Bioleaching, the addition of bacteria to geological materials, has been applied to sludge to remove metals and improve upon sludge dewaterability. This paper investigates the effect of using different quantities of inoculum (bacteria) during bioleaching on sludge dewaterability. The analysis was based on bioleaching experiments conducted in a 20 L bio-reactor using different quantities of inoculum (20%, 10%, 5%, 2%, 0%). Changes in pH, oxidation reduction potential (ORP), capillary suction time (CST), specific resistance to filtration (SRF) and extracellular polymeric substances (EPS) were determined to gauge sludge dewatering. Results indicate that sludge dewaterability during the 2%, 10%, and 20% inoculum experiments declined through time. Decreased dewaterability is attributed to increases in the quantity of proteins and polysaccharides in slime EPS. Dewaterability improved during the 5% inoculum experiment, and reached a maximum when pH was 2.3. During this latter experiment, CST and SRF were reduced by 74% and 62%, respectively, in comparison to control conditions, while total EPS content decreased by 71%. The decrease in total EPS was primarily due to a decrease in proteins associated with tightly bound EPS (TB-EPS). Thus, changes in the amount of proteins in TB-EPS and sludge pH played a crucial role in sludge dewaterability.
In response to the recent, rapid growth in industrialization and population, wastewater treatment plants have produced a large amount of activated sludge. Traditional mechanical dewatering methods can only reduce the moisture content of sludge by about 20%. The high water content of the processed sludge cake makes it difficult and expensive for disposal by incineration or landfilling. It also hinders its re-use as a composting material (Nomeda et al. 2008). Thus, there is considerable interest in developing technologies that can enhance the dewatering of sludge (i.e., its dewaterability), thereby reducing the costs of its disposal and/or reuse.
Bioleaching refers to the inoculation of ores, sludge, and other geological materials with microorganisms (bacteria) to remove selected metals. Recent studies have shown that bioleaching can efficiently remove metals from sewage sludge while improving its dewaterability. The bacteria Acidithiobacillus thiooxidans and Acidithiobacillus ferrooxidans play a significant role in bioleaching processes. A. thiooxidans produces H+ by the bio-oxidation of S0, a process that leads to a decrease in pH. Bioleaching, in comparison to many other physical and chemical methods, will not dramatically decrease the organic matter (OM), total nitrogen or total phosphorus content of the dewatered sludge cake, enhancing its use as compost (fertilizer) (Liu et al. 2012a). The potential use of bioleaching to enhance sludge dewaterability has also been suggested by other studies. Song & Zhou (2008), for example, have observed that the specific resistance of sludge to filtration (SRF) could be decreased from 1.83 × 109 m/kg to 0.39 × 109 m/kg after bioleaching. Biogenic flocculants produced by A. ferrooxidans have been used to improve the dewaterability of anaerobically digested sludge; the capillary suction time (CST) and SRF of sludge were decreased by 74% and 89%, respectively (Kurade et al. 2016).
A potential disadvantage of using bioleaching is that dissolved organic matter (DOM) in activated sludge negatively impacts the growth of A. thiooxidans (Gu & Wong 2004; Ren et al. 2009). For instance, Fang and Zhou demonstrated that the presence of DOM was toxic to microorganisms used in bioleaching when its concentration was higher than 150 mg DOC/L (the dissolved organic carbon (DOC) as representative of sludge DOM was determined by using a total organic carbon analyzer) (Fang & Zhou 2006, 2007). Hence, to improve such negative effects and enhance the efficiency of bioleaching, many acid-tolerant yeast strains have been increasingly employed in bioleaching to reduce DOM in the processed sludge, thus significantly improving the activities of Acidithiobacillus species (Zhou et al. 2013). In addition, the use of the filamentous fungus Mucor circinelloides ZG-3, which is capable of dissolving DOM, positively improves sludge dewaterability; sludge SRF was reduced by 25.9% under its optimum sequential inoculation with M. circinelloides ZG-3 and A. thiooxidans LX5 (Zheng et al. 2016).
