Presently, most research on FAC is generally carried out under the liquid phase and focused on the nuclear power plant pipeline. [23][24][25][26][27][28] However, FAC in the natural gas pipeline usually occurs in the wet gas medium, and its corrosion environment is very different from that in the liquid phase. [29][30][31] Traditional laboratory liquid phase corrosion simulation methods may not be able to obtain the actual corrosion resistance and corrosion resistance mechanism of the material in the wet gas medium.
Therefore, in this study, the corrosion resistance and corrosion mechanism of carbon steel and 3Cr steel under high ow rate and high wall shear stress conditions were studied by combining the traditional experimental method and the selfbuilt wet gas ow loop. The results of this research have essential engineering signicance for the extension of 3Cr steel and the control of FAC in the natural gas pipeline.

RCE test
The RCE is an ideal tool for studying turbulent ow. Higher turbulent velocities are easily accessible at higher rotation rates. Fig. 1 shows a schematic diagram of the RCE. All electrochemical measurements were performed in a glass cell with a traditional three-electrode system using a Gamry Refer-ence600+ electrochemical workstation. The volume of the solution used was 1 L. The working electrodes were carbon steel and 3Cr steel rotating cylinders, 12 mm diameter, 8 mm long, and 3 cm 2 electrode surface area. A platinum sheet was used as a counter electrode, and a saturated calomel electrode (SCE) was used as a reference electrode.
The electrochemical impedance spectroscopy (EIS) measurements were performed using the AC signals of the 10 mV peak-to-peak amplitude at the OCP in the frequency range of 100 kHz to 5 mHz.

Flow loop test
A schematic diagram of the ow loop is shown in Fig. 2. The ow loop is a 25/50 mm diameter, high-pressure system. The entire ow loop is manufactured from 316L stainless steel. The ow loop was driven by a circulating fan. The solution was dosed into the ow loop by a continuous dosing device and mixed with the high ow rate CO 2 gas. The coupon was a disc with a diameter of 16 mm and a thickness of 3 mm. Three coupons were installed in the test section. A cooling casing is installed outside the test section to control the temperature of the system. Downstream of this test section, a gas-liquid separator was used to separate the liquids and gas. Aer gasliquid separation, the gas phase continues to circulate in the loop, and the liquid phase is discharged to the outside. The parameters of the ow loop tests are shown in Table 4.
The corrosion rate was measured using the weight-loss method. The morphology and energy dispersive X-ray spectroscopy (EDS) analysis of the corrosion scale were investigated using a JSM-6510A SEM and a JED-2300 EDS. The X-ray photoelectron spectroscopy (XPS) of the corrosion scale was examined using a Thermo Escalab 250Xi instrument.

