Reductive dechlorination of trichloroethylene by combining autotrophic
hydrogen-bacteria and zero-valent iron particles
Shang-Ming Wang
*, Szu-kung Tseng
Graduate Institute of Environmental Engineering, National Taiwan University, No. 71 Chou-Shan Road, Taipei 106, Taiwan, ROC
a r t i c l e
i n f o
Article history:
Received 19 February 2008
Received in revised form 22 May 2008 Accepted 23 May 2008
Available online 7 July 2008 Keywords: TCE Zero-valent iron Dechlorination Autotrophic hydrogen-bacteria
a b s t r a c t
The objective of this study was to evaluate the dechlorination rate (from an initial concentration of 180lmol l1) and synergistic effect of combining commercial Fe0and autotrophic hydrogen-bacteria in the presence of hydrogen, during TCE degradation process. In the batch test, the treatment using Fe0 in the presence of hydrogen (Fe0/H
2), showed more effective dechlorination and less iron consumption than Fe0utilized only (Fe0/N
2), meaning that catalytic degradation had promoted transformation of TCE, and the iron was protected by cathodic hydrogen. The combined use of Fe0and autotrophic hydro-gen-bacteria was found to be more effective than did the individual exercise even though the hydrogen was insufficient during the batch test. By the analysis of XRPD, the crystal of FeS transformed by sulfate reducing bacteria (SRB) was detected on the surface of iron after the combined treatment. The synergistic impact was caused by FeS precipitates, which enhanced TCE degradation through catalytic dechlorina-tion. Additionally, the dechlorination rate coefficient of the combined method in MFSB was 3.2-fold higher than that of iron particles individual use. Results from batch and MFSB experiments revealed that, the proposed combined method has the potential to become a cost-effective remediation technology for chlorinated-solvent contaminated site.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
Halogenated compounds, such as trichloroethylene (TCE) and perchloroethylene (PCE), are ubiquitous priority groundwater pol-lutants that pose a risk to public health. Most of these solvents are highly toxic; therefore, the remediation of polluted sites is desir-able. The recalcitrance of halogenated compounds depends on number, position and type of the halogen substituents. Owing to the electro-negativity of the halogen atom, these organo-halogens
are difficult to degrade via oxidative processes (Magnuson et al.,
1998). Many of the halogenated solvents exist in the subsurface
as dense non-aqueous phase liquids (DNAPLs) because of their densities and low water solubility; these physi-chemical character-istics make them difficult to remove, via the pump-and-treat method. Permeable reactive barriers (PRBs) containing zero-valent
iron (Fe0) are becoming increasingly popular for in situ
remedia-tion of groundwater contaminated with chlorinated solvents (Matheson and Tratnyek, 1994; Burris et al., 1995; Orth and Gill-ham, 1996; Roberts et al., 1996; Wüst et al., 1999; Choe et al., 2001; Cervini et al., 2002). There are three major reaction
path-ways involved in the dechlorination of Fe0(Matheson and
Trat-nyek, 1994; Burris et al., 1995; Orth and Gillham, 1996): (A)
direct electron transfer from iron surface (Eq. (1)); (B) catalyzed
hydrogenolysis by hydrogen (H2), which is produced by reduction
of H2O during anaerobic iron corrosion (Eq.(5)); (C) reduction by
Fe2+, which is from corrosion of Fe0(Eq.(4))
Fe2þþ 2e$ Fe0 ð1Þ 2H2O þ 2e$ H2þ 2OH ð2Þ Fe0 þ 2H2O $ Fe2þþ H2þ 2OH ð3Þ 2Fe2þþ RH þ Hþ! 2Fe3þþ RH þ X ð4Þ H2þ RX ! catalyst RH þ Hþþ X ð5Þ
A potential limitation of the Fe0remediation technology is the
dete-rioration of the iron materials by corrosion and the subsequent pre-cipitation of minerals, which may cause cementation and decreased
the permeability of PRBs (Reardon, 1995; Gu et al., 1999; Heuer and
Stubbins, 1999; Phillips et al., 2000; Agrawel et al., 2002; Ritter et al., 2002).
Since most of the contaminated sites are anaerobic, stimulation of reductive dehalogenation is a very promising strategy for bio-remediation aquifers contaminated with halogenated compounds. Reductive dechlorination is either a cometabolic or a respiratory process (halorespiration), which is carried out by organochlorine-reducing bacteria. The dechlorination rates of the cometabolic
pro-cesses are much lower than halorespiration (Yager et al., 1997),
and their contributions to dehalogenation observed in natural 0960-8524/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2008.05.033
* Corresponding author. Tel.: +886 918382492; fax: +886 2 23632637. E-mail address:[email protected](S.-M. Wang).
Contents lists available atScienceDirect
Bioresource Technology
environments are negligible. In contrast, organochlorine-reducing bacteria, which utilize chlorinated hydrocarbons as electron accep-tors in energy metabolism and hence for growth are likely to be major contributors in anaerobic environments. Unfortunately, many laboratory and in situ observations show that most halores-pirating microorganisms produce cis-dichloroethene (cis-DCE) and vinyl chloride (VC) as terminal products. Only a few microorgan-isms can completely dechlorinate PCE and TCE to ethene or ethane (Maymó-gatell et al., 1995; Magnuson et al., 1998; Holliger et al.,
1999). The fate of VC in field sites is of particular concern because
the chlorinated ethene is a human carcinogen.
