Chemicals, including Hydrochloric acid (AR, HCl), Sodium hydroxide (AR, NaOH), it is mainly used to adjust the pH of the reaction system; Glycerol (AR, C₃H₈O₃), used for the preservation of microbial bacteria; Tetrachloroethylene (AR, CCl₄), used to extract petroleum hydrocarbons from groundwater; Sodium sulfate (AR, Na₂SO₄), used to remove moisture from the extract; Calcium chloride (AR, CaCl₂), it is used to simulate background ions in groundwater. The 0# diesel was purchased from CNPC gas stations with a density of 0.8 kg/L, serving as the degradation substrate. The sterile-filtered diesel was added to the inorganic salt medium at the concentration of 0.5% to obtain the enrichment culture. The isolated strains were purified and activated in LB agar plate medium (BR), then cultured in LB liquid medium (BR). The Gram staining solution Kit, purchased from China National Pharmaceutical Group Limited, were used in the Gram straining experiment. Quartz sand of different sizes were used to establish one-dimensional soil columns in the migration ability experiment. Quartz sand (Particle size: 0.5–1 mm, 0.25–0.5 mm, 0.125–0.25 mm) was purchased from Hong Andan Building Materials Co., Ltd (Beijing, China).

The experimental bacterial strains were screened from the contaminated soil after a diesel oil pipeline leak in Xinjiang, China. Specifically, 5 g contaminated soil samples were added to the enrichment culture and cultivated at 30 °C and 150 rpm for 7 d. Then, 1 mL of the culture was transferred to a fresh enrichment medium and cultivated as above to obtain a sub-culture. The cultivation from the third sub-culture medium were diluted and spread on solid plate of enrichment medium. The plates were cultivated at 30 °C and every colony was screened to obtain pure strains.

The isolated bacterial strains were identified using physiological and biochemical methods (morphography and standard protocol of Gram staining) and gene sequencing. PCR reaction system components included 10×Ex Taq buffer, 2 μL; 5 U/μL Ex Taq, 0.2 μL; 2.5 mmol/L dNTP Mix, 1.6 μL; 5 μmol/L Primer1, 1 μL; 5 μmol/L Primer2, 1 μL; DNA, 0.5 μL; and ddH₂O, 13.7 μL. The primer sequences used were 27F (AGAGTTTGATCCTGGCTCAG) and 1492R (TACGGYTACCTTGTTACGACTT). The PCR amplification was performed as follows: initial denaturation at 95 °C for 5 min, 25 cycles of denaturing at 95 °C for 30 s, annealing at 56 °C for 30 s, followed by extension at 72 °C for 90 s. The final extension was carried out at 72 °C for 10 min.. The PCR products were then sent for Sanger sequencing at Shanghai Majorbio Technology Co., Ltd., and the assembled sequences were analyzed by BLAST against the GenBank database. There were six strains identified.

The growth curve of diesel oil degrading bacteria was measured using the turbidity method. Specifically, the absorbance of bacterial liquid at 600 nm wavelength (The colorimetric dish is quartz colorimetric dish with seam width of 2 mm.) of ultraviolet spectrophotometer was measured using a ultraviolet spectrophotometer, thereby indirectly reflecting the growth amount of degrading bacteria.

The relationship between the number of microorganisms and the absorbance of OD 600 was obtained using the method of plate counting. First, 100 μL bacterial solution was added into a 2 mL centrifuge tube containing 900 μL sterile water, mixed well, diluted according to the gradient of 10⁻¹, 10⁻², 10⁻³, 10⁻⁴, 10⁻⁵, 10⁻⁶, 10⁻⁷, 10⁻⁸ and 10⁻⁹, and coated on a plate. Subsequently, the absorbance of the bacterial solution with different dilution times was measured using a OD 600 ultraviolet spectrophotometer. After measuring the absorbance, the bacterial solution with different dilution times was evenly coated on LB nutrient agar plate medium with a coating rod, and cultured in a high-precision combined shock incubator at 30 °C for 48 h. After culture, the relationship between OD 600 and biomass was clarified through counting statistics, and the results are shown in Fig. S1.

The six identified strains were inoculated onto LB plate medium in an adjacent manner and cultured in incubator at 30 °C for 7 d. The antagonistic effect between strains was judged according to whether the colonies grew together, and the combination of mutually antagonistic strains was removed. Then the three-bacterium combination scheme was adopted, and there were eight groups of compound bacteria obtained.

