Bioresource Technology 179 (2015) 284–290

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Fluidized-bed denitrification of mining water tolerates high nickel concentrations G. Zou a,⇑, S. Papirio a, E.D. van Hullebusch b, J.A. Puhakka a a b

Department of Chemistry and Bioengineering, Tampere University of Technology, P.O. Box 541, FIN-33101 Tampere, Finland Université Paris-Est, Laboratoire Géomatériaux et Environnement (EA 4508), UPEM, 77454 Marne-la-Vallée, France

h i g h l i g h t s  Ni impact on denitrification was revealed in two continuous FBRs.  Denitrification tolerated soluble Ni concentrations as high as 500 mg/L.  Ni speciation was investigated by using both modeling and XRD analysis.  Ni3(PO4)2 precipitates were detected by XRD on activated carbon at FBR termination.  The high tolerance of Dechloromonas to Ni successfully maintained denitrification.

a r t i c l e

i n f o

Article history: Received 3 November 2014 Received in revised form 11 December 2014 Accepted 12 December 2014 Available online 19 December 2014 Keywords: Denitrification Fluidized-bed reactor Nickel Denitrifying communities X-ray diffraction

a b s t r a c t This study revealed that fluidized-bed denitrifying cultures tolerated soluble Ni concentrations up to 500 mg/L at 7–8 and 22 °C. From 10 to 40 mg/L of feed Ni, denitrification resulted in complete nitrate and nitrite removal. The concomitant reduction of 30 mg/L of sulfate produced 10 mg/L of sulfide that precipitated nickel, resulting in soluble effluent Ni below 22 mg/L. At this stage, Dechloromonas species were the dominant denitrifying bacteria. From 60 to 500 mg/L of feed Ni, nickel remained in solution due to the inhibition of sulfate reduction. At soluble 60 mg/L of Ni, denitrification was partially inhibited prior to recover after 34 days of enrichment by other Ni-tolerant species (including Delftia, Zoogloea and Azospira) that supported Dechloromonas. Subsequently, the FBR cultures completely removed nitrate even at 500 mg/L of Ni. Visual Minteq speciation model predicted the formation of NiS, NiCO3 and Ni3(PO4)2, whilst only Ni3(PO4)2 was detected by XRD. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Nitrogenous compounds and heavy metals are commonly present in mining and industrial wastewaters (Papirio et al., 2014b). Contamination of water bodies by nitrogen and heavy metals poses environmental challenges due to eutrophication and toxicity to aquatic species. Ammonium and nitrate can be removed from metal-containing mining waters by using nitrification and denitrification (Papirio et al., 2014a,b; Zou et al., 2014). Denitrification is

Abbreviations: DGGE, denaturant gradient gel electrophoresis; DOM, dissolved organic matter; ESEM, environmental scanning electron microscopy; EPS, extracellular polymeric substances; FBR, fluidized-bed reactor; FID, flame ionization detector; GC, gas chromatograph; HRT, hydraulic retention time; ICP–OES, inductively coupled plasma–optical emission spectrometry; JCPDS, joint committee on powder diffraction standards; SMP, soluble microbial products; XRD, X-ray diffraction. ⇑ Corresponding author. Tel.: +358 40 1981267; fax: +358 33 641392. E-mail address: gang.zou@tut.fi (G. Zou). http://dx.doi.org/10.1016/j.biortech.2014.12.044 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.

primarily performed by heterotrophic bacteria that use organic carbon as source of energy and electrons (Park and Yoo, 2009). In the presence of heavy metals, the activity of microorganisms could be repressed according to the environmental conditions and chemical behavior of the metal species (Bartacek et al., 2008). Nickel is a common metal in mining environments that can occur in soluble, sulfidic (i.e. NiS, NiS2), elemental (Ni) and oxidic (i.e. NiO, Ni(OH)2, NiCO3) forms (Schaumlöffel, 2012). Nickel has been reported to exhibit both stimulatory and inhibitory effects on microbial processes at low and elevated concentrations, respectively (Gikas, 2008; Zandvoort et al., 2006). Ni speciation determines its bioavailability and toxicity to microorganisms (YebraBiurrun and Castro-Romero, 2011). However, Ni toxicity can be mitigated by the formation of Ni complexes with dissolved organic matter (DOM) or soluble microbial products (SMP) (Doig and Liber, 2007; Kuo and Parkin, 1996) and bio-sorption of Ni on microorganisms (Fomina and Gadd, 2014).