In recent years, several articles have been devoted to the study
of combining Fe0 and cells to degrade chlorinated hydrocarbons
(Weathers et al., 1997; Lee et al., 2001). These studies propose
that the cathodic hydrogen produced from Fe0 can support
dechlorination by microorganism. In other words, the Fe0 is
merely used as a time-released hydrogen source and the cathodic hydrogen is utilized as an electron donor for respiratory
dechlori-nation. Although the results reveal that the combined use of Fe0
and cells showed more effective dechlorination than did the
indi-vidual use, the Fe0still corroded after reaction. The sustainability
of Fe0remains a potential limitation for in situ remediation and
mineral clogging would occur. This research is intended as an
investigation of TCE dechlorination using Fe0 and autotrophic
hydrogen-bacteria under hydrogen supplied environment. In this combined system, autotrophic hydrogen-bacteria utilized
hydro-gen to carry out halorespiration, and Fe0 will be protected due
to the high affinity with hydrogen (there will be a hydrogen
coat-ing around Fe0particles). When the TCE is adsorbed on the surface
of the Fe0, catalytic dechlorination will occur, completely and
rap-idly, in the system (Eq. (5)), therefore, dechlorination could be
achieved in this combined system. The objectives of this paper were; (1) to determine if zero-valent iron could accelerate TCE dechlorination via catalyzed hydrogenolysis; (2) to identify the reduction of iron consumption during this combined system; (3) to evaluate the dechlorination efficiency of TCE in a bioreactor with continuous hydrogen input. A serum bottle batch test was utilized to evaluate the dechlorination efficiency and corrosion
of Fe0
; a membrane feeding substrate bioreactor (MFSB) was set up to examine the removal of TCE with continuously supplied hydrogen.
2. Methods
2.1. Sludge, medium and growth conditions
The digest sludge, utilized in this research, was gathered from a livestock breeding wastewater treatment plant (mixed culture) and was incubated in the growth medium containing the follow-ing; (final concentration in grams per liter): NaHCO3, 2.0; KH2PO4,
2.0; MgSO4 7H2O, 0.148; N-(2-hydroxyethyl)-N-2-ethanosulfonic
acid (HEPES), 11.2, and 1 ml of a trace metal solution comprising
(grams per liter) CaCl2 2H2O, 7.0; MnCl2 4H2O, 2.5; FeCl3, 1.8;
CoCl2 6H2O, 0.5; (NH4) Mo7O24 2H2O, 0.5; CuCl2, 0.1. Before
incubation, the growth medium was purged with N2to avoid the
interference of dissolved oxygen. Digest sludge was incubated in a sealed, continuously stirred reactor, which had a working vol-ume of 2.4 l at 25 °C, purged with hydrogen, via a cylinder (10 ml/min). Pure TCE was injected into the incubation reactor,
the initial TCE concentration was 180
l
mol l1. The pH of mediumsolution was adjusted by 20% H3PO4, to approximately 7.0. The di-gest sludge was maintained by weekly transfer into a 250 ml ser-um bottle, which contained the same growth mediser-um as the reactor, and the subculture had undergone 3 serial transfers before use.
2.2. Batch experiments
Commercial granular iron (purity > 99.5%, 100 mesh, Hognas) was chosen in this experiment. Because there is an oxidation layer on iron particles, before reactions, we immersed the iron into 1 M
HCl solution for 24 h to remove impurities of the iron surface (
Ar-nold and Roberts, 2000). Initially, the reactants (Fe0or sludge) were put into 125 ml serum bottles and then filled with medium solu-tion. Following this, the vials were subsequently sealed with Tef-lon-lined, butyl-rubber septa, and aluminum crimp caps. Then, the serum bottles were flushed with hydrogen or nitrogen gas through the septa to displace the medium solution (60 ml). Finally,
TCE was added to the vials from Teflon septa with a 2
l
l GCsam-pling syringe and sealed with wax. The initial TCE concentration of
the vials was 180
l
mol l1. All the well-sealed vials were incubatedon a circular action shaker at 250 rpm (25 °C). With the exception of the sludge, all the materials (medium, serum bottles, Teflon sep-ta, and sampling syringe) were autoclaved at 120 °C for 20 min be-fore the experiment. A description of the batch experiment is
shown inTable 1.
2.3. Membrane feeding substrate bioreactor (MFSB)
The bioreactor was constructed of Pyrex with an effective vol-ume of 2.4 l. A gas-permeable silicone tube (1.5 mm [i.d.] and 2.5 mm [o.d.] by 11.5 m long; Fuji System Co., Japan) was wound around the glass pillar in the bioreactor. Zero-valent iron particles and hydrogen-bacteria were added to the reactor, which was com-pletely stirred by a propeller. Hydrogen injected into the silicone tube was supplied with gas cylinders via mass controllers. When hydrogen flowed through the silicone tube, it would diffuse the growth medium solution. The dissolved hydrogen provides the en-ergy source to autotrophic hydrogen-bacteria as well as forming a protective layer around the iron surface. Catalytic hydrolysis would
proceed when TCE adsorbs on the iron (Eq.(5)). A membrane
feed-ing substrate bioreactor (MFSB) was utilized to determine dechlo-rination efficiency of different treatments, when continuous hydrogen was supplied. When the reactor was running, it was
Table 1
The constituents of batch experiments Components
Headspace (ml) TCE (lmol l1) Contents of
immobilization Fe0 (g) Sludge (ml) Set 1(Fe0 + H2) 60 180 1.5 0 Set 2(Fe0 + N2) 60 180 1.5 0 Set 3 (sludge + H2) 60 180 0 10 Set 4 (Fe0 + sludge + H2) 60 180 1.5 10 Set 5 (control) 60 180 0 0
Final mass of TCE in each serum bottle (125 ml) is 22.5lmol and the volume of headspace is 60 ml. Table 2 Descrption of MFSB treatments Components Headspace (ml) TCE (lmol l1) Contents of immobilization Fe0(g) Sludge (ml) Set A (Fe0+ H 2) 5 180 20 0 Set B (Fe0+ N 2) 5 180 20 0 Set C (sludge + H2) 5 180 0 100 Set D (Fe0 + sludge + H2) 5 180 20 100 Set E (control) 5 180 0 0
filled with a medium solution (headspace is 5 ml) and kept in a water bath to maintain the temperature at 25 °C TCE was injected through Teflon septa on the head of the reactor, the TCE
concentra-tion was 180
l
mol l1. Properties of the reactors are summarizedinTable 2.
