Enhanced Biodesulfurization with a Microbubble Strategy in an Airlift Bioreactor with Haloalkaliphilic Bacterium Thioalkalivibrio versutus D306.

Biodesulfurization under haloalkaline conditions requires limiting oxygen and additional energy in the system to deliver high mixing quality control. This study considers biodesulfurization in an airlift bioreactor with uniform microbubbles generated by a fluidic oscillation aeration system to enhance the biological desulfurization process and its hydrodynamics. Fluidic oscillation aeration in an airlift bioreactor requires minimal energy input for microbubble generation. This aeration system produced 81.87% smaller average microbubble size than the direct aeration system in a bubble column bioreactor. The biodesulfurization phase achieved a yield of 94.94% biological sulfur, 84.91% biological sulfur selectivity, and 5.06% sulfur oxidation performance in the airlift bioreactor with the microbubble strategy. The biodesulfurization conditions of thiosulfate via Thioalkalivibrio versutus D306 are revealed in this study. The biodesulfurization conditions in the airlift bioreactor with the fluidic oscillation aeration system resulted in the complete conversion of thiosulfate with 27.64% less sulfate production and 10.34% more biological sulfur production than in the bubble column bioreactor. Therefore, pleasant hydrodynamics via an airlift bioreactor mechanism with microbubbles is favored for biodesulfurization under haloalkaline conditions.

Introduction

Hydrogen sulfide (H2S) is a common poisonous gas found in natural gas, biogas, and industrial waste streams.[1,2] Poisonous gas generally deactivates catalysts by chemisorption on catalyst surfaces; this process is known as catalyst poisoning. This is an important problem, resulting in high costs of catalyst regeneration and replacement. Moreover, H2S separation and control processes are necessary to reduce the pollution before releasing it into the environment.[3−5] Generally, there are systematic approaches, such as chemical processes, to remove and treat H2S in the industry, but these processes result in environmental issues and large amounts of energy consumption. An alternative route is biological desulfurization, also known as biodesulfurization; it is a more environmentally friendly option that reduces production costs and currently is generating considerable interest from the industry.[6−9]

The biodesulfurization process uses biological and engineering techniques to solve the abovementioned issue by converting H2S to elemental sulfur. The inorganic sulfur compounds can be diminished by introducing sulfur-oxidizing bacteria (SOB) such as Thioalkalivibrio versutus in the biological process. The SOB can remove sulfide from the sulfide-contaminated sources by converting it in solid sulfur via a limited oxygen route in the biological sulfur-oxidizing mechanism under haloalkaliphilic conditions.[10−13] Sulfide can then be oxidized and defecated from SOB as elemental sulfur via flavocytochrome c sulfide dehydrogenase, a larger flavoprotein subunit (fccB). The produced sulfur is surrounded by sulfur globule protein B (sgpB).[14] Thiosulfate in a sulfur-oxidizing system with a SoxAXYZ pathway also can be oxidized to sulfane sulfur and defecated out of the cell as elemental sulfur. Sulfane sulfur from SoxAXYZ and fccB can be cytoplasmically trafficked and oxidized to sulfite using dsrE-rhd-tusA sulfur transferases and hdr, respectively.[15] The thiosulfate-SoxAXYZ complex with SoxB can be further oxidized to sulfate; however, thiosulfate without the SoxAXYZ pathway also can be oxidized to sulfane sulfur via the tsdBA pathway.

The biodesulfurization process requires optimal operating conditions to be efficiently processed. SOB are able to grow under haloalkaliphilic conditions and survive under ambient temperature and ambient pressure without difficulty.[16−18] Thus, culturing T. versutus bacteria is suitable for fast cultivation and reproduction; moreover, the biodesulfurization process using SOB is significantly controlled by the oxidation–reduction potential (ORP) to keep the biological system steady in the bioreactor.[13,19,20] However, hydrodynamic parameters have a greater influence on the biodesulfurization system as mass transfers and mixes between the gaseous, liquid, and microbial states.

