Available online at www.sciencedirect.com

www.elsevier.com/locate/chemosphere

Chemosphere 70 (2008) 1492–1499

Application of in situ biosparging to remediate apetroleum-hydrocarbon spill site: Field and microbial evaluation

C.M. Kao a, C.Y. Chen a, S.C. Chen b, H.Y. Chien a, Y.L. Chen b,*

a Institute of Environmental Engineering, National Sun Yat-Sen University, Kaohsiung, Taiwanb Department of Biotechnology, National Kaohsiung Normal University, Kaohsiung, Taiwan

Received 30 April 2007; received in revised form 12 August 2007; accepted 14 August 2007Available online 22 October 2007

Abstract

In this study, a full-scale biosparging investigation was conducted at a petroleum-hydrocarbon spill site. Field results reveal that nat-ural attenuation was the main cause of the decrease in major contaminants [benzene, toluene, ethylbenzene, and xylenes (BTEX)] con-centrations in groundwater before the operation of biosparging system. Evidence of the occurrence of natural attenuation within theBTEX plume includes: (1) decrease of DO, nitrate, sulfate, and redox potential, (2) production of dissolved ferrous iron, sulfide, meth-ane, and CO2, (3) decreased BTEX concentrations along the transport path, (4) increased microbial populations, and (5) limited spread-ing of the BTEX plume. Field results also reveal that the operation of biosparging caused the shifting of anaerobic conditions inside theplume to aerobic conditions. This variation can be confirmed by the following field observations inside the plume due to the biospargingprocess: (1) increase in DO, redox potential, nitrate, and sulfate, (2) decrease dissolved ferrous iron, sulfide, and methane, (3) increasedtotal cultivable heterotrophs, and (4) decreased total cultivable anaerobes as well as methanogens. Results of polymerase chain reaction,denaturing gradient gel electrophoresis, and nucleotide sequence analysis reveal that three BTEX biodegraders (Candidauts magnetobac-

terium, Flavobacteriales bacterium, and Bacteroidetes bacterium) might exist at this site. Results show that more than 70% of BTEX hasbeen removed through the biosparging system within a 10-month remedial period at an averaged groundwater temperature of 18 �C.This indicates that biosparging is a promising technology to remediate BTEX contaminated groundwater.� 2007 Elsevier Ltd. All rights reserved.

Keywords: Biosparging; Natural attenuation; BTEX; PCR; DGGE

1. Introduction

Accidental releases of petroleum products from pipe-lines and fuel-oil storage tanks are among the most com-mon causes of groundwater contamination. Petroleumhydrocarbons contain benzene, toluene, ethylbenzene,and xylene isomers (BTEX), the major components of fueloils (especially gasoline), which are hazardous substancesregulated by many nations. At many BTEX spill sites,the residual amount of BTEX exists in a pure liquid phase(commonly referred as non-aqueous-phase liquids) withinpore spaces or fractures. The slow dissolution of residualBTEX results in a contaminated plume of groundwater.

0045-6535/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.chemosphere.2007.08.029

* Corresponding author. Tel.: +886 7 717 2930; fax: +886 7 525 4449.E-mail address: dan1001@ms31.hinet.net (Y.L. Chen).

Given that it is often not possible to locate and removethe residual BTEX, remediation must focus on preventingfurther migration of the dissolved contamination. Thisplume control must be maintained for a long period oftime. Therefore, more economic approaches are desirablefor groundwater remediation to provide for long-term con-trol of contaminated groundwater.

Bioremediation is an attractive remediation optionbecause of its economic benefit (Sutherland et al., 2004;Chen et al., 2005; Chen et al., 2006). Recently, intrinsic bio-remediation has been considered as one of the potentialmethods for the cleanup of petroleum-hydrocarbon con-taminated sites. If the intrinsic bioremediation rate is limitedby in situ environmental factors (e.g., oxygen, nutrients, andelectron acceptors), enhanced in situ bioremediation canbe applied to stimulate contaminants biodegradation

N

Fig. 1. Site map showing the contaminant source area, groundwater flowdirection, biosparging wells, soil vapor sampling wells, and the soil andgroundwater sampling locations.

