Environ. Sci. Technol. 2006, 40, 1956-1961
Uptake of Iron Cyanide Complexes into Willow Trees MORTEN LARSEN AND STEFAN TRAPP* Institute of Environment and Resources, Technical University of Denmark, DK-2800 Kongens Lyngby, Denmark
The uptake of iron cyanide into willows was studied. Trees were grown in solutions with Prussian blue, ferricyanide, or ferrocyanide. Iron cyanide speciation in solution was determined by HPLC during the experiment. Total cyanide and total iron in solution and trees were measured at the end of the experiments. Ferrocyanide was the dominating species in most solutions at the end. Ferricyanide was preferably taken up from solutions. Between 20 and 83% of the cyanide was lost from the solutions, and up to 28% could be recovered from the plants, mainly from roots. Cyanide could also be detected in stems and leaves of most exposed trees. Uptake was increased when no other nitrogen source but cyanide was present in solutions. Iron contents in exposed trees, compared to controls, increased significantly. The ratio of iron to cyanide remained rather stable in solution, but changed to higher values inside the plants. This indicates that iron and cyanide were taken up together as a complex, which was dissolved inside plants, and then cyanide was metabolized. No toxic effects could be seen. The study shows that trees can take up and metabolize iron cyanide complexes, making phytoremediation of iron cyanide waste a feasible option.
The speciation of CN found by leaching experiments with MGP waste at various pH conditions was dominated by Fe(CN)64-, which indicates dissolution of Prussian blue (3). According to equilibrium calculations, iron cyanide complexes should be thermodynamically stable only under extreme soil conditions with high pH and low redox potential, but since Prussian blue and Turnbull’s blue are still present in soil several decades after disposal, the speciation seems to be governed by dissociation kinetics rather than by chemical equilibrium (5). Cyanide is also naturally produced by many organisms. All vascular plants produce cyanide as a byproduct in the synthesis of ethylene (6). Subsequently, vascular plants possess a metabolic pathway for cyanide, which converts this toxic byproduct to beta-cyanoalanine (7) and, subsequently, to asparagine (8). Larsen et al. (9) determined the metabolic capacity of five European woody plants for potassium cyanide (which dissociates in solution to free cyanide, HCN and CN-) and found all were capable of metabolizing cyanide rapidly. Yu et al. (10) tested 12 Chinese plant species. All could metabolize free cyanide. The 12 fitted rates were normally distributed. Cyanide metabolism by plants increased with temperature (11). Uptake, metabolism, and accumulation of free cyanide could be predicted with a nonlinear mathematical model (12). Ebbs et al. (13) found evidence that ferrocyanide is also taken up and probably metabolized by plants. Thus, if plants can metabolize free cyanide and take up iron cyanide complexes, they could be used for phytoremediation of CN polluted soil. The scope of this study was to determine toxicity, uptake, and metabolism of iron cyanide in willow trees. The impact of ion composition and nitrogen deficiency on iron cyanide uptake was studied by exposing the trees to ferricyanide in distilled water and in a nutrient solution with and without nitrate. The difference between uptake of ferro- and ferricyanide was also determined.
Introduction
Materials and Methods
Cyanide (CN) is well-known for its toxic effects, and it is a high volume production compound. CN is also a common pollutant in industrial waste. CN occurs in waste from manufactured gas plants (MGP), where the toxic byproduct, hydrogen cyanide (HCN), was removed from the produced gas by precipitation with iron. The waste containing iron cyanide complexes was often deposited on site (1). CN in soils at MGP sites is mainly found as metal-cyanide complexes. Concentrations up to 60 g CN/kg have been reported (2). Theis et al. (3) found that iron cyanide complexes comprised >97% of the total CN in two different MGP site samples. Cyanide forms complexes with FeII, e.g., ferrocyanide FeII(CN)64-, and with FeIII, e.g., ferricyanide FeIII(CN)63-. The most common CN complexes at MGP sites are ferriferrocyanide FeIII4[FeII(CN)6]3, also known as Prussian blue, and ferroferricyanide FeII3[FeIII(CN)6]2, known as Turnbull’s blue. Those give the characteristic blue soil found at former MGP sites. Prussian blue is the stable complex under oxidizing conditions, while Turnbull’s blue dominates under reducing conditions with an excess of Fe2+ over Fe3+. Other iron cyanide complexes, such as Prussian brown, Berlin green, and Berlin white, are also described in the literature but are generally found to be unstable in aqueous systems (4). The solubility constants for dissolution of iron cyanide complexes are given in the Supporting Information (SI).