According to the limitations of biodesulfurization (the issues of mixing and origination under haloalkaline conditions), the oxygen mass transfer is impeded by the high salinity of the medium solution.[21] Currently, microbubbles are an interesting option for the biological process. Large numbers of industries worldwide have used and developed their process to incorporate the microbubble system.[22−24] Microbubbles have gas–liquid mixing advantages that play an important role in the biodesulfurization process compared with the standard-sized bubbles.[25−27] Studies[28−30] reported the mass transfer coefficient and the diffusion coefficient of small bubbles and reported that they have two-thirds or higher power over the large bubbles. The combination of film and surface renewal theories by Dobbins[31] was supported.

Conventional approaches for generating microbubbles are compressed jet flow to a micropore distributor, ultrasonication, and fluidic oscillation.[32] The compressed jet flow method consumes the highest amount of energy for generating microbubbles. The ultrasonication method requires expensive equipment but consumes less energy than the compressed jet flow method. The fluidic oscillation method consumes the least amount of energy compared with the abovementioned methods; however, the difficulty with this method is in designing and developing the fluidic oscillator.[33] Therefore, the fluidic oscillation method is suitable for achieving the desired microbubble generation results.

The fluidic oscillator comprises a stationary chamber device and a feedback loop connected to the middle tubes of the stationary device to generate the fluidic oscillation. The mechanism uses the oscillation frequency in the chamber and channels to generate a unique pattern of microbubbles. The length of the feedback loop controls the desired microbubble size, for example, high harmonic oscillation in the chamber generates miniature microbubbles.[34] Microbubble generation via an oscillation aeration system has a shoot expulsion (similar to a bullet discharging). However, the expulsion from the direct aeration system, also defined as the conventional aeration system, is similar to a slow pushing out from a pore. Moreover, the size of the microbubbles is influenced by the wetting properties of the sparger surface.

The conventional bioreactors for biodesulfurization are the bubble column and airlift bioreactors. The bubble column bioreactor is a classic pneumatic reactor in which the liquid mixing is random. However, the optimized design of the airlift bioreactor takes advantage of a gaseous stream injection to provide well mixing, mediate the transfer of gaseous substances to the liquid phase, and support higher mass transfer rates per amount of energy input. The transfer efficiency is considerably less affected by the energy input compared to the classic pneumatic reactors.[35] The airlift bioreactor provides liquid circulation between the interconnected zones (the riser zone and downcomer zone). Typically, by injecting the gas in the bottom of the reactor, below the riser section, a mean density gradient is created causing the liquid broth to circulate. The gradient’s circulation between the average fluid densities provides no focal point for the energy dissipation and has homogenous shear forces in each section. This causes less cellular stress and is more suitable for biochemical processes.[36−39] Moreover, the gas–liquid circulation provides the fraction of gas, known as the gas holdup, to be introduced in the downcomer section, increasing the mass transfer efficiency. The gas holdup significantly affects fluid dynamics and bioreactor performance.[39−41] Furthermore, the bioreactor performance depends on the structural design and operational variables.

Airlift bioreactor structures are separated into two categories: the external loop and internal loop airlift reactors.[42] The external loop airlift reactors have separated parts for riser and downcomer zones. This provides higher liquid circulation velocities but lower mass transfer rates. The external loop system is a less versatile design and has supported fewer applications compared to the internal loop.[43−45] In biological processes, the low shear stress and the high mass transfer rates of gas and mediums are required, and the external loop airlift reactor could not provide these.[46] Thus, for the biodesulfurization process, the use of the internal loop airlift bioreactor seems highly important and more suitable.

This study develops a promising progressive biodesulfurization strategy by reducing the required energy input and enhancing the mixing mechanism of the airlift bioreactor. The results obtained from this study are valuable for providing novel insights about the airlift bioreactor hydrodynamics and biodesulfurization clarifications for additional research and commercialization.

Materials and Methods

Inoculum and Medium

The SOB, T. versutus D306, isolated from Soda Lake, was used in this work. It was cultured in a sulfur-oxidizing medium, called thiosulfate-defined medium (TD medium); this medium was composed of Na2S2O3·5H2O (19.82 g/L), NaHCO3 (58.80 g/L), NaOH (5 g/L), NH4Cl (0.268 g/L), KNO3 (0.505 g/L), K2HPO4 (2 g/L), MgCl·6H2O (0.1 g/L), and MSM (200 μL/L).[13] The seed culture was created by adding 5 mL of the bacteria to 100 mL of TD medium. NaHCO3 and NaOH acted as buffer solutions for this medium to stabilize the pH range. Biological cultivation took 72 h in an incubator at 180 rpm and 30 °C.