By comparing the degradation effect of each single bacteria and the eight combinations, the bacteria were inoculated in oily waste-water with diesel oil concentration of 1 g/L at room temperature. The concentration of diesel oil in water was determined by referring to ‘Determination of Petroleum and Animal and Vegetable Oils in Water by Infrared Spectrophotometry ′(HJ 637-2018). The calculation formula of diesel degradation rate was as Eq. (1):

where, η denotes the degradation rate of diesel oil; C₀ represents the content of diesel oil in the blank sample system, mg/L; and C_i refers to the remaining diesel oil content in the system after microbial degradation, mg/L.

The group of compound bacteria with the best degradation effect were selected, and the degradation pattern with time course was further investigated. The components of diesel oil were analyzed by gas chromatography, which was performed on HP-5 quartz capillary column (30 m×0.32 mm, film thickness 0.25 μm). Gas chromatographic conditions were set as follows: the inlet temperature was 320 °C, the column flow rate was 2.0 mL/min, and the column temperature chamber temperature was initially maintained at 60 °C for 1 min, then increased to 290 °C at 8 °C /min, and then increased to 320 °C at 30 °C /min for 7 min. The temperature of FID detector was 330 °C, the hydrogen flow rate was 40.0 mL/min, the air flow rate was 350.0 mL/min, and the tail blow flow rate was 30.0 mL/min. The injection method involved non-split injection initially, followed by split injection after 0.75 min, with a split ratio of 30: 1, and an injection volume of 1 μL.

In order to obtain an efficient TPH-degrading bacterial solution, three strains of C1A, B29 and NK4 were cultured in LB liquid medium at 30 °C for 30 h and mixed in a super clean workbench with a volume ratio of 1:1:1. Quartz sand with different particle sizes (coarse sand: 0.5–1 mm, medium sand: 0.25–0.5 mm, fine sand: 0.125–0.25 mm) was panned with dilute nitric acid solution with a concentration of 10% and soaked for 36 h. After soaking, the sand was washed repeatedly with deionized water until a neutral pH was reached, and then placed in an autoclave for autoclaving. Subsequently, the coarse sand, medium sand and fine sand media were filled into different plexiglass columns, and the filling process was continuously compacted to avoid the emergence of advantageous channels. Then, the quartz sand column was fully saturated with water, and the volume of the water was considered the pore volume. The experimental device is shown in Fig. S2, and the related sand column parameters are shown in Table S1.

After screening, a total of 6 strains of diesel oil degrading bacteria were obtained, named NK1, NK2, NK3, NK4, B29 and C1A, respectively. The colony morphology and streak culture results are shown in Table S2 and Fig. S3 respectively. Some studies have shown that under the same conditions, Gram-positive bacteria have a lower surface charge compared to Gram-negative bacteria [26], potentially influencing their migration behavior in aquifers. So the Gram staining of the six strains was conducted and the results are shown in Table S2 and Fig. S4. All of the six strains showed red color after Gram staining, indicating that they were all Gram-negative bacteria. Under microscope, NK1, NK3 and NK4 were rod-shaped or rod-shaped bacteria, while NK2, B29 and C1A were globular or short rod-shaped bacteria. The strains were identified using 16S rRNA gene sequencing and analyzed by BLAST against the GenBank database. The strains were identified as Acinetobacter venetianus NK1, Pseudomonas aeruginosa NK2, Acinetobacter venetianus NK3, Pseudomonas guguanensis NK4, Alcaligenes faecalis B29 and Pseudomonas aeruginosa C1A (Fig.1).

The six strains were paired with each other to carry out antagonistic experiment. The antagonistic effect of the obtained 15 combinations were determined by measuring the inhibition zones between colonies, and the result was shown in Table S3. The colonies of the four groups of strains, NK1 and NK2, NK1 and NK3, NK1 and B29, and NK2 and NK3, had antagonistic effect between each other. And there were no antagonism between the other 11 groups of strains. The combination of antagonistic strains was removed according to the combination of three strains. In total, eight combinations were involved, including NK2/NK4/C1A, NK2/NK4/B29, NK2/C1A/B29, NK3/NK4/C1A, NK3/C1A/B29, NK4/B29/C1A, NK1/NK4/C1A, and NK3/NK4/B29. Subsequently, by comparing the degradation ability of the eight conbinations using diesel oil as the substrate, the efficient degradation conbination was selected.