2.4. Analytical methods
TCE, cis-DCE, VC, ethane, ethene, acetylene and hydrogen were detected by gas chromatography (GC) using headspace analysis.
A 500
l
l side-port, gas-tight GC sampling syringe (Hamilton) wasused to withdraw the gas sample from the headspace of the vials and bioreactors and then injected into the GC. TCE and the reaction products during the experiment (cis-DCE, VC, ethane, ethene, acet-ylene and methane) were analyzed using a Hewlett–Packard 5890 gas chromatography equipped with a flame ionization detector (GC/FID) and a GS-GasPro capillary column (60 m 0.32 mm i.d., 0.25 mm film thickness GS-GasPro capillary column; J&W Scien-tific, Folsom, CA). The temperature of injector and detector were 200 and 300 °C, respectively. The temperature program was as
fol-lows: 28 °C hold 3.5 min, 15 °C min1to 115 °C no hold, 7 °C min1
to 180 °C, hold 2 min. The column flow rate was 2.5 ml min1and
the detention time of each compound was: CH4, 1.23 min; C2H4,
1.93 min; C2H6, 2.33 min; C2H2, 2.78 min; trans-DCE, 12.83 min; cis-DCE, 14.73 min; TCE, 16.41 min. Hydrogen was analyzed on a Hewlett–Packard 5890 gas chromatography equipped with a
ther-mal conductivity detector (GC/TCD) (Weathers et al., 1997; Lee et
al., 2001) using a packed column (3.2 mm by 3.05 m stainless steel packed with 100/120 carbon sieve-S-II, Supelco, Inc.), and the oper-ation temperature was 150 °C isothermally.
Total iron consumption was quantified by a modified
1,10-phe-nanthroline test, as described byMarchand and Silverstein (2002).
A 1-ml sample of suspension was collected, and then 9 ml of 10% HCl was added to the gathered samples to dissolve the colloid and precipitates. After 30-min digestion at 150 °C, 20 ml
NH2OH HCl (50 g l1) was injected into the samples to reduce
Fe3+to Fe2+. Finally, 10% (v/v) of a 2.0 g l1solution of
1,10-phenan-throline was added to react with Fe2+. The color was measured at
510 nm in a spectrophotometer (Thermo, Spectronic 20D) after a 15-min reaction time. The analysis of sulfate was measured using a Dionex 120 ion chromatography (IC) system with a Dionex
Ion-pac AS-16 column after filtration (0.22
l
m filter, Millipore).Sam-ples, which reacted with microorganisms, required centrifuge before filtration and IC analysis.
Scanning electron microscopy energy dispersive X-ray (SEM-EDX) and X-ray powder diffraction (XRPD) were introduced to identify the chemical constituents and crystal phases of the precip-itates, which was gathered from bioreactors and vials. Samples were stored under He atmosphere at 20 °C until analysis of XRPD and EDX. After drying, samples were shaken onto amorphous car-bon film and supported by a copper adapter to avoid electrical charging during analysis. The samples were observed using the bright-field imaging mode of SEM (Topcon ABT-150s) with an
acceleration voltage of 15 kV, at a take-off angel of 47.76o. XRPD
(Regaku D/max-II B) was performed with 35 kV and 20 mA current
from 5oto 90o(2h) at a scanning speed of 4omin1to identify the
crystal phases of the precipitates.
3. Results and discussion 3.1. Dechlorination of the batch test
The plot of TCE concentration against reaction time in the batch
vials under different incubation conditions are shown inFig. 1. Sets
1 and 2 were treated with identical concentrations of iron powder,
but supplemented with different source gases in the headspace of each set of vials. Hydrogen gas was added to set 1 rather than set 2 in order to investigate if TCE dechlorination would be enhanced by catalyzed hydrogenolysis in the presence of hydrogen. Initially, the
TCE in both vials containing only Fe0obtained high removal
effi-ciency; however, after two days of incubation, the degradation rate
in Fe0/N2vials decreased slightly, whereas TCE transformation in
Fe0/H2vials continued to react rapidly. The efficiency of
dechlori-nation varied in the later period based on different types of gas-filled headspace. The TCE degradation mechanism in set 2 was
mainly concerned with iron corrosion (Eqs. (1)–(4); while set 1
was correlated most strongly with catalytic dechlorination, be-cause set 1 was prevented from the severe corrosion of iron with the hydrogen-filled headspace. In comparison, catalytic dechlori-nation would be the main reaction mechanism in set 1. It was ob-served that, greater than 99% of the added TCE were degraded in
set 1 (Fe0/H2) within 13 days, however, the same degradation
was achieved in set 2 (Fe0/N
2) in 18 days. These results suggested
that the treatment with Fe0could be dechlorinated much more
ra-pid in the presence of hydrogen.
