Biofilm-induced corrosion inhibition of Q235 carbon steel by anaerobic Bacillus cereus inoculum in simulated cooling water

In this study, the corrosion behavior of Q235 carbon steel (CS) under a Bacillus cereus (B. cereus) inoculum in simulated cooling water was evaluated. The weight loss study proved B. cereus inoculum possessed anticorrosion efficiencies of 92.84% and 73.88% for 3-day and 14-day rotation tests, respectively. The electrochemical measurements indicated that the added B. cereus inoculum increased the charge transfer resistance and reduced corrosion current density. B. cereus cells with strong biofilm-forming capacity were able to adhere onto the Q235 CS surface to form compact biofilms and cause biomineralization. Surface characterization analysis demonstrated that the presence of the B. cereus inoculum reduced the amount of Fe2O3 and simultaneously increased the amount of CaCO3 in corrosion products. The corrosion inhibition mechanisms of the B. cereus inoculum involve forming biofilm, generating a biomineralized layer, and consuming dissolved oxygen. Thus, B. cereus inoculum provides a biological strategy for industrial cooling water anticorrosion application.


Introduction
A cooling water system is an essential part in power plants that functions to remove excess heat and ensure safety production; the air-contact water with constant temperature makes cooling water suitable for microorganism growth (Hu et al. 2022a). Long-living microbiota including metal-reducing bacteria (MRB), sulfatereducing bacteria (SRB), and acid-producing bacteria (APB) are easy to attach on metallic surfaces to accelerate corrosion reaction, leading to microbially influenced corrosion (MIC) (Lekbach et al. 2019). A microorganism was also found to inhibit metal corrosion. It has been observed that metal corrosion was alleviated in the presence of a living biofilm (Mansfeld 2007). The first application in microbial-induced corrosion inhibition was raised by Pedersen et al. (Amelie et al. 1988) in 1988. Until now, a variety of microorganisms have shown their capacities on metal corrosion inhibition such as Desulfovibrio alaskensis (Wikieł et al. 2014), Streptococcus mutans , Vibrio neocaledonicus sp. (Moradi et al. 2015), Marinobacter salsuginis (Khan et al. 2020), Bacillus subtilis (Shen et al. 2020), and Bacillus cereus (Li et al. 2019a, b). Mechanisms of microbiologically influenced corrosion inhibition (MICI) are multifaceted, and they involve processes like (1) formation of microbial clusters and biofilms on metal surfaces consume oxygen by microbial respiration (Wu et al. 2016), (2) protective biofilm-released antibiotics inhibit the growth of corrosive microorganisms (Videla & Herrera 2009), and (3) some microorganisms form a compact biomineralized film to prevent contact between metal surfaces and corrosive media (Guo et al. 2021). Such complex processes make MICI an attractive research area.
Based on this, many attempts have been employed to use microbial biofilms as a mean for metal corrosion inhibition. Among which, Bacillus species were usually Responsible Editor: Diane Purchase * Chuanmin Chen hdccm@126.com 1 found to have the function of corrosion resistance. In the research of Qu et al. (2015), corrosion rates of cold rolled steel in simulated artificial seawater were accelerated in the first 16 h and then reduced with Bacillus subtilis C2 inoculation. The gradually compact biofilm acted like a barrier between the corrosive medium and metal surfaces. In a previous study (Aïmeur et al. 2015), a significant reduction in corrosion current densities of A 60 steel was observed under the presence of B. cereus, and the authors attributed this to the formation of protective biofilms. Recently, Li et al. (2019a, b) reported on using a soluble extracellular polymeric substance (s-EPS) of B. cereus in achieving 316L stainless steel corrosion inhibition. It was observed that the s-EPS of B. cereus adsorbed onto the SS surface and formed a thin film to protect the metal from corrosion. Their further study verified the fact that the B. cereus biofilm clearly impeded the electron transfer of 316LSS (Li et al. 2019a, b). Thus, it is easy to draw a conclusion from bodies of literature that the structure of the B. cereus biofilm plays an important role in the corrosion inhibition process.
As far as the current research progresses, the B. cereus biofilm has a potential to be applied as a corrosion inhibitor or anticorrosive coating, which is suitable for usage in cooling water systems. However, current research studies laid most emphases on seawater environment, and very few studies focused on cooling water environment. What's more, most studies used a medium culture of B. cereus, which is not favorable as practical application for engineering environment. Therefore, it is necessary to dedicate the active B. cereus dry powder for industrial cooling water corrosion protection. In our previous study, B. cereus inoculum was found to possess a certain corrosion inhibition property (Hu et al. 2022b); nevertheless, detailed performances as well as some possible influencing factors are still not explored. Under this premise, this paper investigated the corrosion behavior of Q235 CS in simulated cooling water under the presence of B. cereus inoculum. The active B. cereus dry powder with wide availability, low cost, and environmentally friendly properties possesses a great potential to be used as a green corrosion inhibitor. It is of great significance to investigate its corrosion behavior and mechanisms.
In this paper, the hypothesis of using the B. cereus inoculum as a Q235 CS corrosion inhibitor in cooling water media was investigated. First, 3-day rotatory experiments were carried out for preliminary screening of the optimal activation methods, and the 8-h substrate (brown sugar + urea) activation was selected as the optimal activation. Then, a 14-day rotation was conducted as the main platform for further corrosion studies on the electrochemical behaviors. Surface characterization analysis, microbial species composition, and water quality changes were all investigated to explore the corrosion behavior and mechanisms of Q235 CS under the presence of the B. cereus inoculum.

