Introduction

The combustion of fossil fuels generates polluting compounds such as sulfur and nitrogen dioxides (SO2, NO2). These compounds are harmful to human health and to the environment. Therefore, environmental regulations around the world have been demanding a reduction in the sulfur and nitrogen contents, especially for sulfur, requiring concentrations lower than 15 ppm [1]. The attainment of fuels with ultra low polluting content is becoming a problem for the petrochemical industry, since the existing catalytic hydrogenation processes cannot achieve the required removal levels because it is not capable of removing completely the heterocyclic sulfur and nitrogen compounds [1–6]. Another issue is that the main part of the crude oil extracted worldwide is progressively exhibiting higher levels of sulfur content and heavier density values [2].

An alternative technology, called biodesulfurization (BDS), has been developed in the latest decades to replace and/or complement the conventional process. In BDS technique, degradation of sulfur occurs through the action of microorganisms that act as catalysts removing selectively the sulfur from heterocyclic compounds under mild temperatures and pressures [2, 7, 8].

The industrial application of BDS depends on several factors; one of them is the selection of microbial strains with the ability to remove sulfur from fossil fuels without decreasing the energetic capacity, while industrially and environmentally stable. Numerous microorganisms belonging to different genera have demonstrated the ability to remove the sulfur present in heterocyclic compounds such as Brevibacterium sp. [9], Corynebacterium sp. [10], Rhodococcus sp. [11], Paenibacillus sp. [12], Pseudomonas sp. [13], Gordonia sp. [14], and Bacillus sp. [15]. However, the products obtained at the end of the BDS process are not always the same and almost none being economically advantageous [16, 17].

The strain Rhodococcus sp. has been widely applied to BDS process because it is able to remove sulfur from heterocyclic compounds, turning it into sulfate, without breaking the aromatic rings, thus, keeping fuels energetic capacity [7, 18, 19]. Another great advantage of this bacterium is its hydrophobic characteristics which increases the contact between the cell and the sulfurous compounds in the fuels [20]. When compared with other strains, Rhodococcus sp. presents a greater specific degradation rate and stability during BDS process [11, 16, 21–27].

Similarly, the biodenitrogenation (BDN) is a process where nitrogen is removed from the heterocyclic compounds through the action of microorganisms. However, the development of this process did not advance because the microorganisms capable of nitrogen removal use a degradation pathway where only liberate nitrogen in the course of completely degrading the substrates [8].

For practical purposes, development of BDS and BDN requires the use of real products and byproducts from petroleum distillation such as diesel, gasoline, and gas oils [11, 22–24, 27–30]. The heavy gas oil (HGO) is an intermediate fraction obtained from the vacuum distillation used in the production of diesel and some lubricants. The desulfurization and denitrogenation of this fraction could result in the reduction of subsequent treatment processes, increasing the industrial and economic viability of the BDS and BDN [31]. Based on these aspects, the main objective of this work was to evaluate the desulfurization and denitrogenation ability of Rhodococcus erythropolis ATCC 4277 in a batch reactor with different HGO concentrations.

Materials and methods

Materials

The heavy gas oil used was kindly denoted by Petrobras S.A., from the Marlim Sul P-56 platform, Rio de Janeiro, Brazil.

Bacterial strain and maintenance medium

Rhodococcus erythropolis ATCC 4277 used in this study was acquired from Tropical Foundation of Research and Technology André Tosello, Campinas, São Paulo State, Brazil. Following the recommendation of the supplier, it was maintained on Petri dishes containing Yeast Malt Agar medium (YMA) comprised 3.0 g·L−1 of yeast extract, 3.0 g·L−1of malt extract, 5.0 g·L−1of bactopeptone, 10.0 g·L−1of glucose, 2.0 g·L−1 of calcium carbonate, and 20.0 g·L−1of agar, and stored at 4 °C for further use [31].

Minimum inhibitory concentration (MIC)

Diffusion method

In Petri plates containing YMA medium, previously inoculated with a microbial suspension of 4.0 (±0.5) g·L−1, wells of 1.0-cm diameter along the medium were made. The wells were filled with HGO and other compounds to be tested (diesel, crude oil, and atmospheric residue), and then the plates were left in the incubator for 48 h at 28 °C. The MIC was considered that concentration and/or pure compound are capable of developing an inhibition halo greater or equal to 1.4-cm diameter on the microbial growth [32, 33].

Broth dilution MIC—macrodilution

The broth macrodilution method considers the relationship between the growth of the microorganism in the liquid medium and the concentration of the tested substance. The evaluation is performed by comparing microbial growth for each concentration tested against a biological standard. The assays were conducted in flasks containing different concentrations of YMA medium and HGO, as shown in Table 1. Afterward, it was added 5 mL of a Rhodococcus erythropolis preculture with an absorbance of 0.75 at 600 nm (OD600). The flasks remained in an orbital shaker for 24 h at 28 °C and 200 rpm.

