Cadmium and Mercury phytostabilization from soil using Miscanthus × giganteus.

The determination of the effects of cadmium and mercury on the growth, biomass productivity and phytoremediation potential of Miscanthus × giganteus (MxG) grown on contaminated soil was the main aim of this paper. The use of bioenergy plants as an innovative strategy in phytotechnology gives additional benefits, including mitigation and adaptation to climate change, and soil remediation without affecting soil fertility. An experiment was set up as a randomized complete block design with the treatments varied in concentrations of Cd (0, 10 and 100 mg kg-1 soil) and Hg (0, 2 and 20 mg kg-1 soil) added to the soil. Three vegetative years were studied. Yield values ranged from 6.3-15.5 tDM ha-1, cadmium concentration in plants varied from 45-6758 µg kg-1 and Hg varied from 8.7-108.9 µg kg-1. Values between treatments and years were significantly different. MxG can accumulate and remove very modest amount (up to 293.8 µg Cd and 4.7 µg Hg) per pot per year in aboveground biomass. Based on this data it can be concluded that MxG, as a valuable energy crop, is a potential candidate for the phytostabilization and biomass production on soils contaminated with Cd and Hg moderately.

rapid immobilization and no need for biomass disposal, while major disadvantage is a fact that pollutants still remain in the soil 13 or in the root system, generally in the rhizosphere. In comparison to phytoextraction, where pollutants are accumulated in the biomass, which is consider a possible problem afterwards. Phytostabilisation has proved to be useful for the treatment of Pb, As, Cd, Cr, Cu and Zn contaminated soils 9 . Phytoremediation is a very complex biotechnology, which is under the significant influence of (I) plant morphology (growth rate, biomass yields) and physiology (accumulation potential, stress tolerance) 14,15 , (II) agro-ecological conditions of cultivated land (soil type and environment) [16][17][18] , (III) agronomy practices (cultivars, planting density, soil amendments application) [19][20][21] , (IV) origin of contamination [22][23][24] . To increase the potential of phytoremediation, which is usually limited by low above ground biomass and/or a shallow root system [25][26][27] more and more attention has been given to perennial plants with high biomass yield potential 20,28 . The use of fast growing energy crops for purifying polluted lands is an innovative strategy to derive additional benefits from such remediation activities [29][30][31] , it may have important role from ecological and energy point of view 7 . Due to the competition for arable land, water and nutrient resources, an implementation of energy crops in the phytoremediation strategy directly avoids the potential conflict between food and fuel production [31][32][33] . The potential soil amendments and phytoremediation stimulants, including (I) mineral fertilizers 28 , (II) farm manure 34 , (III) organic wastes/biosolids (sewage sludge, compost) 35,36 , (IV) solid bioproducts (biochar) 19 , (V) organic substances/biostimulants (mycorrhizal fungi) 21 can be used to increase the biomass yield, the absorption potential of the plants, the amount of soil organic matter and to immobilize the metals in soil. Due to its morphological and physiological characteristics, one of the most investigated bioenergy plant for the purposes of remediation is Miscanthus × giganteus (MxG). Soils that have suffered from physical, biological and/or chemical degradation (i.e. soils contaminated with trace elements), or are uncultivated or/and adversely affected by climate conditions could be defined as marginal lands 37 . Elbersen et al. mapped 29% of agricultural land in EU being marginal 38 . Production of large quantities of biomass, thus providing the effective phytoremediation showed good potential of using Miscanthus sp. commercially on marginal sites in the regions of Central and Eastern Europe, and United States 39,40 . Khalid reported that the most efficient remediation could be achieved with high biomass plants utilization 11 20 by investigating the influence of sewage sludge fertilization applied at different rates to M. sacchariflorus and compared with plant treatment by mineral fertilizers for uptake of different metals into stems and leaves during two years of observation. Cadmium was not detected in M. sacchariflorus biomass in the first year whereas large amounts of the metal were recorded in the second year (6-9 mg kg −1 ) 39,48-50 . The increased accumulation of Cd in MxG shoots with increasing Cd concentrations in the soil induced a reduction in plant height and shoot dry weight 40 . Miscanthus showed low tolerance to Hg toxicity in terms of biomass productivity. It can be grown in fields contaminated with Hg only for soil remediation purposes, since economically might not be feasible due to decreased productivity. Effectively, with the increased Hg concentrations in the soil, biomass showed significantly higher accumulation of Hg, with lower biomass production, in comparison with the control 49 . The average typical value of Cd in Miscanthus crops was reported 0.1 mg kg −1 and for Hg was 0.03 mg kg −1 and for grass in general 0.2 mg Cd kg −1 and <0.02 mg Hg kg −1 51 . The main aim of this paper was to determine the effects of cadmium and mercury on the growth, biomass productivity and phytoremediation potential of MxG grown on contaminated soil. preparation of contaminated soil. Four treatments (C, L 1 , L 1 + SS, L 2 ) varied in concentration levels of Cd (0, 10 and 100 mg kg −1 ) and Hg (0, 2 and 20 mg kg −1 ) were applied to the soil. The first control group (C) consisted of pure soil. The soil in the second group was treated with lower level (L 1 ) of contaminants: 10 mg Cd kg −1 (in CdO (s) form) and 2 mg Hg kg −1 (in HgCl 2 (s) form). A third group (L 1 + SS) was treated with identical concentrations of Cd and Hg applied to soil as in L 1 , but with an addition of sewage sludge in an equivalent of maximal 1.66 t DM ha −1 according to Croatian legislative 52 . Soil in the fourth group was treated with a higher level (L 2 ) of contaminants: 100 mg Cd kg −1 and 20 mg Hg kg −1 of soil. Contaminants were applied as p.a. salts in solid phase to dry soil before first vegetative year. Subsamples of clean soil were mixed with adequate amounts of salts to achieve homogeneity and then were vigorously mixed for a long time with the whole pot volume mass (~18 kg).