Fig. 2a shows the degradation rates of both the compound agents and single bacterial strains. The degradation rates of NK1, NK2, NK3, NK4, B29 and C1A to oil-bearing water at 1 g/L were 59.99%, 45.61%, 23.49%, 51.95%, 33.53% and 62.66%, respectively. The combination of NK2/NK4/C1A, NK2/NK4/B29, NK2/ C1A/ B29, NK3/ NK4/ C1A, NK3/ C1A/ B29, NK4/ B29/ C1A, NK1/ NK4/ C1A, and NK3/ NK4/ B29 were 80.71%, 51.73%, 49.46%, 69.79%, 47.45%, 85.91%, 46.76% and 51.21%, respectively. Among them, the NK4/ B29/ C1A group had the highest degradation efficiency. The percentage of diesel degradation was higher than that reported by Ganesh and Lin [27], they reported a maximum percentage of diesel degradation of 80% using isolated microorganisms. Though the degradation rate of our bacteria is slightly lower than the consortia reported by Tiralerdpanich P [28], it important to consider that the inoculation amount of diesel was half in our study. This result demonstrated the compound bacteria had an enhanced degradation ability than single bacterial strain. In specific cases, combining multiple strains could enhance the degradation performance of bacterial strains by leveraging synergistic effects, where different strains promote each other in the process of degrading pollutants [29]. It could also be understood that there existed a complementary relationship between the metabolic forms of strains [30]. The metabolic function of compound bacterial relied on both the metabolic pathway of single strains and the metabolite exchange among the strains. The compound bacteria that can self-regulate and adapt during the degradation have a better degradation efficiency, which may explain why the NK4/ B29/ C1A had a better performance in the diesel oil degradation than the other compound bacteria in the study.

The residual products of diesel oil degraded by the compound bacteria were detected by GC-FID, and the TPH of the degradation products was divided into C10–C22. In this experiment, C10–C22 was divided into two groups for analysis, i.e., C10–C16 (low molecular semi-volatile) and C17–C22 (stubborn non-volatile). Among them, C17–C22 was polymer hydrocarbons and was not easy to biodegrade. Fig. 2b shows the change of diesel component proportion at different degradation times. With the increase of degradation time, the proportion of low carbon components in degradation products gradually decreased to zero, and the proportion of long carbon chain components increased. This indicated that the short chain TPH components were degraded prior to the long chain components during the diesel oil degradation. In addition, changes in the proportion of C10–C16 and C17–C22 components were calculated, and the results are shown in Table S4. As degradation time increased, the proportion of low-carbon components in the degradation products gradually decreased to zero, while the proportion of long-carbon chain components increased. The biodegradation of diesel oil pollutants involve sequential metabolic reactions catalyzed by enzymes. The degradation efficiency is depended on both the enzymatic activity, and the solubility, mobility and bioavailability of pollutants. Thus, our compound bacteria may preferentially degrade the short-chain TPH as the solubility and bioavailability of volatile components are stronger than that of non-volatile components. This result further confirmed that the compound bacteria degraded short-chain TPH components of diesel oil pollutants first [31].

Microbial degradation of diesel oils is a complex reaction process involving both microbial and environmental factors. Efficient functional TPH-degrading strains alone are insufficient; favorable environmental conditions for microbial growth, reproduction, and metabolic activities are also essential. Therefore, in order to improve the efficiency of bioremediation of TPH-contaminated sites, the influence of environmental factors on the degradation of TPH by the compound bacteria (NK4/ B29/ C1A) should be necessarily explored. Fig. 3a presents the effect of pH on the degradation of diesel oil by compound bacteria. As shown in the figure, under different pH conditions, the degradation of diesel oil increased with time. At the same time, when the pH was 7, the residual amount of diesel oil after 28 d was 513.7 mg/L, and the corresponding degradation rate was 90.02%. However, the degradation rate of diesel oil significantly decreased when the pH was below 6 or above 9. This effect became more pronounced over time, as the growth of bacterial strains and the enzymatic activity may be inhibited in acidic and alkaline conditions. This result indicating the preferred pH of compound bacteria was pH 7–8.

