CORROSION BEHAVIOR OF CARBON STEEL INFLUENCED BY SULFATE-REDUCING BACTERIA IN SOIL ENVIRONMENTS

SeonYeob Li*, YoungGeun Kim, YoungTai Kho Korea Gas Corporation (KOGAS) 638-1, Il-Dong Ansan, Kyunggi-Do, Korea Tak Kang Seoul National University (SNU) Seoul, Korea

ABSTRACT This work is a comprehensive study of microbiologically influenced corrosion (MIC) of carbon steel in soil environments, induced by sulfate-reducing bacteria (SRB). Four experimental phases were involved in this research: (1) field study, (2) anaerobic corrosion study, (3) localized corrosion study, (4) mathematical modeling of cathodic protection (CP) current/potential distribution. Based on the field survey and statistical approach, the predicting equation for the maximum corrosion depth of steel in soil was presented and it was proved that the predicted values were well matched with the field data. The contribution of microbial factors on soil corrosion was also discussed. It was also concluded that the presence of SRB and the resultant biogenic iron sulfide film on steel surface greatly changed the metal/electrolyte interface properties, and therefore the corrosion behavior of steel from anaerobic study. Using the thin-film electrical resistance (TFER) sensors, it was possible to detect and monitor the localized corrosion behavior induced by SRB. The mathematical model using the boundary element method (BEM) is capable of predicting (1) local potential; and (2) local cathodic current density inside crevice for a given applied holiday potential and soil resistivity. The model predicted a very short depth of current penetration depending on the crevice geometry. From the results of modeling and the 1

following regression analysis, the optimization plots for cathodic protection inside c revice under the disbonded coating was proposed, which makes it possible to predict the current penetration depth. Keywords: microbiologically influenced corrosion (MIC), sulfate-reducing bacteria (SRB), anaerobic soil, thin-film electrical resistance (TFER) sensor, cathodic protection (CP), disbonded coating, current distribution INTRODUCTION It is widely recognized that microorganisms attach to, and influence the corrosion of metals and alloys exposed to the environments. Microorganisms are omnipresent in nature and their ability to grow and reproduce at rapid rates accounts for their presence in soil environments. Buried steel pipeline is designed to have a lifetime of about 30 to 50 years. In general, the protective organic coating and cathodic protection (CP) are applied together. Despite these protection measures, however, failure cases of underground pipelines due to corrosion have been continuously reported1-3. Microbiologically influenced corrosion (MIC) has been identified as one of the major causes of corrosion failures of underground pipelines 4. The first MIC case was identified in 19345, where sulfate-reducing bacteria (SRB) resulted in failure of underground cast iron pipes. Since this observation, numerous works have been made concerning the effect of bacteria, especially SRB, on MIC in soil environments. According to recent survey program performed by authors, almost all corrosion of underground gas pipeline has occurred under the disbonded coating6,7. The corrosion mechanism has been revealed mainly as the action of SRB. SRB are known as one of the key microbes in the MIC process. SRB are obligately anaerobic bacteria utilizing sulfate as a terminal electron acceptor and organic substances as carbon sources. During the metabolic process, sulfate is reduced to sulfide. These biogenic sulfides react with hydrogen ions produced by metabolic activities or by cathodic reaction of corrosion process to form corrosive hydrogen sulfide and iron sulfide precipitate on metal surface. The detailed mechanism of corrosion by SRB is reported elsewhere8,9. The risk of MIC must be exactly evaluated and the proper control methods prepared to mitigate MIC problems. The relationship between the activities of MIC-causing microbes, e.g., SRB, and various soil environmental factors should be investigated and assessed. In view of economic and engineering points, quantitative assessment of MIC is also important. The key point is, therefore, how much the activity of SRB does contribute the entire amount of corrosion phenomena in soil environments, and how the proper assessment techniques are set up. Moreover, MIC is a localized, pitting-type corrosion in most cases. Therefore, proper test methods for localized corrosion are needed. However, the understanding the risk of MIC is only half the solution. The other is to provide suitable protection of control method. Generally, underground pipelines are 2