Previous studies aimed at determining the mechanisms controlling the dewatering of bioleached sludge have shown that many factors influence the efficiency of sludge dewatering, including particle size, the quantity of extracellular polymeric substance (EPS), water content, and pH (Karr & Keinath 1978; Mikkelsen & Keiding 2002; Neyens & Baeyens 2003). Chen et al. (2001) found that sludge dewaterability was improved with a reduction in sludge pH; water content of the sludge reached its lowest value at pH 2.5. In addition, a reduction in pH and the death of microorganisms in the sludge both influence the amount of EPS in the sludge (Chen et al. 2001; Raynaud et al. 2012), the latter of which is a major factor impacting sludge dewaterability (Neyens et al. 2004). Wong et al. (2015) found that the concentration of Fe2+ when the sludge was inoculated with Acidithiobacillus could enhance its dewaterability. At optimum Fe2+ concentrations, there was a 96% and 88% reduction in CST and SRF, respectively. Total EPS was also reduced. Similarly, Liu et al. (2016) reported that sludge dewaterability was improved by bioleaching because it decreased its EPS content; protein and polysaccharide concentrations were reduced by 97.42% and 76.00%, respectively. Huo et al. (2014a) found that the dominant factors impacting sludge dewaterability were changes in microbial, slime EPS (S-EPS), and bound water contents. Their contributions to improved dewaterability were 32.50%, 22.37% and 24.24%, respectively. Kang et al. (1989) investigated the effect of the EPS content of sludge on dewatering by removing or adding the EPS. The result indicated that EPS could exert a negative impact on the dewatering process. Yang & Li (2009) revealed that EPS in sludge flocs exerted a significant influence on the dewaterability of sludge. Excessive EPS in the form of loosely bound EPS (LB-EPS) had a negative influence on sludge dewaterability. Sun et al. (2017) researched the influence of the quantity of inoculum used in bioleaching on sludge dewaterability, and found that using 10% inoculum was optimum for the dewaterability of sludge; SRF decreased by 81.82%, whereas the tightly bound EPS (TB-EPS) content decreased. The TB-EPS content correlated better with sludge dewaterability than did LB-EPS.
Despite the work conducted to date, bioleaching experiments remain in the laboratory stage. In addition, the results of previous studies defining the role of EPS on sludge dewaterability are often inconsistent, and need to be further understood. The objective of the present study was to investigate the effect of different proportions of inoculum on sludge dewaterability.
MATERIALS AND METHODS
Municipal sewage sludge sampling
Sewage sludge was collected from the Hewen Lake Wastewater Treatment Plant, Jiujiang City, Jiangxi Province, China. The plant receives mostly domestic sewage, although it also receives some industrial wastewater. Sludge pH, SRF and CST were determined immediately after sampling. Sludge solid content was measured after oven-drying at 105 °C for 2 h. The dry weight OM content of the sludge was measured according to APHA (2005). The characteristics of the raw sludge are shown in Table 1.
|Parameters .||Value .|
|pH||6.72 ± 0.02|
|Total solid content (%)||4.20 ± 0.02|
|CST (s)||40.10 ± 1.05|
|SRF (1011 m/kg)||8.47 ± 1.05|
|Organic matter content (%, dry weight)||28.00 ± 0.14|
|Parameters .||Value .|
|pH||6.72 ± 0.02|
|Total solid content (%)||4.20 ± 0.02|
|CST (s)||40.10 ± 1.05|
|SRF (1011 m/kg)||8.47 ± 1.05|
|Organic matter content (%, dry weight)||28.00 ± 0.14|
Microorganisms and the preparation of bioleaching inoculum
The A. thiooxidans JJU-1 (GenBank No. KM101109) used in this study was isolated from municipal sewage sludge in our previous study (Yang et al. 2015) and grown using Waksman medium. The medium was autoclaved at 121 °C for 15 min, and adjusted to pH 3.5–4.0 with 9 M H2SO4. It was then supplemented with 5 g/L elemental sulfur as the energy source. The inoculums were cultured in 500 mL Erlenmeyer flasks shaken at 180 rpm, and then held at a constant temperature of 28 °C for 3–4 d until the bacterial cell density reached ∼108 cells/mL.
The Rhodotorula mucilaginosa JJU-2 bacterium was isolated from sludge during a previous study (Wang et al. 2010). R. mucilaginosa is capable of degrading sludge DOM and boosting the growth of A. thiooxidans JJU-1. The inoculums of R. mucilaginosa JJU-2 were grown in flasks containing potato dextrose agar at its optimal pH, which for this strain was 5 (although it can survive in the extremely low pH condition of 1) (Qiu et al. 2017). Flasks were shaken using a gyratory shaker at 28 °C and 120 rpm for 3–4 d until a cell density of 1012 cells/mL was reached.