Experimental Section

Airlift and bubble column bioreactors had similar structures except that the latter does not possess a draft tube in the bubble column bioreactor; both bioreactors had 8.5 L of working volume and 11 L of total volume. The reactors comprised of translucent plastic, and their inner diameter and height were 19 and 40 cm, respectively. The draft tube of the airlift bioreactor was positioned above the bottom of the reactor and had the following dimensions: a diameter of 14 cm, a height of 14 cm, and a thickness of 0.5 cm. The section contained four small legs to attach the draft tube with the reactor. Figure a shows the dimensions of the model. Figure b shows the 3D model of the fluidic oscillator, which was made of translucent stereolithography grade plastic using the 3D printing method.

Figure 1

Bioreactor specifications and the fluidic oscillator model. (a) Airlift bioreactor model dimensions for biodesulfurization via T. versutus D306. The inlet is attached to the sparger for microbubble generation, the inner-middle section of the airlift bioreactor is the riser section, and the outer side is the downcomer section for gas–liquid recirculation. The term H is height, DR,0 is the outer diameter of the bioreactor, DR,i is the inner diameter of the reactor, Ddt,0 is the outer diameter of the draft tube, and Ddt,i is the inner diameter of the draft tube. (b) Fluidic oscillator model for microbubble generation via fluidic oscillation aeration. The details of a fluidic oscillator chamber have been reported in a study.[34] The inner diameters of each inlet, outlet, and control terminal are 6, 6, and 5 mm, respectively.

The length of an oscillation feedback loop was 100 cm, and the stationary part had the following dimensions: an inlet diameter of 0.6 cm, an outlet diameter of 0.6 cm, control terminals of 0.5 cm, and a thickness of 6 cm, and the complex details were reported by Tesař et al.[34] A fluidic oscillator was installed in the system to generate microbubbles via fluidic oscillation aeration. The following measuring equipment was used: an oxidation–reduction potential (ORP) probe (Pt4805-DPAS-SC-K8S/255, Mettler Toledo, Switzerland), a dissolved oxygen (DO) probe (InLab 605-ISM, Mettler Toledo, Switzerland), a pH probe (405-DPAS-SC-K8S/225, Mettler Toledo, Switzerland), and a conductivity probe (InLab 731-ISM, Mettler Toledo, Switzerland). All probes were attached at the top to measure the conditions at the middle of reactors. Furthermore, the probes were connected to the bioreactor controller (BioFlo/CelliGen 115, New Brunswick Eppendorf, Germany) and multi-channel transmitter (M300, Mettler Toledo, Switzerland).

Table provides an overview of the desired system conditions of the experimental setup. The aeration system of the process was generated via an air compressor, rotameters, a fluidic oscillator, and a microporous ceramic flat plate sparger for microbubble generation. Figure shows the experimental setup scheme.

Table 1

Overview of the Process Conditions in the Experimental Setupa

parameter	value
thiosulfate loading, (g/L)	6.0769
thiosulfate loading rate (ten-fold), (L/h)	0.15
salinity, (M Na+)	1.0
carbonate alkalinity, (M)	3.0
pH set-point	9
temperature, (°C)	25
ORP set-point, (mV)	53
DO, (mg/L)	<0.02
conductivity, (mS/cm)	53

The biodesulfurization conditions in this study were set up and operated under a haloalkaline condition via an oxygen limiting route.

Figure 2

Experimental setup schematic throughout the biodesulfurization process, including the preparation and analysis. T. versutus D306 with TD medium in an incubator is the culturing process for the inoculum and the bioreactor energy source. Compressed air flows via the fluidic oscillator and rotameters for an oscillation aeration system. Probes for ORP, DO, pH, and conductivity are connected with the bioreactor controller. Ion chromatography is utilized for sample analysis of the biodesulfurization process.