C.M. Kao et al. / Chemosphere 70 (2008) 1492–1499 1493

(Schirmer et al., 2003; Schmidt et al., 2004). BTEX are bio-degradable under both aerobic and anaerobic conditions.Nevertheless, rates of BTEX biodegradation under aerobicconditions are higher than those under anaerobic conditions(Deeb et al., 2003; Moreels et al., 2004). Organic compoundsare removed more completely under aerobic biodegrada-tion. Moreover, microbiological investigations of aquifersediments have revealed the presence of microbial commu-nities capable of degrading a broad range of naturally occur-ring and xenobiotic compounds under a broad range ofenvironmental conditions (Kao et al., 2003, 2005).

Based on the above discussion, in situ aerobic bioremedi-ation is a feasible technology to clean up BTEX contami-nated sites if oxygen can be provided to the subsurfaceeconomically. Biosparging is an effective mechanism forremoval of volatile organic compounds (VOCs) includingBTEX (Adams and Reddy, 2003; Wu et al., 2005; Braret al., 2006). Biosparging functions by injecting air at alow rate into the aquifer below the zone of contamination.At a relatively close well spacing, the injected air promotesoxygenation of the aquifer as necessary to promote aerobicbiodegradation. In this study, the purpose of the biospar-ging system is to stimulate aerobic biodegradation of BTEX.The objectives of the field-scale study were to (1) evaluate theeffectiveness of in situ biosparging as a method for control-ling of BTEX at the fuel-oil spill site, and (2) determine thedominant native microorganisms at different locations ofthe contaminated aquifer through microbial identification.

The selected site in this study is an abandoned petro-chemical manufacturing facility where petroleum productswere produced. Leakage of a petrochemical pipeline andhas resulted in site groundwater contamination with BTEX.Approximately 125 soil gas samples, 67 soil samples, 27one-time GeoprobeTM groundwater samples, and 23 ground-water samples from monitor wells were collected and ana-lyzed to determine the local hydrogeology and delineatethe BTEX plume during a two-year site investigation period(data not shown). Site investigation results show that thecomponents of site soils are mainly sands and silty sands.The water table is generally found at depths ranging from2 to 2.5 m below ground surface (bgs). The groundwaterflows from northeast to southwest at a velocity of6.5 cm d�1 and a hydraulic conductivity of 0.001 cm s�1.Fig. 1 presents the site map showing the contaminant sourcearea, groundwater flow direction, biosparging wells, andsoil and groundwater sampling locations used in this study.Results from previous site investigation studies reveal thatthe contaminants resulted in an approximately 780 m longand 270 m wide plume from source area to downgradient.

2. Materials and methods

2.1. The performance of biosparging system and sampling

condition

The biosparging system consisted of six biospargingwells (injection points) (well screen at 5.7–6 m bgs), air

compressors, flow indicators, inline regulators, and pres-sure gauges. The air flow was approximately 0.06–0.17 m3 min�1 (2–6 cfm) for each biosparging well.Although a total of 23 monitor wells have been installedat this site, three monitor wells (MW1, MW2, and MW3)were selected as the representative wells to assess the poten-tial of BTEX biodegradation by native microorganisms viabiosparging processes. MW1 was located at the sourcearea, MW2 was located at the downgradient area alongthe groundwater flow path, and MW3 was located in thebackground area (Fig. 1). Twelve soil vapor sampling wells(screened in the unsaturated zone) located within the plumewere selected to collect the soil gas samples to evaluate theamount of BTEX loss due to volatilization (Fig. 1). Thecollected soil vapor was analyzed periodically by anorganic vapor analyzer with a photoionization detector(Model TVA1000B, Thermo Environmental InstrumentsInc., USA). All selected wells were sampled bimonthly dur-ing the 10-month investigation period. The first samplingevent was conducted before the start of the biospargingprocess (day 0). For each sampling event, four 40-mLVOC vials were filled with the groundwater collected fromeach monitor well via the flow cell unit and micro-purgetechnique following the standard sampling procedure(NIEA, 2003). Groundwater sample from one of the vials

1494 C.M. Kao et al. / Chemosphere 70 (2008) 1492–1499

was analyzed for BTEX concentrations. The second vialwas analyzed for verification. The third and fourth vialswere used for dilution purposes for samples with highBTEX concentrations. Groundwater samples from thethree monitor wells were collected and analyzed for organiccompounds and geochemical indicators, including BTEX,methane, CO2, inorganic nutrients, anions, pH, redoxpotential (Eh), groundwater temperature, and DO (Kaoand Wang, 2000; APHA, 2001). Organic compound analy-ses were performed in accordance with US EPA Method502.2, using a Varian 3800 GC. Methane was analyzedon a Shimadzu GC-9A GC using headspace techniques(Kao and Wang, 2000). Ion chromatography (Dionex)was used for inorganic nutrients and anions ðNO�3 , NO�2 ,SO�2

