Remediation of sediment and water contaminated by copper in smallscaled constructed wetlands: effect of bioaugmentation and phytoextraction D. Huguenot, P. Bois, J. Y. Cornu, K. Jezequel, M. Lollier & T. Lebeau
Environmental Science and Pollution Research ISSN 0944-1344 Environ Sci Pollut Res DOI 10.1007/s11356-014-3406-6
1 23
Your article is protected by copyright and all rights are held exclusively by SpringerVerlag Berlin Heidelberg. This e-offprint is for personal use only and shall not be selfarchived in electronic repositories. If you wish to self-archive your article, please use the accepted manuscript version for posting on your own website. You may further deposit the accepted manuscript version in any repository, provided it is only made publicly available 12 months after official publication or later and provided acknowledgement is given to the original source of publication and a link is inserted to the published article on Springer's website. The link must be accompanied by the following text: "The final publication is available at link.springer.com”.
1 23
Author's personal copy Environ Sci Pollut Res DOI 10.1007/s11356-014-3406-6
RESEARCH ARTICLE
Remediation of sediment and water contaminated by copper in small-scaled constructed wetlands: effect of bioaugmentation and phytoextraction D. Huguenot & P. Bois & J. Y. Cornu & K. Jezequel & M. Lollier & T. Lebeau
Received: 23 May 2014 / Accepted: 30 July 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract The use of plants and microorganisms to mitigate sediment contaminated by copper was studied in microcosms that mimic the functioning of a stormwater basin (SWB) connected to vineyard watershed. The impact of phytoremediation and bioaugmentation with siderophore-producing bacteria on the fate of Cu was studied in two contrasted (batch vs. semi-continuous) hydraulic regimes. The fate of copper was characterised following its discharge at the outlet of the microcosms, its pore
Responsible editor: Elena Maestri Electronic supplementary material The online version of this article (doi:10.1007/s11356-014-3406-6) contains supplementary material, which is available to authorized users. D. Huguenot : P. Bois : J. Y. Cornu : K. Jezequel : M. Lollier : T. Lebeau Equipe Dépollution Biologique des Sols, Université de Haute Alsace, EA 3991 LVBE (Laboratoire Vigne Biotechnologies et Environnement), 33 rue de Herrlisheim, BP 50568 68008 Colmar, France D. Huguenot (*) Laboratoire Géomatériaux et Environnement (EA 4508), UPEM, Université Paris-Est, 77454 Marne-la-Vallée, Paris, France e-mail:
[email protected] P. Bois Département de Mécanique, Equipe MécaFlu, Laboratoire ICube (UMR 7357 CNRS/Unistra/ENGEES/INSA), 2 rue Boussingault, 67000 Strasbourg, France J. Y. Cornu INRA, UMR 1391 ISPA, CS 20032 – 33882 Villenave d’Ornon, France T. Lebeau (*) UMR 6112 LPGN (Laboratoire de Planétologie et Géodynamique de Nantes), LUNAM, Université de Nantes, 2 rue de la Houssinière, BP 92208 44322 Nantes, France e-mail:
[email protected]
water concentration in the sediment, the assessment of its bioaccessible fraction in the rhizosphere and the measurement of its content in plant tissues. Physico-chemical (pH, redox potential) and biological parameters (total heterotrophic bacteria) were also monitored. As expected, the results showed a clear impact of the hydraulic regime on the redox potential and thus on the pore water concentration of Cu. Copper in pore water was also dependent on the frequency of Cu-polluted water discharges. Repeated bioaugmentation increased the total heterotrophic microflora as well as the Cu bioaccessibility in the rhizosphere and increased the amount of Cu extracted by Phragmites australis by a factor of ~2. Sugar beet pulp, used as a filter to avoid copper flushing, retained 20 % of outcoming Cu and led to an overall retention of Cu higher than 94 % when arranged at the outlet of microcosms. Bioaugmentation clearly improved the phytoextraction rate of Cu in a small-scaled SWB designed to mimic the functioning of a full-size SWB connected to vineyard watershed. Highlights - Cu phytoextraction in constructed wetlands much depends on the hydraulic regime and on the frequency of Cu-polluted water discharges - Cu phytoextraction increases with time and plant density - Cu bioaccessibility can be increased by bioaugmentation with siderophore-producing bacteria Keywords Bioaccessibility . Copper . Phytoremediation . Phragmites australis . Siderophore-producing bacteria Abbreviations 3,4-DCA 3,4-Dichloroaniline ANOVA Analysis of variance BI Batch-inoculated BNI Batch-non-inoculated CEC Cation exchange capacity
Author's personal copy Environ Sci Pollut Res
CFU CI CNI EDTA HDPE ICP-AES LB medium NPI NPNI PGPR PI PNI PVC SBP SWB THM
