Introduction

Long-distance pipelines, which permanently served in complex soil environments, are the main transportation channels for oil and natural gas1. Corrosion failure of pipeline steel is one of the important threats to the safe operation of long-distance pipelines2,3. It was reported that more than 20% of pipeline system failures were caused by microbiologically influenced corrosion (MIC)4,5,6,7. MIC is the corrosion processes of metals enhanced by the synergistic interaction of microorganisms, biofilms and metabolic products8,9,10. The soil environment that the pipeline steel served in is usually anaerobic or hypoxic, which provides necessary conditions for the survival of anaerobic sulfate-reducing bacteria (SRB)11,12. Therefore, SRB corrosion has become the main MIC type of pipeline steel, and has received extensive attention from scholars and engineers throughout the world13,14,15.

In environments with abundant organic carbon sources, SRB can obtain electrons by lactate oxidation to complete the sulfate reduction as follows16,17:

$$\begin{array}{l}{{{\mathrm{Lactate}}}}\;{{{\mathrm{oxidation}}}}\;\left( {{{{\mathrm{intracellular}}}}} \right):{{{\mathrm{CH}}}}_3{{{\mathrm{CHOHCOO}}}}^ - + {{{\mathrm{H}}}}_2{{{\mathrm{O}}}} \to {{{\mathrm{CH}}}}_3{{{\mathrm{COO}}}}^ -\\\qquad\qquad\qquad\qquad\qquad\qquad\qquad\quad +\, {{{\mathrm{CO}}}}_2 + {{{\mathrm{4H}}}}^ + + {{{{4e}}}}\end{array}$$
(1)
$${{{\mathrm{Sulfate}}}}\;{{{\mathrm{reduction}}}}\;\left( {{{{\mathrm{intracellular}}}}} \right):{{{\mathrm{SO}}}}_4^{2 - } + 9{{{\mathrm{H}}}}^ + + 8{{{{e}}}} \to {{{\mathrm{HS}}}}^ - + 4{{{\mathrm{H}}}}_2{{{\mathrm{O}}}}$$
(2)

However, organic carbon sources are always scarce under mature biofilm and it is difficult for SRB cells in it to obtain the organic carbon source from the medium18,19. Thus, these SRB cells turn to obtaining energy through the reduction of intracellular sulfate (Reaction (2)) and the oxidation of extracellular iron (Reaction (3))20.

$${{{\mathrm{Fe}}}}\;{{{\mathrm{oxidation}}}}\;\left( {{{{\mathrm{extracellular}}}}} \right):4{{{\mathrm{Fe}}}} \to 4{{{\mathrm{Fe}}}}^{2 + } + 8{{{{e}}}}$$
(3)

This process requires the extracellular electrons to transfer cross the cytomembrane. The electron transfer from the outside to inside of the cell can be achieved directly by the redox proteins (c-type cytochromes, conductive nanowires) on the bacteria surface, or indirectly under the help of electron carriers21,22,23. SRB utilizes electrons donated by the iron matrix to reduce sulfate to sulfide and obtain the required energy24,25.

It can be clearly seen from the reactions (2) and (3) that the iron matrix acts as an electron donor under mature biofilm, inducing corrosion of the iron matrix26. This is the extracellular electron transfer (EET) mechanism of SRB corrosion for carbon steel27. Besides, it was reported that SRB become more aggressive under organic carbon source starvation, accelerating the corrosion of carbon steels. Xu et al.16 confirmed that the SRB corrosion rate of C1018 carbon steel was accelerated and the depth of corrosion pits on the steel surface was increased in organic carbon source starvation medium. Pre-established SRB biofilms on carbon steel become more corrosive under organic carbon source starvation. Li et al.28 also obtained similar results through organic carbon source starvation experiments, and confirmed that SRB uses L245 carbon steel as an electron donor to maintain metabolism under starvation conditions. Dou et al.29 compared carbon steel SRB corrosion in standard medium and under organic carbon starvation conditions. The results showed that the weight loss of carbon steel was the highest (6.6 mg cm−2) under the level of 20% carbon source, while the weight loss of carbon steel in the standard medium was the lowest (3.3 mg cm−2).