Materials and methods
Preparation of B. cereus inoculum, culture inoculum, and simulated cooling water environment.
B. cereus inoculum in this study was the wettable brown dry powder of B. cereus (Fig. s1), and its bioactive cell number was 1 × 10 10 CFU/g. Before usage, B. cereus inoculum was applied in two ways of activation: one was inoculated into an LB medium (tryptone 10 g, yeast extract 5 g, NaCl 10 g, pH 7.2) for fermentation. Specific steps are as follows: B. cereus inoculum was first mixed in sterile water for 1 h, subinoculated into a fresh LB medium with a 5%/3%/3% inoculum (v/v%) for 3 passages, and the last broth was stored for following experiments. The growth curve (Fig. s2) was depicted by monitoring its absorbance at OD 598 . Moreover, B. cereus LB broth after 8-h, 24-h, and 72-h incubation was selected to represent B. cereus bacteria in the log phase, stable phase, and decline phase, respectively. The other activation method was by dissolving the B. cereus inoculum in sterile water with or without the substrate (a certain ratio of brown sugar + urea). The group without substrate was incubated for 1 h, and the group with substrate was incubated for 8 h. All the activations were carried out at 36.9 ℃. B. cereus culture after activation was practiced according to the inoculum (v/v%) and mass concentration.

Biofilm formation assay
The film-forming ability of B. cereus inoculum was investigated by the tube assay method as described before (Maurya et al. 2022). Briefly, a 2-mL nutrient LB medium was mixed with an activated B. cereus bacterial suspension for adjusting the absorbance to 0.005 and then placed in 36.9℃ incubation. After a certain incubation time, the medium was decanted, and the tubes were gently washed with sterile PBS (phosphate-buffered saline). Afterwards, biofilm attached on tube wall was first fixed by 4% paraformaldehyde and then stained by a 2 mL crystal violet (0.02% w/v) for 30 min. PBS wash was subsequently repeated to remove excess stain, and the optical density of the eluted crystal violet was determined at a wavelength of 600 nm. The LB medium without bacteria served as control, and each treatment carried out three replicates.

Biocorrosion studies: weight loss and electrochemical measurements
The biocorrosion studies were conducted under the platform of an RCC-II rotating coupon corrosion tester (Liu et al. 2013). A simple representation of the experimental steps is illustrated in Fig. 1. The steps of weight loss measurements referred to China National Standard (CNS) GB/T 2014. Q235 CS coupons with dimensions of 50 mm × 25 mm × 2 mm were used throughout the tests (C, 0.14; Si, 0.12; Mn, 0.41; P, 0.02; S, 0.011; Al, 0.045; balance, Fe). Before being suspended on a PTFE holder, the Q235 CS coupons were first scraped with acetone and ethanol, and holders with coupons were subsequently immersed in a beaker containing 1.6 L simulated cooling water to started rotation tests. The initial simulated cooling water quality is shown in Table s1, and deionized water as make-up water was filled every 6 h to cope with constant water evaporation. All the rotation tests were performed at 35℃ with 75-rpm speed. After the whole experimental cycle (3 days for activation screening tests and 14 days for further corrosion tests), the Q235 CS coupons were thoroughly rinsed with distilled water, sequentially washed with an acid and alkali solution, dried, and accurately weighed. Each set of experiment was repeated three times to ensure reproducibility. The corrosion rate (v) was calculated by Eq. (1): where m is the coupon weight loss (g), m 0 is the coupon weight loss in the acid cleaning test (g), s is the coupon surface area (cm 2 ), is the coupon density (g/cm 2 ), t is the test time (h), 8760 is a constant which represents the hours in a year (365 × 24 h), and 10 is corresponding millimeters (mm/ cm). The corrosion inhibition efficiency ( X ) was calculated using Eq. (2): where v 0 and v are the corrosion rates in the absence and presence of inhibitors, respectively. The electrochemical measurements were operated with a three-electrode system, and Q235 CS with a copper wire connected was sealed with epoxy resin leaving a 1-cm 2 area of exposure to serve as a working electrode. A platinum plate and a saturated calomel electrode were used as the counter and the reference electrodes, respectively. After the same pretreatments as weight loss coupons, the prepared working electrodes were fastened on the holder together and rotated with weight loss coupons. Before each EIS measurement, open circuit potential (OCP) was performed under a 30-min stabilization time, and the EIS was then operated by applying 5-mV voltage over frequencies ranging from 0.01 to 10 5 Hz. Obtained EIS data were analyzed by the Zview software. The polarization curves were measured at a scan rate of 1 mV/s, and the potential applied was in the range of − 500 to 500 mV with respect to the OCP.

