Engineering Failure Analysis 36 (2014) 372–378
Contents lists available at ScienceDirect
Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal
Corrosion failure analysis of L485 natural gas pipeline in CO2 environment Lihong Shi a, Changquan Wang b, Changjun Zou a,⇑ a b
College of Chemistry and Chemical Engineering, Southwest Petroleum University, Chengdu 610500, PR China College of Petroleum Engineering, Yangtze University, Jingzhou 434023, PR China
a r t i c l e
i n f o
Article history: Received 8 August 2013 Received in revised form 6 November 2013 Accepted 19 November 2013 Available online 27 November 2013 Keywords: Carbon dioxide Natural gas Corrosion Pipeline Failure analysis
a b s t r a c t The corrosion failure analysis of Qinhuangdao–Shenyang L485 natural gas pipeline in CO2 environment has been investigated in the present study. The morphology and composition of the corrosion scale are characterized by scanning electron microscopy, X-ray powder diffraction and energy-dispersive spectroscopy. The weight loss experiment is conducted to estimate the actual corrosion rate of L485 pipeline steel. A conceptual model is developed to illustrate the corrosion behavior by electrochemical test of potentiodynamic polarization curve. Results show that corrosion scale mainly consists of FeCO3, and the corrosion rate is in the range of 1.413–1.978 mm/a. The manganese content of L485 pipeline steel exceeds the standard compared with API Spec 5L. In addition, the corrosion behavior of L485 pipeline steel can undergo the periods of bare metal, scale-covered and reaction destruction. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction As a source of clean energy, natural gas has been applied extensively for heating, cooking and fueling [1]. The demand of natural gas is expected to increase to 53% of the energy consumption structure by 2020 [2]. Over the past decade, China had launched the West–East Gas Transmission and Sichuan–East Gas Pipeline Projects to meet the needs of natural gas, and their pipeline lengths were 4200 km and 1700 km, respectively. However, the impurities especially acidic gases such as hydrogen sulfide (H2S) and carbon dioxide (CO2) in the natural gas stream, present a vital problem on the corrosion of the pipeline [3,4]. The length of Chinese Qinhuangdao–Shenyang gas pipeline is 60 km, and the acidic gas in the pipeline transportation only contains CO2. Although the pipeline was well designed by using the corrosion-resistant L485 steel material, now there still exist some serious corrosion problems by CO2 corrosion (Fig. 1). Particularly, localized corrosions such as pitting corrosion and mesa attack occur. The CO2 corrosion makes the pipeline steel failure prematurely, and leads to great economic loss [5]. In practice, the CO2 corrosion behavior of the L485 steel is dominated by the precise environmental conditions such as temperature, CO2 partial pressure (Pco2) and corrosion scale etc. In the present work, the corrosion failure analysis of L485 low-alloy steel that was used in the Qinhuangdao–Shenyang transmission pipelines has been investigated through weight loss experiments and electrochemical tests. The microstructure and composition of the corrosion scale were investigated by scanning electron microscope (SEM), X-ray power diffraction (XRD) and energy-dispersive spectroscopy (EDS). Additionally, the corrosion behavior of the L485 steel immersed in the static simulated condition saturated with CO2 was also described by potentiodynamic polarization curve.
⇑ Corresponding author. Address: School of Chemistry and Chemical Engineering, Southwest Petroleum University, No. 8 Xindu Road, Chengdu 610500, PR China. Tel.: +86 02883037327; fax: +86 02883037305. E-mail address:
[email protected] (C. Zou). 1350-6307/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.engfailanal.2013.11.009
373
L. Shi et al. / Engineering Failure Analysis 36 (2014) 372–378
Fig. 1. Corrosion macro-morphology of L485 pipeline steel.