Following completion of the above procedures, a volume of 2.1 L of A. thiooxidans JJU-1 inoculum and 700 mL of R. mucilaginosa JJU-2 inoculum were added to a 20 L automatic bioreactor containing 14.2 L of municipal sludge. The mixture was then supplemented with 8 g/L S0, and an appropriate proportion of nutrients. The reaction was carried out at 28 °C, an agitator speed of 200 rpm and an aeration of 15 L/min until the sludge pH decreased to about 2.0. The above procedures were repeated three times. The bioleached sludge was then employed as bioleaching inoculum in the following bioleaching experiments.
Bioleaching using 20% inoculum was conducted in a 20 L automatic bioreactor containing 3.8 L acidified bioleaching inoculum, 12.2 L raw sludge, and 8 g/L S0 as an energy source. The sludge was mixed by a continuous stirring system in the bioreactor at 200 rpm at a constant working temperature of 28 °C (maintained with an automated control system). The pH of the sludge was adjusted to about 4.8–5.0 with 2.0 mol/L H2SO4. During the incubation process, air was supplied at 15 L/min from the bottom of the bioreactor by an air compressor. After the sludge pH had dropped to about 2.0, the sludge was stored at 4 °C in a refrigerator for the next bioleaching experiments.
The 10%, 5%, and 2% inoculum experiments contained 1.6 L acidified bioleaching inoculum and 14.4 L raw sludge, 0.8 L acidified bioleaching inoculum and 15.2 L raw sludge, and 0.32 L acidified bioleaching inoculum and 15.68 L raw sludge, respectively. The other experimental conditions were identical to those used in the 20% experiment described above.
For the experimental control, 16 L of raw sludge and 8 g/L S0 were added to the bioreactor; no acidified bioleaching inoculant was added. The other experimental conditions were identical to those used for the other experiments.
During each of the above bioleaching experiments, 600 mL of material was collected from the bioreactor at 24 h intervals, after which pH, oxidation reduction potential (ORP), CST, SRF, and SO42− concentrations were determined along with S-EPS, LB-EPS, and TB-EPS contents according to the methods described below.
Both pH and ORP were determined using a pHS-3C digital pH-meter with the Pt-Ag/AgCl electrode system. Dewaterability was determined through CST using a capillary suction timer (Model 304 M CST, Triton, UK), and SRF was determined using the Buchner funnel test (Lo et al. 2001). Moisture content of the filter cake was measure by oven-drying at 105 °C for 2 h. The dried digestate sample was then measured for OM content according to APHA (2005).
EPS extraction and chemical analysis
A heat extraction method was modified to extract different EPS fractions from the activated sludge (Li & Yang 2007). First, a 100 mL sludge sample was collected and placed into a flask. The samples were then washed twice with a 0.05% (w/w) NaCl solution brought to its original volume of 100 mL. Sixty millilitres of the sludge sample was subsequently centrifuged at 600 rpm for 15 min. The resulting sludge supernatant was collected as S-EPS; the precipitant remaining in the tube was used for additional EPS layer extractions. The solution of 0.05% NaCl was prepared. The sludge pellet in the tube was then resuspended into 60 mL of 0.05% NaCl to its original volume of 60 mL and treated by ultrasound at 20 kHz and 40 W for 2 min. The aliquot was centrifuged at 9,000 rpm for 20 min to separate solids and supernatant. The collected supernatant was regarded as the LB-EPS. Finally, the residual sludge pellet left in the centrifuge tube was re-suspended in a 0.05% NaCl solution to its original volume of 60 mL, then heated at 80 °C for 30 min, and finally centrifuged at 20,000 rpm for 20 min to collect the TB-EPS. The supernatant samples were filtered through a 0.22 μm membrane before chemical analysis.
Polysaccharide was measured by the anthrone method (Gaudy 1962) using glucose as the standard. The protein content was determined with a modified Lowry method using bovine serum albumin as a standard (Frolund et al. 1995). Lastly, the DNA content was determined with the diphenylamine colorimetric method using sodium DNA as a standard (Sun et al. 2009).