Surface characterization
For morphology observation, Q235 CS coupons after 1-day, 7-day, and 14-day rotation tests were taken out, gently washed with PBS to remove plankton cells, and then immersed in 2.5% (v/v %) glutaraldehyde for 8 h at 4 ℃ to immobilize the biofilm. Afterward, the Q235CS coupons were dehydrated in ethanol solutions of different concentrations (25, 50, 75, and 100% by volume) for 15 min each, gold sputtered, and subjected to SEM analysis. Corresponding element composition was analyzed by EDS mapping. Before SEM observation, Q235CS coupons were coated with gold.
In addition, a sterile scraper was used to scrape the corrosion products from the Q235CS surfaces after the 14-day rotation tests without and with B. cereus inoculum. The corrosion products were freeze-dried before characterization. The composition of the corrosion products was analyzed by XPS and XRD. In addition, Fourier transform infrared (FTIR) spectroscopy was also used to characterize the compounds in the corrosion products. The samples were pressed into a KBr pellet and analyzed under the frequency range of 400-4000 cm −1 .
The live and dead sessile cells and biofilm morphology attached on the metal surface were detected by confocal laser scanning microscopy (CLSM), and the Q235CS coupons after 1-day, 7-day, and 14-day rotation tests were first retrieved, rinsed briefly with PBS, and then stained with SYTO-9 and propidium iodide dyes in a completely dark environment. The living cells showed green fluorescence at an excitation wavelength of 488 nm, and the dead cells showed red fluorescence at an excitation wavelength of 559 nm.

Bacterial 16S rRNA sequencing
The commercial B. cereus inoculum was sealed in lyophilized tubes for 16S rRNA sequencing, and the total DNA was extracted using a commercial kit (Real Biotech Corporation). Universal primer sets 16sf (5′ CAG CAG CCG CGG TAA TAC 3′) and 16Sr (5′ TAC GGC TAC CTT GTT ACG 3′) were used to amplify the 16S rRNA gene. Agarose gel electrophoresis confirmed the PCR amplified product of the 16S rRNA gene of bacteria. In addition, a QIA gel extraction kit was used to purify the PCR product in the gel, which was processed for sequencing. The sequences were saved in single FASTA format files and further submitted to the NCBI using BLAST. A phylogenetic tree was constructed using the MEGA-6.0 software, and the results revealed that the commercial B. cereus (Bacillus-b1) inoculum is most closely related to B. cereus (WXZP3 OL468526.1) (Fig. s3). Moreover, for the 14-day rotation experiment, 400 mL of water samples at 1 day, 7 days, and 14 days were, respectively, filtered through a 0.22-mm pore size and 45-mmdiameter membrane filters. Afterward the membrane filters were sealed at − 20 °C until the 16S rRNA sequencing was carried out.

Water quality analysis
During the 14-day rotation tests, water quality indexes including pH, dissolved oxygen (DO), ammonia nitrogen (NH 3 -N), and chemical oxygen demand (COD) were analyzed in order to explore the role of the B. cereus inoculum in Q235 CS corrosion process in simulated cooling water. A pH meter along with a DO meter was used to monitor DO and pH values, and the COD and NH 3 -N values were determined according to China National Standards HJ/T 399-2007 and HJ 535-2009, respectively. Each sample was measured at least three times, and average values were taken as the represented results.

Biofilm formation assay
Biofilm formation plays a vital role in metal corrosion. In this study, we assessed the biofilm-forming capability of the promising anticorrosion B. cereus inoculum in polystyrene tubes. The biofilm formation of the B. cereus inoculum after different incubation times is presented in Fig. 2a, and the inset corresponds to its digital photographs. As can be seen, the biofilm attached on the polystyrene tube wall accumulated gradually, and the OD 598 increased from 0.02 to 0.81 during the 48-h incubation, demonstrating that the B. cereus inoculum exhibited a strong biofilm-forming capacity. The inset view displayed gradually deepening staining, which illustrated that the biofilm increases (Xu et al. 2020). Apparently, the biofilm induced a violet color that gradually became deeper with time.