protected from corrosion by CP as mentioned above. It is well known that CP could prohibit the activity of SRB and therefore SRB-induced corrosion effectively if the protection potential is maintained under 0.95V/CSE10-12 . However, for coated steel pipeline, all SRB-corrosion occurs under the disbonded coating where CP current cannot penetrate and remains as free-corroding conditions 6. Therefore, it is important to study the effect of CP under various conditions, such as disbonded area, resistivity and the degree of applied CP current, etc. In this study, various field survey, laboratory techniques and mathematical tools for the monitoring, detection, assessment and the protection of SRB-induced corrosion are briefly summarized. FIELD STUDY An attempt was made to investigate the effect of environmental factors on corrosion of underground steel pipeline. Field survey was carried out, and the relationship between maximum pit depth (Pmax ) and the measured environmental factors were analyzed statistically. The field survey of soil parameters and the analysis of sampled soil around the cathodically protected, polyethylene (PE) coated steel pipes (API X65 Gr.E) was performed. Detailed procedures are found elsewhere13. These survey results provide the useful information on the risk of MIC. Chemical and microbiological analysis of soil samples and the combination of the statistical methods such as multiple regression analysis are useful tools for this purpose. It is well known that the rate at which corrosion pits grow in the soil under a given set of conditions tends to decrease with time and follows a power-law equation (1)14.
P = kt n

(1)

Where P is the maximum pit depth in time t, and k and n are constants. If k and n can be determined using proper statistical methods, it becomes possible to predict the progress of corrosion in depth. Therefore, this equation (1) was adopted as the basic model. The predicting equation was obtained by multiple regression analysis. Figure 1 shows the dependence of the maximum corrosion depth on the key environmental factors. Considering the effect of environmental factors and the burial time, the final predicting equation for the maximum corrosion depth of steel pipeline was obtained as follows15;
Pmax,cal = 1. 573Pc t 0. 372

(2)

wherein Pmax ,cal: the predicted value of Pmax

LogPc = 0.700 + 0.069Log(SRB ) + 0.749P / S + 0. 203Log(Cl ) 0. 050Eh Clay 0.014pH Log() (3)

SRB: the population of SRB (cells/g of soil) 3

P/S: the pipe-to-soil potential (V/CSE) Cl-: the level of chloride (ppm) Eh: the reduction-oxidation potential (V/NHE) Clay: the clay content (%) ?: the soil resistivity (O? cm) Figure 2 shows the scatter diagram of Pmax and Pmax,cal. The correlation coefficients between P0 and P0,cal was 0.947. Data fell closely onto the straight line with the slope of 0.860. It is also found in this study that the underground corrosion of steel was affected mainly by three factors as shown in equation (3): (1) chemical factors such as Cl-, pH Log (?), (2) biochemical (microbial) factors such as Log (SRB) and Eh Clay and (3) CP factors such as P/S. This means that the corrosivity of soil can be evaluated quantitatively by analyzing the chemical and biochemical properties separately if ignoring the interaction effect of each factor. Figure 3 shows the relative contributions of each factor. The contribution of microbial, chemical, CP factors are in a range of 45 to 75%, 20 to 45% and 2 to 7%, respectively. It is evident that the contribution of microbial factor (about 45 to 75%) is most important in anaerobic environment as shown in Figure 3. Therefore, it can be concluded that the risk of underground corrosion can be successfully predicted quantitatively and the main cause for the corrosion is the activity of anaerobic microorganism, i.e., SRB. LABORATORY STUDY Corrosion potential and DC polarization techniques Because of its simplicity, the measurement of the corrosion potential (Ecorr) and linear polarization resistance (LPR) method has been used for MIC studies for many years16. A rapid and easy interpretation of the results is achievable using these methods together. Figure 4 shows the trends for Ecorr and uniform corrosion rate of carbon steel obtained by LPR method exposed to anaerobic soil for 148 days in laboratory. It is apparent that the existence of SRB greatly influences the corrosion behavior of carbon steel. In case of active growth of SRB, the potential increased slightly for the first 6 days and then maintained around -740 mV/SCE, but the potential fluctuated -600 mV to -800 mV/SCE after 50 days until the experiment ended. The potential in control (biocide-added) case was around -600mV and always more positive than that in SRB-active cases. One of the mechanisms for SRB-induced MIC of carbon steel has been thought to be the galvanic action between precipitated iron sulfides and metal2. After 50 days exposure, Ecorr of carbon steel in SRB-active soil fluctuated until the experiment ended with the continuous increase of uniform corrosion rate in Figure 4. It seems that this potential fluctuation was attributed to the repetitious production and 4