Materials and Methods
Soil characteristics. Soil used in the experiment was characterized as silt-loam texture (66.3% silt, 21.3% sand and 12.4% clay; sieving and sedimentation method were used 53 ) with acid reaction (pH KCl = 5.12; obtained in 1 M KCl in 1:2.5 (m/v) 54 ) and having low content of organic matter (OM = 2.26%; determined by wet combustion method with sulfochromic oxidation 55 ). Soil was classified as Stagnosol 56 . Soil was well supplied with total nitrogen (0.12%, determined by dry combustion (Dumas) method 57 ). All light elements (C, H, N and S) were analyzed by dry combustion method on Vario Macro CHNS analyzer, Elementar, 2006. Soil was low supply with plant available potassium (74 mg kg −1 ) and phosphorous (26 mg kg −1 ); (AL method; extraction with ammonium lactate acetic acid in 1:20 (m/v) ratio 58 ). CEC was 18.4 cmol + kg −1 (determined using barium chloride method in 1:40 (m/v) ratio 59 ). Total Cd and Hg in soil were measured in aqua regia extract 60 on AAS equipped with graphite and hydride technique (SOLAR AA Spectrometer M Series, Thermo Scientific, 2008 with Graphite Furnace and Cold Vapour System; see plant analysis, Table 1). Measured Cd concentration was 119 μg kg −1 , which was far below MAC (maximal allowable concentration) for agricultural soils (MAC = 1500 µg Cd kg −1 for soils with pH value between 5 and 6) and measured Hg concentration was 66 μg kg −1 , which was also far below MAC value for agricultural soils (MAC = 1000 µg Hg kg −1 ) according to Croatian legislative 61 .
Sewage sludge characteristics. Wastewater sewage sludge (SS) was characterized having neutral pH value (pH KCl = 7.54 in 1 M KCl in 1:2.5 (m/v) ratio 54 ) and 63% of water 62 . Total carbon content was 22.7% (determined by dry combustion 63 ). Content of hydrogen was 9.52%, nitrogen 2.24% 57 and sulphur 0.36% 64 . A total phosphorous content was 1.35% (extraction in aqua regia 60 ; and detection by ICP-OES 65 , ICP-OES, Vista MPX Axial, Varian, 2004). The concentration of total cadmium in municipal waste water sewage sludge was 349 μg kg −1 which was approximately 3 times higher than in soil, and concentration of mercury was 299 μg kg −1 , which was 4.5 times higher than in soil. Still, concentrations of Cd and Hg in sewage sludge were far below the permitted content of heavy metals in the sludge prescribed by Croatian law while used in agriculture (5000 μg kg −1 for Cd and Hg 52 ).
Biomass sampling and growth parameters. The sampling of MxG was conducted at the beginning of March in 2015, 2016 and 2017 for each experimental pot. The biomass harvest was carried out by manual cutting of the plants at the height of 5 cm from the soil. Whole above ground biomass represents the sample with all dead leaves which were collected, if were any. Yield parameters, including plant height, shoot numbers per rhizome and mass of biomass with natural moisture content were determined on the site. Afterwards samples were cut to smaller pieces and transported to the Lab. The dry matter yield was determined gravimetrically after drying at 60 °C to the constant mass. Afterwards samples were milled to a powder and proceeded to digestion and metals analysis.