Digested sludge and hydrogen gas were introduced into set 3 for examination of the effect of hydrogen-bacteria on TCE
biodeg-radation, via halorespiration. As shown inFig. 1, the TCE
degrada-tion rate was significantly enhanced by the sludge incubated in set 3 after a lag period of two days. Unfortunately, the biodegradation of TCE ceased after seven days of reaction, because of a hydrogen shortage in the headspace (more details on hydrogen concentra-tion will be provided in the later secconcentra-tion). The cells showed that
42% of the added TCE (9.3
l
mol) was degraded during 21 days ofincubation; yet complete biodegradation was almost obtained
after only 7 days of incubation, as illustrated inFig. 1, indicating
that the later reaction came close to a halt. Accordingly, it was as-sumed that the treatment with the cells might be degraded in a better manner with a sufficient supply of hydrogen.
Set 4 was added with Fe0, digested sludge, and a hydrogen
sup-plement to the headspace. The results as shown inFig. 1suggest
that the combined use and concurrent treatment of Fe0and
auto-trophic hydrogen-bacteria exhibited the fastest degradation rate
to remove most of the TCE within 10 days. In addition,Fig. 1also
shows that the length of the reaction time would not reduce the efficiency of set 4, in contrast, TCE degradation in set 2 might be-come negligible, a result of iron corrosion in the post period. As for control vials containing medium solution with hydrogen-filled headspace (set 5), the emission concentrations of TCE revealed less than 5% total over the duration of the batch test. Abiotic loss and degradation of indigenous population in medium solution or vials might be the major reason for TCE decrease, which is too slight to be significant.
reaction time (days)
0 10 15 20 mass of TCE ( μ mol) 0 5 10 15 20 25 set 1 (Fe+H2) set 2 (Fe+N2) set 3 (cells + H2) set 4 (Fe + cells + H2) control
5
Fig. 1. The mass of TCE against reaction time in batch test under different incubation conditions. The initial TCE mass in serum bottle was 22.5lmol (180lmol l1).
When TCE concentrations, as shown inFig. 1, were plotted as;
ln(Ct/C0) versus t (where Ctis TCE concentration at reaction time
t, and Corefers to the initial concentration of TCE), the data
pro-duced linear plots, indicating that TCE degradation followed first order kinetics (non-linear data in lag-phase and steady-state data are not included). The slopes of these plots were equivalent to the pseudo-first-order rate coefficient in each set. The TCE degra-dation rate coefficient for the treatment of each set is as follows:
0.36 day1 (r2= 0.97) for Fe0 and hydrogen (set 1), 0.24 day1
(r2= 0.98) for Fe0 and nitrogen (set 2), 0.08 day1(r2= 0.95) for
cells only (set 3), and 0.48 day1(r2= 0.99) for combined use of
Fe0and cells in presence of hydrogen (set 4).
The dechlorination intermediates, products types, and TCE concentrations against reaction time in each set are shown in
Figs. 2–4. FromFig. 2, it is revealed that, cis-DCE was the most
com-mon intermediate in all dechlorination treatments during the experiment. In set 2, the accumulation of cis-DCE rose to a peak
of 1
l
mol on the 3rd day, and then gradually degraded to ethane,or ethene, until the concentration of cis-DCE fell below the detec-tion limit on 15th day, indicating that cis-DCE was a reactive
inter-mediate during Fe0 treatment, and hence could be completely
dechlorinated. Although similar trends were observed in set 1, it
revealed far less accumulated cis-DCE (about 0.5
l
mol), and a morerapid disappearance of cis-DCE in the catalytic dechlorination sys-tem; thus, no additional cis-DCE accumulation could be found on 7th day in set 1. That is to say, the transformation of cis-DCE would be accelerated by catalyzed hydrogenolysis, as compared with
tra-ditional Fe0treatments, as in set 2). Formation and accumulation of
cis-DCE were observed in the biological vials of set 3, which exhib-ited a dramatic increase in the concentration of cis-DCE, rising to
the maximum on the 9th day (over 3
l
mol), and being persistentin vials to the end of incubation. The results showed that auto-trophic hydrogen-bacteria reacted worse to the dechlorination intermediates, with the possibility of being influenced by biologi-cal toxicity, referring to the difficulty of complete dechlorination proceedings, via simple cells. Rapid formation and disappearance
of cis-DCE was observed in the vials containing Fe0and cells with
hydrogen (Fig. 2), with highest concentration (0.24
l
mol) aftertwo days and complete disappearance after five days. In
compari-son with Fe0-only in a biological system, these evidences support
the hypothesis of the higher reactivity of the combined system on dechlorination, and transformation, of chlorinated metabolites. In each set of batch experiments, ethane and ethene accumula-tion were shown as end products of TCE transformaaccumula-tion, in addi-tion, a low-concentration of methane was produced. The concentration-versus-time profiles of ethane and ethene in batch
experiments are displayed inFigs. 3 and 4. Treatments with Fe0
and hydrogen-bacteria revealed the highest extent in accumula-tion of ethane and ethene among the five sets of vials. This repre-sents that the most complete dechlorination occurred within the combined system, namely, the disappeared TCE in these vials was totally converted to non-chlorinated hydrocarbons (ethane and ethene). However, the composition of the products in biologi-cal treatment (set 3) was largely different from in the combined system. In set 3, merely 55% of the disappeared TCE was converted into non-chlorinated hydrocarbons (ethane and ethene), and nearly 33% of that transformed to cis-TCE; undeniably, it could be the lack of hydrogen in the vials that counts for the lower reductive capability of biodechlorination.