breakdown of iron sulfide film and is indicative of the onset of localized corrosion. Figure 5 is the Tafel polarization curves of carbon steel exposed to anaerobic soil. The growth of SRB had reduced continuously the polarization resistance of carbon steel, i.e., the increase of corrosion rate as shown in Figure 4. Electrochemical impedance spectroscopy (EIS) The electrochemical impedance spectroscopy (EIS) is particularly useful in the presence of nonconducting and semi-conductive films such as organic coating and inorganic precipitates. Many of biofilms adsorbed on metal surface are non-conductive and EIS techniques are potentially useful in their presence16. EIS plot in Figure 6 also shows the decrease of polarization resistance in the presence of SRB in anaerobic soil. In the first 7 days, it is thought that only one mechanism dominated the reactions in both cases from the phase angle-vs-frequency curves. After 148 days, however, there were significant changes in SRBactive case. It was observed that an increase in phase angle to a positive value, the appearance of a new maximum value of phase angle at a lower frequency (<10mHz) - which were thought to be related to the diffusion gradients by the combination of strongly adhered microbial films and corrosion products 17 and a decrease in impedance. These are well matched with the results of DC polarization tests (Figure 4 and Figure 5). However, the interpretation of EIS data depends largely on the development of equivalent circuits for modeling of the solid/electrolyte interface. These equivalent circuits are difficult to formulate for complex interfaces such as partial coverage by microbial films17. Surface analysis Surface analysis technique complemented with the microscopical examination such as scanning electron microscopy (SEM) is useful in monitoring the chemical nature of biological and inorganic deposits. Figure 7 shows the results of SEM observation of the change of surface morphology and energy dispersive X -ray analysis (EDXA) of carbon steel specimens exposed to SRB-active anaerobic soil in laboratory. In SRB-active soil, the whole surface of the steel coupon was dense iron sulfide films at 14 days (Figure 7 (a)). However, this sulfide film became porous at 148 days (Figure 7 (b)). The bacterial cells were observed on the surface (Figure 7 (c)). After removal of these films, a large pit was easily found (Figure 7 (d)). This porous film and base metal set up the galvanic couple, and the corrosion rate increased dramatically by galvanic action as shown in Figure 4. It is also remarkable that the S/Fe ratio in iron sulfide film increased at 148 days. Many researchers reported that the sulfide film became porous as the sulfur content in iron sulfide increased 5
18,19