4 , and PO�34 Þ analyses (APHA, 2001). Total iron and

ferrous iron were analyzed by Hach DR/400 Spectropho-tometer using US EPA Method 8008 and Method 8146,respectively. DO, Eh, pH, CO2, and temperature were mea-sured in the field. Two MP120 pH Eh�1 meters (Mettler-Toledo) were used for pH and Eh measurements. AWTW DO meter (Oxi 330) was used for DO and tempera-ture measurements, and a Hach digital titrator cartridgewas used for CO2 measurements.

2.2. Manipulation of aquifer sediments

Aquifer sediments were collected from soil borings SB1,SB2, and SB3, which were located adjacent to MW1,MW2, and MW3, respectively (Fig. 1). Soil borings SB1,SB2, and SB3 were located adjacent to monitor wellsMW1, MW2, and MW3, respectively, both spatially andhorizontally. All SBs were collected below the groundwatertable in the saturated zone. Sediments were gathered at thesame time while monitor wells were sampled. Aquifer sed-iments from SB1, SB2, and SB3 were collected on day 0and 300. Collected aquifer sediments were used for soilorganic content analysis and microbial enumeration andidentification study. Soil organic content was analyzedquarterly and was determined by burning the samples at550 �C and calculating the organic content as the preburnweight minus the postburn weight (NIEA, 2002). Microbialenumeration was performed to determine the number oftotal cultivable heterotrophs, total cultivable heterotrophicanaerobes (total anaerobes), and methanogens. Total platecounts were conducted using plate count agar (Difco) toassess the approximate size of the total cultivable hetero-trophic bacterial in soil samples using the spread platemethod (APHA, 2001). Prepared plates were incubated at30 �C for 48 h, then counted for colony forming unit(CFU). The analytical methods for total cultivable hetero-trophic anaerobes and methanogens are described by Kaoand Wang (2000) and enumerated using five-tube MPNassay. The total anaerobe tubes contained media and werescore positive based on optical density. The methanogentubes contained 20% H2 and 80% CO2 in the headspace,and were score positive based on the production ofmethane.

2.3. Denaturing gradient gel electrophoresis (DGGE)

Total bacterial DNAs from 1 g of collected soil sampleswere extracted with a Soil Genomic DNA Purification kit(GeneMark Co., Taiwan) for detecting the communitydynamics in the process of BTEX degradation. Bacterial200-bp fragments of 16S rDNA V3 region for subsequentdenaturing gradient gel electrophoresis (DGGE) analysiswere amplified with the primer sets (341f, forward: 5 0-CCTACGGGAGGCA GCAG-3 0 containing a GC clampof 40-nucleotide GC-rich sequence; 534r, reversed: 5 0-ATTACCGCGGCTGCTGG-3 0) (Chen, 2005). The mix-tures of polymerase chain reaction (PCR) contained10 ng of DNA extract, 4 pmol of each primer, and 5 U ofTaq polymerase (Takara, Shiga, Japan) in final concentra-tions of 2.5 mM of MgCl2 and 0.12 mM of deoxyribonucle-oside triphosphates in PCR buffer. The PCR amplificationwas conducted for 35 cycles: denaturation at 94 �C for1 min, an initial annealing temperature of 65.8 �Cdecreased by 1 �C per cycle until it reached 55.8 �C, fol-lowed by 25 additional cycles at 55.8 �C, and extension at72 �C for 2 min. The equal concentration of each amplifiedPCR products (2.5 lg) was furthermore performed withDGGE using a Bio-Rad DCode system (Bio-Rad, Hercu-les, CA, USA), as described by the manufacturer. The10% polyacrylamide gel with a 30–60% denaturant gradientwas used and electrophoresis was performed at 60 �C and70 V for 14 h. The gels were then stained with SybrGreenI-and photographed.