On the other hand, MIC of welded joints has also attracted the increasing attention throughout the world. Nandakumar et al.30 studied the corrosion of AISI 304 L-type stainless steel by Pseudomonas sp.. It was found that the weld zone (WZ) with many grains and grain boundaries showed more bacterial attachment, while the base metal (BM) had the least number of bacteria. Therefore, changes in the micro-structure after welding can affect bacterial attachment, resulting in selective corrosion of welded joints. Zhu31 studied the SRB corrosion of manual TIG welded joints of X80 steel in artificial seawater and verified that the most serious selective corrosion occurred in the WZ. Recently, we investigated the SRB corrosion behavior of X80 pipeline steel welded joints and found that SRB tended to adhere to the BM and WZ, and accelerate the localized corrosion process in these two zones. However, SRB did not like to adhere to the heat affected zone (HAZ) surface, reducing the SRB corrosion tendency in the HAZ32. Obviously, SRB corrosion of different zones of carbon steel welded joints is quite different from each other. Unfortunately, the study and exploration on the related mechanism are still not enough, thus it certainly looks worthy of further investigation.

From the above, EET is the main corrosion mechanism of carbon steel under the action of SRB metabolism. In the organic carbon source starvation environment, SRB use carbon steel as an electron donor and accelerate its corrosion process. Besides, the micro-structure and electrochemical activity of different zones of welded joints are significantly different, which affects the microbe adhesion and the electron transfer. Therefore, two questions emerge: When carbon-starved SRB encounter a welded joint, do they use all zones or a specific zone of the welded joint as electron donor(s)? What is the difference in corrosion behavior between the selected electron donor zone and other zones of the welded joints under organic carbon source starvation condition? These questions are very interesting, and they also are important for in-depth understanding of the SRB corrosion mechanism of metals. However, to our knowledge, they are still unknown. To discover the answers of these questions, the SRB corrosion of an X80 steel welded joint under organic carbon source starvation conditions was studied in this work. The results of the work are beneficial for us to reveal the behavior and mechanism of SRB corrosion for carbon steel welded joints.

Results

Mature biofilms on the sample surface after pre-incubation

The biofilm morphology on different zones of the welded joint after 3 d, 7 d, and 14 d pre-incubation in the medium with 100% carbon source containing SRB are shown in Fig. 1. After 3 d pre-incubation, thin layer of mature biofilms were respectively attached on the three zones of the sample surfaces, and the threadlike SRB were inlaid on the sample surfaces to form a loose network of biofilms. Obviously, no microbial tubercles formed on the three zones and the difference between them was very slight. After 7 d pre-incubation, the biofilm on the three zones became markedly different. Cluster-like microbial tubercles containing a large number of SRB and flocculent corrosion products appeared on the BM and WZ zones of the welded joint. However, after 14 d pre-incubation, the number of microbial tubercles on the BM and WZ zones continued to increase. According to our previous study, localized corrosion rightly generated under the microbial tubercles32. Therefore, the mature biofilms after 3 d pre-incubation in the medium with 100% carbon source containing SRB were selected as the original biofilms for the organic carbon source starvation experiment in this work.

Fig. 1: Biofilm on the surface of the welded joint.
figure 1

a–c Show biofilm morphology on different zones after 3 d pre-incubation. d–f Show biofilm morphology on different zones after 7 d pre-incubation. g–i Show biofilm morphology on different zones after 14 d pre-incubation.