. These facts were again proved by

SEM/EDXA analysis. LOCALIZED CORROSION STUDY As mentioned above, MIC is in general localized, pitting-type corrosion. The authors have applied multilined thin-film electrical resistance (TFER) probes for the study of this localized corrosion induced by SRB. The ER method is widely used as corrosion monitoring techniques especially in industrial field environments. At constant temperature, corrosion of the probe, which causes a decrease in crosssectional area of probe, can be monitored by periodically measuring the electrical resistance. The corrosion rate can be determined over very short time16. In general, this technique is used for the measurement of uniform corrosion rate. However, TFER probe shown in Figure 8 can be used to assess the localized corrosion due to its multi-sensing line. The details of fabrication, applications are previously described environments. It is apparent that the responses of each probe were quite different. The corrosion behavior of carbon steel can be divided as three categories in anaerobic condition, i.e., (1) anaerobic inorganic corrosion which depends on the ability to utilize the cathodic reactants, which may be water or hydrogen ion. As shown in Figure 9, the anaerobic, inorganic corrosion rates are in a range of 0.016 mm/y to 0.026 mm/y, and showed the uniform corrosion characteristics regardless of the types of electrolytes. (2) the precipitation of protective films (carbonate/bicarbonate); in this case, hard and protective film precipitated on the metal surface and remained for a long time, which resulted in no decrease of electrical resistance, i.e., no start of corrosion, (3) MIC induced by SRB; initially transient protective iron sulfide film prohibited the progress of corrosion for about 4 days, however, the film became ruptured and the localized corrosion developed (stepwise increase of resistance). The responses of TFER sensor buried in real soil environments is presented in Figure 10, which shows the relationship between the time to the onset of localized corrosion, that is, the time to the first resistance-decrease step detected, and the population of SRB; the larger the population of SRB, the faster the initiation of localized corrosion. Some scattering of data might be attributed to the incubation time for the activation of MIC or the influence of other factors such as oxygen or stray current detected at some sites. This result also confirms the applicability of TFER sensor in real situations. Though quantitative information on the corrosion is insufficient, TFER sensor can be used for the assessment of MIC risk, especially in terms of localized corrosion behavior within a relatively short period. EFFICIENCY OF CP
20

. Figure 9 shows the response of thin-film

ER probes immersed in various electrolytes. Nitrogen was purged in all cases to maintain anaerobic

CP effectively inhibits the progress of MIC in soil environments. This may be due to the increase of pH induced by hydroxyl ion produced by cathodic reaction suppresses the growth of microbes. It is more important, however, that the efficiency of CP should be considered. Figure 11 is the typical example of the disbondment of heat-shrink sleeve applied at the girth weld joint of carbon steel pipeline and the resultant corrosion morphology under the disbonded sleeve4,6. It is very difficult for CP current to penetrate inside this disbonded region because of defect geometry. Table 1 is the measured potential inside this disbonded coating by introducing the reference electrode (saturated copper/copper sulfate electrode) and platinum (Pt) electrode. CP current could not penetrate into the disbonded area, which resulted in the pipe surface exposed to anaerobic soil under the disbonded region being in the freely corroding condition (-610 mV to -550 mV/CSE) even though the pipe-to-soil potential was in a range of -1200 mV - -1400 mV/CSE. Eh under the disbonded area was about -160 mV/NHE. When SRB is cultivated in culture medium, the recommended initial Eh should around -100 mV/NHE [12]. Therefore, the suitable anaerobic condition for the active growth of SRB was provided under the disbonded coatings and MIC occurred in this region mostly. This insufficient CP condition was confirmed by numerical modeling of current/potential distribution inside the crevice under the disbonded sleeve, calculated by two-dimensional boundary element method (BEM) at various applied CP potential and the soil resistivity15. Figure 12 is the schematic representation of the model simulating the disbondment case in Figure 11. Figure 13 is the general trend for the CP potential distribution expressed in constant equipotential lines at the resistivity of 3000 Ocm, P/S = -1.2 V/CSE. A sharp potential drop was observed adjacent to the opening region of crevice, which results in the poor potential condition inside crevice. Figure 14 shows the potential/current distribution at metal/solution interface for crevice formed under the disbonded sleeve at the crevice opening potential (E0) of -0.85V/CSE. It can be seen that a significant shift on the crevice potential to more positive values occurred as the distance from the holiday opening (position = 20 cm) along the crevice increased, which results in the freely-corroding condition at the crevice center (position = 0 cm). This trend is related to the current density available along the crevice as shown in Figure 14. Of course, as the soil resistivity decreases and the applied CP potential shifts to more negative values, the effective penetration depth which satisfies the CP criteria inside crevice increases as shown in Figure 15. Considering the overprotection problems such as hydrogen embrittlement21,22, however, the optimization plot for CP of the crevice inside the disbonded sleeve is given in Figure 16. As shown in Figure 16, the optimum applied opening potential should be maintained in a range of -0.85 V/CSE to -1.2 V/CSE, i.e., the hatched areas should be avoided because of underprotection or overprotection. Therefore, the effective penetration depth of CP into the crevice is in a range of 1 cm to 12 cm depending on the soil conductivity (resistivity) and applied potential (or current) in this case and there is a limitation of CP to protect the crevice region, which result sin the MIC in the crevice under the disbonded coating. This disbonded coating should be prepared as quickly as possible to avoid the 7