2.4. Banding analysis and phylogenic analysis

The relative intensity of amplified bands in gels was ana-lyzed with Phoretix 1D software (Nonlinear Dynamics,Newcastle upon Tyne, NE1, UK). The PCR-amplifiedproducts were electro-eluted from gel and then sequencedby MdBio, Inc. in Taiwan. Those sequences were evaluatedby using the basic local alignment search tool to determinethe closest relatives in the GenBank databases (http://www.ncbi.nlm.nih.gov). Alignment of nucleotide sequencesof PCR-amplified products generated a matrix of similaritycoefficients with Neighbor-Joining method (Saitou andNei, 1987). The dendrogram based on these similarity coef-ficients was plotted with UPGMA (unweighted pair-groupmethod with arithmetic mean) method for clustering. Clus-tal X software and Jukes and Cantor distances model wereapplied for the Phylogeny Tree analysis (Watanabe et al.,2000).

3. Results and discussion

3.1. Change of BTEX concentration after the operation of

biosparging

Groundwater samples were collected from monitor wellsMW1, MW2, and MW3, which were located at the sourcearea, downgradient area, and background area of the

0.1

1

10

100

1000

10000

0 60 120 180 240 300

Time (d)

ben-MW1tol-MW1ethylben-MW1xyl-MW1ben-MW2tol-MW2ethylben-MW2xyl-MW2C

onc.

(μg

l-1)

Fig. 2. Variations in BTEX concentrations (log scale) versus the samplingtime after the onset of the biosparging process at MW1 and MW2 (ben:benzene; tol: toluene; ethylben: ethylbenzene; xyl: xylene isomers).

C.M. Kao et al. / Chemosphere 70 (2008) 1492–1499 1495

plume, respectively. Fig. 2 presents variations in BTEXconcentrations (log scale) versus the sampling time atMW1 and MW2 after the onset of the biosparging process.Results show that the BTEX concentrations were signifi-cant decreased after the operation of biosparging mainlydue to the enhanced aerobic biodegradation. Part of theBTEX loss might be also due to the vaporization. Becausethe measured soil gas results from twelve soil gas samplingwells were insignificant (<46 ppm) (data not shown) beforeand after the biosparging process, the amount of BTEXloss due to volatilization would be insignificant. Table 1shows the results of groundwater analyses in MW1,MW2, and MW3, and microbial enumeration in SB1,

Table 1Concentrations of BTEX and indicating parameters in monitor wells on day 0

Monitor Well MW1 MW1Location Sourcea Source

Days after biosparging 0 300Benzene (lg l�1) 190 41Toluene (lg l�1) 6430 124Ethylbenzene (lg l�1) 125 38Xylenes (lg l�1) 244 66DO (mg l�1) 0 2.1Nitrate (mg l�1) 0.1 3.3Total iron (mg l�1) 17.0 0.9Ferrous iron (mg l�1) 15.5 0.4Sulfate (mg l�1) 9 27Sulfide (lg l�1) 18 10Carbon dioxide (mg l�1) 221 265Methane (mg l�1) 12.2 0.12pH 7.2 6.8Redox potential (mV) �254 �18Ammonia nitrogen (mg l�1) 2.1 1.3Phosphate (mg l�1) 0.92 0.14Temperature (�C) 24.3 25.2Soil organic content (%) 0.51 (SB1) 0.47 (SB1)Total cultivable heterotrophs (cell g�1) 7.6 · 105 (SB1)e 8.2 · 107 (SB1)Total anaerobes (cell g�1) 1.3 · 105 (SB1) 7.4 · 102 (SB1)Methanogens (cell g�1) 3.4 · 102 (SB1) 2.4 · 101 (SB1)

a Source: sample collected at the source area.b Down: sample collected at the downgradient area.c BK: sample collected at the background area.d BDL: below detection limit.e SB: microbial enumeration was performed using soil sediments.f ‘‘–’’: not detected.

SB2, and SB3 on day 0 and 300 after the operation of bio-sparging. The occurrence of aerobic biodegradation ofBTEX due to the air injection can be confirmed by theincreased population of total cultivable heterotrophs anddecreased population of total anaerobes as well as metha-nogens in sediments SB1 and SB2.