2. Materials and methods 2.1. Background of failure The material of Qinhuangdao–Shenyang gas pipeline was L485 low-alloy steel, and the pipe specification was D914 19.1. Natural gas was the transported medium in the pipeline, and the chemical composition of the natural gas was listed in Table 1. Based on the data of natural gas component mined each year, the CO2 partial pressure was calculated in the range of 0.2346–0.3765 MPa. The temperature of the natural gas in Qinhuangdao–Shenyang gas pipeline was maintained at 55 °C. 2.2. Characterization The chemical composition of the pipe was analyzed using a direct reading spectrometer (ARL 4460, Beijing) on the basis of ASTM A751-2008. To identify the chemical composition of the corrosion scale, EDS (Oxford Inca 300, England) was employed at 10 kV accelerating voltage. XRD was carried out by an X-ray diffractometer (DX-2000, Dadong) using Cu Ka X-ray at 40 kV and 40 mA, and the XRD diffractogram was recorded in the angle range between 20° and 70° at a step size of 0.05°. The surface morphology of the corrosion scale was observed with SEM (PHLIPS-XL30, Holland). Before EDS, XRD and SEM characterizations, the specimen was cleaned in acetone to eliminate the surface contamination. The electrochemical measurement of potentiodynamic polarization curve was conducted in a conventional glass cell using a CH1604D electrochemical workstation (Shanghai Chenhua Instrument Co., Ltd., China). 2.3. Weight loss experiment The rectangular test specimen with dimensions of 50 mm 10 mm 3 mm was made of the L485 pipeline steel. The surface of the specimen was first polished with silicon carbide papers progressively up to 1200 grit, next degreased with acetone and rinsed with absolute alcohol, then dried with cold air, and finally weighed to the precision of 0.0001 g [6]. Static corrosion tests were conducted in a 10L autoclave at some a pressure to investigate corrosion rate in CO2-saturated formation water. The test system was deoxygenated by pure nitrogen for 1 h, and then it was set to the required partial pressure and temperature. After being corroded for 10 days without a stir, all specimens were taken out and immediately cleaned by deionized water and absolute alcohol, then dried in air. After that, they were weighed again to obtain the final weight. The weight loss of the specimen was determined, and the average corrosion rate was calculated by the following equation: [6,7]
V ¼ 8:76 104 Dm=ðS d TÞ
ð1Þ 2
where V is the corrosion rate, mm/a; Dm is the weight loss, g; S is the surface area of specimen, cm ; d is the density of L485 pipe steel, g/cm3; T is the immersion time, h. 2.4. Electrochemical test Electrochemical polarization tests were conducted to confirm the corrosion behavior of L485 pipeline steel immersed in static simulated condition saturated with CO2. The corrosion cell was a three-electrode electrochemical cell assembly with a saturated calomel reference electrode, a platinum counter electrode and a cylindrical working electrode made of L485 steel with a surface area of 1 cm2. The corrosion medium was CO2-saturated formation water. After the electrode exposed to the
Table 1 Chemical composition (mol%) of natural gas in gas pipeline. Component
C1
C2
C3
C4
Cþ 4
N2
CO2
He
H2O
Content, mol%
89.09
3.78
0.64
0.49
0.20
0.93
2.27
0.02
2.58
374
L. Shi et al. / Engineering Failure Analysis 36 (2014) 372–378
corrosion medium for a period of time at the desired temperature, then the polarization measurements were carried out over a potential automatically from +250 mV to 250 mV at stable open circuit potential with a scan rate of 0.5 mV/s, and the results were recorded and analyzed using the ZView software. Prior to the electrochemical tests, the pre-treatment of working electrode was the same as that of weight loss experiment. 3. Results and discussion 3.1. Chemical composition of L485 pipeline steel The chemical composition of the L485 natural gas pipeline steel was presented in Table 2, which was compared with the parameter requirements of API Spec 5L standard. The results demonstrated that the content of C, S and P of the L485 steel pipe all met technical requirements of API Spec 5L, while the content of manganese (Mn) exceeded standard. This would increase the brittleness of L485 pipeline steel, and once it was subjected to the tensile stress, the pipeline steel was prone to stress corrosion cracking [8]. 