Chemicals. All chemicals used were reagent grade and were obtained from Bie & Berentsen A-S, 2610 Rødovre, DK, or from the laboratory storage. Solutions. All concentrations of iron cyanide complexes and of total cyanide are referred to as mg CN/L Five different solutions were prepared for the cyanide uptake experiments: (1) distilled water containing Fe(CN)63- at 1, 5 and 10 mg/L; (2) nutrient solution containing Fe(CN)63- at 5 and 10 mg/L as the only iron source; (3) nutrient solution containing Fe(CN)63- at 5 and 10 mg/L as the only iron and nitrogen source; (4) nutrient solution containing Fe(CN)64- at 5 and 10 mg/L as the only iron and nitrogen source; (5) nutrient solution containing Prussian blue at 5 mg/L as the only iron and nitrogen source. The nutrient solution, a modified ISO 8692 standard nutrient solution (Table 1), was prepared from reagent-grade chemicals. The iron cyanide complexes were added to the solutions as K3Fe(CN)6 (s), K4Fe(CN)6 × 3H2O (s), and Fe4[Fe(CN)6]3 (s). The speciation in all solutions was modeled with the chemical equilibrium program PHREEQC prior to the start of the experiments to provide information about speciation and solubility of the different CN species. To prevent precipitation of iron cyanide complexes, the only iron source in the nutrient solution was the iron CN complexes. Nitrate was omitted from solutions 3-5 to see
* Corresponding author phone: +45 4525 1622; fax: +45 4593 2850; e-mail
[email protected]. 1956
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 6, 2006
10.1021/es051224q CCC: $33.50
2006 American Chemical Society Published on Web 02/07/2006
TABLE 1. Composition of the Nutrient Solution Used in the Plant Uptake Experiments macronutrients (µmol/L) NaNO3a
MgCl2×6H2O CaCl2×2H2O MgSO4×7H2O KH2PO4 NaHCO3 a
2823.9 59.0 122.4 60.9 246.0 1785.5
micronutrients (nmol/L) H3BO3 MnCl2×4H2O ZnCl2 CoCl2×6H2O CuCl2×2H2O Na2MoO4×2H2O
2992.1 2097.0 22.0 6.3 0.1 28.9
Only in solution 2.
whether nitrogen deficiency would influence uptake and metabolism of CN. pH and pe. The pH was measured with a PHM 210 standard pH meter from Radiometer Analytical. The pe was measured with an Ingold Pt4805-60-88TE-S7/120 combination ORP/redox electrode, with pe being defined by
Eh )
2.303RT pe F
where Eh is the redox potential (V), R is the universal gas constant (8.314 J K-1 mol-1), T is the absolute temperature (K), and F is the Faraday constant (96 484.56 C mol-1). The pe value was recorded when the value was stable for at least 10 s. Total Cyanide Analysis. The final concentration of total CN in solution and plant was analyzed according to the standard method ISO 11262 (14). Total CN in roots, stem, and leaves was released as HCN by reflux distillation of the sample for 2 h at pH < 2. The HCN was collected in a sodium hydroxide scrubber and measured as free cyanide. To avoid interference from sulfate, the sample was tested for sulfate on lead acetate test paper (15). Free Cyanide Analysis. Free CN (HCN and CN-) in solution reacts with chloramine-T to form cyanogen chloride. This reacts with pyridine-4-carboxylic acid and 1,3-dimethylbarbituric acid to form a colored complex, which was determined photometrically at a wavelength of 606 nm (14). Total Iron Analysis. Total iron in plant material was analyzed using Atomic Absorption Spectrometry following the ISO 15586:2004 standard for sediment (16). Iron Cyanide Analysis. Quantification of ferro- and ferricyanide was conducted on a Perkin-Elmer LC 235, equipped with a C18 column (Phenomenex Luna 5 µm, 250 × 4.6 mm), a C18 precolumn (Phenomenex Security Guard), a Perkin-Elmer LC 235 diode array detector (λ ) 215 nm), and a Perkin-Elmer binary LC pump. All measurements were carried out at 35 °C using a mobile-phase flow-rate of 1.0 mL/min and a sample volume of 50 µL. The mobile phase was a modification of the eluent used by Giroux and Barkley (17): 35% (v/v) acetonitrile and 65% Milli-Q water (Millipore Corporation, MA) with 12.5 mM NaH2PO4 and 4 mM tetrabutylammonium hydrogensulfate (C16H36NHSO4). Further details can be found in the SI. All concentrations of cyanide and iron in plant material are given for fresh weight. Plant Uptake Experiments. The plant uptake and phytotoxicity experiments were conducted according to Trapp et al. (18). Willow cuttings (length 30 ( 5 cm) were placed at the window in buckets with 10-15 cm depth of tap water and pre-grown for 2-3 weeks. The cuttings were transferred to 500-mL Erlenmeyer flasks covered with aluminum foil to prevent algal growth and photodegradation. Cork stoppers, with a hole in the middle, were fitted into the necks of the flasks, to avoid evaporation from the system. Further sealing of the flasks was done with plasticine around the cuttings. The systems were placed in a climate chamber under a rack