The results fromFig. 1revealed that, 42% TCE was removed in
Set 3, however, the TCE removal mechanisms involved halorespira-tion and cell-adsorphalorespira-tion, meaning that only a part of removed TCE was reduced through biodegradation. From the above, the TCE was decomposed to cis-DCE, ethane and ethane; therefore, the amount of TCE adsorbed on sludge could be estimated by mass balance of
residual TCE and dechlorination products. According toFigs. 2–4,
the mass of cis-DCE, ethane, and ethane were 3.0, 3.9, and 0.9
l
mol,respectively. The total amount of carbon contained in these
prod-ucts was 15.6
l
mol (each mole product contains two moles ofcar-bon).The amount of TCE degraded in set 3 was 9.3
l
mol (18.6l
molof carbon), meaning that the difference between mass balance and
degradation was 1.5
l
mol of TCE (3.0l
mol of carbon). Hence, theTCE adsorbed by sludge was approximately 1.5
l
mol (6.7%). Inother words, only 35.3% TCE (7.8
l
mol) was actually degraded bycells.
3.2. Impacts of hydrogen concentration on batch tests
Hydrogen is an excellent energy source in a variety of auto-trophic anaerobes, such as methanogens, sulfate reducers and
halorespirers (Holliger et al., 1999; Gu et al., 1999). Therefore,
suf-reaction time (days)
0 10 15 20 mass of c is-DCE ( μ mol ) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 set 1 ( Fe + H2) set 2 ( Fe + N2) set 3 ( cells + H2 ) set 4 ( Fe + cells + H2) set 5 ( control ) 5
Fig. 2. The mass of cis-DCE against reaction time in batch test under different incubation conditions.
reaction time (days)
0 10 15 20 mass of ethene ( μ mol) 0 2 4 6 8 10 12 14 set 1 ( Fe + H2) set 2 ( Fe+ N2) set 3 ( cells + H2) set 4 ( Fe+ cells + H2) set 5 ( control )
5
Fig. 3. The mass of ethene against reaction time in batch test under different incubation conditions.
reaction time (days)
0 10 15 20 mass of ethane ( μ mol) 0 1 2 3 4 5 set 1 ( Fe + H2 ) set 2 ( Fe + N2 ) set 3 ( cells + H2 ) set 4 ( Fe + cells + H2 ) set 5 ( control ) 5
Fig. 4. The mass of ethane against reaction time in batch test under different incubation conditions.
ficient supply of hydrogen would be a primary factor for halorespi-ration. In this study, in addition to being regarded as an electron donor during biological dechlorination, hydrogen also played the important roles of reducer, in the catalytic dechlorination (Eq.
(5)), as well as reducing gas in protection of iron oxidation. From
the results of hydrogen concentration analysis, as detected by GC-TCD, show that 30% of the added hydrogen was preserved in vials of set 1 to the end of the treatment, meaning that the hydro-gen supply in the headspace was abundant for catalytic hydrohydro-gen- hydrogen-olysis (not shown in figures). The treatment containing cells only (set 3) had exhausted the supply of hydrogen in vials within 8 days, and consequently, the biological dechlorination ceased along with
a great accumulation of cis-DCE (Fig. 2). FromFigs. 3 and 4, it was
obvious that the autotrophic hydrogen-bacteria introduced in this experiment could degrade TCE to non-chlorinated metabolite (eth-ane and ethene), hence the lower efficiency of TCE degradation and the accumulation of cis-DCE during the latter stage of reaction in set 3, were strongly dependent on the lack of hydrogen. The
hydro-gen consumption in the combined use of Fe0and microorganisms
(set 4) would be higher than set 1 or 3 because of hydrogen supply in biodegradation and catalytic hydrogenolysis. From the results of hydrogen analysis, we found that hydrogen was almost exhausted within 7 days in set 4; nevertheless, extremely low levels of
hydro-gen (3.4–4.5
l
mol) in the headspace of vials was detected duringthe 7th–21st day of treatment. The production of hydrogen via fer-mentation could be neglected because there were no other organic substances in this reaction system, except for TCE and its
metabo-lites. Several previous studies have reported that Fe0corrosion in
water could support continuous release of cathodic hydrogen, which provide an energy source of microorganisms for
halorespira-tion (Weatherset al., 1997; Lee et al., 2001). In our experiments,
Fe0began to corrode when the hydrogen protection layer
disap-peared (set 4 had almost exhausted hydrogen within 7 days), therefore, it seems reasonable to suppose that the produced
hydro-gen after 7days in set 4 was released from Fe0corrosion. Supposing
the residual TCE in set 4 after 7 days was degraded by Fe0corrosion
and halorespiration (hydrogen is supplied by Fe0corrosion), then
accumulation of cis-DCE and a slower rate of degradation in the
ini-tial stage could be expected. FromFigs. 1 and 2, however, the
com-bined system showed rapid dechlorination with no cis-DCE accumulation after 7 days of incubation. This observation sug-gested that TCE was probably dechlorinated via another
mecha-nism, in addition to Fe0corrosion and halorespiration in the late
reaction period. The specific reaction mechanism able to achieve rapid, and complete dechlorination, will be discussed in the follow-ing section.