Before the operation of biosparging, natural attenuationmechanisms were the major causes of the decrease ingroundwater contaminant concentrations through mixedphysical, chemical, and biological processes. Field resultsshow that the detected Eh and DO near the source areawere low, which reflects the reduced conditions in the mostcontaminated zone. Moreover, high CO2 concentrationswere observed in the plume, which indicates that significantmicrobial activity and natural bioremediation occurred inthis area. The averaged phosphate concentrations rangedfrom 0.8 to 3.2 mg l�1 during the investigation period.Thus, the observed natural occurring phosphate in the sub-surface would not be the limiting factor for the growth ofbacteria. The lower nitrate and sulfate concentrationswithin the plume reveal that both nitrate and sulfate wereused as the electron acceptors after the depletion of oxygenin the contaminated zone. The production of sulfide inMW1 also confirmed the occurrence of sulfate reductionprocess. High ferrous concentrations were detected inMW1 indicating that ferric irons might have also been usedas the electron acceptors around the source area.

and 300 after the operation of biosparging

MW2 MW2 MW3 MW3Downb Down BKc BK

0 300 0 30025 3 BDLd BDL657 14 0.4 0.217 0.8 BDL BDL33 0.5 BDL BDL0.3 1.1 2.4 2.90.5 4.6 6.2 4.73.2 0.5 1.9 1.22.8 0.3 0.1 0.219 23 31 2812 8 2 4158 184 124 982.5 0.02 0.02 0.017.1 6.7 7.4 7.3�81 54 189 2381.2 0.28 4 3.80.8 1.23 1.16 1.724.6 25.1 24.1 250.42 (SB2) 0.33 (SB2) 0.28 (SB3) 0.35 (SB3)8.8 · 106 (SB2) 7.1 · 107 (SB2) 1.2 · 105 (SB3) 4.2 · 105 (SB3)8 · 104 (SB2) 8.2 · 102 (SB2) 2.3 · 102 (SB3) 1.7 · 102 (SB3)2.1 · 102 (SB2) 1.3 · 101 (SB2) –f (SB3) – (SB3)

1496 C.M. Kao et al. / Chemosphere 70 (2008) 1492–1499

Moreover, relative higher concentration of methane wasalso detected in MW1. This indicates that mixed anaerobicbiodegradation processes occurred within the most con-taminated zone. The decrease in BTEX concentrationsfrom MW1 to MW2 suggests the occurrence of intrinsicattenuation of BTEX. Results show that significantamount of total cultivable heterotrophs and total anaer-obes (>105 cells per g of soil) were detected in SB1 and

Fig. 3. DGGE profiles of the PCR-amplified 16S rDNA for soils collectedfrom SB1, SB2, and SB3 on day 0 and 300.

Table 2Variations in the intensities of the selected 22 strains versus time in SB1

Intensity

Strain D 0 D 300 Strain D 0 D 300

1 0 21008 12 17445 332582 0 12134 13 52441 145793 0 16584 14 13519 171514 0 32445 15 14213 164735 23712 15447 16 0 173746 0 27836 17 7354 174997 11302 40227 18 16432 51278 7249 50776 19 0 301979 41349 10442 20 10258 990310 11230 9847 21 0 1024811 22587 21305 22 11243 9972

SB2 soil samples collected from the contaminated areas.Compared to the background soil sample (SB3), SB1 andSB2 contained more bacterial population. The increasedbacterial population might be due to the supplement of car-bon sources (petroleum hydrocarbons) to the subsurfacemicroorganisms. The observed soil organic content waslow (<0.51%) (Table 1), and thus, the natural organic car-bon would not be the major cause of the significantincrease in microbial population in the soil samples afterbiosparging process. This indicates that higher BTEX con-centration caused the increased bacterial population. Fieldresults also indicate that the anaerobic biodegradation pat-terns were the dominant intrinsic biodegradation processes.Because bioremediation rates under aerobic conditions aregenerally higher those under anaerobic conditions, airinjection into the subsurface would enhance the BTEX bio-degradation rates.

Results from the field investigation also reveal that thebiosparging process caused significant changes in environ-mental conditions. Increased DO in the groundwater acti-vated the aerobic microorganisms and enhanced theBTEX removal rates. Effects of biosparging on the varia-tions in indicating parameters and enhancement of aerobicbiodegradation include: (1) increase in DO, CO2, redoxpotential, nitrate, and sulfate within the plume; (2) decreasein pH, dissolved ferrous iron, sulfide, and methane withinthe plume; (3) increased total cultivable heterotrophswithin the plume; and (4) decreased total anaerobes andmethanogens within the plume.