G. Zou et al. / Bioresource Technology 179 (2015) 284–290

285

Denitrifying microorganisms have been found in a variety of taxonomic groups (Zumft, 1997). Bioprocess operating conditions influence the composition of denitrifying microbial communities and denitrification potential (Papirio et al., 2014a; Rasool et al., 2014). Modern molecular techniques have allowed revealing the effects of environmental conditions (Cao et al., 2008) and heavy metals (Magalhães et al., 2011) on denitrifying microorganisms in soil and sediments. However, the understanding of the evolution of denitrifying communities in biofilm reactors remains limited. The present work investigated the effects of increasing Ni concentrations on heterotrophic denitrification and microbial population in two fluidized-bed reactors (FBRs), operated at 7–8 and 22 °C under neutral conditions. Denaturing gradient gel electrophoresis (DGGE) was used to reveal the evolution of the FBR denitrifying microbial communities under gradual feed Ni increases. Ni speciation and solid phase characterization were studied in both FBRs.

bands in the DGGE gel were excised and the DNA was sequenced by MacroGen (Seoul, Korea). Sequencing data were compared to the database of the National Center for Biotechnology Information. At FBR termination, the precipitates on the biofilm-coated activated carbon were characterized by ESEM, XRD and ICP–OES analyses. Environmental scanning electron microscopy analysis (ESEM) was performed on an Electroscan (Wilmington, USA) Type II LaB6 gun microscope. X-ray diffraction (XRD) analysis was performed on a Bruker D8 Advance diffractometer equipped with an energy dispersion Sol-X detector with copper radiation (CuKa, k = 0.15406 nm). The acquisition was recorded between 10° and 80°, with a 0.02° scan step and 1 s step time. The precipitates were dissolved in 1 M HNO3 and the solubilized elements were analyzed by inductively coupled plasma–optical emission spectrometer (ICP–OES) (Perkin-Elmer Optima 8300). Phosphorus concentration was determined according to the standard colorimetric method (Rodier et al., 2009). The analysis of the elemental composition of the precipitates was performed in triplicate.

2. Methods

2.3. Thermodynamic modeling by Visual MINTEQ

2.1. FBR operation

The chemical equilibrium code Visual MINTEQ v3.1 (Visual MINTEQ is a freeware chemical equilibrium model that can be downloaded from http://vminteq.lwr.kth.se/- accessed on December 4th, 2014) was used to identify the possible phases controlling Ni solubility in the bioreactor. The concentrations of the elements used in the feed were input into the model. Besides, the temperature was kept constant at 8 and 22 °C, the pH was set to 7.0, the oversaturated solids were allowed to precipitate, and the alkalinity measured in the effluent was specified.

Two FBRs (FBR1 and FBR2) were used to study denitrification in presence of soluble Ni under neutral conditions. The FBR scheme and characteristics were as reported by Papirio et al. (2014a). Prior to this, steady denitrification with ethanol as organic electron donor was initially maintained for 415 days in FBR1 and FBR2, operated at 7–8 and 22 °C, respectively, with the aim of investigating the effects of feed pH, HRT, ethanol/nitrate ratio and temperature on the process (Papirio et al., 2014a). Subsequently, Ni was injected twice to FBR1 and FBR2 between days 416 and 567, in order to study Ni impacts at feed pH of 2.5 (Zou et al., 2014). Finally, from day 568 to 632, feed pH was increased from 2.5 to 5.5 and Ni was not supplemented to FBRs, in order to recover the denitrifying process in both reactors. In this work, the feed Ni concentration was gradually increased from 10 to 500 mg/L in both FBRs between days 633 and 887. Nitrate and ethanol concentrations were maintained at 200 and 123 mg/L, respectively. Feed pH was kept at 5.5 and HRT was 5.4 h. The composition of the mineral medium was as reported by Papirio et al. (2014a). From day 864 to 887, feed nitrate was increased from 200 to 300 mg/L in order to investigate FBR denitrifying potential at feed Ni 500 mg/L. Ethanol concentration was accordingly increased to 185 mg/L to maintain the ethanol/nitrate ratio constant (0.84 mol/mol). Both FBR1 and FBR2 were terminated on day 887. Effluent samples were taken every three days for pH, nitrate, nitrite, ethanol and Ni measurements. 2.2. Sample analysis Samples of FBR effluents were filtered through 0.2 lm ChromafilXtra PET-20125 membranes (Macherey–Nagel, Germany). pH, nitrate, nitrite and Ni concentrations were measured in both FBRs as reported by Zou et al. (2014). Ethanol was analyzed with a gas chromatograph (GC-2010 Plus, Shimadzu, Kyoto, Japan) equipped with a ZB-WAX plus column (Phenomenex, USA) and a flame ionization detector (FID). The alkalinity in the effluent was measured as reported by Papirio et al. (2014b). The microbial communities were studied both in absence and in presence of nickel at different feed Ni concentrations. Samples of biofilm/activated carbon were taken four times during FBRs operation (on days 607, 704, 788 and 887). PCR-DGGE analysis was performed as reported by Papirio et al. (2014b). The unambiguous

Fig. 1. Feed Ni and nitrate and nitrite profiles in FBR1 (A) and FBR2 (B). Explanation of symbols: influent nitrate (closed circles); effluent nitrate (open circles); effluent nitrite (open triangles); feed Ni (solid line).