breakdown of the pipeline. CONCLUSIONS (1) It is possible to assess the activity of SRB, i .e., the risk of MIC from the relationship with key environmental factors qualitatively and this risk is directly related to the corrosion rate by field survey and the statistical approach. It was also found that the main factor affecting underground corrosion is the action of SRB. (2) Microbial colonization of carbon steels can drastically alter their corrosion behavior. Electrochemical measurements can be made under MIC condition and showed good results. However, all electrochemical methods are averaging techniques that work best when the chemical and electrochemical conditions on the metal surface are uniform and steady state and provide no information on the localized corrosion. (3) The thin film ER probe test makes it possible to shorten the experimental periods and to study the localized corrosion behavior, which is a characteristic of MIC. (4) The optimization plot for the CP of crevice under the disbonded coating was also presented using the multiple quadratic regression analysis by considering the numerical modeling results and the overprotection problems. It is possible to predict and evaluate the CP status under the disbonded coating using this plot. According to the optimization plot for CP, it was inevitable to avoid MIC under the disbonded sleeves and therefore the proper and quick repair should be made. (5) The study for MIC of steel in soil environments requires a multi-disciplinary approach comprising different combination of microbiological, electrochemical and microscopic measurement as well as the consideration of CP that may be the only protective measure in soil environments. ACKNOWLEDGEMENT The authors gratefully acknowledge the Korean Ministry of Science and Technology for their financial support through the National Research Laboratory (NRL) program. REFERENCES 1. F. Kajiyama and Y. Koyama, Field Studies on Tubercular Forming Microbiologically Influenced Corrosion of Buried Ductile Cast Iron Pipes: Proceedings of International Conference on Microbially Influenced Corrosion, P. Angell, S.W. Borenstein, R.A Buchanan, S.C. Dexter, N.J.E. Dowling, B.J. Little, D.C. Lundin, M.B. McNeil, D.H. Pope, R.E. Tatnall, D.C. White and H.G. Ziegenfuss eds., (Houston, TX: NACE International, 1995): p. 30 2. T.R. Jack, M.J. Wilmott, R.L. Sutherby, MP, November, 19 (1995) 8

3. 4. 5. 6. 7. 8. 9.

S.Y. Li, K.S. Jeon and Y.T. Kho, Kr. J. Corros. Sci., 26, 321 (1997) (in Korean) S.Y. Li, Y.G. Kim, K.S. Jeon, Y.T. Kho and T. Kang, Corrosion, 57(9), 815 (2001) C.A.H. von Wohlzogen Khr and L.S. van der Vlugt, Water, 18, 147 (1934) S.Y. Li, Y.G. Kim, K.S. Jeon and Y.T. Kho, Metals and Materials, 6, 281 (2000) S.Y. Li, Y.G. Kim, J.Y. Jeon, Y.T. Kho and T. Kang, Corrosion Science and Technology, 31(1), 52 (2002) H.A. Videla, Biologically Induced Corrosion, S.C. Dexter, eds., (Houston, TX: NACE International, 1985): p. 162 W. Lee and W.G. Characklis, Corrosion, 49, 186 (1993)