Full size image

Planktonic, sessile cell counts and corrosion rates

The number of planktonic cells in the mediums with different carbon source levels are shown in Fig. 2a. Three similar typical growth curves in limited environments were observed in the three mediums. During the initial phase of the experiment, the number of planktonic SRB cells exponentially increased and reached a peak on 3th day. It remained in the range of 108–1010 cell ml−1 on the 3th to 8th days and then began to decline sharply. On the 15th day the SRB number was only about 104 cell ml−1 and slowly decreased at the end of the experiment. The growth curves of planktonic SRB cells in the mediums with 0 and 1% carbon sources were very similar, but the peak of the SRB number in the mediums with 100% carbon source was 1–2 orders of magnitude larger than those in the mediums with 0 and 1% carbon source. The abundant nutrients in the medium with 100% carbon source supported the propagation of the SRB cells in the stable phase.

Fig. 2: SRB cell count and corrosion rate of the steel.
figure 2

a Shows the number of planktonic SRB cells in the mediums with different levels of organic carbon source. b1–b6 Show the macroscopic morphologies of the BM surfaces before and after removal of the corrosion product. c Shows the number of sessile cells on the BM surfaces in the mediums with different levels of organic carbon source. d Shows the corrosion rate of the BM of the welded joint in mediums with different levels of organic carbon source (error bars stand for the standard deviations from three independent sample).

Full size image

The macro-morphologies of the BM surface after 30 d immersion in the three mediums are presented in Fig. 2b1–b3. A large number of microbial tubercles were distributed on the BM surfaces in the mediums with 0 and 1% carbon sources, and the number of microbial tubercles in the latter was more than that in the former (Fig. 2b1, b2). However, the BM surface was relatively flat in the medium with 100% carbon source (Fig. 2b3). The facts show that the reduction of carbon source significantly affects the adhesion of SRB cells on the sample surfaces. After removal of the corrosion product, the macro-morphologies of the naked BM surface are presented in Fig. 2b4–b6. Dense corrosion spots can be distinctly seen on the BM surface in the mediums of 0 and 1% carbon sources (Fig. 2b4, b5), and the positions of the corrosion spots are in keeping with that of the microbial tubercles as shown in Fig. 2b1, b2. However, no obvious corrosion spot was found on the BM surface in the medium with 100% carbon source (Fig. 2b6).

The numbers of sessile cells on the BM surfaces after 30 d immersion in the three mediums are presented in Fig. 2c. At the end of the experiment, the number of sessile cells on the BM surface was the largest in the medium with 1% carbon source, followed by with 0% carbon source, while that in the medium with 100% carbon source was the smallest one. In the medium with 100% carbon source, SRB exponentially multiplied and thus consumed sharply, resulting into the least SRB cell in the end of the experiment. In the medium with 0% carbon source, SRB cells only relied on sulfate reduction and Fe oxidation to obtain energy, so the survival of the SRB cells was difficult and thus the number of SRB was relatively less. However, in the medium with 1% carbon source, SRB cells can obtain simultaneously electrons from the carbon source and iron matrix. Therefore, more sessile cells existed on the steel surface in this medium. Besides, the numbers of planktonic cells in the three medium was far less than those of sessile cells in the end of the experiment, indicating that bacteria tend to grow on the steel surface at this time. What’s more, no reasonable correlation was found between the numbers of planktonic and sessile cells. The fact showed that the number of planktonic cells was not the decisive factor affecting the adhesion of SRB cell on the steel surface.

The average corrosion rate of the BM after 30 d immersion in the three mediums are presented in Fig. 2d. The corrosion rate of the BM was the largest in the medium with 1% carbon source, followed by that with 0% carbon source, while that with 100% carbon source was the lowest. The results showed that organic carbon source starvation made SRB more aggressive16. In the environment with organic carbon source starvation, SRB tended to utilize Fe matrix as an electron donor, promoting the corrosion process of the matrix. In addition, the corrosion rates of the steel were completely identical with the number of sessile SRB cells in the three mediums (Fig. 2c, d), indicating that the number of sessile SRB cells on the steel surface was the decisive factor of the aggressive SRB corrosion for the steel.