Colony forming unit Semi-continuous-inoculated Semi-continuous-non-inoculated Ethylenediaminetetraacetic acid High-density polyethylene Inductively coupled plasma–atomic emission spectroscopy Luria Bertani medium Non-planted-inoculated Non-planted-non-inoculated Plant-growth-promoting rhizobacteria Planted-inoculated Planted-non-inoculated Polyvinylchloride Sugar beet pulp Stormwater basin Total heterotrophic microflora
Introduction Copper sulphate applied as Bordeaux mixture is widely used in viticulture since more than a century to control powdery mildew. As a consequence, Cu accumulates in vineyard topsoils (Mackie et al. 2012) up to concentrations sometimes higher than 1,000 mg Cu kg−1 (Flores-Vélez et al. 1996). Furthermore, during storms, runoff and soil erosion may occur (Ribolzi et al. 2002). Copper is then recovered at the outlet of some vineyard watersheds (Banas et al. 2010) with concentrations exceeding the French limit authorised in freshwaters (50 μg L−1) (JORF 2007). Stormwater basins (SWBs) were built to ensure temporary storage of runoff water and to retain dissolved and bound-toparticles contaminants (Meyer 1985). In these structures, metals like copper are mainly recovered in settling sediments (Banas et al. 2010) and the remaining part is discharged in the outlet water flux. SWBs are exposed to strong variable environmental conditions and are subjected to highly variable hydraulic and chemical retention times. Most often, retention times are not long enough to sorb all copper (Kadlec 2000). Increasing the sorbing capacities of SWB using low-cost materials such as agricultural wastes (Bailey et al. 1999; Huguenot et al. 2010) can be a way to manage the variability of the copper discharge. It is generally assumed that SWBs have mitigation capacities (Gregoire et al. 2009). These devices are generally colonised by macrophytes able to extract and accumulate metal in their tissues, e.g. copper (Bragato et al. 2006; Ye et al. 2003) up to 261 mg kg−1 of dry matter (Özdemir and Sagiroglu 2000). Unfortunately, this technique known as phytoextraction is most of the time too slow to be considered
as efficient (Van Nevel et al. 2007). Phytoextraction-assisted bioaugmentation is an innovative ecological way to improve phytoextraction (Lebeau et al. 2008; Sessitsch et al. 2013) and was shown for Cu in several studies (Andreazza et al. 2010; Chen et al. 2005; So et al. 2003), without any toxicity contrary to what is commonly reported for synthetic chelates (Doumett et al. 2008). Bioaugmentation may indeed increase phytoextraction by promoting plant biomass production, through the inoculation of plant-growth-promoting rhizobacteria (PGPR) (Andreazza et al. 2010; Płociniczak et al. 2013) and/or by increasing the phytoavailability of metals in soil with, e.g. siderophore-producing bacteria as potential candidates. Indeed, bacterial siderophores are known to efficiently complex Fe (III) and also divalent cations like Cu (II) (Hoegy et al. 2009). Moreover, siderophores are also suspected to enhance metal translocation from roots to aboveground tissues (Braud et al. 2009). To our knowledge, no work has already been performed on the implementation of such a remediation process coupling bioaugmentation and phytoextraction in a constructed wetland context. Indeed, low redox conditions most often prevailing in SWBaccumulating sediments do not favour copper bioaccessibility (Pulford and Flowers 2006) and thus its accumulation in plants. On the contrary, Cu-containing water passing through the SWB may be released at the SWB outlet without time enough for Cu to be extracted by plants. This work aimed at studying the coupling of bioaugmentation and phytoextraction in a small-scale SWB for the removal of copper by Phragmites australis. A previously welldescribed vineyard SWB located in Rouffach (Alsace, France) (Wanko et al. 2009) was used as an example in our study and was already validated in a previous study (Bois et al. 2013: i.e. same design and functioning by submitting the small-scale basins to stochastic polluted water inputs and sediment accumulation). A bacterial mixture (Bois et al. 2011) previously isolated from the sediment of the aforementioned SWB and known to produce siderophores (Bois et al. 2011) was used in bioaugmentation assays to potentially maximise Cu phytoextraction. Several operating conditions were tested (hydraulic regime, bioaugmentation, plant density and processing time). To avoid any release of copper at the outlet of the SWB, sugar beet pulp (SBP) already tested as a valuable sorbent in previous studies (Cornu et al. 2013; Huguenot et al. 2010) was used. Eventually, copper mass balance was assessed to evaluate the performance of our system.