Thioalkalivibrio versutus D306 Mechanisms and Biodesulfurization Procedures

The biodesulfurization using T. versutus D306 was performed under haloalkaliphilic conditions via the limiting oxygen route of the thiosulfate-oxidizing mechanism.[18,47,48] In biological investigation studies and articles, sulfur salts are an established representative reactant of H2S to avoid inhaling dangerous chemicals and maintaining safety practices.[49−51] Thiosulfate was selected for this work because of the previously mentioned reasons. The limited oxygen pathway is shown in Figure . The biological reactions started to convert thiosulfate to the primary products such as elemental sulfur and sulfate via the SoxAXYZ and SoxAXYZ with SoxB pathways, respectively. Tetrathionate was generated by the thiosulfate-oxidizing system via tsdBA.[10,12]

Figure 3

Biological desulfurization on the limited oxygen route pathway scheme. This pathway shows three routes to biodesulfurize the energy source as thiosulfate (S2O32–) via the SOB, T. versutus D306. The first route is the direct SoxAXYZ pathway for S0 production. The second route is the SoxAXYZ with SoxB for sulfate (SO42–) production, and the third route is the sulfane sulfur or tetrathionate (S4O62–) production by the thiosulfate-oxidizing system as tsdBA.

First, 8.5 L of the TD medium solution was prepared as per this study’s formula. Then, the medium was stabilized and added to the bioreactor. The Thialkalivibrio versutus D306 seed culture was prepared and used to inoculate the bioreactor. The aeration system was added via an air inlet at 0.5 L/min and controlled by a flow meter to achieve the desired oxygen supply. The air flowed through the microporous ceramic flat plate sparger for microbubble generation to perform bacterial culturing in the system. After 36 h of bacterial culturing, the TD medium with highly concentrated thiosulfate (ten-fold) was added at a rate of 2.5 mL/min via a centrifugal pump to the bioreactor to perform the biodesulfurization. The microbubbles flowed upward to the riser section where the biodesulfurization reactions occur. The circulation pattern of the microbubbles occurred between the riser section and the downcomer section because of the stronger drag force, affecting the buoyancy force at the draft tube section. The drag force drives the microbubbles downward in the downcomer section. Thus, the recirculation of microbubbles in an airlift bioreactor increases the mixing rate and contact rate between the oxygen and microbial cultures in the liquid phase, enhancing the performance of the biodesulfurization process.

Analytical Methods

Measurement of Microbubble Generation

Microbubbles were observed during the experiments. The optical capturing method, using a backlight technique, was performed using a CMOS camera (5600D, Nikon, Japan) with an attached macro lens of 50–500 mm (D365, Tamron, Japan). The camera stand was placed in front of the riser and downcomer sections for collecting the bubble graphic data. The bubble’s size, shape, and distribution were computed using biological Fiji ImageJ.[52−54]

Measurement of Biodesulfurization via Thioalkalivibrio versutus D306

The biodesulfurization controlling process measurements such as pH, ORP, DO, and conductivity are listed here. The pH was measured and controlled in the favorable pH range of 8.5–10 via a pH probe placed in the middle of the reactor. The ORP was measured via the ORP probe during the biodesulfurization process. A DO probe was used to measure the DO in the range of 0.02–1 mg/L. The stability of the system was measured by using the conductivity probe. The abovementioned probes were placed at the downcomer section in the bioreactor to avoid direct contact with the microbubbles and thereby the influence from the microbubbles attaching to the riser section; the microbubble impacted the oxygen detection for the ORP and DO measurements.

The biodesulfurization result was observed via ion chromatography (Dionex model ICS-900, Dionex, United States of America). To track the thiosulfate and sulfate concentrations, the samples were examined every 6 h during the 90 h of the biodesulfurization process. The experimental data were statistically analyzed and repeated several times for accuracy and precision.

The yields and performance indicators of the biodesulfurization were calculated with the following equations

Theoretical yield was calculated as if the reaction would be converted either to sulfate or sulfur.

Results and Discussion

Microbubbles with Oscillation Aeration Effects

Microbubble generation via oscillation and conventional aeration under the actual flow rate was investigated. Figure shows the size and distribution of the microbubbles interpreted in the results. Figure a,b shows that the microbubble distribution of the oscillation aeration was an improvement over the conventional aeration system in both bioreactors.