286

G. Zou et al. / Bioresource Technology 179 (2015) 284–290

Table 1 DGGE bands associated with 16SrRNA in the biomass samples from FBR1. Sample

Bands

Most similar genus or species

Accession number

Similarity (%)

Phylogenetic group (class/order/family)

On day 607

1 2, 3 4, 5 6

Uncultured bacterium Comamonadaceae bacterium Dechloromonas sp. Desulfovibrio magneticus

FM992023 JF694750.1 KF800710.1 NR_074958

98.9 99.2–100.0 97.5–98.3 100.0

a-Proteobacteria

On day 704

15 10, 12, 13–14 11

Zoogloea resiniphila Dechloromonas sp. Delftia acidovorans

AJ505854.1 NR_042819.1 KJ806439.1

100.0 98.9–99.8 99.5

b/Rhodocyclales/Rhodocyclaceae b/Rhodocyclales/Rhodocyclaceae b/Burkholderiales/Comamonadaceae

On day 788

27 22–23, 25–26 24

Zoogloea resiniphila Dechloromonas sp. Azospira sp. Tagus Dechlorosoma suillum

AJ505854.1 KF826886.1 KC247691.1 NR_074103.1

99.8 98.5–98.9 100.0 100.0

b/Rhodocyclales/Rhodocyclaceae b/Rhodocyclales/Rhodocyclaceae b/Rhodocyclales/Rhodocyclaceae

On day 887

32 33–38 39

Sediminibacterium salmoneum Dechloromonas sp. Delftia sp. Cs1–4

NR_044197.1 KF826886.1 NR_074626.1

99.4 98.9–100.0 99.8

Sphingobacteriia/Sphingobacteriales/Chitinophagaceae b/Rhodocyclales/Rhodocyclaceae b/Burkholderiales/Comamonadaceae

3. Results and discussion 3.1. Effects of Ni on denitrification in FBRs Fig. 1 shows nitrate and nitrite profiles in FBRs at 7–8 and 22 °C. At feed Ni concentrations from 10 to 40 mg/L, denitrification resulted in nitrate and nitrite below the detection limit in both FBRs. Nitrate removal rate was 890 mg/L d. When feed Ni was increased to 60 mg/L, effluent nitrate and nitrite increased at both 7–8 and 22 °C. Effluent nitrate reached 107 mg/L on day 695 and 101 mg/L on day 707 in FBR1 and FBR2, respectively. Nitrite increased up to 42 and 16 mg/L at 7–8 and 22 °C, respectively. After 34 days of operation at 60 mg/L of feed Ni, denitrification recovered resulting in complete nitrate and nitrite removals on day 725. This was likely due to the enrichment of Ni-tolerant microbial species, as demonstrated by the higher number of denitrifying strains revealed on day 704 (Tables 1 and 2). Moreover, the high tolerance of denitrifiers to Ni may be due to the long-term acclimation. Holtan-Hartwig et al. (2002) observed the enhancement of metal tolerance of denitrifying species after 2 months of exposure to Cu, Cd and Zn in soil. A similar behavior was reported in activated sludge and anammox cultures which recovered after adaptation to 60 mg/L Ni for 40 days and 20 mg/L Zn for 20 days, respectively (Wang et al., 2010; Daverey et al., 2014). From day 725 on, the gradual increases of feed Ni from 80 to 500 mg/L did not affect denitrification. Nitrate and nitrite remained below the detection limit in both FBRs, even when feed

b/Burkholderiales/Comamonadaceae b/Rhodocyclales/Rhodocyclaceae D/Desulfovibrionales/Desulfovibrionaceae

nitrate increased from 200 to 300 mg/L at the highest feed Ni concentration (Fig. 1). Under these conditions, nitrate removal rate reached 1330 mg/L d. Ethanol remained below detection limit throughout the experimentation (data not shown). No effect of the operating temperature was observed. Effluent pH remained stable at approximately 7.2 throughout the experimentation in both FBRs (Fig. 2). The profiles of dissolved Ni in FBR1 and FBR2 were as reported in Fig. 3. When Ni was added up to 20 mg/L, the soluble Ni remained below the detection limit in both FBRs, prior to gradually increase to 22 mg/L when Ni supplementation was 40 mg/L. During this period (days 633–685), nickel sulfide precipitation played an important role in the occurrence of soluble Ni in both FBRs, as also reported by Zou et al. (2014). Approximately 10 mg/L of biogenic sulfide was produced by the reduction of 30 mg/L of sulfate present in the feed. This was also confirmed by the presence of the sulfate-reducing bacterium Desulfovibrio magneticus in FBR cultures (Fig. 4, Tables 1 and 2). At higher feed Ni (60–500 mg/L), sulfate reduction was no longer observed and Ni was mostly removed with soluble effluent in both FBRs (Fig. 3). 3.2. Evolution of denitrifying microbial communities Fig. 4 shows the DGGE profiles of microbial cultures in absence (day 607) and in presence of Ni (days 704, 788 and 887) in FBRs. The presence of many bands indicates the diversity of microbial communities at both 7–8 and 22 °C.