Biomass analysis. Table 1  Statistical analysis and quality control. Statistical analysis was done with the use of statistical software SAS 9.1 (SAS Inst. Inc.), One-Way ANOVA and post-hoc (Fisher LSD) test were used for processing of data. The threshold of significance was 5% for all tests. Quality control was included. Measurement accuracy and method precision for Cd and Hg determination were checked using reference materials (IPE 171 and IPE 186 for plant and ISE 865 for soil, Wageningen University) and were satisfactory. Absolute error for Cd measurements was up www.nature.com/scientificreports www.nature.com/scientificreports/ to maximal 8% and for Hg up to 5%, respectively. Relative standard deviation (RSD) or repeatability of measurement for Cd was up to maximum 7% and for Hg 8%, respectively.

Results and Discussion
Growth parameters. Results of the study including yield, length of plants, number of shoots regarding the treatment options and vegetative years are presented in Figs. 1 to 3. Statistical analysis of the results shows the influence of vegetative years on the yield for MxG by different treatment. The significant difference for C, L 1 +SS and L 2 treatments has been determined between the years of investigation (Fig. 1). In comparison to the first year of the study, the decrease of the MxG's yield was found in the second and third year, both on the control treatment and on the contaminated soil, in the range of 37% up to 55%. It is highly unlike that the shoots will accumulate any significant amount of heavy metals from the soil in the first year of growth and have a significant impact on the yield, thus obtained values were expected. However, only in the control treatment the increase in the yield was observed in the third year. An interaction between MxG yield and treatments displays the statistically significant difference only in the third year of the research. It can be noticed that the highest yield was determined  in the control treatment C (7.3-15.5 t DM ha −1 ), and the lowest in the treatment L 2 (6.3-11.4 t DM ha −1 ). Even application of municipal sewage sludge, as a soil amendment was not resulting in biomass increase because the yield on sewage sludge treatment compared to L 1 was not statistically significant. Thus, no impact on soils slightly contaminated with Cd and Hg during three-year investigation period was observed. The negative impact of Cd and/or Hg on biomass yield was also determined by Arduini et al. and Fernando and Oliveira 42,45 . Antonkiewicz et al. found a positive impact of sewage sludge on the yield, in five years long research 66 . However, they applied up to 36 times higher doses of municipal sewage sludge (0-60 t DM ha −1 ) in comparison to our study. The average yield in their study was reported 15.3 t DM ha −1 in control treatment and 16.6 t DM ha −1 in treatment with sewage sludge, respectively. The length of the plant was not statistically influenced by the treatments, while significant differences can be seen between years of research for treatments C, L 1 + SS and L 2 (Fig. 2). If we compare first and third year, the increasing length of the plant in treatment C could be noticed; while this was not the case for other treatments. Length of the plant in this study ranged from 103 cm up to 172 cm. Fernando and Oliveira, and Arduini et al. determined the reduction of plant length in relation to the increase of Cd concentration in the soil 42,67 . Zhang et al. noted that growth of Miscanthus sacchariflorus was significantly inhibited when Cd concentration in the soil was above 50 mg kg −1 compared with control 68 . Kocoń and Jurga were investigating shoot numbers and shoot length of MxG and Sida hermaphrodita on two different soil textures, including sandy and loamy soil contaminated with Cd, Cu, Ni, Pb and Zn of the second year of cultivation. They determined that MxG had greater number of shoots and lower shoot length compared to Sida hermaphrodita regardless of soil texture 28 . However, Fernando and Oliveira did not observe the negative impact of Hg on the length of MxG plant 42 . There were no statistically significant differences between treatments and years for the number of MxG shoots (Fig. 3).   (Fig. 5). Statistically significant differences of Hg in the MxG are observed between years. For the first and second year of investigation, we noted statistically significant differences of Hg in the MxG between treatments too. All measured values of Hg in the MxG except those revealed at L 2 , 2016 (108.9 μg kg −1 ) were below typical (30 μg kg −1 ) of Hg in the MxG according to HRN ISO 17225-1:2014 51 . Fernando and Oliveira investigated Hg concentration in the MxG aboveground biomass, cultivated on soils with two different levels of contamination (5000 and 6700 μg Hg m −2 ), and determined that contamination with lower Hg dose resulted with Hg in biomass below limit of detection of method (<LOD), while contamination of soil with higher Hg dose resulted with Hg concentration in the biomass of 4 μg kg −1 42 . Pérez-Sanz et al. investigated mercury uptake by Silene vulgaris, grown on contaminated (5.5 mg Hg kg −1 ) spiked soils (alkali and neutral pH) and observed that S. vulgaris retains more Hg in the root (3700 and 2900 μg kg −1 ) than in aerial part (550 and 980 μg kg −1 ) 71 . Still, plants grew healthy and showed good appearance throughout the study without significantly decrease in the biomass production. Hg values in the aboveground biomass of the MxG in this