Table 3Comparison of the nucleotide sequences of 16S rDNA of 22 specificmicroorganisms in SB1 with the database from Gene Bank

Strain Microorganisms Similarity(%)

1 Ensifer sp. (AF229863) 902 Geitlerinema sp. (AF132780) 873 Uncultured bacterium (AY917343) 1004 Uncultured bacterium (AB195911) 965 Uncultured bacterium (AJ133615) 1006 Uncultured green non-sulfur bacterium

(AJ441228)96

7 Sulfurihydrogenibium sp. (DQ906006) 988 Boyliae praeputiale (AY101388) 969 Pseudomonas saccharophila (AF368755) 92

10 Variovorax paradoxus (AF451851) 9611 Uncultured bacterium (AJ289998) 9912 Rubrivivax gelatinosus (AF487435) 9313 Methylobacterium sp. (AY436812) 9314 Sulphate reducing bacterium (AJ300515) 9715 Green non-sulfur bacterium (AF027032) 9516 b-Proteobacterium (U46748) 9417 Methylobacterium sp. (M95655) 9818 Rhodothermus marinus (AF217499) 9519 Bacterium Ellin6089 (AY234741) 9820 Candidauts Magnetobacterium bavaricum

(X71838)99

21 Uncultured bacterium (AJ441230) 9822 Aquificales str. (AF255597) 96

Fig. 4. The Phylogeny tree for illustrating relationships among 22 microbial strains collected from SB1. The bootstrap value, as determined from 1000bootstrap samples, is presented at each node (in percent). Only bootstrap values of P50% are presented.

C.M. Kao et al. / Chemosphere 70 (2008) 1492–1499 1497

Fig. 5. The UPGMA dendrogram for illustrating relationships among microorganisms collected from SB1, SB2, and SB3 at different time points.

1498 C.M. Kao et al. / Chemosphere 70 (2008) 1492–1499

3.2. Change of microbial community structures after the

operation of biosparging

To determine if microbial community patterns in envi-ronment were changed due to the biosparging process,the PCR-DGGE was performed to investigate the domi-nant microorganisms on BTEX biodegradation. On day300, the numbers of DGGE bands in SB1 or SB2 sedimentswere greater than that in the background sediments SB3(Fig. 3). Results of the microbial enumeration of soil sedi-ments reveal that the total number of cultivable hetero-trophs in SB1 and SB2 was substantially increased onday 300 (Table 1). As shown in Table 2, the amounts anddiversity of organisms were dramatically changed on day300 after the operation of biosparging. In the instance ofSB1 sediment sample, a total of 22 bands on DGGE profil-ing were clear in an appearance of their intensities after a300-day operation (Fig. 3). Results also show that theintensities of eight bands (5, 10, 11, 12, 14, 15, 20, and22) had slight changes (<2-fold). Moreover, eight novelbands (1, 2, 3, 4, 6, 16, 19, and 21) were observed. Intensi-ties of three bands (7, 8, and 17) were significantlyincreased (>2-fold), and intensities of another three bands(9, 13, and 18) were significantly decreased (>2-fold).Results also show that bands 5, 9, 11, and 13 were amongthe dominant bands on day 0 under the anaerobic condi-tions (appearance of anaerobic condition in highly contam-inated areas). However, bands 1, 4, 6, 7, 8, 11, 12, and 19became dominant on day 300 under aerobic conditionsafter the onset of biosparging process. Band 11 was theonly one, which was able to remain high intensity duringthe shifting of redox conditions from anaerobic to aerobicenvironment. Although the band intensity provides usinformation about the dominance of the bacteria in the soilenvironment, it might not be necessarily a valid estimate ofchanges in population density.

To determine the meaning of representatives for bacte-rial species, the bands of DGGE profiles in SB1 were elutedand then amplified and sequenced for their nucleotidesequences of 16S rDNA variable V3 regions. As shown in