Table 2 DGGE bands associated with 16SrRNA in the biomass samples from FBR2. Sample

Bands

Most similar genus or species

Accession number

Similarity (%)

Phylogenetic group (class/order/family)

On day 607

7 8 9

Uncultured bacterium Dechloromonas sp. Desulfovibrio magneticus

FM992023 AF170355 NR_074958

98.9 99.1 100.0

a-Proteobacteria b/Rhodocyclales/Rhodocyclaceae D/Desulfovibrionales/Desulfovibrionaceae

On day 704

20–21 18–19 17 16

Dechloromonas sp. Azospira sp. Flavobacterium sp. Uncultured bacterium

AB769215.1 KC247691.1 KC22779.1 KF875616.1

99.0 98.1–100.0 99.6 100.0

b/Rhodocyclales/Rhodocyclaceae b/Rhodocyclales/Rhodocyclaceae Flavobacteriia/Flavobacteriales/Flavobacteriaceae Found in a denitrification system

On day 788

31 30

Zoogloea sp. Chao13 Azospira sp. Dechlorosoma suillum Dechloromonas sp Flavobacterium sp.

KC473458.1 KC247691.1 NR_074103.1 KF826886.1 AJ508709.1

99.8 99.6 99.6 100.0 98.6

b/Rhodocyclales/Rhodocyclaceae b/Rhodocyclales/Rhodocyclaceae

Sediminibacterium salmoneum Dechloromonas sp Azospira sp. Tagus Dechlorosoma suillum

NR_044197.1 KF826886.1 KC247691.1 NR_074103.1

99.4 99.8–100.0 100.0 100.0

Sphingobacteriia/Sphingobacteriales/Chitinophagaceae b/Rhodocyclales/Rhodocyclaceae b/Rhodocyclales/Rhodocyclaceae

29 28 On day 887

40 41–44, 46, 47 45

b/Rhodocyclales/Rhodocyclaceae Flavobacteriia/Flavobacteriales/Flavobacteriaceae

G. Zou et al. / Bioresource Technology 179 (2015) 284–290

Fig. 2. Effluent pH in FBR1 (A) and FBR2 (B). Influent pH (closed circles); effluent pH (open circles); feed Ni (solid line).

287

operated at 7–8 °C implying their potential to grow even at low temperature. DGGE revealed the presence of a sulfate-reducing bacterium, Desulfovibrio magneticus (bands 6 and 9), in both FBRs, confirming the formation of biogenic sulfide capable of precipitating heavy metals (Villa-Gomez et al., 2011, 2012). By day 704, at 60 mg/L feed Ni, microbial population had changed. The increase of feed and soluble Ni concentrations resulted in the enrichment of a higher number of denitrifying species. Dechloromonas species were still the main denitrifiers at both 7–8 (bands 10, 12, 13 and 14) and 22 °C (bands 20 and 21). In addition, Delftia (band 11), belonging to the family Comamonadaceae, and Zoogloea (band 15) genera were also observed at 7–8 °C, whereas Azospira (bands 18 and 19) and Flavobacterium (band 17) genera were revealed at 22 °C. Bacteria belonging to these genera are all common denitrifiers involved in different steps of denitrification (Horn et al., 2005; Xie and Yokota, 2006). Desulfovibrio magneticus remained below DGGE detection in both FBRs, indicating that sulfate reduction was inhibited at higher soluble Ni concentrations. On day 788, at feed Ni of 100 mg/L, microbial community did not significantly change. Dechloromonas species were detected in both FBRs (Fig. 4, Tables 1 and 2). At 7–8 °C, the genera Zoogloea (band 27) and Azospira (Fig. 4 and Table 1) were still present, whereas bacteria belonging to Comamondaceae were below DGGE detection. At 22 °C, bacteria belonging to Azospira (band 30) and Flavobacterium (band 28) were revealed (Fig. 4 and Table 2). On day 887, at FBR termination, Dechloromonas species again dominated the microbial communities at both 7–8 (bands 33–38) and 22 °C (bands 41–44, 46 and 47) (Fig. 4, Tables 1 and 2). The number of denitrifying species significantly decreased at 500 mg/ L of feed Ni in both FBRs. Besides Dechloromonas, only the genus Azospira was still detected at 22 °C (band 45). In summary, Dechloromonas species were the dominant denitrifying bacteria throughout FBRs operation. The supplementation of Ni at 60–100 mg/L selectively enriched other Ni-tolerant species that maintained denitrification. Subsequently, the increase of feed Ni up to 500 mg/L resulted in a lower community diversity with Dechloromonas as main species. This is in agreement with the results obtained by Sobolev and Begonia (2008) who observed

Fig. 3. Effluent Ni concentration in FBRs at feed Ni from 10 to 500 mg/L. Internal graph (feed Ni from 10 to 100 mg/L); soluble Ni in FBR1 (open circles) and FBR2 (open triangles).

Tables 1 and 2 report the microbial species based on the sequenced bands in FBR1 (7–8 °C) and FBR2 (22 °C), respectively. Before Ni supplementation, Dechloromonas species (bands 4 and 5) and one bacterium belonging to Comamonadaceae (bands 2 and 3) were responsible for denitrification at 7–8 °C. Only one Dechloromonas species (band 8) was detected at 22 °C. Microbes belonging to the genus Dechloromonas oxidize nitrate and nitrite in anoxic conditions using organic electron donors (Coates et al., 2001; Horn et al., 2005). Comamonadaceae have also been reported to oxidize organic compounds using nitrate as electron acceptor at 28 °C (Khan et al., 2002). Comamonadaceae were present in FBR1

Fig. 4. DGGE profiles of microbial cultures in FBR1 and FBR2 on days 607, 704, 788 and 887.