Table 1 Test conditions for MIC determination by macrodilution method
Full size table

The MIC was calculated using Eq. (1), where the final cell concentration in each flask was measured throw turbidimetry at 600 nm (OD600) and compared to a standard cell concentration without addition of HGO [33].

$$\% {\text{ cells growth}} = \frac{{[X]_{\text{final}} - [X]_{\text{standard}} }}{{[X]_{\text{final}} }} \times 100$$
(1)

Culture conditions

The R. erythropolis cells were grown in 0.125-L flasks containing 0.05 L of YM medium for 18 h on an orbital shaker at 150 rpm and 28 °C. Then, 5 % (v/v) of this preculture, at an absorbance of 0.85 at 600 nm (OD600), was transferred to 0.5-L flasks containing 0.1 L of YM medium. Those flasks remained on an orbital shaker during 18 h at 150 rpm and 28 °C. The cells were recovered by centrifugation at 5000×g for 20 min at 4 °C, discarding the supernatant and washed with 0.1 M phosphate buffer (pH 7.0) [34].

Biodesulfurization and biodenitrogenation

The BDS and BDN assays were carried out in 0.5-L flasks containing 1.0 g·L−1 of cells with a total volume of reaction medium of 0.2 L. The reaction medium consisted of two phases: an aqueous (YM medium) and other organic (HGO) at different concentrations, as described in Table 2. The biocatalytic system remained on an orbital shaker for 18 h at 28 °C and 200 rpm [11]. All experiments were performed in triplicate so that the values presented in tables and figures are in fact average of three replicates.

Table 2 Test conditions for desulfurization and denitrogenation of HGO by Rhodococcus erythropolis ATCC 4277 whole cells
Full size table

Sulfur and nitrogen contents determination

After submitted to BDS and BDN processes, the content of the flasks was centrifuged at 5000×g at 4 °C for 20 min and then the sulfur and nitrogen contents in the HGO were determined by an elementary analysis CHNS. A flash elemental analyzer (C.E. Elantech, Inc., model 1112) is used for bulk CHNS weight percent determination of the HGO with triplicate samples employed.

Optical microscopy

At the end of BDN and BDS processes, samples were collected from the reaction media containing 40 and 100 % (v/v) of HGO and analyzed by microscopy (BIOVAL) by increasing 400 times the extent of the objective lens (Olympus, model EA40) and ocular (Olympus, model PL 100/1.25).

Results and discussion

Minimum inhibitory concentration (MIC) of HGO

The MIC assays by diffusion method served to determine qualitatively the ability that R. erythropolis ATCC 4277 has to grow in the presence of HGO and other petroleum fractions. When tested by the diffusion method, inhibition halo was not observed for any of the compounds, as noted in Fig. 1.

Fig. 1
figure 1

Determination of MIC through diffusion method for a diesel; b atmospheric residue; c petroleum; d heavy gas oil

Full size image

The adaptability of R. erythropolis to high concentrations or pure HGO is due to the fact that these bacteria modulate the viscosity of cell membrane lipids to maintain or increase the fluidity of the same by reducing the degree of saturation, increasing the ratio cis/trans of fatty acids and by increasing the relative amount of branched fatty acids [35, 36].

Tests of MIC by macrodilution method aimed to determine quantitatively the concentration of HGO that inhibits the growth of R. erythropolis. Results shown in Fig. 2 confirm the data obtained by the diffusion method, hence demonstrating that the growth of R. erythropolis suffers no inhibition in the presence of HGO.

Fig. 2
figure 2

Determination of MIC through macrodilution method for different HGO concentrations

Full size image

Rather, the presence of HGO in the culture medium promotes a higher cell growth, especially for HGO concentrations from 10 to 50 % (v/v) (Fig. 2). Other authors also had a good cell adaptation when using approximately 20 % (v/v) of organic phase [12, 15, 24, 25]. The presence of hydrocarbons in abundance, as well as nitrogen and sulfur compounds, contributes to a better nutrition of the microorganism acting as a substrate and a co-substrate. The slight decrease in cell concentration for higher concentrations of HGO is due to the resistance of the microorganism decreases with the increase of the toxicity of the system, thus, making it impossible to maintain the cell viability under more severe conditions [4, 37, 38].

Biodesulfurization

Heavy gas oil samples, after subjected to desulfurization process, were analyzed by CHNS, with results of this elementary analysis reported in Table 3.