3.2. XRD and EDS analysis of corrosion scale Fig. 2 shows the XRD spectrum of the corrosion scale on L485 pipeline steel. It can be seen that the characteristic diffraction peaks of FeCO3 appeared at the 2h of 24.3°, 31.7°, 37.9°, 41.9°, 51.9° and 64.9° [9]. In additional, the bands at 44.3° and 45.6° were assigned to the diffraction peak of Fe and Fe3C. The Fe and Fe3C were the original phase from the matrix, and they were wrapped in the corrosion scale when the matrix was being corroded. Apparently, there was not any evidence exhibiting the characteristic diffraction peak of iron oxide such as FeO or Fe3O4 in the spectrum, which revealed that the corrosion scale was mainly composed of FeCO3. In order to further clarify the composition of the corrosion scale, EDS depth profiles were acquired, and the results were presented in Fig. 3. It could be found that the element such as Fe, C, O, Mn, Si, and Ni appeared in the survey spectrum of the corrosion scale surface, and the atomic ratio of Fe, C and O was approximately 1:1:3, indicating that the composition of the corrosion scale was FeCO3. The evidence was in accordance with that previously revealed by XRD. Localized corrosion usually occurred once the deposited FeCO3 was damaged locally, and it was the principal cause resulting in the failure of steel pipelines. Thus, it can be concluded that the corrosion failure of the L485 pipeline was attributed to the formation of FeCO3 leading to the thinning and localized corrosion perforation of the pipeline steel. 3.3. Corrosion mechanism CO2 corrosion was reported as early as 1940s. The dry CO2 cannot corrode steel, but the carbon steel and low alloy steel can be significantly eroded in a wet CO2 environment. The CO2 corrosion of the L485 pipeline steel occurred in oxygen-free condition, which was the electrochemical corrosion, and it was mainly controlled by the kinetics of hydrogen evolution. The
Table 2 Chemical composition (wt.%) of L485 pipeline steel. Element
C
Mn
P
S
Si
Ni
Ti
V
L485 API Spec 5L
0.16 60.28
1.7 61.4
0.025 60.030
0.02 60.030
0.45 –
0.06 –
0.06 –
0.1 –
Fig. 2. XRD spectrum of the corrosion scale on L485 pipeline steel.
L. Shi et al. / Engineering Failure Analysis 36 (2014) 372–378
375
Fig. 3. EDS analysis of the corrosion scale on L485 pipeline steel.
schematic diagram of the CO2 corrosion on the L485 pipeline steel surface was illustrated in Fig. 4. In the wet CO2 environment, the steel surface adsorbed water vapor to form a thin water film. Then, an electrolyte solution was formed due to the CO2 dissolved in the water film. Thus, the numerous tiny primary cells was generated, in which the iron was the cathode; the carbon or iron carbide was the anode; and the acidic water film was acted as the electrolyte solution. The corrosion mechanism can be described by the cathodic and anodic reactions [10]. At the anode, the dissolution of CO2 in water produced carbonic acid and released the hydronium ion. The hydronium ion is a kind of strong depolarization ion, which was easy to be reduced by capturing electrons. The reduction procedure promoted the dissolution of Fe at cathode leading to the aggravated corrosion, while the hydrogen atoms penetrated into the steel resulting in the cracking of the metal components. The anodic reactions were summarized into five sub-reactions Eqs. (2)–(7), where the subscript ‘‘sol’’ and ‘‘ad’’ refer to the solution state and adsorption state, respectively.
CO2ðsolÞ ! CO2ðadÞ
ð2Þ
CO2ðadÞ þ H2 O ! H2 CO3ðadÞ
ð3Þ
H2 CO3ðadÞ þ H2 O ! H3 Oþ þ HCO3
ð4Þ
HCO3 þ H2 O ! H3 Oþ þ CO2 3
ð5Þ
H3 Oþ þ e ! HðadÞ þ H2 O
ð6Þ
HðadÞ þ HðadÞ ! H2
ð7Þ
The electrochemical reaction at the cathode was the iron dissolution which can be supported by Eq. (8). Based on the previous results such as EDS and XRD characterization, it has been confirmed that the corrosion scale is FeCO3. So the FeCO3 precipitation can be described by following Eqs. (9)–(11).
Fe ! Fe2þ þ 2e
ð8Þ
Fig. 4. Schematic diagram of the CO2 corrosion on L485 pipeline steel.