288

G. Zou et al. / Bioresource Technology 179 (2015) 284–290

the reduction of the number of denitrifying species in soil under increasing Pb concentrations. 3.3. Ni speciation and solid phase characterization 3.3.1. Visual Minteq modelling Tables 3 and 4 show the fate of Ni in FBR1 and FBR2, respectively, predicted by Visual Minteq speciation model. Initially, the fate of Ni was simulated in both FBRs at feed Ni of 10 mg/L (0.17 mM) and in presence of sulfate reducing activity as shown by the microbial diversity study (Tables 1 and 2). The modelling results confirmed that almost 100% Ni precipitated as nickel sulfide at both temperatures, explaining why nickel was not detected in the soluble effluent as shown in Fig. 3. Nickel sulfide precipitation occurred when Ni was supplemented from 10 to 40 mg/L. Under these conditions, approximately 30 mg/L of feed sulfate (data not shown) was biologically reduced to 10 mg/L of sulfide according to reaction 1:   þ 2CH3 CH2 OH þ 3SO2 4 ! 3HS þ 4HCO3 þ H þ H2 O

ð1Þ

Theoretically, 10 mg/L of sulfide reacts with 18 mg/L of Ni (reaction 2), confirming the highest soluble Ni concentration of approximately 22 mg/L observed in both FBRs when Ni was fed at 40 mg/L (Fig. 3). 2þ

Ni

þ H2 S ! NiS ðsÞ þ 2Hþ

ð2Þ

Nickel sulfide formation was estimated by Visual Minteq modelling, whilst NiS precipitates were not identified by XRD and ICP–OES analyses. Previous studies have demonstrated the

amorphous and poorly crystalline of metal sulfides (i.e. NiS) precipitated with biogenic sulfide (Sampaio et al., 2010; Villa-Gomez et al., 2012). Moreover, the nickel sulfide precipitates may have been removed with the effluent as reported by Kaksonen et al. (2003) and Zou et al. (2014). Simulations were also performed at initial Ni concentration of 500 mg/L (8.52 mM). In this case, biological sulfate reduction was supposed to be inhibited as observed in the microbial community results (Tables 1 and 2). Visual Minteq predicted that 87.7% of Ni precipitated mainly as Ni3(PO4)2(s) and Ni(OH)2 in colloidal state in FBR1 (Table 3) and 94.4% of Ni precipitated mainly as Ni3(PO4)2(s) and NiCO3 in FBR2 (Table 4). 59.6 and 26.9 mg/L were the predicted free Ni2+ concentrations in FBR1 and FBR2, respectively. These differences were especially due to the two operating temperatures used. Phosphate was added to the FBRs in the mineral med2 ium. The presence of HCO was due to the production of 3 /CO3 alkalinity during denitrification. Alkalinity was averagely 182.5 and 132.5 mg/L CaCO3 in FBR1 and FBR2, respectively (data not shown). The Visual Minteq simulation results are in contrast with what observed in both FBRs, where Ni was mostly remained in the soluble effluent (Fig. 3). Visual Minteq predicts metal speciation under thermodynamic equilibrium conditions. In this study, the precipitate formation kinetics was likely much longer than the HRT (5.4 h) used, indicating that the predicted amount of Ni precipitates was overestimated. Similarly, Guillard and Lewis (2001) reported that the equilibrium model was unable to predict the Ni removal efficiency since it did not take into account the Ni fines formation that occurred in the outlet stream. Moreover, Visual Minteq modelling does not consider the possible formation of Ni

Table 3 Visual Minteq modelling of Ni speciation and precipitation in FBR1 at initial 10 and 500 mg/L of Ni. FBR1 at 8 °C

Ni dissolved

Ni precipitated

– pH = 7.0 – Feed Ni = 10 mg/L (0.17 mM) – Sulfate reduction resulting in 10 mg/L of S2 (0.31 mM) – Alkalinity = 182.5 mg/L CaCO3

– Almost equal to 0% of total Ni – NiHS+ as dominant dissolved species – Free Ni2+ ion concentration of 5.4  1013 mM (7.8% of the dissolved Ni)

– Almost 100 % of total Ni – Main Ni precipitate : NiS (gamma) (0.17 mM) – Other precipitates: s CoS (beta) (2.1  104 mM) s MnHPO4(s) (8.8  103 mM) s Ca5(PO4)3(OH) (7.6  102 mM)

– – – –

– 12.3% of total Ni (61.5 mg/L of Ni) – Free Ni2+ ion concentration of 59.6 mg/L (1.0 mM) (96.9% of the dissolved Ni)