Table 3, the identities of the nucleotide sequences of 22dominant bands are shown to be in a range of 87–100%of specific microorganisms as compared to database ofGeneBank (Table 3). Using the similarity coefficients in16S cDNA gene sequences, an UPGAM dendrogram allo-cated 22 specific microorganisms in this population intotwo major separate phylogenetic clusters (A and B)(Fig. 4). In Cluster A, the representatives of bands 2, 3,5, 6, 7, 8, 20, 21, and 22 were closely related to Candidauts

magnetobacterium. In Cluster B, strains 4 and 11 were clo-sely related to Flavobacteriales bacterium and Bacteroidetes

bacterium. C. magnetobacterium, F. bacterium, and B. bac-

terium have been reported to able to biodegrade petroleumhydrocarbons under aerobic conditions (Greene et al.,2000; Prince, 2000; Watanabe et al., 2000; Duarte et al.,2001; Fiorenza and Rifai, 2003). This indicates that BTEXdegrading bacteria in SB1 could be positively selected aftera 300-d operation of biosparging process. The dominantmicroorganisms involving in BTEX degradation could beexploited and isolated for their application on the bioreme-diation of BTEX-contaminated sites.

Fig. 5 presents the UPGMA dendrogram for illustratingrelationships among microorganisms collected from SB1,SB2, and SB3 at different time points (0, 60, 120, 180,240, or 300 d after the onset of the biosparging). Resultsshow that the microbial species in those three areas couldbe grouped into two major phylogenetic clusters. SB1and SB2, which were located at source and downgradientareas, contained similar genetic information in DNA andare somewhat related. Moreover, the microbial species inSB3 were not closely related to SB1 and SB2. This indicatesthat the released BTEX have caused the variations in thedominant microbial species within the BTEX plume.Results also show that both sampling location and sam-pling time would cause the shifting of microbial species.

4. Conclusions

Results from this field-scale study indicate that naturalattenuation was the major cause of the decrease in BTEX

C.M. Kao et al. / Chemosphere 70 (2008) 1492–1499 1499

concentrations in groundwater before the operation of bio-sparging system. Evidence of the occurrence of naturalattenuation includes the following: (1) decrease of DO,nitrate, sulfate, and Eh within the plume, (2) productionof dissolved ferrous iron, sulfide, methane, and CO2 withinthe plume, (3) decreased BTEX concentrations along thetransport path, (4) increased microbial populations withinthe plume, and (5) limited spreading of the BTEX plume.Field results also reveal that the operation of biospargingcaused the shifting of anaerobic conditions inside theplume to aerobic conditions. This variation can be con-firmed by the following field observations inside the BTEXplume due to the biosparging process: (1) increase in DO,Eh, nitrate, and sulfate, (2) decrease in dissolved ferrousiron, sulfide, and methane, (3) increased total cultivableheterotrophs, and (4) decreased total anaerobes as well asmethanogens. Results show that the aerobic biodegrada-tion was the dominant degradation processes of BTEXafter the operation of biosparging. According to the resultsfrom GeneBank, three microorganisms, C. magnetobacteri-

um, F. bacterium, and B. bacterium, which can biodegradeBTEX under aerobic conditions might exist at this site.Results also reveal that DGGE and nucleotide sequencetechniques provide a guide for microbial ecology, whichcan be used as an indication of the trend of the biodegra-dation process. Moreover, the significant decrease (morethan 70%) in BTEX concentrations within the plume alsoindicates that biosparging might be a very promising tech-nology to remediate BTEX contaminated groundwater.Further field investigation is a necessity to confirm theremoval efficiency and removal mechanisms of BTEX ingroundwater via the biosparging process. Results from thisstudy provide us insight into the characteristics of aerobicbiodegradation of BTEX. The knowledge and comprehen-sion obtained in this study will be helpful in designing abiosparging system for the bioremediation of BTEX-con-taminated site.

Acknowledgements

This study was funded by National Science Council inTaiwan. Additional thanks to Mr. C.Y. Hsieh and L.W.Wang of National Sun Yat-Sen University for their assis-tance throughout this project.

References

Adams, J.A., Reddy, K.R., 2003. Extent of benzene biodegradation insaturated soil column during air sparging. Ground Water Monit.Remediat. 23, 85–94.

APHA, 2001. Standard methods for the examination of water andwastewater. American Public Health Association, 21th ed. APHA-AWWA-WEF, Washington, DC, USA.

Brar, S.K., Verma, M., Surampalli, R.Y., Misra, K., Tyagi, R.D.,Meunier, N., Blais, J.F., 2006. Bioremediation of hazardous wastes – a

review. Pract. Period. Hazard. Toxic. Radioact. Waste Manage. 10,59–72.

Chen, K.F. 2005. Evaluation of the biodegradability of MTBE ingroundwater. Ph.D. Dissertation, Institute of Environmental Engi-neering, National Sun Yat-Sen University, Taiwan.