Figure 4

Microbubble distribution: cases of airlift and bubble column bioreactors with and without an oscillation aeration system. (a) Distribution of microbubbles at the riser section in the airlift bioreactor and the bubble column bioreactor with and without the oscillation aeration system. (b) Distribution of microbubbles at the downcomer section in the airlift bioreactor and the bubble column bioreactor with and without the oscillation aeration system.

The fluidic oscillator plays an important role in controlling the size and distribution of microbubbles.[55] The microbubble distribution sizes in the bioreactors with the oscillation aeration system were substantially smaller than those of the conventional aeration method. The average microbubble size via fluidic oscillation is <100 μm. Furthermore, at the airflow of 0.5 L/min, the average size of microbubbles was reduced to 80.43 and 75.33% at the riser and to 79.92 and 69.17% at the downcomer section with the installation of the fluidic oscillator on the airlift bioreactor and the bubble column bioreactor, respectively. The microbubbles were noted to be 81.87 and 76.19% smaller on the airlift reactor than on the bubble column reactor in the conventional aeration system at the riser and downcomer sections, respectively; furthermore, the microbubbles were 26.49 and 22.77% smaller on the airlift bioreactor than on the bubble column bioreactor in oscillation aeration systems at the sections of riser and downcomer, respectively.

However, the microbubble sizes in both bioreactors with the installed oscillation system differed slightly. Figure shows the microbubbles’ sizes and shapes in the different aeration systems. The shape of the microbubbles generated via fluidic oscillation was identically spherical in both the riser and downcomer sections.

Figure 5

Photomicrograph of microbubbles in the airlift bioreactor via different aeration systems for microbubble generation. The photomicrograph is obtained via a CMOS camera using a backlight technique at the actual flowrate of 0.5 L/min in both (a) fluidic oscillation aeration system and (b) conventional aeration system.

Nevertheless, the conventional aeration demonstrated a mix of spherical and oval shapes. The size and distribution of the microbubbles in the conventional aeration system were large and diverse, contributing to chaos in controlling the system and the preferred products. Moreover, the haloalkaline condition in the biodesulfurization system inhibited the oxygen mass transfer in the medium solution, considerably reducing the oxygen used for large bubbles. However, considering a single microbubble, the contacting surface was large; it rapidly rose through the riser section with less usage of captured oxygen. The oscillation aeration generated small microbubbles, and this affected the mixing and gas holdup escalation because of the slow rising. The small microbubbles with larger total contacting surface area enhanced the oxygen distribution and hydrodynamic mixing in the bioreactors.

Bioreactors with Microbubble Effects

The mixing mechanisms of the airlift and bubble column bioreactors differed. In the airlift bioreactor, the liquid recirculated between the inner and outer annular of the draft tube. However, random mixing appeared in the bubble column bioreactor as the random clockwise and anti-clockwise liquid circulation in the upper and lower sections. The generation of microbubbles resulted in the abovementioned mechanisms. They demonstrated an entire area of microbubble distribution in the riser region of both bioreactors. There was a limitation of microbubbles at the bottom outer annular region of the bubble column bioreactor.

However, the airlift bioreactor reported a well microbubble distribution in both the downcomer and riser regions. Figure a,b shows the experimental scheme of the microbubble generation in both bioreactors. The results demonstrated that the microbubble distribution in the bubble column bioreactor decreased 23.18% of the bioreactor’s working volume during the reacting period. Figure c,d shows the comparison of liquid recirculation between the airlift bioreactor and the bubble column bioreactor by focusing on the interconnected section of the inner and outer annular region.

Figure 6

Grand scheme of microbubble generation via the fluidic oscillation aeration system and a comparison of the liquid recirculation between the airlift bioreactor and the bubble column bioreactor. This image shows the overall recirculation mechanism of both bioreactors in the frontage aspect of the (a) airlift bioreactor and the (b) bubble column bioreactor. The comparison shows the lagging of microbubbles in the bubble column bioreactor in the focus aspect at the interconnected section of the inner and outer annular region on both bioreactors: the (c) airlift bioreactor and the (d) bubble column bioreactor.

Microbubbles accumulated at the downcomer of the airlift bioreactor because of the liquid recirculation. The distribution of microbubbles was less in this region of the bubble column bioreactor, and this confirmed the lack of microbubbles in the other aspects. Therefore, the gas holdup built up better in the airlift bioreactor systems and achieved advantages from the recirculation mechanism with the microbubble aeration system via a fluidic oscillator.