– 87.7 % of total Ni – Main Ni precipitates: s Ni3(PO4)2(s) (1.2  102 mM) s NiMoO4(s) (4.0  103 mM) s Colloidal Ni(OH)2 (7.1 mM) – Other precipitates: s MnHPO4(s) (8.8  103 mM) s Ca5(PO4)3(OH) (3.0  102 mM)

pH = 7.00 Feed Ni = 500 mg/L (8.52 mM) No sulfate reduction Alkalinity = 182.5 mg/L CaCO3

Table 4 Visual Minteq modelling of Ni speciation and precipitation in FBR2 at initial 10 and 500 mg/L of Ni. FBR 2 at 22 °C

Ni dissolved

Ni precipitated

– pH = 7.00 – Feed Ni = 10 mg/L (0.17 mM) – Sulfate reduction resulting in 10 mg/L of S2 (0.31 mM) – Alkalinity = 132.5 mg/L CaCO3

– Almost equal to 0% of total Ni – Free Ni2+ ion concentration of 2.1  1013 mM (5.0% of the dissolved Ni)

– Almost 100 % of total Ni – Main Ni precipitate : NiS (gamma) (0.17 mM) – Other precipitates: s CoS (beta) (2.1  104 mM) s MnHPO4(s) (8.8  103 mM) s Ca5(PO4)3(OH) (8.2  102 mM)

– – – –

– 5.6% of total Ni (28.0 mg/L) – Free Ni2+ ion concentration of 26.9 mg/L(0.46 mM) (96.0% of the dissolved Ni)

– 94.4 % of total Ni – Main Ni precipitates: s Ni3(PO4)2(s) (6.5  102 mM) s NiCO3(s) (7.8 mM) s NiMoO4(s) (4.0  103 mM) – Other precipitates: s MnHPO4(s) (8.8  103 mM) s Ca5(PO4)3(OH) (6.0  102 mM)

pH = 7.00 Feed Ni = 500 mg/L (8.52 mM) No sulfate reduction Alkalinity = 132.5 mg/L CaCO3

G. Zou et al. / Bioresource Technology 179 (2015) 284–290 Table 5 Mineral precipitates elemental composition. The concentrations are expressed in mg per gram of granulated activated carbon harboring precipitates. Elements

FBR1 (mg/g)

FBR2 (mg/g)

Fe Ca Mg Mn Ni S P

1.37 (±1.10) 1.40 (±0.09) 0.82 (±0.02) 0.43 (±0.01) 55.09 (±1.00) 0.73 (±0.00) 11.66 (±0.62)

0.45 (±0.24) 1.43 (±0.31) 1.04 (±0.13) 0.42 (±0.02) 58.15 (±1.15) 0.43 (±0.07) 15.50 (±0.08)

complexes with extracellular polymeric substances (EPS), DOM and SMP that maintain metals in solution (Doig and Liber, 2007; Kuo and Parkin, 1996; Guibaud et al., 2005). 3.3.2. Ni precipitate characterization (XRD + ICP–OES) and inhibition on denitrification Ni precipitates onto the surface of granular activated carbon were similar for both FBR1 and FBR2 (ESEM analysis not shown). The XRD analysis showed that the main crystalline structure present in the precipitates was Ni3(PO4)28H2O (JCPDS 33-951) (Wu et al., 2010) (Fig. S1), in agreement with the elemental composition (Table 5). However, the actual molar Ni/P ratio of the precipitates in FBR1 and FBR2 was 2.5 and 1.98, respectively, and higher than the theoretical value for Ni3(PO4)28H2O (Ni/P equal to 1.5). This indicates that the precipitates contained more nickel than phosphorus. Nickel was therefore retained by other mechanisms such as sorption onto biofilm (Fomina and Gadd, 2014), precipitation as Ni(OH)2 or NiCO3, as proposed by the Visual Minteq modelling, or surface complexation by Fe and Mn oxides (Hitchcock et al., 2009). To the best of our knowledge, this is the first study reporting the tolerance of denitrifying cultures to such high soluble Ni concentrations in biofilm reactors. In contrast, Lawrence et al. (2004) reported that only 0.5 mg/L Ni exhibited negative effects on denitrifying activity by repressing the nirS (nitrite reductase) gene in a biofilm community cultivated from river water. Li et al. (2011) observed that 240 mg/L of Ni totally repressed the metabolic activity of activated sludge in a sequencing batch reactor (SBR). Heavy metal toxicity is generally due to the presence of significant concentrations of free metal ions in solution (Bartacek et al., 2008). In this work, at feed 500 mg/L of Ni, nickel mostly remained in solution but the predicted free Ni2+ was only 59.6 and 26.9 mg/L in FBR1 and FBR2, respectively. The actual free Ni2+ concentration in both FBRs was presumably higher as the HRT used was shorter than the time needed to achieve the thermodynamic equilibrium and the predicted Ni precipitation. On the other hand, the possible complexation of Ni with SMP decreased Ni2+ concentration and reduced Ni bioavailability and toxicity (Kuo and Parkin, 1996) resulting in a higher tolerance of the denitrifying cultures. The use of continuous FBRs permitted a better adaptation of the enrichment cultures to increasing concentrations of nickel and the likely formation of EPS in the biofilm that further enhanced the resistance of microbial cells (Guibaud et al., 2005). Nickel precipitation, adsorption and complexation with Fe and Mn slightly occurred in both FBRs and can only explain the resistance of denitrifiers to Ni to a small extent. 4. Conclusions Continuous-flow denitrification was maintained in FBRs even at soluble 500 mg/L Ni and a nitrate loading rate of 1330 mg/L d. At 60 mg/L of feed Ni, denitrifying cultures required 34 days of acclimation before completely removing nitrate. PCR-DGGE revealed the dominance of Dechloromonas throughout FBR operation.