Weight loss tests
Figure 2b-d shows the Q235 CS corrosion inhibition performances of B. cereus on 3-day rotation tests determined by weight loss measurement. Overall, the corrosion inhibition performances of the activated B. cereus inoculum supernatant were superior than those of the B. cereus broth, especially at high concentration levels. Figure 2b comprehensively investigates the Q235 CS corrosion inhibition performances of the B. cereus broth with respect to different growth periods. In general, the anticorrosion capacities of the B. cereus broth were common. The corrosion inhibition efficiencies of log, stable, and decline of the B. cereus broth all decreased gradually with the increment of inoculation dosage. In the inoculum range of 0.025-1%, the corrosion inhibition efficiency of log phase B. cereus broth decreased from 54.32 to 32.31%, and this efficiency range reduced to 54.95-16.32% for the stationary phase broth and 52.19-23.73% for the decline phase broth. The corrosion inhibition efficiencies of the sterilized stationary phase B. cereus broth were stable, maintained at about 50% in the inoculum range. In addition, the contribution of the LB medium was also considered, the inhibition performances of the LB medium increased slightly with the increase of inoculum, and the highest inhibition rate was 54.58% at 1% inoculum. The corrosion inhibition efficiencies of the LB medium were higher than those of the B. cereus broth, which indicated that inoculated B. cereus did not enhance the Q234 CS anticorrosion properties. Figure 2c shows the corrosion inhibition performances of the B. cereus inoculum in another activation approach. As it can be seen, the B. cereus inoculum had excellent corrosion inhibition capacities on Q235 CS; regardless of 1 h or with substrate 8-h activation, their corrosion inhibition efficiencies increased continuously. In the tested concentration range of 50-2000 mg/L, the anticorrosion efficiencies increased from 40.94 to 88.59% for 1-h activation and 63.09 to 92.84% with substrate 8-h activation. In contrast, the B. cereus inoculum after sterilization showed first a decreasing and then a weakly increasing anticorrosion performance. The anticorrosion efficiencies of the sterilized 1-h activated B. cereus inoculum were 57.69% at 50 mg/L, decreased to 39.09% at 1000 mg/L, and then slightly increased to 52.16% at 2000 mg/L. The corresponding anticorrosion efficiencies of the sterilized 8-h activated B. cereus inoculum was 68.53% at 50 mg/L, slightly reduced to 60.38% at 1000 mg/L, and later rose to 83.54% at 2000 mg/L. The substrate (brown sugar + urea) used for the 8-h activation was also examined, and its corrosion inhibition efficiencies were maintained at about 55% (+ − 5%). The above results demonstrated that at higher concentrations, the survival of B. cereus contributes to Q235 CS corrosion protection.
The 3-day rotation tests were conducted for primary activation screening. According to the results, the 8-h activation with substrate exhibited the optimal corrosion inhibition performance and therefore was selected for subsequent 14-day rotation tests (Fig. 2c). The compared 3-day and 14-day corrosion performances of the B. cereus inoculum after activating 8 h with substrate are shown in Fig. 2d. The results showed that the higher the concentration of the B. cereus inoculum added, the higher the corrosion inhibition efficiencies. The anticorrosion effects were similar between the 3-and 14-day rotation tests at lower concentrations (50-600 mg/L) but obviously distinct at higher concentration (2000 mg/L), of which the 3-day corrosion inhibition efficiency (92.84%) was significantly higher than that at 14 days (70.38%). For this phenomenon, we believed that the 14-day had a longer run time than the 3-day experiment, so the nutrient-free state under the prolonged time caused the death or inactivation of our B. cereus inoculum and ultimately led to a lower corrosion inhibition efficiency.