RCE test
For typical RCE devices, the transition from laminar to turbulent ow occurs when the Reynolds number exceeds 200. Because this transition occurs at a relatively low rotation rate, the RCE is considered an ideal tool for studying turbulent ow at a low velocity, which is a condition frequently observed in pipeline infrastructures. Moreover, higher turbulent velocities are easily achievable at higher rotation rates.  The corrosion of 3Cr steel and carbon steel under different rotation rates was periodically investigated through EIS. EIS measurements can provide insight into the change of corrosion lm characteristics and the corrosion process occurring at the interface. Fig. 3a presents the Nyquist diagrams of carbon steel obtained at 2000 rpm aer immersion for 1, 3, 5, 9, 12, 18, and 24 h, with the characteristics of a capacitance loop and an inductive loop. The impedance increased slowly, which means the product lm grows slowly, and the corrosion rate decreases slowly. A distinct inductive loop in the low frequency was always present. The low-frequency inductive loop is associated with the adsorption of the intermediate product on the substrate surface. It indicated that a corrosion lm was barely formed on the carbon steel surface at a high ow rate. The adsorption and desorption of intermedia products on the steel surface were extremely high. Fig. 3b presents the Nyquist diagrams of 3Cr steel obtained at 1000 rpm aer immersion for 1, 3, 5, 9, 12, 18, and 24 h, with the characteristics of two typical capacitance loops and an inductive loop. The impedance gradually increased, which was attributed to corrosion scale growth on the substrate. Initially, the inductive loop expanded and then shrank. The inductive loop was closely related to the adsorption and desorption of intermediate products. In the initial stage, the corrosion lm coverage was low, and the formation of intermediate products such as FeOH ad and CrOH ad increased with time. The adsorption and desorption of intermediate products on the steel surface became stronger. Subsequently, as corrosion proceeded, the coverage of the corrosion lm increased, and the adsorption and desorption of intermediate products on the steel surface weakened. Fig. 3c presents the Nyquist diagrams of 3Cr steel obtained at 2000 rpm aer 1, 3, 5, 9, 12, 18, and 24 h of immersion, with the characteristics of two typical capacitance loops and an inductive loop. Initially, the impedance increased and then shrank, and the inductive loop gradually increased. In the initial stage, the corrosion products deposited on the steel surface and corrosion lm coverage increased. When corrosion proceeded, the corrosion lm was destroyed because of the high ow rate. Moreover, the formation of intermedia products increased with time, and the adsorption and desorption of intermedia products became stronger. Fig. 4 illustrates equivalent circuits for EIS tting. R s , CPE lm , and R pore denote the electrolyte solution resistance, constant phase element (CPE) used to t the corrosion lm capacitance, and resistance of the pores in the corrosion lm, respectively. Electrochemical processes at the interface are represented by R ct and CPE dl , which denote the charge transfer resistance and electric double-layer capacitance, respectively. Y o and n were the parameters obtained using the constant phase element. R L and L were the inductive resistance and inductance, respectively. Table 5 presents the results. The polarisation resistance (R p ), that is, the sum of R pore and R ct , was inversely proportional to the corrosion rate. For carbon steel or low Cr-bearing alloy steel, the corrosion medium can react with steel through pores present on a protective scale, and dissolved Fe 2+ can exit the pores. Thus, protecting the corrosion scale is closely associated with the R pore . Fig. 5 presents R pore and R p obtained from Table  5.
The R p value of carbon steel at 2000 rpm was considerably low and slowly increased with time, indicating that the corrosion lm may form only negligibly on the steel surface. For 3Cr steel at 2000 rpm, R p and R pore increased rst and then decreased, which was attributed to the formation and destruction of the corrosion lm. The R p of 3Cr steel was considerably higher than that of carbon steel, which revealed that the corrosion resistance of 3Cr steel at a high ow rate was substantially stronger than that of carbon steel. When the rotation speed decreased to 1000 rpm, R p and R pore sharply increased with time. In the initial stage (1-12 h), the value was lower than that at 2000 rpm, which indicated that the high ow rate could accelerate the formation of considerably more protective corrosion lms on the 3Cr steel surface. Aer 12 h, the value at 1000 rpm rapidly exceeded that at 2000 rpm. This nding revealed that the rotation speed of 1000 rpm posed no threat to the corrosion lm. At a high ow rate, 3Cr steel exhibited substantially better ow accelerated corrosion resistance than carbon steel did.
The EIS results indicated that the addition of Cr improved the corrosion resistance of steel, mainly because Cr positively inuences the deposition of a corrosion lm on steel at a high ow rate. Therefore, to further conrm the aforementioned conjugation, SEM and EDS analyses were conducted, and the results are presented in Fig. 6.
For carbon steel, the SEM results revealed no prominent corrosion lm on the steel surface. The EDS results indicated that the O and Fe concentrations were considerably lower and higher, respectively, in carbon steel than in 3Cr steel. This  Fig. 3. Table 5 Values of the elements of the equivalent circuit in Fig. 4 to fit the impedance spectra of Fig. 3 Conditions nding indicated that FeCO 3 rarely deposited on the surface of steel under the condition of the high ow rate. The results are consistent with those presented in Fig. 3a. For 3Cr steel, the surface was covered with the cracked corrosion lm deposited at 1000 and 2000 rpm. The EDS results revealed that the corrosion lm was Cr-enriched, indicating that the main components of the corrosion product lm were Cr(OH) 3 and FeCO 3 . The corrosion lm can provide excellent protection for 3Cr steel at a high ow rate. In the produced lm, the ratio of Cr/Fe at 1000 rpm was slightly higher than that at 2000 rpm, and the oxygen content was slightly higher at 2000 rpm than that at 1000 rpm. This may be because the corrosion rate at 2000 rpm was higher than that at 1000 rpm, which led to the generation of more Fe 2+ . Finally, the content of FeCO 3 deposited on the corrosion lm increased, and the corrosion lm became thicker.
The corrosion lm on the surface was complete without any damage at 1000 rpm, and at 2000 rpm, most of the corrosion lm on the surface was complete, but some areas of the corrosion lm were damaged.
This nding is consistent with the EIS results, that is, when the ow rate was 1000 rpm, the corrosion lm on the surface grew steadily with time and exhibited excellent ow accelerated corrosion resistance, which was not damaged by uids under this working condition.    Table 4.
When the rotation rate increased to 2000 rpm, the severe turbulence damaged the corrosion lm. However, because most areas of the steel surface were covered with the complete corrosion lm, the corrosion resistance of 3Cr steel remained considerably higher than that of carbon steel, which cannot form a continuous corrosion lm.