289

Desulfovibrio magneticus was responsible for concomitant sulfate reduction and NiS precipitation at 10–40 mg/L feed Ni, prior it became inhibited at higher Ni concentrations. Visual Minteq model confirmed NiS precipitation at the lowest feed Ni concentration. At the highest Ni feed concentration, Ni solubility was controlled only by Ni3(PO4)2 as shown by XRD analysis. Acknowledgements The current research was conducted as part of the Green Mining program (project number 230229) funded by the Finnish Funding Agency for Technology and Innovation (TEKES). The authors would like to thank Dr. David Huguenot for ICP–OES technical support, Dr. Rossano Combes for ESEM analysis, Prof. Stéphanie Rossano for XRD technical support and Jean-Michel Riom for phosphorus colorimetric analysis. This work was supported by a grant from Région Ile de France. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2014.12. 044. References Bartacek, J., Fermoso, Gonzalez F., Baldó-Urrutia, A.M., van Hullebusch, E.D., Lens, P.N.L., 2008. Cobalt toxicity to methylotrophic methanogenesis in anaerobic granular sludge: role of speciation. J. Ind. Microbiol. Biotechnol. 35, 1465–1474. Cao, Y., Green, P.G., Holden, P.A., 2008. Microbial community composition and denitrifying enzyme activities in salt marsh sediments. Appl. Environ. Microbiol. 74, 7585–7595. Coates, J.D., Chakraborty, R., Lack, J.G., O’Connor, S.M., Cole, K.A., Bender, K.S., Achenbach, L.A., 2001. Anaerobic benzene oxidation coupled to nitrate reduction in pure culture by two strains of Dechloromonas. Nature 411, 1039– 1043. Daverey, A., Chen, Y., Sung, S., Lin, J., 2014. Effect of zinc on anammox activity and performance of simultaneous partial nitrification, anammox and denitrification (SNAD) process. Bioresour. Technol. 165, 105–110. Doig, L.E., Liber, K., 2007. Nickel speciation in the presence of different sources and fractions of dissolved organic matter. Ecotoxicol. Environ. Saf. 66, 169–177. Fomina, M., Gadd, G.M., 2014. Biosorption: current perspectives on concept, definition and application. Bioresour. Technol. 160, 3–14. Gikas, P., 2008. Single and combined effects of nickel (Ni (II)) and cobalt (Co (II)) ions on activated sludge and on other aerobic microorganisms: a review. J. Hazard. Mater. 159, 187–203. Guibaud, G., Comte, S., Bordas, F., Dupuy, S., Baudu, M., 2005. Comparison of the complexation potential of extracellular polymeric substances (EPS), extracted from activated sludges and produced by pure bacteria strains, for cadmium, lead and nickel. Chemosphere 59, 629–638. Guillard, D., Lewis, A.E., 2001. Nickel carbonate precipitation in a fluidized-bed reactor. Ind. Eng. Chem. Res. 40, 564–5569. Hitchcock, A.P., Dynes, J.J., Lawrence, J.R., Obst, M., Swerhone, G.D.W., Korber, D.R., Leppard, G.G., 2009. Soft X-ray spectromicroscopy of nickel sorption in a natural river biofilm. Geobiology 7, 432–453. Holtan-Hartwig, L., Bechmann, M., Risnes Høyås, T., Linjordet, R., Reier Bakken, L., 2002. Heavy metals tolerance of soil denitrifying communities: N2O dynamics. Soil Biol. Biochem. 34, 1181–1190. Horn, M.A., Ihssen, J., Matthies, C., Schramm, A., Acker, G., Drake, H.L., 2005. Dechloromonas denitrificans sp. nov., Flavobacterium denitrificans sp. nov., Paenibacillus anaericanus sp. nov. and Paenibacillus terrae strain MH72, N2Oproducing bacteria isolated from the gut of the earthworm Aporrectodea caliginosa. Int. J. Syst. Evol. Microbiol. 55, 1255–1265. Kaksonen, A., Riekkola-Vanhanen, M., Puhakka, J., 2003. Optimization of metal sulphide precipitation in fluidized-bed treatment of acidic wastewater. Water Res. 37, 255–266. Khan, S.T., Horiba, Y., Yamamoto, M., Hiraishi, A., 2002. Members of the family Comamonadaceae as primary poly(3-hydroxybutyrate-co-3-hydroxyvalerate)degrading denitrifiers in activated sludge as revealed by a polyphasic approach. Appl. Environ. Microbiol. 68, 3206–3214. Kuo, W., Parkin, G.F., 1996. Characterization of soluble microbial products from anaerobic treatment by molecular weight distribution and nickel-chelating properties. Water Res. 30, 915–922. Lawrence, J.R., Chenier, M.R., Roy, R., Beaumier, D., Fortin, N., Swerhone, G.D., Neu, T.R., Greer, C.W., 2004. Microscale and molecular assessment of impacts of nickel, nutrients, and oxygen level on structure and function of river biofilm communities. Appl. Environ. Microbiol. 70, 4326–4339.