Electrochemical measurements
The Nyquist and Bode plots of Q235 CS during the 14-day rotation tests under different bacterial groups are described in Fig. 3. In general, during the first 4 days of the experiments, the Nyquist plot semicircle radius in the biotic group was obviously larger than that in the abiotic group, reflecting that adding the B. cereus inoculum significantly enhanced the Q235 CS corrosion resistance in the first 4 days. Also, the radius showed different tendencies for the two groups. In the abiotic group (Fig. 3a), the immersed Q235 CS gave an unoptimistic corrosive condition, and the radius of the Nyquist plot continuously declined over the 14-day experiments and corresponded to a continuously accelerated corrosion rate. However, in the biotic group, Q235 CS coupons experienced a first inhibited and then accelerated corrosion rates. As Fig. 3c shows, the Nyquist plots exhibited an enlarged semicircle radius in the first 4 days; the maximum plot radius that emerged in 1 day indicated the highest corrosion resistance appeared at 1 day after the 1000 mg/L B. cereus inoculum inoculated. Then in 2-4 days, the radius was slightly reduced but still larger than 8 h. In the following 7-14 days, the radius of the Nyquist plots dropped dramatically (Fig. 3c illustration), implying that the B. cereus inoculum had lost most of its corrosion protective function.
The corresponding Bode plots are shown in Fig. 3b and d. The phase angle of the Bode plots showed only one peak at low frequency in our experiments, which was attributed to the formation of a passive film on the surface of the working electrode (Giorgi-Pérez et al. 2021). The larger the phase ankle values, the better the protection (Dong et al. 2021). Furthermore, a corresponding equivalent circuit was designated for this transfer charge process as shown in Fig. 3e, where Rs represents the solution resistance, R ct is the charge transfer resistance at the interface of the substrate/film, and the capacitance for electrical double layer at the metal/solution interface is denoted by Q dl . The EIS results are fitted using Zview and are shown in Table 1. It can be seen from Table 1 that the impedance of Rs fluctuated with immersion time, but the R ct of the biotic group exhibited first a rising and then a falling trend. The R ct first rose to a maximum value of 20,961 (Ω·cm 2 ) at 1 day and then gradually decreased to only 961.6 (Ω·cm 2 ) at 14 days, demonstrating the first inhibited and then accelerated corrosion of Q235 CS coupons in the presence of 1000 mg/L B. cereus inoculum. In contrast, the R ct values of the abiotic medium were far lower than those of the biotic medium, which kept a declining trend from 3160 to 814.4 (Ω·cm 2 ) throughout the experiments, confirming the continuous accelerated corrosion rates in the abiotic group. Figure 3f and g displays the potentiodynamic polarization curves of Q235 CS during 14 days of rotation in the abiotic and biotic groups. The corresponding parameters are listed in Table 2. As can be seen, the Q235 CS coupons in the biotic medium had lower corrosion current densities (I corr ) compared to the coupons in the abiotic medium. This suggested that B. cereus protected Q235 CS from corrosion to a certain extent. The I corr of the two groups gradually increased during the experimental cycle, indicating that the corrosion condition was gradually deteriorating from 1 to 14 days, which is consistent with the results of EIS.

SEM along with EDS analysis
The surface morphology and elemental composition of tested Q235 CS were analyzed by SEM-EDS analysis. The surface morphology of pristine Q235 CS without immersion is displayed in Fig. s4, and it showed regular vertical strips of texture and elemental compositions of Fe (92.37%) and C (7.63%). Figure 4 presents the surface morphologies and EDS results of Q235 CS coupons in rotation tests inoculated with the B. cereus inoculum. Figure 4a-c respectively represents Q235 CS after 1-, 7-, and 14-day rotations. It can be seen from Fig. 4a that a large number of B. cereus bodies were densely attached to the Q235 CS surface, indicating that B. cereus could complete the attachment to a metal surface within 1 day after addition, and similar morphologies of B. cereus were also found in a previous study (Aïmeur et al. 2015;Qu et al. 2015). In addition, some crystal particles were intertwined around some bacteria, and the EDS results revealed that the location of crystal particles (point 1: O 34.05%, Fe 46.12%) had an increase in O ratio and a decrease in Fe ratio compared to the location of bacteria (point 2: O 19.58%, Fe 65.83%), indicating that more oxidation reactions occurred at the location of crystal particles, while the corrosion due to iron oxidation may be less severe at the location where the bacteria attached. At 7 days in Fig. 4b, the surface of Q235 CS distributed mainly dense homogeneous and agglomerated spherical crystals, which were identified as iron oxides in the EDS pattern (point 2). Some slender strips were attributed to bacterial metabolites due to a relatively higher C content (point 1: C 5.73%), while bacteria are no longer visible on the surface. At 14 days in Fig. 4c, nonuniform corrosion products appeared on the surface, and the Fe element content had decreased compared with 1 and 7 days. The upper layer of iron oxides emerged agglomerates of massive materials, which contained a