study are considerably lower than those expected. However, it is not surprising due to Hg as a specific element and its behavior being a bit different than all other heavy metals. Mercury has been known as an environmental pollutant for over a century and it is well known that it may evaporate (volatilization) from some compounds and be released to various ecosystems. When added to the soil, whether in elemental, inorganic or organic form, it is likely to be strongly bond. Generally, 97-99% of total Hg is in complex form, and behavior of Hg species in the soil is controlled by soil factors, especially temperature, pH,  www.nature.com/scientificreports www.nature.com/scientificreports/ texture, organic matter content but also the concentration of all other ions. Phytoavailability and toxicity of Hg in the soil-plant system depend on the forms in soil 72 . Lomonte et al. study has shown that Hg accumulated by C. zizanioides via root uptake is mainly present in the root epidermis and exodermis and its translocation to the aerial parts is insignificant 73 . In contrast, another study done by Lomonte et al. shows that some species (Atriplex conodocarpa and Australodanthonia caespitose) can be good candidates for mercury phytoextraction because of their ability to translocate mercury from roots to the aboveground tissues 74 . Lomonte et al. also study Hg behavior in soils and Hg efflux to atmosphere 75 . They applied biosolids (3.5-8.4 mg Hg kg −1 ) from waste water treatment plant to soil and investigated potential for Hg remediation. They observed that 59% of the total mercury was complexed with organic ligands and that the influence of water content and irradiation on the emission of gaseous elemental mercury are the main factors affecting this emission with flux values up to 132 ngm −2 h −1 . Lomonte et al. revealed that some ions mobilize Hg in the soil, creating chelate-assisted phytoextraction for some species and increase its uptake in the plant shoots 76 . Pogrzeba et al. also studied Hg behavior in contaminated soil with addition of granular sulphur and observed that in Hg stressed environment, plant (Poa pratensis) developed the defense mechanism resulting in the reduction of Hg evaporation and higher S content in plant tissue 77 . Those authors recommended this technology for soil remediation heavily contaminated with mercury. cadmium and mercury biomass removal. Cadmium and Hg biomass removals by MxG according to treatments and studied years are presented in Figs. 6 and 7. Statistically significant interaction between years of research and Cd removal is determined for all treatments, with the highest biomass removal determined in the second year. In the first year of investigation, Cd biomass removal between the treatments is not statistically significant and values determined below 10 µg pot −1 . In the second and third year of investigation, significant increase in Cd biomass removal is noted and is in a complete positive correlation with Cd concentration in soil (Fig. 6) 70 . Values of Cd removal ranged from 70.8 up to 293.8 µg per pot (11.8-49.0 g ha −1 ) observed on contaminated treatments (L 1 and L 2 ), in the second and third year of investigation, respectively. This is in accordance  to Barbu et al. who determined an uptake of 35-55 g Cd ha −1 47 . Bang et al. noted limited Cd accumulation by Miscanthus in a marginally contaminated ecosystem, although they observed 100% of Cd removal from contaminated water after 40 days 78 . Yao et al. observed Cd accumulation in plant tissue of 2.2 mg Cd m −2 (22 g ha −1 ) in Miscanthus sacchariflorus and 700 kg Cd per year where it was accumulated by aboveground organs and removed from the lake (Dongting Lake wetlands, China) through harvesting for paper manufacture 79 . Hg biomass removals by the MxG according to treatments and studied years varied up to 4.7 µg per pot (0.79 g ha −1 ) in our study. Statistically significant influence of years on Hg removal has been determined for treatments L 1 + SS and L 2 . In terms of interaction between treatments, significant Hg removal is observed only in the second year of research in treatment L 2 (Fig. 7).

conclusion
Values of Cd in the MxG in first investigated year were low in all treatments, much lower than the typical value when MxG is used as a biofuel (100 μg kg −1 ). The significantly higher concentrations of Cd in the MxG were observed in the second and third year of investigation due to increased doses of Cd in the soil, and correlations were completely positive.
The detected mercury concentration in MxG plants was very low. The whole measured values of Hg in MxG except those of treatment L 2 , 2016 (108.9 μg kg −1 ) were below typical for MxG used as biofuel (30 μg kg 1 ).
Finally, it can be concluded that MxG, as a valuable energy crop, could be a good candidate for the Hg and Cd phytostabilization, due to the low metal accumulation in aboveground biomass. This could be benefit for biomass production of MxG on soils moderately contaminated with Cd and Hg, where contamination still not significantly affected the yields amounts.
Except phytostabilization of the contamination, the MxG can also be used in locations where some other remediation strategies of ecosystem/agroecosystem need to be achieved, like prevention of soil erosion due to the high biomass above but also below ground.

Data availability
The row datasets generated and/or analyzed during this study are available from the corresponding authors upon request.