Material and methods General settings General settings are described and detailed in Bois et al. (2013). Briefly, microcosm experiments were performed in
Author's personal copy Environ Sci Pollut Res
high-density polyethylene (HDPE) rectangular boxes (Garhin, France) with a 6.55-L (l×w×h, 39 cm×24 cm×10 cm) working volume corresponding to the SWB usable volume at the 1:150,000 scale. Microcosms were filled with a mix (20:80) of sediment (sampled from the SWB) and sand (purchased by Holcim, France), hereafter named “sediment”. Hydraulic regimes used for the supply of the microcosms were closed to those observed in the reference SWB (Wanko et al. 2009), by taking into account the scaling ratio. A mixture of Cu, glyphosate, diuron and 3,4-DCA (as main contaminants recovered in the SWB water and sediment (Maillard et al. 2011)) was used for each microcosm experiments. Concentrations are given in the “Microcosm experiments” section. Plants P. australis was chosen for microcosm experiments as the most representative plant in the SWB to which it is referred (70–80 % of the total vegetation cover) (Maillard et al. 2011). Nine-month-old plants of P. australis (Aquatic’bezançon, France) were arranged in microcosms at different densities (see the “Microcosm experiments” section for details). Inoculum preparation and sediment bioaugmentation The bacterial inoculum, hereafter named “106”, consists of four identified strains (Arthrobacter sp., Pseudomonas putida, Delftia acidovorans and Brevundimonas sp.) and two unidentified ones (Bois et al. 2011). The preparation of the inoculum and the sediment bioaugmentation is described in Bois et al. (2013). Briefly, the bacterial mixture stored at −80 °C in a sediment extract medium spiked with copper and herbicides was grown for 4 days at 28 °C and 200 rpm in a non-polluted Luria Bertani (LB) medium. Bacterial suspensions were then thoroughly washed with sterile distilled water prior to bioaugmentation at 1.1.1011 CFU kg−1 dry sediment. The bacterial inoculum and the sediment surface were homogenised together.
located at the outlet of each microcosm except for the control (non-planted-non-inoculated (NPNI)). SBP was packed in an 80-cm-long and 4-cm-diameter polyvinylchloride (PVC) column. This specific location was chosen based on previous results (Cornu et al. 2013). Microcosm experiments Experiments I, II and III were performed successively; each new experiment being an improvement of the previous one (i.e. by repeating bioaugmentation, by increasing plant density or by modifying hydraulic regimes and time). Table 1 shows a synopsis of the experiments (replicated three times each, i.e. three microcosms per treatment). Experiment I: effect of plant and bioaugmentation on the fate of Cu in semi-continuous regime (single bioaugmentation) Four treatments were studied: “planted-non-inoculated” (PNI), “planted-inoculated” (PI), “non-planted-non-inoculated” (NPNI) and “non-planted-inoculated” (NPI). In each planted microcosm, three plants were lined up in the direction of the water flux 2 weeks before the first simulated rainy event to allow plant colonisation. The bacterial mixture “106” was inoculated 1 week before the first simulated polluted event. 1.6 L of the artificial runoff water, i.e. deionised water with Cu (37.5 mg.L−1), glyphosate (50 mg.L−1), diuron and 3,4-DCA (10 mg.L−1 each), was added to each microcosm in 1 h 30 corresponding to an inlet water discharge of 18 mL min−1. Water was stored in microcosms for 4 h before an approximate 1-h 30 drainage. The discharge rate value as well as the hydraulic retention time was obtained from rainy events occurring on the reference SWB and then adapted at the microcosm scale. This simulated artificial runoff water was supplied three times fortnightly. All these treatments were studied under hydraulic regime considered as semi-continuous. Experiment II: effect of bioaugmentation on Cu phytoextraction in semi-continuous versus batch regimes
Sugar beet pulp characteristics and use Sugar beet pulp (SBP) was obtained from a sugar refinery (Erstein, France), dried into a furnace at 105 °C until constant weight and used in its raw form: No additional pre-treatment was performed. SBP consisted of coarse particles of approximately 1 mm in width and 5 mm long. Cation exchange capacity (CEC) assessed by acid–base titration was 0.360 meq g−1 and equilibrium pH averaged 6.0 at 20 °C (Cornu et al. 2013). In experiments I and II, 8 and 32 g of dried SBP were respectively spread on the sediment surface of each microcosm whereas, in experiment III, 100 g of dried SBP was
Four treatments were performed in planted microcosms supplied with artificial runoff water in a semi-continuous (C) versus batch (B) hydraulic regimes: non-inoculated (CNI vs. BNI) and inoculated (CI vs. BI) microcosms. The same volume of artificial runoff water, pollutant concentrations and water discharge was applied as for experiment I. Semicontinuous conditions were those of experiment I except that time elapsed between each simulated polluted event was reduced to 1 week. Batch hydraulic regime means that polluted water was stored in microcosms during the whole experiment. In each microcosm, four plants were planted 2 weeks before the first simulated rainy event. The bacterial mixture