290

G. Zou et al. / Bioresource Technology 179 (2015) 284–290

Li, J., Liu, Y., Zhang, T., Wang, L., Liu, X., Dai, R., 2011. The effect of Ni(II) on properties of bulking activated sludge and microbial analysis of sludge using 16S rDNA gene. Bioresour. Technol. 102, 3783–3789. Magalhães, C.M., Machado, A., Matos, P., Bordalo, A.A., 2011. Impact of copper on the diversity, abundance and transcription of nitrite and nitrous oxide reductase genes in an urban European estuary. FEMS Microbiol. Ecol. 77, 274–284. Papirio, S., Ylinen, A., Zou, G., Peltola, M., Esposito, G., Puhakka, J., 2014a. Fluidizedbed denitrification for mine waters. Part I: low pH and temperature operation. Biodegradation 25, 425–435. Papirio, S., Zou, G., Ylinen, A., Di Capua, F., Pirozzi, F., Puhakka, J., 2014b. Effect of arsenic on nitrification of simulated mining water. Bioresour. Technol. 164, 149–154. Park, J.Y., Yoo, Y.J., 2009. Biological nitrate removal in industrial wastewater treatment: which electron donor we can choose. Appl. Microbiol. Biotechnol. 82, 415–429. Rasool, K., Ahn, D.H., Lee, D.S., 2014. Simultaneous organic carbon and nitrogen removal in an anoxic–oxic activated sludge system under various operating conditions. Bioresour. Technol. 162, 373–378. Rodier, J., Legube, B., Merlet, N., 2009. L’analyse de l’eau, ninth ed. Dunod, Paris. Sampaio, R.M.M., Timmers, R.A., Kocks, N., André, V., Duarte, M.T., van Hullebusch, E.D., Farges, F., Lens, P.N.L., 2010. Zn–Ni sulfide selective precipitation: the role of supersaturation. Sep. Purif. Technol. 74, 108–118. Schaumlöffel, D., 2012. Nickel species: analysis and toxic effects. J. Trace Elem. Med. Biol. 26, 1–6. Sobolev, D., Begonia, M., 2008. Effects of heavy metal contamination upon soil microbes: lead-induced changes in general and denitrifying microbial communities as evidenced by molecular markers. Int. J. Environ. Res. Public Health 5, 450–456.

Villa-Gomez, D.K., Ababneh, H., Papirio, S., Rousseau, D.P.L., Lens, P.N.L., 2011. Effect of sulfide concentration on the location of the metal precipitates in inverse fluidized bed reactors. J. Hazard. Mater. 192, 200–207. Villa-Gomez, D.K., Papirio, S., Van Hullebusch, E., Farges, F., Nikitenko, S., Kramer, H., Lens, P., 2012. Influence of sulfide concentration and macronutrients on the characteristics of metal precipitates relevant to metal recovery in bioreactors. Bioresour. Technol. 110, 26–34. Wang, L., Liu, Y., Li, J., Liu, X., Dai, R., Zhang, Y., Zhang, S., Li, J., 2010. Effects of Ni2+ on the characteristics of bulking activated sludge. J. Hazard. Mater. 181, 460–467. Wu, H., Gao, Y., Li, H., 2010. Controlled synthesis of nickel phosphate hexahedronal and flower-like architectures via a simple template-free hydrothermal route. CrystEngComm 12, 3607–3611. Xie, C., Yokota, A., 2006. Zoogloea oryzae sp. nov., a nitrogen-fixing bacterium isolated from rice paddy soil, and reclassification of the strain ATCC 19623 as Crabtreella saccharophila gen. nov., sp. nov. Int. J. Syst. Evol. Microbiol. 56, 619– 624. Yebra-Biurrun, M.C., Castro-Romero, J.M., 2011. Speciation of dissolved trace nickel in environmental waters by on-line sonodigestion-flow injection solid phase extraction coupled to flame atomic absorption spectrometry. Am. J. Anal. Chem. 2, 116–125. Zandvoort, M., van Hullebusch, E., Fermoso, Gonzalez F., Lens, P.N.L., 2006. Trace metals in anaerobic granular sludge reactors: bioavailability and dosing strategies. Eng. Life Sci. 6, 293–301. Zou, G., Papirio, S., Ylinen, A., Di Capua, F., Lakaniemi, A., Puhakka, J., 2014. Fluidized-bed denitrification for mine waters. Part II: effects of Ni and Co. Biodegradation 25, 417–423. Zumft, W.G., 1997. Cell biology and molecular basis of denitrification. Microbiol. Mol. Biol. Rev. 61, 533–616.