Biodesulfurization via Thioalkalivibrio versutus D306 with Microbubble Strategy

The biodesulfurization experiments in the bubble column bioreactor and airlift bioreactor were performed over a period of 90 h. Figure shows the measured conditions of the biodesulfurization.

Figure 7

Biodesulfurization conditions such as pH, DO, ORP, and the conductivity of the airlift and bubble column bioreactors with an installed oscillation aeration system for 90 h. The experimental results are the biodesulfurization conditions of the (a) airlift bioreactor and the (b) bubble column bioreactor with the actual flow rate.

The pH and conductivity were stable in the range of 9–9.5 and 53–72 mS/cm in the bubble column bioreactors and the airlift bioreactors, respectively. During the culturing phase (0–36 h), the DO and ORP were interrupted by bacterial growth. The ORP slightly increased until reaching 118.9 and 112.5 mV in the bubble column bioreactor and airlift bioreactor, respectively. Then, after 36 h, a high concentration of thiosulfate was added to the biodesulfurization phase of the experiments (36–90 h). The stage of limited oxygen route biodesulfurization occurred, and the utilization of oxygen in the system was increased due to the high oxidation level of excess reductive sulfur compounds; this decrease the DO concentration and ORP in both bioreactors.[47,56,57] Moreover, a high level of soluble matters decreased the oxygen level. Results showed that both DO and ORP significantly decreased during the biodesulfurization process; furthermore, the oxygen concentration in the TD medium was near zero for the DO measurement. ORP reached −220.51 mV, and it was slightly stable close to −160 mV on the bubble column bioreactor system. The airlift bioreactor system showed a deeper ORP at −256.73 mV that stabilized close to −160 mV, occurring in both bioreactor systems. The negative ORP condition in the airlift bioreactor system may support the favorable condition of biodesulfurization and hypothesis of pathway selection via hydrodynamics. This hypothesis preferred the direct SoxAXYZ biodesulfurization pathway over the SoxAXYZ with the SoxB pathway.

Figure shows the biodesulfurization results of the bubble column and airlift bioreactors via ion chromatography analysis of the liquid samples from every 6 h during the biodesulfurization process.

Figure 8

Biodesulfurization results of the airlift and bubble column bioreactors with the installed oscillation aeration system via ion chromatography. The analyzed results by the ion chromatography of the biodesulfurization samples during 90 h of the (a) airlift bioreactor and (b) bubble column bioreactor.

The consumption of the thiosulfate concentration in the bubble column bioreactor system took ∼22 h to reach near zero. Similarly, the airlift bioreactor system took ∼12 h to complete the thiosulfate usage for culturing. In the biodesulfurization phase, the airlift bioreactor system demonstrated the stability of biodesulfurization; the complete conversion of concentrated thiosulfate medium was achieved in the experiment. However, the bubble column bioreactor system retained an amount of thiosulfate. This result demonstrated the favored biodesulfurization with Thialkalivibrio versutus D306 in an airlift bioreactor system. The production of sulfate concentration in the bubble column bioreactor and airlift bioreactor was 24.82 and 17.96 g/L, respectively, during the 90 h of the biodesulfurization process. These results demonstrated 27.64% less sulfate production in the airlift bioreactor system compared to the bubble column bioreactor in this study. The total biological sulfur accumulation was 40.05 and 35.91 g/L at the end of the biodesulfurization phase, respectively, for the airlift and bubble column bioreactors. The airlift bioreactor with the oscillation aeration system produced 10.34% of biological sulfur beyond the bubble column bioreactor. The abovementioned results confirmed that the direct SoxAXYZ pathway was preferred over the SoxAXYZ with the SoxB pathway in the airlift bioreactor system. Moreover, the results reported the superior performance of biodesulfurization in the airlift bioreactor system over the bubble column bioreactor system. The biodesulfurization in the airlift bioreactor has an advantage in hydrodynamics as per its mixing mechanism with the microbubble generation strategy and afforded a greater biodesulfurization performance than the bubble column bioreactor.