Corrosion product on the surface of welded joint

The micro-morphologies of the corrosion products on the surface of BM, HAZ, and WZ after 30 d immersion in the three mediums are shown in Fig. 3, and the corresponding EDS results are listed in Table 1. The surfaces of all samples were covered with a complete mixture of SRB and corrosion products in the three mediums. The distribution of corrosion products on the sample surface in the mediums with 0 and 1% carbon sources was uneven, and corrosion products and SRB cells varied in different zones. The corrosion products on the sample surface in the medium with 0% carbon source tended to be distributed in flakes, while those with 1% carbon source were mainly distributed in clusters. However, the distribution of corrosion products and SRB on the sample surface in the medium with 100% carbon source was relatively uniform, only with a small amount of prominent corrosion products inlaid on the BM and WZ surfaces. EDS results showed that the relatively higher S content existed in the corrosion products in the medium with 100% level, while the Fe and O contents increased in the corrosion products in the environments with organic carbon source starvation.

Fig. 3: Micro-morphologies of corrosion products on welded joints after 30 d immersion.
figure 3

a–c Show the micro-morphologies of corrosion products on the surfaces of BM, HAZ, and WZ zones in the medium with 0% organic carbon sources level. d–f Show the micro-morphologies of corrosion products on the surfaces of BM, HAZ, and WZ zones in the medium with 1% organic carbon sources level. g–i Show the micro-morphologies of corrosion products on the surfaces of BM, HAZ, and WZ zones in the medium with 100% organic carbon sources level.

Full size image
Table 1 EDS results of corrosion products on the surface of BM, HAZ, and WZ for the welded joints (at.%).
Full size table

The number and morphology of SRB cells on the sample surfaces in the three mediums are of great differences. The number of SRB cells in the product film in the environments with organic carbon source starvation was relatively higher than that in the medium with 100% carbon source, which was consistent with the results of sessile cell counts (Fig. 3c). Besides, the sessile SRB cells in the environments with organic carbon source starvation were mainly slender and stubby, but most of the sessile SRB cells were stubby in the medium with 100% carbon source. The facts certainly indicated that the general metabolism of starving SRB cells was relatively higher.

The high-resolution spectra of Fe, N, and S orbitals of the corrosion products on the sample surfaces after 30 d immersion in the three medium are presented in Fig. 4. In the medium with 100% carbon source, the corrosion products on the sample surfaces of the welded joint are mainly FeOOH33, and there is no obvious difference in chemical composition between the three zones (Fig. 4g). However, the species of Fe oxides on the sample surfaces increased in the environments with organic carbon source starvation, containing FeOOH, FeO34,35, Fe3O433,36 and Fe2O336,37,38 (Fig. 4a, d). In the same environment, the species of Fe oxides in the BM and WZ were more than those in the HAZ. Besides, many sulfur-containing compounds, such as FeS39, FeS240,41, and FePS342, were detected on the sample surfaces in the three medium, among which FeS and FeS2 are typical metabolic sulfides of SRB43. In addition, various organic bond pairs44,45 also existed in the corrosion products in the sample surface, which may come from the amino acids, constituting proteins, polypeptides, and enzymes in the SRB biofilm28.

Fig. 4: XPS spectra of corrosion products on welded joints after 30 d immersion.
figure 4

a–c Show the XPS spectra of corrosion products on the welded joints in the medium with 0% organic carbon sources level. d–f Show the XPS spectra of corrosion products on the welded joints in the medium with 1% organic carbon sources level. g–i Show the XPS spectra of corrosion products on the welded joints in the medium with 100% organic carbon sources level.

Full size image

Corrosion morphologies of different zones of welded joints

The corrosion morphologies of the surface of BM, HAZ, and WZ after 30 d immersion in the three mediums are shown in Fig. 5. Obviously, the main form of SRB corrosion of the welded joints in the three mediums was localized corrosion, but the degree of corrosion pits were different for different welded zones in different mediums. In the same medium, the diameter and density of corrosion pits in the BM and WZ were always larger than those in the HAZ, indicating that SRB was more inclined to promote localized corrosion in the BM and WZ. SRB corrosion for the welded joint was distinctly selective32. Besides, for the same zone of the welded joint, the diameter and density of corrosion pits on the sample surface were the largest in the medium with 1% carbon source, followed by that with 0% carbon source, with it was the smallest in the medium with 100% carbon source. Especially, in the medium with 100% carbon source, there was almost no corrosion pit existing in the HAZ (Fig. 5h). The aggressiveness of the SRB cells to the welded joint was strongest in the medium with 1% carbon source, significantly promoting the localized corrosion of the welded joints. Therefore, SRB corrosion of the welded joint was selective, and organic carbon source starvation increased its selective characteristics.