Author's personal copy Environ Sci Pollut Res Table 1 Parameters applied for the three different implemented remediation experiments
Experiment I
Experiment II
Experiment III
Experiment term Hydraulic load Bioaugmentation strategy Plant number Polluted event number
8 weeks Semi-continuous Initial 3 3
7 weeks Batch/semi-continuous Every week 4 3
27 weeks Semi-continuous Every week 5 2
Sugar beet pulp location
Surface spreaded
Surface spreaded
Microcosm outlet
106 was inoculated at the beginning of the experiment and then weekly, a few hours after each simulated rainy event. Experiment III: effect of plant and bacterial inoculation on the fate of Cu in semi-continuous regime (repeated bioaugmentation) Four treatments were tested: planted-non-inoculated (PNI), planted-inoculated (PI), non-planted-non-inoculated (NPNI) and non-planted-inoculated (NPI). All these treatments were studied under the above described semi-continuous hydraulic regime. In each microcosm, five plants were planted 2 weeks before the first simulated rainy event was applied. The bacterial mixture 106 was inoculated at the beginning of the experiment and then weekly, a few hours after each simulated rainy event. Distilled water of 1.6 L, whose volume and pollutants were the same as those of experiment I, was stored in microcosms for 6 h before an approximate 1 h 30 emptying (semi-continuous regime). A second water supply spiked with only copper was achieved 2 weeks later under the same conditions as for the first water supply. The copper concentration (56.2 mg L−1) was calculated to keep the same total input of copper in all experiments in spite of the change in the number of simulated rainy events (two vs. three). SBP was packed as described in the “Sugar beet pulp characteristics and use” section and was located at the outlet of the appropriate microcosms. Sediment analysis The sediment moisture and Eh as well as the assessment of the total heterotrophic microflora (THM) were determined and performed as described in Bois et al. (2013). pH of the sediment was determined in water with the ratio 1:2.5 (w/v) using a PHM 210 pH meter (Radiometer Analytical, France) at 20 °C from liquid samples stored at 4 °C. For semi-continuous treatments, rhizospheric sediment was separated from bulk one by shaking the roots. This procedure was not feasible in batch conditions, so the whole sediment was considered as bulk. Bioaccessible Cu was extracted after plant harvesting according to Lebourg et al. (1996): 0.01 M CaCl2 with a liquid/solid ratio of 5 and a 25-
min shaking period. Sediment was dried at 105 °C prior to copper liquid extraction. Liquid extraction samples were then centrifuged at 10,400g for 5 min and filtered on cellulose acetate at 0.45 μm prior to analysis. Pore water was extracted from each microcosm using 2.5mm Rhizon® MOM soil moisture samplers (Rhizosphere Research Products, the Netherlands). All samplers were previously cleaned with 5 % HNO3 and rinsed with deionised water. Three samplers were inserted into each microcosm (in each third of length and at the bottom of each microcosm), just as the sediment was being packed, and left in place throughout the experiment. Each sampler was connected to a polypropylene tube (Rhizosphere Research Products, the Netherlands). A 10-mL syringe suction was applied. Sediment solution was collected immediately after each simulated rainy event and 7 days after. Samples were pooled (for each microcosm) and stored at 4 °C prior to analysis. All samples were 5 % acidified with HNO3 65 % prior to Cu analysis. Plant analysis Aerial parts of P. australis were rinsed with deionised water and roots were washed in 20 mM EDTA and rinsed with deionised water to remove apoplasmic Cu. Prior to weighing, samples were dried at 70 °C during 48 h (Bragato et al. 2006) and then stored at room temperature. Samples of more than 1 g were crushed in a grinder (RETSCH ZM200, Germany). An amount of 0.25 g was mixed with 8 mL HNO3 (65 % w/w) and 2 mL H2O2 (30 % w/w) in a closed Teflon® vessel. Samples were then mineralised at 180 °C for 10 min in a microwave ETHOS (Milestone, USA) prior to copper analysis. Water recovered at the outlet of the microcosms Water recovered at the outlet of the microcosms was retrieved and collected in polyethylene flasks at 4 °C prior to analysis. Copper analysis Copper was analysed by inductively coupled plasma–atomic emission spectroscopy (ICP-AES) (Ultima, Jobin Yvon Horiba Group, Kyoto, Japan) at the wavelength of
Author's personal copy Environ Sci Pollut Res
324.754 nm. Composite solutions were analysed by graphite furnace atomic absorption spectroscopy (Solaar M6, Thermo Fischer). Statistical analyses ANOVAs were performed with StatBox Pro software (Grimmer Soft, USA) after basic hypotheses (i.e. normal distribution of the residues) checked by the same software.