3.3. The synergistic effect of combined use of autotrophic
hydrogen-bacteria and Fe0
Results of batch studies demonstrated that TCE could be dechlo-rinated and transformed to ethane and ethene in very low hydro-gen concentrations during the late reaction period in this combined system (set 4). This may be attributed to different hydrogen concentrations and changes in the chemical properties of an iron surface; and any variations in hydrogen levels in the combined system should be the most noticeable contributing fac-tors. In natural ecosystems, hydrogen is an important intermediate in the anaerobic degradation of organic matter. In addition, compe-tition exists due to the limited supply of hydrogen produced by fer-mentation. The production of hydrogen in a microbial system is generally believed to be the rate-limiting step. In the initial stage of batch tests, hydrogen in headspace was sufficient for catalytic hydrogenolysis and TCE biodegradation, via halorespiration (set 4). However, after 7 days of incubation, heterogeneous catalytic dechlorination and halorespiration had utilized most of the
hydro-gen in the headspace, and then Fe0began to corrode with the
last-ing liberation of cathodic hydrogen. Therefore, the rate of hydrogen production became the rate-limiting step, and competition for the limited supply of hydrogen produced from iron corrosion existed in all types of hydrogen-consuming microorganisms in the com-bined system, such as iron reduction, sulfate reduction, chloroeth-ylene reduction, and methanogenic. Vials, which degraded TCE to cis-DCE and non-chlorinated hydrocarbons (ethane and ethene) after treatment of autotrophic hydrogen-bacteria (set 3), were ana-lyzed by ion chromatography. The results showed that 95% of the sulfate initially added was preserved in vials after 21-days treat-ment (data not shown in figures). As illustrated in set 3, the pri-mary biological reaction was halorespiration rather than sulfate reduction. However, there was a significant difference in the
sul-fate degradation rate between set 3 and set 4. From Fig. 5, we
can find that sulfate in set 4 on the 5th day began to decrease as soon as the hydrogen concentration was gradually made inade-quate. Moreover, a sudden drop in sulfate concentration was ob-served after 7 days of reaction when the hydrogen level was
below 5
l
mol (Fig. 5). The decrease of sulfate in set 4 showed thatsulfate reduction had replaced dechlorination, and played a pre-dominant role in the biological reaction during the later period of incubation. The distinction between sets 3 and 4, in the post peri-od, existed in the low-level hydrogen produced from iron corrosion in set 4, a concentration that makes sulfate reducers reactive.
After reduction of sulfate, a large amount of sulfide was pro-duced in the vials. The ferrous iron from iron corrosion precipitated
with S2, and then FeS settled on the surface of the iron particles
(Eqs. (6) and (7)). Precipitates of FeS, or other sulfides, may be
deposited on the surface of iron particles.
SO24 þ 4H2! S2þ 4H2O ð6Þ
Fe2þþ S2! FeS # ð7Þ
Hassan (2000)has proposed that TCE would be dechlorinated at a
faster rate by combined used of ferrous sulfide and Fe0, rather than
Fe0individually utilized. It also illustrated that, metallic iron acted
as a source of molecular hydrogen, through its reaction with water, while iron sulfide behaved as the reaction site (catalyst) in this sys-tem – i.e. molecular hydrogen is necessary for dechlorination to oc-cur. Ethene and ethane were the major products via the treatment
with FeS/Fe0 and results showed that, FeS that degraded TCE in
the presence of hydrogen, exhibited approximately 20-fold higher
reaction rate than Fe0did. Furthermore, it also confirmed the
rela-tionship between sulfur content and its dehalogenation capacity. In the present study, SEM-EDX was utilized to analyze the sulfur content of the iron surface. The EDX spectrum of iron particles incu-bated in set 4 for 21days indicated that the sulfur content reached to 28.4% (the initial sulfur content of the iron particles was less then
reaction time(days) 0 10 15 20 mass o f T C E (μ mol) 0 5 10 15 20 25 concentration of slfate (mg/l) 100 110 120 130 140 150 hydrogen ( μ mol ) 0 500 1000 1500 2000 TCE in set 4 sulfate in set 4 hydrogen in set 4 5
Fig. 5. The contents of TCE, hydrogen and sulfate concentration against reaction time in set 4 (Fe + cells + H2).
0.5%). Application of XRPD was adopted in this research to identify the crystal phase of iron particles. The iron particles were ground into powders in an agate mortar and then analyzed with XRPD.
The XRPD pattern showed a clear peak at 45.12o (2h), indicating
the presence of FeS on the iron surface (data not shown in figures). All observations indicated that dechlorination, via FeS and cathodic hydrogen released by iron corrosion occurred in the combined system. For this reason, the dechlorination rate did not slow when hydrogen was nearly exhausted in the headspace. A low level
of hydrogen produced by Fe0corrosion was sufficient for catalytic
hydrogenolysis to occur (FeS/H2).
3.4. The consumption of Fe0
The oxidative dissolution of iron particles occurred at neutral pH according to the characteristic reaction of iron corrosion (Eq.
(3)). In the absence of oxygen, iron corrosion occurs, with water
as an oxidant, under anaerobic conditions according to Eq.(3). As
mentioned above, it is easy to know that a large amount of ferrous
and hydroxyl ion are released with a conventional treatment of Fe0.
The degree of iron corrosion could be estimated by analyzing the
accumulation of iron species (Fe2+, Fe3+, colloid and precipitates)
in a medium solution, after incubation.Fig. 6shows profiles of iron
consumption in vials with reaction time under different treatment.