10. J. Hovath and M. Novak, Corrosion. Sci., 4, 159 (1964) 11. K.P. Fischer, MP, 20(10), 41 (1981) 12. F. Kajiyama and K. Okamura, Corrosion, 55, 74 (1999) 13. S.Y. Li, J.Y. Jeon and Y.T. Kho, Statistical Approach to Underground Corrosion of Carbon Steel Pipeline: Proceedings of Asian-Pacific Corrosion Control Conference (APCCC), Vol.2, 2001: pp.1239-1248 14. H.H. Uhlig and R.W. Revie, in Corrosion and Corrosion Control, 3rd ed., John Wiley & Sons, New York, NY (1985): p.183 15. S.Y. Li, Corrosion Behavior of Carbon Steel Influenced by Sulfate-reducing Bacteria in Soil Environments, Ph.D. Thesis, Seoul National Univ. (SNU), Seoul, Korea (2002) 16. S.C. Dexter, D.J. Duquette, O.W. Siebert and H.A. Videla, Corrosion, 47(4), 308 (1991) 17. C.A.C. Sequira, in Bioextraction and Biodeterioration of Metals, C.C. Gaylarde and H.A. Videla, eds., (Cambridge, UK: Cambridge University Press, 1995): p.307 18. 40. R.A. King, J.D.A. Miller, nature, 223, 491 (1971) 19. B.D. Craig, Corrosion, 35, 136 (1979) 20. Y.G. Kim, D.S. Won, H.S. Song, S.M. Lee, Y.T. Kho, I.H. Bang, J.Y and Kim, in Proceedings of 194th Meeting of the Electrochemical Society (Boston, NJ: the Electrochemical Society, 1998): No. 320 21. W.H. Hartt, C.C. Kumira and R.J. Kessler, Corrosion, 49(5), 377 (1993) 22. P. Andrews, M. McQeen and N. Millwood, Corrosion, 57(8), 721 (2001)

TABLE 1. RESULTS OF POTENTIAL MEASUREMENTS Types Pipe-to-Soil Potential Potential inside disbonded region Reduction-Oxidation Potential (Eh) Potential -1430 - -1200 mV/CSE -610 - -550 mV/CSE -160 mV/NHE

FIGURE 1 - Relationship between P 0 (Pmax/Pmax_average) and environmental factors

10

FIGURE 2 - Relationship between P max and Pmax,cal

11

FIGURE 3 - Relative contribution of environmental factors to underground corrosion

FIGURE 4 - Effect of SRB on the corrosion potential and uniform corrosion rate of carbon steel exposed to anaerobic soil in laboratory

12

FIGURE 5 - Polarization curves for steel exposed to SRB-active soil for specified periods

FIGURE 6 - Comparison of EIS plot obtained after exposure to anaerobic soil (a) for 7 days, (b) for 145 days (A: SRB-active, B: control)

13

20m

(a)

B
20 m

(b)

2m

(c)

14

(d) FIGURE 7 - Surface morphology and the results of EDXA of carbon steel exposed to SRB-active soil; (a) for 14 days, (b) for 148 days, (c) SRB mixed with iron sulfide film after 148 days, (d) pit morphology after removal of surface deposit

Lead wire

Insulation layer

Sensing area

FIGURE 8 - Schematic diagram of thin film electrical resistance (TFER) probe

15

FIGURE 9 - The response s of TFER probes in anaerobic environments

FIGURE 10 - Time to the onset of localized corrosion vs. the population of SRB detected using TFER sensor

16

(a)

(b)

FIGURE 11 - Typical example of (a) disbonded sleeves at the girth weld joints, (b) morphology of corroded area
u=P/S

GROUND

2m
SOIL
q=0

q=0

2m

1cm
q=0

40cm
CREVICE y x 5cm

SLEEVE
q=0

40cm

STEEL

FIGURE 12 - Schematic representation of the model simulating the disbondment in Figure 11.

17

FIGURE 13 - Typical potential distribution expressed in constant equipotential lines at the resistivity = 3000 Ocm (0.00033 S/cm), P/S=-1.2V/CSE

FIGURE 14 - Potential and current profiles for crevice formed under disbonded coating at 3000 Ocm and crevice opening potential of 0.85V/CSE

18

(a)

(b) FIGURE 15 - (a) Response surface and (b) contour plot for the effective penetration depth of CP current

19

FIGURE 16 - Optimization plot for CP of crevice under the dibonded sleeve

20