RESULTS AND DISCUSSION
Variations in pH and ORP as a function of treatment time
As shown in Figure 1, pH decreased and ORP increased at different rates when different proportions of inoculum were used during the bioleaching experiments. More specifically, the data indicate that Acidithiobacilli bacteria have a higher activity, which is consistent with previous research (Liu et al. 2012b).
In the control experiment (Figure 1(a)), pH decreased slowly from 4.97 to 3.36 and ORP increased from 110 mV to 198 mV at day 4. These trends probably resulted from the oxidation of S0 by oxygen in the circulating air and the residual Acidithiobacilli bacteria in the sludge. In the 20% inoculum experiment (Figure 1(b)), pH rapidly decreased from 4.8 to 0.76, while the value of ORP increased from 125 mV to 346 mV during the 4 d bioleaching treatment. When the proportion of inoculum was reduced to 10%, 5%, and 2%, the rate of bioleaching slowed. For instance, during the 10% inoculum experiment (Figure 1(c)), pH declined from 4.92 to 1.56, during which the value of ORP increased from 118 mV to 300 mV at day 4. In the 5% inoculum experiment (Figure 1(d)), pH and ORP changed from 5.05 and 108 mV to 1.74 and 291 mV, respectively, after 4 d of bioleaching. During the 2% inoculum experiment (Figure 1(e)), pH and ORP changed from 4.97 and 110 mV to 2.52 and 253 mV, respectively, after 4 days of bioleaching.
With increases in the proportions of inoculum, the amount of A. thiooxidans was significantly increased, and the rate at which pH decreased and ORP increased was drastically accelerated. These differences may be attributed to the generation of sulfuric acid that varied in response to differences in the rate of microbial oxidation of elemental sulfur (Pathak et al. 2009).
Variations in CST and SRF as a function of treatment time
CST and SRF are widely used parameters to monitor the dewaterability of sludge. Many studies show that longer CST or higher SRF values are indicative of poor sludge dewaterability (Feng et al. 2009).
As shown in Figure 2(a), CST and SRF increased from 46.5 s and 0.78 × 1012 m/kg to 290.0 s and 3.79 × 1012 m/kg, respectively, after 4 d in the control experiment, indicating that sludge dewaterability deteriorated. For the bioleached sludge, the proportion of inoculum used exerted an important influence on the dewaterability of sludge. More specifically, during the 20% inoculum experiment (Figure 2(b)), CST and SRF rapidly increased from 153.2 s and 1.19 × 1012 m/kg to 885.4 s and 10.64 × 1012 m/kg, respectively, at day 4. Similarly, during the 10% inoculum experiment (Figure 2(c)), CST and SRF drastically increased from 90.9 s and 1.22 × 1012 m/kg to 364.9 s and 4.08 × 1012 m/kg, respectively, at day 4. During both of these experiments, CST and SRF unexpectedly increased, indicating that sludge dewaterability deteriorated. In addition, the rate of CST and SRF increase during the 20% inoculum experiment was faster than during the 10% experiment.
As shown in Figure 2(d), during the 5% inoculum experiment, CST and SRF decreased from 151.0 s and 2.60 × 1012 m/kg to 65.0 s and 1.00 × 1012 m/kg, respectively, at 48 h. Subsequently, and unexpectedly, CST and SRF increased to 82.8 s and 5.79 × 1012 m/kg, respectively, at 24 h. The lowest value of CST and SRF was 65.0 s and 1.00 × 1012 m/kg, respectively, at which the corresponding pH was approximately 2.3 (Figure 1(d)). The observed patterns demonstrate that bioleaching using 5% inoculum was optimal for improving the dewaterability of sludge. Moreover, there was an optimum pH (∼2.3) to achieve the highest degree of dewaterability of the bioleached sludge. Sludge characterized by the highest degree of dewaterability exhibited a 74% and 62% reduction in CST and SRF, respectively, as compared to the control within 3 days of bioleaching.
Figure 2(e) shows that during the 2% inoculum experiment, CST and SRF increased from 54.6 s and 0.78 × 1012 m/kg to 343.8 s and 5.50 × 1012 m/kg, respectively, at day 3. During this time, pH declined from 4.97 to 2.87 (Figure 1(e)). CST and SRF then rapidly decreased to 293.7 s and 4.37 × 1012 m/kg at 24 h, respectively, at which time the pH was 2.5 (Figure 1(e)). Data collected during the 2% inoculum experiment show that the dewaterability of the sludge was inferior to the sludge treated with 5% inoculum.