cereus inoculum (c and d).
Equivalent circuit model used to fit the EIS data (e). Potentiodynamic polarization curves of Q235 CS in the abiotic group (f) and the biotic group (g) during 14-day rotation tests higher proportion of organic elements than 1 and 7 days. Point 1 had 22.37% and 36.43% of C and O elements while point 2 had 17.82% and 37.68% of C and O elements, and they should be organic matters with respect to microbial metabolism (Liu et al. 2017a, b).
The EDS line and mapping results of Q235 CS coupons after the 1-, 7-, and 14-day rotation tests inoculated with the B. cereus inoculum are displayed in Fig. 5. The Fe, Ca, C, and O elements were observed to be the major elements of the corrosion products. At 1 day, the O, P, and Na elements were more dense on the surface of Q235 CS in the region where the bacteria adhered, while the Fe element was relatively sparse. At 7 days, the morphology of the Q235 CS surface could be divided into two halves. On the left side of Fig. 5b, the Fe, O, and P elements were more concentrated, while the Ca and C elements were more separated. The stronger peaks of the O and Fe elements proved the presence of iron oxides. However, the right side was on the contrary, where Ca was found as the most dense element accompanied by a certain amount of the O element and a very low intensity of the Fe element. The clear hexagonal rhombic crystals obtained are typical morphology Ca scale precipitates and imply that mineralization had occurred (Duan et al. 2008). Figure 5c represents the EDS analysis of 14-day Q235 CS surface, and the surface showed surrounded iron oxides with an enrichment of the Ca element in the center, where a laminar stacking morphology could be recognized. In addition, a very low content of other elements was observed. EDS mapping and line results demonstrated the elements C, O, P, and Ca were mainly rich in cluster-like material around Fig. 5c, which should be the microbial product or calcium scale induced by biomineralization.

Biofilm observation by CLSM
CLSM studies were carried out to examine the bacterial attachment and biofilm growth on the Q235 CS surface exposed to the simulated cooling water medium. The biofilm samples on the Q235 CS surfaces at three different times behaved differently. The dense surface covered with shiny green color cells at 1 day indicated the complete attachment of B. cereus cells (Fig. 6a). At 7 days, the surface exhibited aggregated green fluorophores with slightly darker fluorescence intensity compared to 1 day.
At 14 days, the biofilm became more heterogeneous with green fluorescence gathered as clusters. The green fluorophore intensity became weak once again, and it appeared that the biofilm on the Q235 CS surface had decayed. The clusters were also found in the SEM results (Figs. 4c and 5c), which demonstrated that at the later stage of the experiment, the biofilm on Q235 CS surfaces existed

s analysis along with water quality indexes
The changes in bacterial community structures in 14 days of rotation tests are monitored by 16S rRNA gene amplicon sequencing. Bacteria composition at the genus level is compared in Fig. 7a for 1, 7, and 14 days. At 1 day, Acinetobacter (36.2%), uncultured_bacterium_f_Enterobacteriaceae (25.8%), Paenibacillus (9.9%), and others (15.2%) were the dominant genera. However, when ran to 7 days, the main microbial community structure changed significantly, Azospirillum accounted for the most (61.3%), and other strains showed varying degrees of decline. The changes in Azospirillum are closely related to ammonia nitrogen in water. In Azospirillum, a well-known nitrogenfixing bacterium, ammonium, nitrate, nitrite, amino acids, and molecular nitrogen can all serve as its nitrogen sources (Steenhoudt and Vanderleyden 2000). The increase in its relative abundance could be caused by the death of dosed B. cereus that released organic nitrogen that acted as its energy source. Finally at 14 days, the bacteria in the simulated cooling water were mostly composed of miscellaneous bacteria from the environment, and Pseudoxanthobacter (15.6%), uncultured_bacterium_p_Armatimonadetes (12.1%), possible_genus_04 (8.8%), and others (42.4%) were found to be more abundant. The B. cereus inoculum we dosed probably belongs to the genus Paenibacillus, and it occupied a relative low abundance of 9.9% at 1 day, 1% at 7 days, and 0.5% at 14 days. This may be due to the lack of nutrients and the open experimental environment of the simulated cooling water where the B. cereus inoculum can easily be defeated by environmental microorganisms and thus could not occupy a dominant population position. Figure 7b and c displays changes of four water quality indexes over time: pH value, DO, NH 3 -N, and COD. During the 14-day rotation tests, the pH and DO values showed continuous increasing trends (Fig. 7b). pH values increased from 5.5 at 1 day to 7.5 at 14 days, which was probably due to the hydrogen precipitation reaction producing a certain amount of OH − (Fig. 7a) (Qu et al. 2017(Qu et al. , 2015. The DO value was maintained at around 0 mg/L for the first 2 days and kept continuously increasing in the rest period. The 0 mg/L DO value at 2 days is closely related to the metabolism of the dosed B. cereus inoculum (Suma et al. 2019), and the death of which led to a subsequent increase in DO. Finally, The DO value reached 3.09 mg/L at 14 days, which is disadvantageous for Q235 CS corrosion protection.
The changes of NH 3 -N and COD are shown in Fig. 7c. NH 3 -N values increased from 5.60 to 6.85 mg/L from B. cereus inoculum dosed to 1 day; this slight increase was because of fractionation of residual substrates. The NH 3 -N values then decreased from 6.85 to 4.33 mg/L in 3 days probably due to the degradation by our dosed B. cereus inoculum. It has been widely accepted that B. cereus is able to reduce ammonia nitrogen (Hlordzi et al. 2020). In the following 4 days, the NH 3 -N values continually increased and reached a maximum value of 10.43 mg/L at 6 days. This is presumably attributed to the inactivation and breakdown of our activated B. cereus inoculum (brown sugar + urea). There are some microorganisms that have the ability of decomposing organic nitrogen to produce ammonia, such as Azospirillum, a free-living N bio-fixer, capable of synthesizing molecular nitrogen into the ammonia state (Steenhoudt & Vanderleyden 2000). The previous 16 s sequencing results had indicated an increase in its relative abundance in between 1 and 7 days (Fig. 7a), which can cause a distinct increase in NH 3 -N. NH 3 -N showed a decline in the last experimental period, indicating that the organic state of NH 3 -N had been gradually degraded by other environmental microorganisms.
The COD in the cooling water showed a decreasing trend from 391.4 to 431.4 mg/L in the first 2 days, and this may also be due to increase in reducing substances induced by residual substrate decomposition. In the remaining tests, the COD values kept declining from 431.4 to 61.7 mg/L, indicating the concentration of reducing substances in the water body gradually decomposed by environmental microorganisms (Chaudhry et al. 2022).