Fig. 5: Corrosion morphology of welded joints after 30 d immersion.
figure 5

a–c Show the corrosion morphology of BM, HAZ, and WZ zones after 30 d immersion in the medium with 0% organic carbon sources level. d–f Show the corrosion morphology of BM, HAZ, and WZ zones after 30 d immersion in the medium with 1% organic carbon sources level. g–i Show the corrosion morphology of BM, HAZ and WZ zones after 30 d immersion in the medium with 100% organic carbon sources level.

Full size image

3D morphologies of the surfaces of BM, HAZ, and WZ after 30 d immersion in the three mediums are presented in Fig. 6. Obviously, the localized corrosion susceptibility of the HAZ was always less than those of the BM and WZ in the same medium. For the same zone in the welded joint, the corrosion degree of the sample in the medium with 1% carbon source was more severe. The facts were in complete agreement with the corrosion morphologies of the samples. The depths of 5 random corrosion pits on the surface of each sample were recorded, and the average depth of their corrosion pits of the samples are shown in Fig. 7. For the same zone in the welded joint, the average depths of the corrosion pits in the mediums with 0 and 1% carbon sources were about 1.5 and 2.5 times of that with 100% carbon source, respectively. Organic carbon source starvation-induced SRB to accelerate the corrosion of the welded joint, and the acceleration effect was more obvious in the medium with 1% carbon source. In all mediums, the average depth of the corrosion pits in the WZ was about 3.5–4.0 times of that in the HAZ, while it became 4.5–6.5 times in the BM. The selectivity of the SRB corrosion for the welded joints varied in different medium.

Fig. 6: 3D ultra-depth micro-morphology of the welded joints after 30 d immersion.
figure 6

a–c Show the 3D ultra-depth micro-morphology of BM, HAZ, and WZ zones after 30 d immersion in the medium with 0% organic carbon sources level. d–f Show the 3D ultra-depth micro-morphology of BM, HAZ, and WZ zones after 30 d immersion in the medium with 1% organic carbon sources level. g–i Show the3D ultra-depth micro-morphology of BM, HAZ and WZ zones after 30 d immersion in the medium with 100% organic carbon sources level.

Full size image
Fig. 7: Average pit depth of welded joints after 30 d immersion.
figure 7

The figure shows the average pit depth of BM, HAZ, and WZ for the welded joints after the 30d immersion in different organic carbon source medium (error bars stand for the standard deviations from three independent sample).

Full size image

Discussion

From Fig. 2b, d, organic carbon source starvation-induced SRB corrosion and promoted the localized corrosion susceptibility of the welded joint. In order to quantitatively describe the effect of organic carbon source starvation on SRB corrosion acceleration of welded joints, the uniform corrosion acceleration factor (κCR) and pitting corrosion acceleration factor (κpit) were defined using the average corrosion rate and the average depth of the corrosion pits. κCR and κpit can be expressed as follows:

$$\kappa _{{{{\mathrm{CR}}}}} = \frac{{\mathrm{CR}}}{{{\mathrm{CR}}_0}}$$
(4)
$$\kappa _{{{{\mathrm{pit}}}}} = \frac{l}{{l_0}}$$
(5)

where, CR and CR0 are the average corrosion rates of the BM in the mediums with organic carbon source starvation and in the medium with 100% carbon source, respectively. l and l0 represent the average depth of the corrosion pits of the samples in the mediums with organic carbon source starvation and in the medium with 100% carbon source, respectively. Obviously, both κCR and κpit were equal to 1 in the medium with 100% carbon source. This method has been widely used in the studies of corrosion and pitting behavior of metallic materials46,47,48,49,50.