The combined results indicate that a significant difference exists between the control and the bioleached sludge. Bioleaching using 5% inoculum achieved optimum results with regards to sludge dewaterability; CST and SRF were reduced by 74% and 62%, respectively, as compared with the control within a period of 3 d. The data also demonstrated that a pH of 2.3 was optimum for enhanced dewaterability of sludge. When sludge pH declined to below 2.3, sludge dewaterability deteriorated, suggesting that excessive bioleaching time negatively affected sludge dewatering, a finding consistent with previous analyses (Feng et al. 2009; Huo et al. 2014b).
Variations in the content and fractionations of sludge EPS with treatment time
Sludge EPS is composed of functional groups, such as hydroxyl, that add to the repulsion between flocs (Merrylin et al. 2013a) and collect a massive amount of bound water (Raynaud et al. 2012). The role of EPS on sludge dewaterability has been widely studied and has been found to significantly impact the dewatering characteristics of sludge (Murugesan et al. 2016). However, the relationship between EPS content and sludge dewatering remains controversial (Huo et al. 2014a; Murugesan et al. 2016). In order to investigate the effect of EPS content on sludge dewaterability, changes in the quantity of total EPS and different EPS fractions were determined for the control experiment as well as for the experiments using different proportions of inoculum during bioleaching.
As shown in Figure 3(a), in the control experiment, sludge dewaterability deteriorated as both the S-EPS and the LB-EPS content slightly increased. In contrast, the TB-EPS content decreased. During the 20% inoculum experiment, sludge dewaterability deteriorated, the total EPS content increased, and S-EPS rapidly increased from 9.29 mg/g SS to 17.83 mg/g SS at day 4 (Figure 3(b)). During the 10% inoculum experiment, sludge dewaterability decreased further, the S-EPS content, which comprised most of the total EPS, dropped from 9.39 mg/g SS to 5.1 mg/g SS at day 2, before increasing to 10.13 mg/g SS (Figure 3(c)). Given the above, it appears that S-EPS content was the dominant factor controlling the deterioration of sludge dewaterability. It is likely that acid-tolerant microorganisms generated large amounts of stress resistance protein to protect themselves from the low pH environment (Cabiscol et al. 2010), thus leading to the following increase of total EPS content.
It was remarkable that variations in different fractions of EPS during the 5% inoculation experiment differed from the 20% and 10% inoculation experiments. Interestingly, total EPS content during the 5% experiment decreased from 21.78 mg/g SS to 6.27 mg/g SS at 3 d, and then increased to 8.37 mg/g SS at 4 d (Figure 3(d)). The lowest value corresponds to a pH of approximately 2.3 (Figure 1(d)). Variations in total EPS content appear to have been synchronized with CST and SRF (Figure 2(d)). Notably, at 48 h of leaching, the content of TB-EPS comprised the majority of the total EPS, the latter of which dropped from 12.10 mg/g SS to 0.38 mg/g SS. Then, when sludge pH dropped below 2.3, sludge dewaterability deteriorated as the content of S-EPS increased and become the predominant component of total EPS (end of day 3). Thus, the decrease in TB-EPS content was the dominant factor controlling the improvement of sludge dewaterability. It is possible that the decrease is due to abscission and decomposition of EPS from the surface of the sludge as it is affected by acidification (Chen et al. 2001). The observation that S-EPS increased at end of day 3 and led to the deterioration of sludge dewaterability during the 5% inoculum experiment was consistent with the 20% and 10% inoculum experiments. The observed changes are probably due to the secretion of EPS by Acidithiobacillus sp. to protect themselves against the exceedingly low pH environment (pH <2.3).
As shown in Figure 3(e), during the 2% inoculum experiment, total EPS content increased from 22.50 mg/g SS to 28.43 mg/g SS at 4 d, decreasing the dewaterability of sludge. The increase in total EPS was mainly due to an increase in S-EPS content from an initial of 3.66 mg/g SS to 18.38 mg/g SS after 4 d of treatment. S-EPS was the predominant component of total EPS. It is noteworthy that the content of TB-EPS rapidly dropped from 9.49 mg/g SS to 5.29 mg/g SS at 3–4 d, during which the dewaterability of sludge improved and pH was 2.5 (Figure 2(e)).