The incubation of set 2 (Fe0/N
2), which degraded TCE analogously
to conventional Fe0treatment, showed the most rapid corrosion
rate of iron, as shown inFig. 6. The iron consumption in set 2
ini-tially increased sharply, then slowed during the late stages, proba-bly owing to the precipitation of iron oxyhydroxides on the iron
surface (Reardon, 1995; Gu et al., 1999); however, treatment of
set 1 presented less oxidation of iron during reductive dechlorina-tion of TCE. This proved that hydrogen did prevent iron from oxidizing, and that TCE was degraded mostly via catalyzed hydro-lysis in set 1. In addition, it demonstrated again that the dechlori-nation mechanism between set 1 and set 2 were entirely different,
as based on distinct corrosion levels. The Fe0in set 2 oxidized
al-most 9-fold faster than in set 1. In contrast to set 1, similar trends of iron corrosion rates were observed in the combined system (set 4) during the first five reaction days, however, this was followed by a sudden increase in corrosion rate after seven days of incubation.
According to the hydrogen profiles inFig. 5, the corrosion rate of
iron might be accelerated by inadequate hydrogen supply in the vials. Furthermore, the increase of iron consumption in the later period of set 4 suggested that, the possibility of the low level of hydrogen was produced by iron corrosion. As compared with set 2, the iron consumption of set 1 and set 4 were 11.5% and 53.1% of set 2, respectively.
3.5. Dechlorination of TCE in the MFSB
In the previous section, we pointed out that the supply of hydrogen was inadequate in the treatment of biological (set 3) and combined system (set 4). The membrane feeding substrate bioreactor (MFSB) was utilized to estimate the efficiency of dechlo-rination with continuous hydrogen production under different treatments. The experiment conditions remained the same as those of batch incubation, in addition to the manner in which
hydrogen was supplied. The reactor containing Fe0with
continu-ous hydrogen production (set A) dechlorinated TCE completely within 30 days of treatment (data not shown in figures). The
dechlorination rate constant for the reaction was 0.15-day1
(r2= 0.97). Compared with set A, a similar reaction rate in set B
was displayed in the earlier stage of incubation, however, it in-creased more slowly during the rest period of reaction, which
may have been caused by passivation of Fe0due to corrosion and
precipitation of minerals. The rate constant for reaction of set B
was 0.09-day1 (r2= 0.97). For the duration of the treatment in
set B, cis-DCE appeared in the first ten days, and then disappeared through the reduction of iron (the cis-DCE data was not displayed in figures). Roughly, 98% of the added TCE was transformed into ethane and ethene till the termination of the reaction.
The performance of biodegradation was greatly improved in MFSB with continuous supply of hydrogen. After a lag phase of three days, the biodegradation of TCE became much faster in set C. Only 4.8% of the added TCE was preserved in the bioreactor within 30 days of incubation, and the rate constant for the reaction
of set C was 0.08-day1(r2
= 0.94). The dechlorination efficiency of cells arose in MFSB, nevertheless, the accumulation of cis-DCE was still detected during the treatment (data not shown). At the end of
the reaction, cis-DCE was accumulated to 28.2
l
mol l1in thebio-reactor. The sludge-adsorption was estimated by the same manner described above, which showed that the amount of TCE adsorption
was 9.3
l
mol l1 (5.2%). The combined use of Fe0 and hydrogenautotrophic bacteria (set D) displayed the best capacity of TCE deg-radation (also obtained in set 4). Complete dechlorination was achieved in set D within 16 days of incubation. The rate of
dechlo-rination in set D was up to 0.29 day1(r2
= 0.89). However, the re-sults of statistics revealed that the data of set D did not fully
conform to pseudo-first-order kinetics (r2= 0.89). The results of
MFSB indicated that, the dechlorination of set D could be divided into two stages. The first stage, in which the rate of dechlorination
was 0.14 day1(r2= 0.98), covered the period from the start of the
reaction to the 5th day. The second stage was from the 5th day to the 13th day of set D, and the rate constant for the reaction was
0.41 day1(r2= 0.98). These observations could be illustrated with
the results of the batch test. In the first stage, TCE was dechlorinat-ed mostly by catalyzdechlorinat-ed hydrogenolysis and biological degradation. Compared with stage 1, in addition to the mechanisms mentioned above, the catalytic dechlorination by FeS, which was produced from the sulfate reduction, could exist in stage 2. It is noteworthy
that, the concentration of sulfate in set D decreased by 34.1 mg l1
within 30 days of treatment (data not shown), and that the sulfur content on the iron surface, as detected by SEM-EDX, was 24.2%. In addition, FeS was found on the surface of iron particles from the qualitative analysis of XRPD (data not shown). These results con-firmed that, the catalytic dechlorination proceeded via FeS and hydrogen should exist in the combined system and do accelerate the reaction rate in the later period of set D.
4. Conclusion
Results showed that the combined use of Fe0and autotrophic
hydrogen-bacteria did have benefits of removing TCE. In summary,
reaction time (μmol)
0 10 15 20 weight loss of Fe (mg) 8 10 12 14 16 18 20 22 hydrogen ( μ mol) 0 500 1000 1500 2000 set 4 set 1 set 2 hydrogen in set 4 5
Fig. 6. The weight loss of Fe in set 1 (Fe + H2), set 3 (cells + H2) and set 4
(Fe + cells + H2) during dechlorination. The dotted line is the hydrogen contents of
the significant achievements obtained in this study were; (a) the
dechlorination efficiency of Fe0was improved in the presence of
hydrogen; (b) the consumption of iron was reduced in virtue of
the protection of hydrogen; (c) the combined use of Fe0and cells
en-hanced the dechlorination of TCE by 320%, as compared with the
conventional treatment of Fe0in MFSB, and no chlorinated
metabo-lite accumulated in this system; (d) the FeS produced from sulfate reduction had positive synergistic impacts on TCE degradation. Acknowledgements
This research was funded by the National Science Council of the Republic of China and the Contrast Number is NSC93-2211-E-002-031.