Figure 8 presents the acceleration factors of SRB corrosion in the BM of welded joints in the three mediums. In the mediums with 0 and 1% carbon sources, κCR of the BM were 1.15 and 1.26, while κpit of the BM were 1.75 and 3.33, respectively. This result shows that the starving SRB cells accelerated simultaneously the uniform corrosion and localized corrosion of the welded joints, and the accelerated SRB corrosion was more obvious in the medium with 1% carbon source. Besides, κpit of the BM was much larger than its corresponding κCR under the same carbon source starvation. The accelerated SRB corrosion was more embodied in κpit rather than κCR. Therefore, κpit was selected to analyze the effect of organic carbon source starvation on SRB accelerated corrosion of the welded joint hereinafter.

Fig. 8: κCR and κpit for BM zone.
figure 8

The figure shows κCR and κpit of organic carbon source starvation on corrosion of BM for the welded joints (error bars stand for the standard deviations from three independent sample).

Full size image

Given that the above results were merely relative to the BM of the welded joint, did the organic carbon source starvation have the same influence on SRB corrosion of the HAZ and WZ? To answer this question, κpit of the HAZ and WZ in the three mediums were calculated according to the Eq. (5), and the results are presented in Fig. 9. Under organic carbon source starvation, κpit of all zones were greater than 1, showing that the organic carbon source starvation accelerated the localized pits in the SRB corrosion of the welded joint. In the environments with organic carbon source starvation, the starving SRB used carbon steel as an electron donor and preferentially attached to the sample. Then, many microbial tubercles will be generated on the sample surface (Fig. 2b), resulting in the increased corrosion rate of the steel (Fig. 2d) and the severe corrosion pits on the sample surfaces (Fig. 5). Besides, κpit of the samples in the medium with 1% carbon source were greater than that with 0% carbon source, revealing the more obvious acceleration effect in the medium with 1% carbon source. In the medium with 0% carbon source, the number of planktonic and sessile SRB cells was less. Thus, the aggressiveness of SRB cells to the iron matrix was relatively weak. However, in the medium with 1% carbon source, SRB can obtain electrons from both carbon source and iron matrix. Thus, the metabolism of SRB was relatively vigorous (Fig. 3), and the number of sessile SRB cells was relatively large (Fig. 2). Under the mature biofilm, the starving SRB cells were more aggressive to the iron matrix and promoted more serious localized corrosion on the sample surfaces.

Fig. 9: κpit for welded joints.
figure 9

The figure shows κpit of organic carbon source starvation on SRB corrosion of BM, HAZ, and WZ for the welded joints (error bars stand for the standard deviations from three independent sample).

Full size image

As shown in Figs. 5 and 9, SRB corrosion and κpit of different zones varied with organic carbon source starvation. In other words, the influence of organic carbon source starvation on the selectivity of SRB corrosion was different for different zones of the welded joint. To quantitatively describe the influence, a selectivity factor (κselect) was defined using the average pit depth. Given that the localized corrosion susceptibility of the HAZ was always the lowest in all mediums, the average pit depth of the HAZ was used as a reference standard and thus κselect can be expressed as follows:

$$\kappa _{{{{\mathrm{select}}}}} = \frac{l}{{l_{{{{\mathrm{HAZ}}}}}}}$$
(6)

where, l is the average pit depth in the BM or WZ and lHAZ is the average pit depth of the HAZ in the same medium, respectively. Obviously, κselect of the HAZ was equal to 1 in all mediums.