Biodesulfurization Performances

The performance of biodesulfurization depends on the oxygen distribution and mixing under haloalkaline conditions. Therefore, a complete understanding of the bioreactor hydrodynamics allows for a better interpretation of the biodesulfurization’s desired products. This study approached the microbubble aeration by the fluidic oscillator to elevate the oxygen distribution via microbubbles under haloalkaline conditions. Figure shows the biodesulfurization cultivation results and performances with T. versutus D306 in each bioreactor.

Figure 9

Performance indicators: sulfate and biological sulfur production of the airlift and bubble column bioreactors with the microbubble strategy for 90 h. The analyzed biodesulfurization results of sulfate and sulfur in the 6 h samples and in the accumulation in the bioreactors; (a) airlift bioreactor and (b) bubble column bioreactor.

In the biodesulfurization phase, the percentage yield of biological sulfur reached 94.94% in the airlift bioreactor; this showed 9.21% increase than that in the bubble column. Complete consumption of thiosulfate was demonstrated in the airlift bioreactor, indicating a removal efficiency of ∼100% in the airlift bioreactor and 98.5% in the bubble column bioreactor. The desired product of the process, biological sulfur (S0), was indicated by selectivity. The airlift bioreactor’s selectivity was 10.9% greater than that of the bubble column bioreactor. In particular, the airlift bioreactor was 84.91% selective, and the bubble column bioreactor was 74.08% selective. Every 6 h during the biodesulfurization phase, the average sulfate produced in the airlift and bubble column bioreactors was 0.71 and 1.37 g/L, respectively, and biological sulfur produced was 3.83 and 3.45 g/L, respectively.

Moreover, in the biodesulfurization under a limited oxygen route, the sulfur oxidation performance demonstrated a remarkable sulfate inhibition for the biodesulfurization process. The sulfur oxidation performance of 5.06 and 10.04% was demonstrated in the airlift bioreactor and bubble column bioreactor, respectively. Biological sulfur was produced in greater quantities than the undesired sulfate by a beneficial oxygen distribution of microbubbles under haloalkaline conditions. Therefore, the mixing mechanism of the bioreactors influenced the oxygen distribution under haloalkaline conditions; consequently, the additional biological sulfur production was achieved, and the sulfate production was inhibited. The total production of sulfate in the airlift and bubble column bioreactors was 17.96 and 24.82 g/L, respectively, and that of sulfur was 40.05 and 35.91 g/L, respectively, at the end of 90 h. To summarize, the performance indicators of the airlift bioreactor with the microbubble strategy have shown the biodesulfurization enhancement because a less amount of sulfate was produced, and additional biological sulfur was produced with higher performances.

Biodesulfurization Conditions and Performances of Recent Studies Using Sulfur-Oxidizing Bacteria

Table shows the comparison of the biodesulfurization results with the microbubble strategy with other related studies. High-performance indicators were reported in this study. The sulfur percent yield, selectivity, sulfur oxidation performance, and removal efficiency were few of the superior characteristics in their categories compared to related studies. The results with the lowest-energy input method using the microbubble strategy were significant. Most recent studies agreed on the highest removal efficiency of the reactant at ambient temperature with the alkaline solution. Flores-Cortés et al.[58] reported a significant sulfur oxidation performance of 8.2 ± 1.2% using a continuously stirred bioreactor for additional oxygen distribution and effective mixing in the biodesulfurization process. The microbubble strategy in this study enhanced the oxygen distribution significantly compared with conventional aeration. However, the selectivity and sulfur oxidation performance results demonstrated the superior enhancement of the airlift bioreactor over the bubble column bioreactor. Therefore, the mixing mechanism of the bioreactor was important for biodesulfurization. Quijano et al.[59] and San-Velero et al.[60] reported high sulfur oxidation performances in a complete oxygenation experimental design with an airlift bioreactor and a bubble column bioreactor for typical biotrickling filter issues. The remarkable selectivity of 96.6% of biological sulfur was reported by De Rink et al.[61] with two bioreactors, a continuously stirred bioreactor and an airlift bioreactor, under different substrate conditions. Moreover, 81% selectivity was reported by Zhang et al.[62] with a packed bed bioreactor via an internal sulfur cycling process. The biological sulfur selectivity of the airlift bioreactor in the study was more acceptable than in recent studies. However, the stream recycling process and additional bioreactors with the microbubble strategy were interesting for additional research. Furthermore, De Rink et al.[61] and Mu et al.[56] reported a sulfur yield of 80–90% and 79.1 ± 1.3%, respectively. Moreover, high sulfide bioconversion to biological sulfur was obtained, while the thiosulfate bioconversion remained mediocre in recent studies.