Chen, K.F., Kao, C.M., Wang, J.Y., Chien, C.C., 2005. Naturalattenuation of MTBE at two petroleum-hydrocarbon spill sites. J.Hazard. Mater. 125, 10–16.

Chen, K.F., Kao, C.M., Chen, T.Y., Weng, C.H., Tsai, C.T., 2006.Intrinsic bioremediation of MTBE-contaminated groundwater at apetroleum-hydrocarbon spill site. Environ. Geol. 50, 439–445.

Deeb, R.A., Chu, K.H., Shih, T., Linder, S., Suffet, I.H., Kavanaugh,M.C., Alvarez-Cohen, L., 2003. MTBE and other oxygenates: envi-ronmental sources, analysis, occurrence and treatment. Environ. Eng.Sci. 20, 433–447.

Duarte, G.F., Rosado, A.S., Seldin, L., de Araujo, W., van Elsas, J.D.,2001. Analysis of bacterial community structure in sulfurous-oil-containing soils and detection of species carrying dibenzothiophenedesulfurization gene. Appl. Environ. Microb. 67, 1052–1062.

Fiorenza, S., Rifai, H.S., 2003. Review of MTBE biodegradation andbioremediation. Biorem. J. 7, 1–35.

Greene, E.A., Kay, J.G., Jaber, K., Stehmeier, L.G., Voodouw, G., 2000.Composition of soil microbial communities enriched on a mixture ofaromatic hydrocarbons. Appl. Environ. Microb. 66, 5282–5289.

Kao, C.M., Wang, C.C., 2000. Control of BTEX migration by intrinsicbioremediation at a gasoline spill site. Water Res. 34, 3413–3423.

Kao, C.M., Chen, Y.L., Chen, S.C., Yeh, T.Y., Wu, W.S., 2003.Enhanced PCE dechlorination by biobarrier systems under differentredox conditions. Water Res. 37, 4885–4894.

Kao, C.M., Huang, W.Y., Chang, L.J., Chien, H.Y., Hou, F., 2005.Application of monitored natural attenuation to remediate a petro-leum-hydrocarbon spill site. Water Sci. Technol. 53 (2), 321–328.

Moreels, D., Bastiaens, L., Ollevier, F., Merckx, R., Diels, L., Springael,D., 2004. Evaluation of the intrinsic methyl tert-butyl ether (MTBE)biodegradation potential of hydrocarbon contaminated subsur-face soils in batch microcosm systems. FEMS Microb. Ecol. 49,121–128.

NIEA, National Institute of Environmental Analysis, 2002. Methods ofSoil Analysis, Taiwan Environmental Protection Administration,Taiwan.

NIEA, National Institute of Environmental Analysis, 2003. Methods ofSoil and Groundwater Sampling, Taiwan Environmental ProtectionAdministration, Taiwan.

Prince, R.C., 2000. Biodegradation of methyl tertiary-butyl ether (MTBE)and other fuel oxygenates. Crit. Rev. Microb. 26, 163–178.

Saitou, N., Nei, M., 1987. The neighbor-joining method: a new method forreconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425.

Schirmer, M., Butler, B.J., Church, C.D., Barker, J.F., Nadarajah, N.,2003. Laboratory evidence of MTBE biodegradation in Borden aquifermaterial. J. Contam. Hydrol. 60, 229–249.

Schmidt, T.C., Schirmer, M., Weib, H., Haderlein, S., 2004. Microbialdegradation of methyl tert-butyl ether and tert-butyl alcohol in thesubsurface. J. Contam. Hydrol. 70, 173–203.

Sutherland, J., Adams, C., Kekobad, J., 2004. Treatment of MTBE by airstripping, carbon adsorption, and advanced oxidation: technicaland economic comparison for five groundwaters. Water Res. 38,193–205.

Watanabe, K., Kodama, Y., Syutsubo, K., Harayma, S., 2000. Molecularcharacterization of bacterial populations in petroleum-contaminatedgroundwater discharged from underground crude oil storage cavities.Appl. Environ. Microb. 66, 4803–4809.

Wu, Y.W., Huang, G.H., Chakma, A., Zeng, G.M., 2005. Separation ofpetroleum hydrocarbons from soil and groundwater through enhancedbioremediation. Energy Sources 27, 221–232.