These results indicate that a decrease in TB-EPS content and sludge pH exerted a significant impact on sludge dewaterability as demonstrated by a decrease in CST and SRF. Moreover, the amount of S-EPS present was the dominant factor controlling the deterioration of sludge dewaterability: the higher S-EPS content, the less the sludge could be dewatered.
Variations in different EPS fractionations with treatment time
To compare the effect of different fractions of EPS on the dewaterability of sludge, protein, polysaccharides and DNA were measured every 24 h during the control experiment and experiments using different proportions of inoculum. As shown in Figure 4(a), during the control treatment, the quantity of proteins, polysaccharides and DNA in S-EPS increased slightly, whereas the amount of proteins in TB-EPS had decreased by day 4. During the 20% inoculum experiment, the amount of proteins, polysaccharides and DNA in S-EPS increased from 2.31 mg/g SS, 1.36 mg/g SS and 5.61 mg/g SS to 4.79 mg/g SS, 5.18 mg/g SS and 7.83 mg/g SS by day 4, respectively (Figure 4(b)). Changes in DNA are linked to the death and lysis of bacteria, processes that release DNA to the sludge. It was apparent that the increase of proteins and polysaccharides in S-EPS had a significant impact on the deterioration of sludge dewaterability. Figure 4(c) shows that during the 10% inoculum experiment, the results had some similarities to the 20% inoculum experiment; more specifically, increases in protein and polysaccharide content within S-EPS were major contributors to the deterioration in sludge dewaterability. In fact, these results demonstrate that the increase of proteins and polysaccharides in S-EPS was the dominant factor controlling the deterioration of sludge dewaterability, which was mainly due to the death and lysis of bacteria that probably released intracellular materials such as protein, carbohydrate, lipids, DNA and RNA (Merrylin et al. 2013b).
During the 5% inoculum experiment, protein content, as a major component of TB-EPS, exhibited a tremendous reduction of 97%, declining from 10.35 mg/g SS to 0.32 mg/g SS (Figure 4(d)). This decline exerted an important control on improvements of sludge dewaterability. It could be due to the decomposition of sludge EPS components by enzymes, such as protease, amylase, DNase, and RNase (Esakki et al. 2013; Kavitha et al. 2013), and the consumption and biodegradation of EPS by acid-tolerant microorganisms (Rani et al. 2012; Yang et al. 2015). Furthermore, the increase in DNA content in S-EPS demonstrated that the death and lysis of bacteria occurred during bioleaching, which may also help improve sludge dewaterability. In addition, as sludge dewaterability decreased during later leaching as sludge pH declined below 2.3 (Figures 1(e) and 2(e)), the increase in S-EPS content became the major component of total EPS. This phenomenon was likely due to the generation of stress resistance proteins by some acid-tolerant microorganisms present within the extremely low pH environment (Cabiscol et al. 2010; Zhu et al. 2012).
As shown in Figure 4(e), during the 2% inoculum experiment, the increase of DNA demonstrated that death and lysis of bacteria occurred during bioleaching. In addition, protein and polysaccharide contents in S-EPS also increased, and exerted an important influence on the deterioration of sludge dewaterability. This was consistent with results from the 20% and 10% inoculum treatments.
Given the above, sludge pH and decreases in protein content within TB-EPS had a significant impact on improvements in sludge dewaterability. Increases in proteins and polysaccharides in S-EPS were the dominant factors controlling the deterioration of sludge dewaterability. Change in EPS is mainly controlled by changes in microbial quantity.
Different proportions of inoculum have a significant impact on the bio-oxidation rates of S0 and bio-acidification during bioleaching. Optimum enhancements in sludge dewaterability were observed during the 5% inoculum experiment during which CST and SRF significantly declined during bioleaching. The enhancement of sludge dewaterability was related to sludge pH; a pH of 2.3 resulted in the largest improvements in sludge dewaterability. Enhancements in sludge dewaterability were also linked to a reduction in proteins within TB-EPS. Furthermore, an increase in proteins and polysaccharides in S-EPS exerted a significant impact on the deterioration of sludge dewaterability. Changes in the composition of EPS were linked to the quantity (proportion) of inoculum used in bioleaching.
This work was supported by the National Natural Science Foundation of China (21367014, 51741805), Natural Science Foundation of Jiangxi Province, China (20151BAB203026), and Science and Technology Program of Education Department of Jiangxi Province, China (GJJ14732).