Table 2

Operating Conditions and Biodesulfurization Performances with the Different Strategies of Recent Studies Using Sulfur-Oxidizing Bacteria and S2O32– or HS– Substratea

experimental conditions						performance indicators
experimental designs	substrate	bioreactor	T [°C]	pH	microorganism	PY (%)	SE (%)	SOP (%)	RE (%)	reference
microbubble strategy via fluidic oscillator	S2O32–	ALR	25	8.5–10	T. versutus D306	94.94	84.91	5.06	∼100	this study
	S2O32–	BBC	25	8.5–10	T. versutus D306	85.73	74.07	10.04	98.5	this study
batch assays	S2O32–		30	10	T. versutus	45	NR	NR	99.9	(12)
continuous-flow fluidized bed bioreactor	S2O32–	FBR	30	∼10	T. versutus	27 ± 2	NR	NR	88–99.9	(63)
coupling absorption column	HS–	ALR	25 ± 3	8.0–9.0	domestic WWTP sludge	NR	NR	52–100	95–100	(59)
two bioreactors with different substrate conditions	HS–	CSTR ALR	34.9–37.9	7.5–8.7 8.4–9.0	domestic sludge	80–90	96.6	NR	NR	(61)
two-stage bioreactor under anoxic denitrifying	HS–	CSTR	25 ± 1	7 ± 0.2	SOB	NR	NR	8.2 ± 1.2	>95	(58)
internal sulfur cycling (ISP) process	HS–	PBR	25	7–8.2	SOB	NR	81	NR	∼100	(62)
haloalkaliphilic CROS in deep treatment	HS–	ALR	25	9.5	T. versutus D306	79.1 ± 1.3	50	NR	∼100	(56)
coupling absorption column	HS–	BBC	25 ± 1.5	7.5–8.5	SOB	NR	NR	>75	>80	(60)

The comparison of recent studies on biodesulfurization performance indicators: biological sulfur percent yield (PY), selectivity (SE), sulfur oxidation performance (SOP), and removal efficiency (RE). ALR = airlift bioreactor, BBC = bubble column bioreactor, CSTR = continuous stirred-tank bioreactor, PBR = packed-bed bioreactor, FBR = fluidized-bed bioreactor, NR = not reported.

The airlift bioreactor with the microbubble strategy demonstrated a significant percent yield of biological sulfur with the thiosulfate substate. Therefore, the biodesulfurization with the microbubble strategy via a fluidic oscillator under haloalkaline conditions is interesting for enhancing the primary thiosulfate treatment processes and other biodesulfurization processes.

Conclusions

The biodesulfurization of thiosulfate via T. versutus D306, under haloalkaline conditions with an oscillation microbubble generation system in an airlift bioreactor, improved the biodesulfurization performance over the conventional bubble column bioreactor. The direct SoxAXYZ pathway was preferable owing to the result of complete consumption of thiosulfate with less sulfate production (17.96 g/L) and additional biological sulfur production in the airlift bioreactor system, indicating 27.64% less sulfate and 10.34% more biological sulfur than in the bubble column bioreactor. The results indicated 94.94% sulfur percent yield, 84.91% biological sulfur selectivity, and 5.06% sulfur oxidation performance in the airlift bioreactor with the microbubble strategy during the biodesulfurization phase. The strategy of microbubble generation via an oscillation system played an important role in influencing the oxygen distribution and hydrodynamics enhancement with minimum oxygen and energy inputs. The results revealed 81.87 and 76.19% smaller average microbubbles at the riser and downcomer sections, respectively, over the conventional aeration of a bubble column bioreactor. Moreover, the enhancement of gas distribution and liquid recirculation was provided on the airlift bioreactor because of its determinant mixing mechanism; this is favorable in the oxygen distribution for biodesulfurization under haloalkaline conditions.