Suitable temperature is important for bacterial growth and activity. On the one hand, the growth of many strains were slowly at low temperature, which was not favor for degradation due to the lack of enough biomass. On the other hand, many strain could not grow under high temperature because of the cell protein denaturing. Fig. 3b shows the influence of temperature on the degradation of diesel oil. At Day 7, when the temperature increased from 15 °C to 40 °C, the degradation rate of diesel oil increased correspondingly from 43.48% to 66.68%. This occurred might because the microbial biomass in the reaction system accumulated with time, and the activity of microorganisms was increased at suitable temperature. Many bacterial strains grew slowly at low temperature. At the same time, the solubility of diesel oil increased with the increase of temperature, and the easily degradable components were degraded by the compound bacteria in the early stage of degradation [32]. However, over time, the degradation rate of diesel oil initially increased and then decreased after 14 d. This trend might be attributed to higher temperatures reducing the activity of microbial enzymes or the inactivation of certain active substances, which hampered microbial growth and metabolism [33].

The varying amounts of inoculation resulted in differentiate loads of contaminated substrate shared by each cell in the system. Fig. 3c presents the degradation rate of diesel oil at different inoculations. With the increase of inoculant inoculations, the degradation efficiency of diesel oil also increased. When inoculations amount was 20%, the degradation rate of diesel oil reached 98.40% after 28 d.

Besides, it should be noted that when the inoculation amount was greater than 6%, the degradation efficiency of compound bacteria were only slightly increased. This occurred might because larger amount of bacteria would consume more nutrient [34]. Although the degradation effect was good in a short period of time, the lack of sufficient nutrients would still affect the degradation effect over time. As shown in Fig. 3d, the degradation amount and degradation rate of the compound bacteria presented different trends under different diesel concentration conditions. With the increase of diesel concentration, the degradation amount of diesel oil increased, but the degradation rate decreased. In other words, with the increasing concentration of diesel oil in the system, the degradation ability of microorganisms was restricted Besides, it might also be possible that the high concentration of diesel oil obstructed the oxygen exchange between the water body and the air, blocking the contact between the bacteria and oxygen [35].

In situ groundwater microbial remediation is conducted as follows: firstly, the compound bacteria is added into the aquifer through the remediation well, then they migrate to the groundwater pollution plume under the combined action of environmental and microbial factors, and finally degrade the pollutants [36]. The information of microbial migration in aquifers can be obtained through one-dimensional sand column experiment. The migration performance of bacteria is evaluated by penetration curve, which is measured by the ratio of bacterial concentration flowing out of sand column to flowing into sand column (C/C₀). The greater value of the C/C₀ indicates the greater the migration capacity of bacteria under this aquifer condition [37]. At the ionic strength of 1 mM CaCl₂, the penetration curve of the compound bacteria under different medium particle sizes and injection flow rates is shown in Fig. 4a. In coarse sand, the plateau C/C₀ of the penetration curve was 0.97, 0.92, and 0.89 at flow rates of 3.75 cm/min, 2.25 cm/min, and 0.75 cm/min, respectively. For medium sand, it decreased to 0.93, 0.90, and 0.88. In fine sand, the plateau C/C₀ dropped to 0.87, 0.83, and 0.80. The penetration curves of the compound bacteria with 50 mM CaCl₂ ion strength under different medium particle sizes and injection flow rates are shown in Fig. 4b. In coarse sand, the plateau C/C₀ of the penetration curve at flow rates of 3.75 cm/min, 2.25 cm/min, and 0.75 cm/min was 0.88, 0.81, and 0.73, respectively. For medium sand, it decreased to 0.77, 0.69, and 0.57. In fine sand, the plateau C/C₀ dropped to 0.67, 0.60, and 0.49. At the ionic strength of 100 mM CaCl₂, the penetration curve of the compound bacteria under different medium particle sizes is shown in Fig. 4c. The penetration curve plateau C/C₀ at pore flow rates of 3.75 cm/min, 2.25 cm/min and 0.75 cm/min was 0.78, 0.71 and 0.63, respectively. In medium sand medium, C/C₀ decreased to 0.70, 0.62 and 0.54, respectively. In fine sand medium, C/C₀ decreased to 0.60, 0.51 and 0.39, respectively.

These exprimental results indicated that medium particle size and flow velocity served as the main factors restricting the migration of compound bacteria in aquifers. The larger the particle size of porous media, the stronger the migration ability of microorganisms in it [38]. Higher flow rates generally result in stronger shear forces between water flow and the medium, providing kinetic energy for microorganisms and enhancing their migration capacity, thus reducing retention in the medium [39].