Surface characterization
The compositions of the Q235 CS corrosion products after 14-day immersion with and without the B. cereus inoculum were examined by XRD, FTIR, and XPS. The XRD spectra in Fig. 8a demonstrate that in the abiotic group, the corrosion products were mainly FeO(OH), Fe 3 O 4 , and a little Fe 2 O 3 , while under the presence of the B. cereus inoculum, corrosion products were mainly FeO(OH) mixed with a small amount of Fe 3 O 4 , Fe 2 O 3 , and CaCO 3 . The proportion of Fe 3 O 4 was obviously reduced, and it appeared to have peaks corresponding to CaCO 3 . This result indicated that B. cereus cells that adhered to metal surfaces may induce calcium carbonate precipitation. Figure 8b and c represents the FTIR spectra of Q235 CS corrosion products after the 14-day rotation tests with and without the B. cereus inoculum. In Fig. 8b, broad peaks at 3149.75 cm −1 were attributed to C-H and N-H groups related to surface adsorption of water molecules or organic matters such as proteins and polysaccharides (Li et al. 2019a, b). The peak at 1790.23 cm −1 was related to stretching modes of the Fe-EPS complex (Ghafari et al. 2013), while peaks at 1018.87 and 572.47 cm −1 were assigned to FeO(OH) and Fe-O (Li et al. 2019a, b). The unique fingerprints containing peaks at 1480.13 and 858.55 cm −1 were associated with calcite structure (Santos et al. 2021). The asymmetric stretching vibration of the C-O-C group from carbohydrates was demonstrated by the peak at 1145.41 cm −1 (Li et al. 2019a, b). For the Q235 CS corrosion product without the B. cereus inoculum in Fig. 8c, the broad peak located at 3099.94 cm −1 was attributed to O-H groups of the adsorbed water molecule. Peaks at 1022.94 and 747.4 cm −1 were ascribed to FeO(OH), while the peak at 471.95 cm −1 was possibly from Fe 2 O 3 (Jasinski 1988). The results were consistent with the XRD results. Figure 8 also shows the XPS spectra analysis of the O 1 s, Fe 2p, and Ca 2p of the Q235 CS corrosion products after the 14-day rotation tests without ( Fig. 8d-f) and with the B. cereus inoculum (Fig. 8g-i). In the sterile cooling water medium, the spectrum of Fe 2p was decomposed into Fe, Fe 2 O 3 , Fe 3 O 4 , FeO(OH), and FeSO 4 (Khan et al. 2020). However in the medium inoculated with the B. cereus inoculum, except for conventional iron oxides, a complex of organic matter and iron Fe(CH 3 C(O)CHC(O)CH 3 ) 3 had also been obtained . The presence of Fe was mostly due to scraping of corrosion products, while FeSO 4 was an impurity caused by the experimental water. In the deconvoluted O 1 s spectra ( Fig. 8e and h), the peaks were related to iron oxides (Fe 3 O 4 and Fe 2 O 3 ), hydrous iron oxides (FeO(OH)), adsorbed water (H 2 O), the bond of carbon oxygen (C-O), and organic ligands (N-O) that resulted from microbial processes (Liu et al. 2017a, b;Yaseen et al. 2019). The two peaks of Ca 2p ( Fig. 8f and i) were identified as Ca2p 3/2 and Ca2p 1/2 (Ni and Ratner 2008). Its peaks' intensities were much stronger in the presence of the B. cereus inoculum than the abiotic group, which indicated that inoculating B. cereus induced the calcium precipitation on the Q235 CS surface.