Table 3 Sulfur content in HGO after submitted to BDS process with Rhodococcus erythropolis ATCC 4277
Full size table

The best results were achieved for HGO ratios of 40, 20, and 60 % (v/v) where the mass of sulfur removed per gram of (266 ± 9) × 10, (246 ± 3) × 10, and (221 ± 2) × 10 µg, desulfurization rate of 148, 137, and 123 mg sulfur kg−1 h−1, and removal percentages of 42.7, 39.4, and 35.5 % (Fig. 3), respectively.

Fig. 3
figure 3

Denitrogenation and desulfurization percentage for the different HGO concentrations tested in a batch reactor containing Rhodococcus erythropolis ATCC 4277 free cells

Full size image

Since the HGO contains various easily desulfurizable compounds, such as thiols and sulfides, besides recalcitrant compounds as DBT [14], the mass of sulfur removed and the desulfurization rate obtained is superior to those reported by other authors for the same ratio of organic phase in the BDS process [11, 21–24, 27, 32]. In those researches, synthetic fuel was used containing only DBT or oils with low sulfur content decreasing this way the offer of sulfur. Others authors who worked with light gas oil and heavy gas oil also observed an increase in the desulfurization activity when compared with model oils [2, 14, 29].

Although R. erythropolis has a hydrophobic characteristic that increases the contact between cells and the sulfurous compounds, the high viscosity of HGO may have difficulted the access to this compounds leading this way to a minor removal percentages [15, 21, 39]. A solution to this problem has been developed employing surfactants to increase the contact between those two phases [2].

Biodenitrogenation

Table 4 shows the results of CHNS analysis performed for HGO samples after being subjected to BDN process.

Table 4 Nitrogen content in HGO after submitted to BDN process with Rhodococcus erythropolis ATCC 4277
Full size table

Similarly to what happened to the sulfur compounds, the best results for nitrogen removal were achieved in the biphasic systems containing 40, 20, and 60 % (v/v) HGO, where the masses of reduced nitrogen were (162 ± 1) × 10, (151 ± 2) × 10, and (117 ± 1) × 10 μg and the percentages of reduction 43.2, 40.2, and 31.2 % (Fig. 3), respectively.

These results are similar to those achieved by other nitrogen removal techniques such as oxidative denitrogenation [31]. However, the removal occurred simply by free cells of R. erythropolis without the need to add other chemical reagents or catalysts.

In order to understand what led the cells of R. erythropolis to remove more sulfur and nitrogen in the experiments containing 40 % (v/v) of HGO, and why they were not capable of removing any quantity of sulfur and nitrogen in the tests with 100 % (v/v) of HGO, both reactions media were analyzed by optical microscopy (Fig. 4).

Fig. 4
figure 4

Fluorescence microscopy of Rhodococcus erythropolis ATCC 4277 cells after 18 h in a batch reactor containing a 40 % (v/v) and b 100 % (v/v) HGO

Full size image

From Fig. 4a, it can be noted the presence of a considerable amount of refringent cells, which is an indication that the cells of R. erythropolis ATCC 4277 were able to adapt well to the reaction medium containing 40 % (v/v). Moreover, cells are dispersed in the medium, which is also an indication that the medium provides favorable conditions for the microorganism. A good adaptation of the cells to this particular reaction medium is also observed in the result of MCI test by macrodilution method (Fig. 2) [35, 37–40].

A nutritional deprivation growth condition increases cell surface hydrophobicity, which acts as an ignition power for a cell–cell junction or biogranulation. Adhesion occurs when the electrostatic repulsion is overcome by van der Waals forces and hydrophobic interactions [40]. Furthermore, the cells form an aggregate to protect cell population against the toxicity of hydrophilic compounds, since this microorganism resistance decreases with the increase of the system toxicity, restraining the cells to maintain their viability under severe conditions. Thus, these factors have influenced the ability of desulfurization and denitrogenation of R. erythropolis ATCC 4277, since the cells are aggregated and there is no evidence of refringent cells, as shown in Fig. 4 [35, 37–40].

Conclusions

The strain R. erythropolis ATCC 4277 was able to grown in every concentration of heavy gas oil tested. Different concentrations of heavy gas oil in the biocatalytic process affected the desulfurization rate, and mass of nitrogen and sulfur removed. The best desulfurization result achieved was with water–oil rate of 40 % (v/v) where the mass of sulfur removed per gram was of 266 (±9) × 10 µg, desulfurization rate of 148 mg sulfur kg−1 h−1, and removal percentage of 42.7 %. Similarly the greatest results to BDN process were in the systems using 40 % (v/v) of HGO where 162 (±11) × 10 μg, and the percentages of reduction 43.2 %. This strain was able to remove this quantity of pollutants from a real fraction of petroleum in the presence of a high concentration thereof. This represents an advance in the development of BDN and BDS processes in order to make it economic, environmental, and industrially viable.