The calculated κselect of SRB corrosion for the welded joints in the three mediums are shown in Fig. 10. κselect for the BM and WZ in the medium with 100% carbon source were far greater than 1, indicating that SRB corrosion of welded joints was intensively selective. κselect for the HAZ and BM were respectively the smallest and the largest in this medium, which is consistent with our previous results32. κselect for the BM and WZ in the mediums with 0 and 1% carbon sources were respectively greater than those in the medium with 100% carbon source, and κselect for the BM and WZ were the largest in the medium with 1% carbon source. The facts showed that organic carbon source starvation improved the selectivity of SRB corrosion of welded joints, and the selectivity was most significant in the medium with 1% carbon source. However, in all mediums, κselect for the HAZ was always smallest, following by that for the WZ, while that for BM was largest. Organic carbon source starvation improved the selectivity of SRB corrosion of welded joints, but did not change the order of this selectivity.

Fig. 10: κselect of SRB corrosion for welded joints.
figure 10

The figure shows κselect of SRB corrosion of the welded joints in the mediums with different organic carbon source starvation (error bars stand for the standard deviations from three independent sample).

Full size image

As shown in Fig. 11, the main micro-structures of the BM are uniformly distributed polygonal ferrite and bainite (Fig. 11a, d, and g), while that in the HAZ is the granular bainite structures with coarse grains and obvious grain boundaries dominate the (Fig. 11b, e, and h). The WZ is composed of fine ferrite with coarse feathered Widmanstatten structures (Fig. 11c, f, and i). It was reported that the grain boundaries in metal have the characteristics of high density dislocation, high intrinsic energy, and low electron work function51,52. Thus, it is easy for bacteria to obtain electrons and hence adhere near or to the grain boundaries30,53. Therefore, SRB prefers to use the BM or WZ of the welded joint as the electron donor. In the mediums with carbon source starvation, SRB turns to iron matrix to obtain electrons. It is a reasonable speculation that the starving SRB more liked to obtain electrons from the zones with high intrinsic energy and low electron work function, leading to more serious SRB corrosion in the BM and WZ of the welded joint in the mediums with organic carbon source starvation. On the other hand, the number and active metabolism of SRB cells may be relatively higher in the BM and WZ with high intrinsic energy and low electron work function (Figs. 2c and 3) under the level of organic carbon source starvation. Thus, SRB were more aggressive against iron matrix. Under the combined action of the above two factors, organic carbon source starvation promoted the selectivity of SRB corrosion of welded joints.

Fig. 11: Microstructures of welded joints.
figure 11

a–c Show the metallographic photos of the welded joints. d–f Show the secondary electron images of the welded joints. g–i Show the backscattered electron images of the welded joints.

Full size image

Methods

Materials

A spirally welded joint of X80 pipeline steel was selected as the experimental sample in this work, with the composition of the pipeline steel (wt.%): C 0.0466, Si 0.204, Mn 7.154, P 0.0082, S 0.0009, Ni 0.206, Cr 0.235, Cu 0.174, Nb 0.524, V 0.0022, Ti 0.0142, Mo 0.125, Al 0.0265, B 0.0004, and Fe Bal. Three zones (BM, HAZ, and WZ) of the welded joint were processed by wire cutting. Three groups of the BM, HAZ, and WZ samples were taken for carbon starvation experiments with different carbon sources. Due to the small zones of the WZ and HAZ in the welded joint, the weight loss measurement and sessile counts were only performed on the BM samples. The working surfaces of all samples were gradually polished from 280# to 1200# with water-resistant sandpaper, rinsed repeatedly with deionized water and anhydrous ethanol, and then blown dry in cold air for use. Before the experiment, all samples were sterilized by ultraviolet light in a vacuum glove box.

The SRB strains selected in this study were consistent with the previous experiments54. The SRB strains were cultivated in the ATCC 1249 medium and the 3rd generation strain was used for the follow-up experiment. The composition of the ATCC 1249 medium was shown in Table 2. In this study, modified ATCC 1249 medium was used as the experiment solution, and the contents of carbon sources including trisodium citrate and sodium lactate were controlled at 0%, 1%, and 100%, respectively. The three mediums with different levels of carbon source were sterilized in an autoclave at 121 °C for 30 min, and then deoxygenated by high-purity nitrogen for 2 h. In these mediums, yeast extract contains a small amount of organic carbon, which can provide growth factors and vitamins for SRB metabolism in the medium with 0% carbon source.