Flow loop test
The results of the RCE experiment indicated that adding Cr can enhance the corrosion resistance of steel at a high ow rate. Furthermore, to conrm the excellent ow accelerated corrosion resistance of low Cr-containing alloy steel at a high ow rate and its corrosion resistance mechanism, ow loop tests were conducted, which can simulate actual conditions in the wet gas pipeline. Fig. 7 presents the corrosion rate of carbon steel and 3Cr steel measured under different conditions by conducting ow loop tests. The ow accelerated corrosion resistance of 3Cr steel was considerably better than that of carbon steel under wet gas conditions at a high ow rate. Fig. 8 illustrates the macro morphologies of coupons with and without the corrosion lm aer 24 and 72 h ow loop tests under conditions 5 and 6 in Table 4. The corrosion lm on carbon steel coupons exhibited obvious damage along the direction of uid ow, resulting in some matrices being exposed, and the surface of 3Cr steel was covered with a complete corrosion product lm. The result indicated that the corrosion lm formed on the 3Cr steel surface exhibited considerably better ow accelerated corrosion resistance than that formed on the carbon steel surface did. Fig. 9 shows the surface morphology of corrosion lm on the carbon steel and 3Cr steel surface aer 72 h corrosion in the ow loop tests under condition 6. The corrosion lm of carbon steel showed an obvious sign of being damaged by uid erosion. However, the corrosion lm of 3Cr steel was intact, and some cracks formed on the entire surface because of dehydration. 15,32   Table 4.
The EDS results indicated that the corrosion lm of 3Cr steel was rich in Cr. Furthermore, to determine the composition of the corrosion product lm, the corrosion lm was analyzed through XPS; Fig. 10 presents the results. The XPS results revealed that the corrosion lm mainly comprises Cr(OH) 3 and FeCO 3 .
Furthermore, the corrosion lms formed on X65 and 3Cr steel in the ow loop under the condition of the highest wall shear stress were analyzed through SEM and EDS. The carbon steel matrix exhibited a crater-like surface, and no prominent corrosion lm was observed, which indicated that it is difficult for FeCO 3 deposited onto the steel surface to form an intact corrosion lm (Fig. 11a). However, a mixture of Cr(OH) 3 and FeCO 3 could still be deposited on the 3Cr steel surface and form an intact and protective corrosion lm even under the conditions of 212 Pa wall shear stress (Fig. 11b), indicating that Cr(OH) 3 can improve the shear resistance of corrosion lms.

Conclusions
The effect of chromium on the corrosion behaviour of low Crbearing steel under a high ow rate or wall shear stress was studied using RCE and ow loop tests. The main conclusions are as follows: (1) The ow accelerated corrosion resistance of 3Cr steel is considerably better than that of carbon steel under a high ow rate, especially in a wet gas environment. Adding chromium to steel can effectively increase its ow accelerated corrosion resistance.
(2) It is difficult for pure FeCO 3 deposited onto the carbon steel surface to form an intact corrosion lm at a high ow rate. However, a mixture of Cr(OH) 3 and FeCO 3 could still be deposited on the 3Cr steel surface and form an intact and protective corrosion lm even under the conditions of 212 Pa wall shear stress, indicating that Cr(OH) 3 can improve the shear resistance of corrosion lms.

Conflicts of interest
There are no conicts to declare.   Table 4. (a) Carbon steel, (b) 3Cr steel.