376
L. Shi et al. / Engineering Failure Analysis 36 (2014) 372–378
Fe2þ þ 2HCO3 ! FeðHCO3 Þ2
ð9Þ
FeðHCO3 Þ2 ! FeCO3 þ CO2 þ H2 O
ð10Þ
Fe2þ þ CO2 3 ! FeCO3
ð11Þ 2+
And the overall reaction for the CO2 corrosion could be explained by Eq. (12). When the concentrations of Fe and CO2 exceeded the solubility of FeCO3, the solid FeCO3 precipitated and deposited onto the steel surface. Then a porous 3 and incompact FeCO3 corrosion scale formed, and finally, it caused the localized corrosion reflecting in the gradual thinning or pitting corrosion perforation of the L485 pipeline steel.
Fe þ CO2 þ H2 O ! FeCO3 þ H2
ð12Þ
3.4. Weight loss experiment 3.4.1. Influence of CO2 partial pressure The corrosion rate of the L485 pipeline steel was affected by Pco2. A series of weight loss experiments was carried out under the various Pco2 at the temperature of 55 °C, and the result was presented in Fig. 5. As shown in Fig. 5, the corrosion rate increased gradually with the enhancement of Pco2. And the relationship between the corrosion rate and Pco2 exhibited the linear equation that was y = 0.4799 + 3.978 with the R2 of 0.9927. Thus, it can be deduced that the corrosion rate of actual L485 natural gas pipeline steel was in the range of 1.413–1.978 mm/a according to the in situ Pco2 of 0.2346–0.3765 MPa. The variation of corrosion rate depended on whether the solubility of corrosion scale FeCO3 was exceeded [7]. The increase of Pco2 enhanced the solubility of CO2 resulting in the low pH, which accelerated the corrosion and facilitated the dissolution of FeCO3 [11]. On the other hand, the high Pco2 promoted the ionization of carbonic acid (Eqs. (4) and (5)) resulting in the increase of H+, and finally, the depolarization of H+ caused the acceleration of corrosion. 3.4.2. SEM analysis of corrosion morphology To figure out the adsorption state of the corrosion scale on the L485 pipeline steel, Fig. 6 shows SEM surface morphologies of corrosion scales on the L485 pipeline steel at different temperatures under the 0.3 MPa Pco2. Generally, the corrosion scale at low temperature was easily exfoliated from the matrix that would expose the fresh steel surface to the corrosive medium [6]. As shown in Fig. 6a, It was found that uniform corrosion and obvious acerose characters with poor adhesion occurred on the surface. With the temperature increasing to 55 °C (Fig. 6b), the electrochemical reactions at both the cathode and the anode were speeded up. The FeCO3 crystal grains grew bulkily and irregularly. The relatively big and loose scale had an ion-passing property, leading to an obvious localized corrosion, especially the pitting corrosion on the surface. And some gaps between the grains apparently occurred (the upper-right illustration of Fig. 6b). The gaps acted as passages of the corrosive ions diffusing into the matrix surface, which resulted in the pitting corrosions. Moreover, the galvanic corrosion with strong self-catalytic effect could be generated among the heterogeneous corrosion regions, which was more beneficial for the formation of localized corrosion. With the increasing temperature, the pH rose due to the less dissolution of CO2, which decelerated the formation of the corrosion scale. As shown in Fig. 6c, the corrosion scale produced at 90 °C was very compact and had good adherence to the matrix. The fine and dense scale could impede the corrosive medium through the corrosion scale, providing a good protection for the matrix. All these indicate that the corrosion failure of L485 natural gas pipeline was due to the corrosion scale formed on the surface was loose and poor protection.
Fig. 5. Influence of CO2 partial pressure on the corrosion rate at 55 °C.
L. Shi et al. / Engineering Failure Analysis 36 (2014) 372–378
377
Fig. 6. SEM surface morphologies of corrosion scales on L485 pipeline steel, (a) 30 °C; (b) 55 °C; and (c) 90 °C.
Fig. 7. Polarization curves of L485 pipeline steel at various test durations in CO2 saturated solution.