References
Agrawel, A., Ferguson, W.J., Gardner, B.O., Christ, J.A., Bandstra, J.Z., Tratnyek, P.G., 2002. Effect of carbonate species on the kinetics of dichloroethene by zero-valent iron. Environ. Sci. Technol. 36, 4326–4333.
Arnold, W.A., Roberts, A.L., 2000. Pathways and kinetics of chlorinated ethylene and chlorinated acetylene reaction with Fe0particles. Environ. Sci. Technol. 34,
1794–1805.
Burris, D.R., Campell, T.J., Manoranjan, V.S., 1995. Sorption of trichloroethylene and tetrachloroethylene in a batch reactive metallic iron–water system. Environ. Sci. Technol. 29, 2850–2855.
Cervini, J., Larson, R.A., Wu, J., Stucki, J.W., 2002. Dechlorination of pentachloroethane by commercial Fe and Ferruginous smectite. Chemosphere 47, 971–976.
Choe, S., Lee, S., Chang, Y., Hwang, K., Khim, J., 2001. Rapid reductive destruction of hazardous organic compounds by nanoscale Fe0
. Chemosphere 42, 367–372. Gu, B., Phelps, T.J., Liang, L., Dickey, M.J., Roh, Y., Kinsall, B.L., Palumbo, A.V., Jacobs,
G.K., 1999. Biological dynamics in zero-valent iron columns: implications for permeable reactive barriers. Environ. Sci. Technol. 33 (13), 2170–2177. Hassan, S.M., 2000. Reduction of halogenated hydrocarbons in aqueous media:
involving of sulfur in iron catalysis. Chemosphere 40, 1357–1363.
Heuer, J.K., Stubbins, J.F., 1999. An XPS characterization of FeCO3films from CO2
corrosion. Corros. Sci. 41, 1231–1243.
Holliger, C., Wohlfarth, G., Diekert, G., 1999. Reductive dechlorination in the energy metabolism of anaerobic bacteria. FEMS Microbiol. Rev. 22, 383–398. Lee, T., Tokunaga, T., Suyama, A., Furukawa, K., 2001. Efficient dechlorination of
tetrachloroethlene in soil slurry by combined use of an anaerobic Desulfitobacterium sp. Strain Y-51 and zero-valent iron. J. Biosci. Bioeng. 92 (5), 453–458.
Magnuson, J.K., Stern, R.V., Gossett, J.M., Zinder, S.H., Burris, D.R., 1998. Reductive dechlorination of tetrachloroethene to ethene by a two-component enzyme pathway. Appl. Environ. Microbiol. 64 (4), 1270–1275.
Marchand, E.A., Silverstein, J., 2002. Influence of heterotrophic microbial growth on biological oxidation of pyrite. Environ. Sci. Technol. 36, 5483–5490.
Matheson, L.J., Tratnyek, P.G., 1994. Reductive dehalogenation of chlorinated methanes by iron metal. Environ. Sci. Technol. 28, 2045–2053.
Maymó-gatell, X., Tandoi, V., Gossett, J.M., Zinder, S.H., 1995. Characterization of an H2-utilizing enrichment culture that reductively dechlorinates
tetrachloroethene to vinyl chloride and ethene in the absence of methanogensis and acetogenesis. Appl. Environ. Microbiol. 61, 3928–3933. Orth, W.S., Gillham, R.W., 1996. Dechlorination of trichloroethene in aqueous
solution using Fe0
. Environ. Sci. Technol. 30 (1), 66–71.
Phillips, D.H., Gu, B., Watson, D.B., Roh, Y., Liang, L., Lee, S.Y., 2000. Performance evaluation of a zerovalent iron reactive barrier: mineralogical characteristics. Environ. Sci. Technol. 34 (19), 4169–4176.
Reardon, E.J., 1995. Anaerobic corrosion of granular iron: measurement and interpretation of hydrogen evolution rates. Environ. Sci. Technol. 29 (12), 2936–2945.
Ritter, K., Odziemkowski, M.S., Gillham, R.W., 2002. An in situ study of the role of surface films on granular iron in the permeable iron wall technology. J. Contam. Hydrol. 55, 87–111.
Roberts, A.L., Totten, L.A., Arnold, W.A., Burris, D.R., Campell, T.J., 1996. Reductive elimination of chlorinated ethylenes by zero-valent metals. Environ. Sci. Technol. 30 (8), 2654–2659.
Weathers, L.J., Parkin, G.F., Alvarez, P.J., 1997. Utilization of cathodic hydrogen as electron donor for chloroform cometabolism by a mixed, methanogenic culture. Environ. Sci. Technol. 31 (3), 880–885.
Wüst, W.F., Köber, R., Schlicker, O., Dahmke, A., 1999. Combined zero and first-order kinetics model of the degradation of TCE and cis-DCE with commercial iron. Environ. Sci. Technol. 33 (23), 4304–4309.
Yager, R.M., Bilotta, S.E., Mann, C.L., Madsen, E.L., 1997. Metabolic adaptation and in situ attenuation of chlorinated ethenes by naturally occurring microorganisms in a fractured dolomite aquifer near Niagara Falls, New York. Environ. Sci. Technol. 31, 3138–3147.