However, based on the experimental results, the change of ionic strength also had a great impact on the migration performance of compound bacteria in aquifers. Thus, the mechanism were further investigated by hydrodynamic diameter, zeta potential, adhesion efficiency and the deposition rate experiments. Malvern laser particle size analyzer was used to test the hydrodynamic diameter of the compound bacteria at three CaCl₂ ionic strengths (1 mM, 50 mM and 100 mM), and the particle size distribution results of the compound bacteria are shown in Fig. 5. The median particle size d (0.5) of the compound bacteria was 0.757 μm at the ionic strength of 1 mM, 0.944 μm at the ionic strength of 50 mM, and 1.031 μm at the ionic strength of 100 mM. The increase in ionic strength from 1 mM to 100 mM led to an increase in the median particle size of the compound bacteria from 0.757 μm to 1.031 μm. This could be attributed to the compression of the double electric layer due to the higher ionic strength, resulting in reduced electrostatic repulsion between microorganisms. As a result, microorganisms were more likely to approach each other, forming larger aggregates. The enhancement of electrostatic effect on the surface of microorganisms contributed to the stability of these aggregates, leading to the adsorption or agglomeration in aquifers. Hence, the migration performance of compound bacteria became weakened at high ionic strength and the diameter of which was increased [40].

The Zeta potential of the compound bacteria at different ionic strengths is shown in Table S5. The results showed that with the increase of ionic strength, the Zeta potential of the compound bacteria gradually decreased. When the ionic strength increased from 1 mM to 100 mM, the Zeta potential of the compound bacteria decreased from −27.7±0.79 mV to −10.5±0.7 mV. This occurred because when the ionic strength increased, these ions were adsorbed on the surface of the microorganisms. Besides, a thick electric double layer was formed, resulting in a lower surface charge density [41].

Further, the adhesion efficiency (α) and deposition rate (K_d) at different ionic strengths were calculated, and the results are shown in Fig. 6. Under high flow velocity conditions, the deposition rate of fine sand reached its peak, and it exhibited the most substantial fluctuations as the ion strength increased. When the ion strength increased from 1mM to 100 mM, the deposition rate of the compound bacteria increased by 0.0039 s⁻¹. On the contrary, the deposition rate of coarse sand was found to be the smallest under slow flow conditions and demonstrated the least responsiveness to variations in ionic strength. When the ionic strength increased from 1 mM to 100 mM, the deposition rate of the compound bacteria only increased by 0.00068 s⁻¹. The α and K_d of the compound increased nonlinearly with the increase of ionic strength. α was introduced to mitigate the adhesion between nanoparticles and porous media, and it was directly related to the potential energy that governed the interaction between the nanoparticles and the aquifer materials [42]. With the increase of ionic strength, the double electric layer between the compound bacteria and the quartz sand medium was compressed, which reduced the electrostatic repulsion and promoted the deposition of the compound bacteria. For K_d in porous media, under a consistent ionic strength environment, fine sand demonstrated a higher deposition rate within the media compared to medium and coarse sands. Moreover, the deposition rate was higher at elevated fluid flow rates than at lower ones [43].

Following the migration experiment, the sand column was sectioned at intervals of 1.5 cm to identify and analyze the types of microbial bacteria retained at various depths within the column. The experimental results are shown in Fig. 7. As shown in the figure, the allocation proportion of the three bacterial strains were different at different locations. Pseudomonas aeruginosa (C1A) had strong migration performance in the aquifer, followed by Alcaligenes_faecalis (B29) and Pseudomonas_guguanensis (NK4), and the retention abundance changed with the medium conditions. This might be attributed to the morphology of the bacteria itself. In fine sand medium, when the ion strength was 1mM, the injection rate was 0.75 cm/min, and the proportion of Pseudomonas_guguanensis decreased from 27.52% to 6.06% after passing through the 15 cm sand column. The proportion of Pseudomonas aeruginosa increased from 22.14% to 77.69%, while that of Alcaligenes_faecalis decreased from 41.11% to 14.05% (Fig. 7a). In the coarse sand medium, the proportion of Pseudomonas_guguanensis decreased from 30.94% to 8.76%, while that of Pseudomonas aeruginosa increased from 34.28% to 56.62 when the ion strength was 100 mM and the injection rate was 2.25 cm/min. The proportion of alcaligenes_faecalis increased from 34.46% to 34.54% (Fig. 7b). In comparison with the fine sand medium, compound bacteria exhibited a more evenly distributed migration pattern in the coarse sand medium. However, their proportion at each location were still differentiated from the original distribution.

Thus, when using compound bacteria for in-situ groundwater remediation, appropriate control measures should be adopted to ensure that various bacterial strains in the compound bacteria could be evenly distributed in the aquifer medium as far as possible, so as to maximize the removal efficiency of pollutants.