Discussion
Recently, biofilm-induced corrosion inhibition has captured considerable attention, and numerous reports have shown that microorganisms are capable of causing corrosion inhibition (Gao et al. 2021;Shen et al. 2020;Zhao et al. 2021). Among which, B. cereus has been featured as a typical and repeated case of corrosion inhibition studies (Li et al. 2019a, b;Li et al. 2019a, b;Qu et al. 2015). Though Liu et al.  reported that the B. cereus biofilm accelerated corrosion in X80 pipeline steel, such conflict corrosion behavior might be attributed to differences in B. cereus strains, coupons, and experimental parameters.
The corrosion of Q235 CS is an electrochemical reaction, which can be expressed as follows: Anodic reaction: Cathodic reaction: The following reactions occur when OH − concentration increases to a certain level at anodic sites: Then, the iron oxides accumulate on the surface of Q235 CS and prevent further corrosion.
On this basis, this study found that the addition of the activated B. cereus inoculum significantly inhibited the corrosion of Q235 CS in the simulated cooling water. The anticorrosion mechanisms are closely associated with compact biofilm formation, consumption of DO, and biomineralization layer.
The B. cereus inoculum possesses strong adhesion and film-forming ability (Fig. 2a). In the growth stage of 8 h to 1 day after inoculation, the macromolecules in B. cereus EPS, especially proteins and polysaccharides with special functional groups, complexed with Fe ions to promote the adsorption of B. cereus cells onto a metal surface (Fig. 8c) (Liu et al. 2017a, b). As a result, B. cereus cells could form a compact biofilm (Figs. 4a and 5a), which acted like a barrier to block the Fe 0 dissolution and achieved a corrosion inhibition effect. This was assigned to the role of biofilms in blocking the extracellular electron transfer (EET) (Li et al. 2019a, b).
It was also found that the presence of the B. cereus inoculum maintained DO levels to nearly 0 mg/L at the earlier stage (Fig. 7c). In view of the higher levels of dissolved oxygen in the cooling water, this is closely related to the anaerobic respiration of microbial metabolism (Moradi et al. 2015). The biotic group exhibited a relatively larger impedance radius with a higher R ct (Fig. 3c), and the corresponding cathodic polarization curves in Fig. 3g further supported the fact that cathodic electrochemical reaction has been inhibited (Eq. (5)). What's more, the XPS and XRD spectra showed an obvious distinction between the corrosion products of the abiotic and biotic groups. Higher proportion of FeO(OH) and lower proportion of iron oxides indicated the presence of B. cereus inoculum retarted the next decomposition reaction of FeO(OH) (Eqs. (9) and (10)). This was attributed to organic acids secreted from B. cereus physiological activities, which will decrease pH values to provide H + , (6) Fe 2+ + 2OH − → Fe(OH) 2 (7) 4Fe(OH) 2 + O 2 + 2H 2 O → 4Fe(OH) 3 Another corrosion inhibition mechanism was related to the biomineralized film. The B. cereus inoculum induced calcite crystal precipitation on Q235 CS during this experiment, which was proved by the EDS and XRD results. It is already known that organic macromolecules with negatively charged functional groups (-COOH) in B. cereus EPS can complex with free Ca 2+ and offer grow sites for CaCO 3 crystals (Wada et al. 2018;Zhuang et al. 2018). Additionally, the carbonic anhydrase enzyme produced by B. cereus activity may also participate in the biomineralization process. CA could catalyze the hydration reaction of CO 2 to release HCO 3 − and further induce calcite precipitation (Xiao et al. 2014). The generated biomineralized layer adhered to the Q235 CS surface and hindered its contact with the corrosive medium ( Fig. 6b and c).

Conclusion
In this study, the corrosion behavior of Q235 CS under B. cereus inoculum in simulated cooling water was investigated. After activation by brown sugar and urea, the B. cereus inoculum significantly inhibited Q235 CS corrosion, increased the impedance, and decreased the corrosion current densities. The biofilm formed by B. cereus adhesion and calcium scale layer was caused by biomineralization, and together with consuming DO through anaerobic respiration may be the main corrosion inhibition mechanism. However, B. cereus is susceptible to losing a dominant population, which may be due to lack of nutrition. Following research studies on long-term stable corrosion inhibition are still necessary. Our work suggested that the B. cereus inoculum possesses a great potential for industrial cooling water anticorrosion application.

Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations
Ethics approval and consent to participate. Additional informed consent was obtained from all individual participants for whom identifying information is included in this article.

Consent for publication
All the authors have read and approved this manuscript and take responsibility for its contents. The participant has consented to the submission of the manuscript to the journal.

Competing interests
The authors declare no competing interests.