3.5. Potentiodynamic polarization measurement To study the electrochemical corrosion behavior of the L485 pipeline steel, potentiodynamic polarization measurements were conducted in CO2 corrosion conditions. Fig. 7 shows the relationship between corrosion potential (EvsSEC) and corrosion current density (I). It can be seen that the polarization curves varied greatly over immersion time. That was to say, the corrosion behavior of the L485 pipeline steel was controlled by the test time. When the immersion time was 1d, the corrosion current density of the L485 pipeline steel specimen was the most negative. With increasing immersion time, the corrosion current density showed varying degrees of ‘‘positive shift’’. During the immersion time of 1–3d, there was a spot of corrosion product on the surface, but the corrosion current density obviously increased and the corrosion potential decreased, indicating that the CO2 corrosion of the L485 pipeline steel proceeded. This stage can be considered as the ‘‘bare metal’’ period. While the immersion time was 3–5d, large amounts of corrosion scales were generated on the specimen surface. Moreover, the corrosion current density almost unchanged but the corrosion potential increased apparently, which demonstrated the corrosion scales had a protective effect against the corrosion of L485 pipeline steel. This stage belonged to the ‘‘scale-covered’’ period. When the immersion time was 5–10d, the corrosion current density sharply increased and the corrosion potential decreased resulting in accelerated corrosion, which was due to the looseness and poor protection of the formed corrosion scale. This stage was called ‘‘reaction destruction’’ period. Thus, it can be concluded that the corrosion behavior of L485 pipeline steel can undergo the periods of bare metal, scale-covered and reaction destruction.
4. Conclusions In the present work, the corrosion failure analysis of L485 natural gas pipeline in CO2 environment was investigated. The morphology and chemical composition of the formed corrosion scales were characterized by SEM, XRD and EDS analysis, indicating that the corrosion scale of L485 pipeline steel was mainly FeCO3. The actual corrosion rate of L485 pipeline steel
378
L. Shi et al. / Engineering Failure Analysis 36 (2014) 372–378
could be calculated as 1.413–1.978 mm/a by a series of weight loss experiments. The chemical composition analysis of L485 pipeline steel also presented the Mn content exceeded the standard compared with API Spec 5L. In addition, the potentiodynamic polarization curve demonstrated the corrosion behavior of L485 pipeline steel underwent the periods of bare metal, scale-covered and reaction destruction. Acknowledgement This work was financial supported by National Natural Science Foundation of China, China National Petroleum Corporation Petrochemical Unite Funded Project (U1262111). References [1] Zou CJ, Zhao PW, Wang M, Liu DL, Wang HD, Zhang W. Failure analysis and faults diagnosis of molecular sieve in natural gas dehydration. Eng Fail Anal 2013;34:115–20. [2] Amirfakhri J, Vossoughi M, Soltanieh M. Assessment of desulfurization of natural gas by chemoautotrophic bacteria in an anaerobic baffled reactor (ABR). Chem Eng Process 2006;45:232–7. [3] Kermani MB, Morshed A. Carbon dioxide corrosion in oil and gas production-a compendium. Corrosion 2003;59:659–83. [4] Zou CJ, Zhao PW, Ge J, Qin YB, Luo PY. Oxidation/adsorption desulfurization of natural gas by bridged cyclodextrins dimer encapsulating polyoxometalate. Fuel 2013;104:635–40. [5] Liu YC, Zhang YL, Yuan JM, Ye MJ, Xu JZ. Research on corrosion perforation on pipeline by media of high salinity acidic oil-water mixture. Eng Fail Anal 2013;34:35–40. [6] Li WF, Zhou YJ, Xue Y. Corrosion behavior of 110S tube steel in environments of high H2S and CO2 content. J Iron Steel Res Int 2012;19:59–65. [7] Wu QL, Zhang ZH, Dong XM, Yang JQ. Corrosion behavior of low-alloy steel containing 1% chromium in CO2 environments. Corros Sci 2013;75:400–8. [8] Sadeghi Meresht E, Shahrabi Farahani T, Neshati J. Failure analysis of stress corrosion cracking occurred in a gas transmission steel pipeline. Eng Fail Anal 2011;18:963–70. [9] Li DG, Feng YR, Bai ZQ, Zheng MS. Characteristics of CO2 corrosion scale formed on N80 steel in stratum water with saturated CO2. Appl Surf Sci 2007;253:8371–6. [10] Nesic S, Nordsveen M, Maxwell N, Vrhovac M. Probabilistic modeling of CO2 corrosion laboratory data using neural networks. Corros Sci 2001;43:1373–92. [11] Ezuber HM. Influence of temperature and thiosulfate on the corrosion behavior of steel in chloride solutions saturated in CO2. Mater Des 2009;9:3420–7.