Results and discussion Global functioning of the mitigation system Phytoextraction performances Copper accumulation in P. australis is first shown in nonbioaugmented conditions, i.e. (i) amount of Cu in plants to the total copper input ratio (Fig. 1), (ii) copper extraction rate by plants (Fig. 2) and (iii) amount of copper extracted by plants (Fig. 3). Copper extracted by P. australis to copper input ratio (Fig. 1) averaged, in the whole plant, 0.47 % at the end of experiment I up to 2.53 % for experiment III, as the maximum value obtained among the three experiments, in accordance with Lesage et al. (2007). These authors found 2 % of copper in P. australis during the whole season in a natural wetland. In such a mitigation system, most of the remaining copper is then sorbed onto soil particles, not extracted by plants, as shown by Sinicrope et al. (1992): 70 % of copper was sorbed onto the sediment (only 7 % in the tissues of P. australis) in similar experimental conditions as ours. Experiment II shows a E xtracte d Cu %
twofold increase in Cu phytoextraction compared to experiment I as a result of both the increased number of plants (four plants instead of three), and the reduced time (1 against 2 weeks) elapsed between two events. Compared to experiment II, the long duration of the experiment III (27 weeks) and possibly the number of plants (five plants in experiments III) explain the large increase observed in copper extraction. On top of that, P. australis accumulated Cu almost in roots, like many plants. In case of aquatic plants like P. australis, it can be expected, however, that the whole plant can be harvested. To analyse phytoextraction performances independently of the processing time, phytoextraction rates were compared (Fig. 2). The extraction rate in experiment II (39.06 μg Cu day−1 for the whole plant) is higher compared to that in experiment I (14.74 μg Cu day−1). This is probably not related to the experimental duration (8 weeks for experiment I vs. 7 weeks for experiment II) but to the time elapsed between two consecutive events (only 1 week in experiment II against 2 in experiment I). This could be attributed to a stronger modification of sediment chemistry within 2 weeks rather than in 1 week. The difference in contact time between copper and plants, i.e. the time elapsed between the plantation and the harvest—7 weeks at the minimum in experiment II and 27 weeks at the maximum in experiment III—means that phytoextraction rate decreases with time. It can be suspected that easily bioaccessible copper is absorbed by plant in the early days of the experiments. Once in the sediment, copper undergoes a chemical speciation and easily bioaccessible forms decreased during time as it generally occurs during ageing processes (Bruus Pedersen et al. 2000). The extraction rate in experiment II (39.06 μg Cu day−1 for the whole plant) is higher compared to that in experiment I (14.74 μg Cu day−1). This is probably not related to the experimental duration (8 weeks for experiment I vs. 7 weeks for experiment II) but to the time elapsed between two consecutive events (only 1 week in experiment II against 2 in experiment I). Indeed, successive drying and rewetting
3.0
a’’
Aerial parts 2.5 2.0
Roots Whole plant
a’
1.5
b’
1.0
c’
0.5
b’’ a
c’’
b
b
0.0 Experiment I
Ex periment II
Experiment III
Fig. 1 Copper extracted in aerial parts, roots and whole plant of P. australis during non-bioaugmented experiments I, II and III in semicontinuous hydraulic conditions. Data are given in percent of the total copper supply. Mean values with different letters are significantly different (at P