Table 2 Chemical composition of 1L ATCC1249 medium.
Full size table

Experimental procedure

The samples pre-polished via water-resistant sandpaper were divided into three groups and pre-immersed in anaerobic flasks with 1 L full carbon source medium containing SRB. After 3 d pre-incubation, the samples with mature biofilms were taken out from the flasks and the planktonic cells on the sample surfaces were washed away with sterile and deoxygenated phosphate-buffered saline (PBS) solution. Then the three groups of samples were respectively immersed into anaerobic flasks with 1 L medium with 0%, 1%, and 100% carbon source. After that, the flasks were immediately sealed using silica gel. During the experiment, planktonic cell counts in the three mediums were measured at intervals. After 30 d immersion, three groups of samples were taken out from the flasks, immersed with 5.0 wt.% glutaraldehyde for 5 h to fix the biofilm. Then, they were dehydrated successively by 30, 50, 70, and 100 vol.% anhydrous ethanol, dried, and placed in a sterile glove box for subsequent experiments. Beside, after 30 d immersion, another three BM samples that were just taken out from the flasks were used for sessile cells count and weight loss experiments, respectively. All the experimental procedures were conducted in a glove box full of high-purity nitrogen.

Bacteria count

In this work, KBC bacteria test flasks were used to measure planktonic cells in medium and sessile cells on the sample surface. The flasks was designed for the most probable number (MPN) method according to GB/T 14643.5-2009. During the experiment, 3 ml of solution was extracted from the anaerobic flasks at intervals to determine the planktonic SRB cell count and the results were reported using the unit of is cell ml−1. After the experiment, three BM samples were taken out from the flasks, and shook in a conical flask with 100 mL of deoxygenated PBS solution for 2 h. Then, 3 ml of the PBS solution was used to determine the sessile SRB cell count and the results were reported using the unit of is cell cm−2.

Weight loss analysis

Before the experiment, an analytical balance with an accuracy of 0.0001 g was used to recorded the initial weight of the BM samples. After the 30 d immersion, the corrosion products on the sample surfaces were removed using a rust remover consisting of 500 mL hydrochloric acid (18 wt.%), 500 mL deionized water, and 3.5 g hexamethylenetetramine. The final weight of the samples were determined again with the same analytical balance, and the corrosion rate was calculated as follows15:

$$v_{{{{\mathrm{corr}}}}} = \frac{{{\it{K}} \times \Delta m}}{{\rho \times S \times t}}$$
(7)

Where vcorr is the corrosion rate (mm y−1), K is the unit constant (87,600), Δm is the average weight loss of three replicate samples (g), and ρ is the density of the sample (7.85 g cm−3), S is the working area of the sample (cm2), t represents the immersing time (h), respectively.

Micro-morphology and composition characterization

After the 30 d immersion, the samples were treated according to the procedure in section 2.3. X-ray photoelectron spectroscopy (XPS, ESCALAB250XI surface analysis system, Thermo VG, USA) was used to determine the chemical state of the elements in the corrosion products. The binding energy in the high-resolution spectra was corrected referring to the C1s sub-peak at 284.6 eV and the sub-peaks in the XPS spectra were separated using XPSPEAK4.1 software. Then the sample surfaces were sprayed with gold for 120 s to increase the electrical conductivity of the corrosion products, and then the morphology of the corrosion products on the sample surfaces was observed by scanning electron microscope (SEM, Zeiss EVO). The ingredients of the corrosion products on the sample surface were analyzed by energy dispersive spectroscopy (EDS, OXFORD X-MaxN) as well. After that, the corrosion products were removed using the same rust remover from the sample surfaces, and the corrosion morphology of the samples was detected by SEM. Finally, the 3D morphology of the corroded surfaces and the corrosion pit on it were reconstructed using 3D ultra-depth microscopy (VHX-2000).