of 36-W fluorescent lights. The lights were elevated 65 cm from ground level, and the distance between each light was 20 cm. The temperature and the relative humidity in the chamber were constant at 24 ( 1 °C and 60 ( 5%, respectively. Tap water (500 mL) was added to the flasks, and the plants were left for 2-3 days to adapt to the conditions. Subsequently, the initial transpiration was determined by daily weighing of the flasks for 2 days. Then, the plants were removed from the flasks, and the solution was discharged. A new solution (500 mL) containing the CN complex was added to the flasks, which were weighed, before the plants were replaced in such a way that only the roots were submerged into the solution to avoid uptake of iron cyanide through the lenticels on the stem. For each concentration, 5 to 6 replicates were prepared. Two sets of controls were made. Control A was flasks with willow trees in solutions 1-5 without addition of cyanide. Control B was solutions 1-5 with cyanide at each concentration, but without trees. Fresh weights of the cuttings ranged from 20 to 65 g at start of the treatments, with an average weight of 41 g. The systems were weighed regularly to determine the transpiration. Every 6-8 days, the flasks were refilled to the initial weight with the initial solution, but without the iron cyanide complex. After mixing, samples of 0.5 mL were taken to measure the concentration of the iron cyanide complexes. The experiments with solution 1 were terminated when severe necrosis and defoliation was seen on the willows. Experiments with solutions 2-5 were stopped earlier to avoid the decay of the trees. After harvest, each plant was divided into roots, stem, and leaves. The stems were ground on a Retsch SM 2000 cutter mill to a diameter < 2 mm. Each plant compartment was analyzed for total CN content. The plants grown in solutions 3-5 were further analyzed for iron content. Typical weights for roots, stem, and leaves for total cyanide analysis were 2, 40, and 3 g, respectively. For all plant parts, approximately 1 g of plant material was used for analysis of iron. Phytotoxicity. Inhibition of transpiration is a rapid measure for the toxic effect of a chemical or a substrate to willow trees (18). To take into consideration that healthy trees grow quickly and, thereby, increase transpiration, the normalized relative transpiration (NRT) was calculated. The NRT is the change of transpiration of treated trees divided by the change of transpiration of control trees (see SI). Experiments with Detached Leaves and Roots. To clarify the uptake mechanism of iron cyanide complexes, additional experiments were performed as described above, but with (1) no leaves on the stem and (2) neither leaves nor roots on the stem. Leaves and/or roots were detached from the stem right before the cyanide solution was added. Solution 3 (500 mL) with 5 mg/L ferricyanide as CN was added to the flasks. Refilling, sampling, and analysis was performed as described above. In (1), the detachment of the leaves should decrease the transpiration radically; thus, any uptake should be by diffusion into roots. In (2), the lower 5 cm of the stem was submerged into the solution to determine diffusive uptake into the stem.
Results Transpiration and Phytotoxicity. None of the solutions used in this study showed any significant effect of iron cyanide on the plants. Plants gained weight, except in solution 1. This solution (distilled water plus ferricyanide) had a clear adverse effect. The absolute transpiration fell to below 50% of the initial value (see the Supporting Information), and plants turned yellow and lost leaves. Plants in solution 2 (with NaNO3 as N-source) first increased their transpiration, but after >2 weeks, it fell again. The effect was also seen for the controls; VOL. 40, NO. 6, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1957
FIGURE 1. Concentration and speciation of iron cyanide complexes when trees are grown in solutions 1-5. Initial concentrations of 10 mg CN/L for solutions 1-4 and 5 mg CN/L for solution 5. Control B is the sum of ferricyanide and ferrocyanide for cyanide solution without trees. Error bars denote 95% C.I. thus, NRT remained practically constant (see the Supporting Information), and the effect cannot be contributed to iron cyanide. pH and pe. In the solutions with trees (both with and without iron cyanide), the pH fell from approximately 8 to 6.2-6.5, while pH in control B (without trees) remained practically stable (see the Supporting Information). The pe in the solutions depended on the concentration and speciation of CN. The addition of ferricyanide (solution 3) increased pe, and it decreased during the experiment. After the addition of ferrocyanide (solution 4), the pe was lowered, but it increased during the experiment. In solution 5 (Prussian blue), the pe remained stable during the experiment. The pe in control A, (with trees) decreased to around 2.6 while the final pe in control B was close to the initial values. Concentration and Speciation in Solution. Both concentration and speciation changed over time with trees growing in solution (Figure 1). Only results with 10 mg/L are shown for solutions 1-4 since the results with 5 mg/L showed the same pattern. When grown in solution 1, the main part of the supplied iron cyanide remained as ferricyanide, and the total iron cyanide concentration decreased to below 40% of the initial concentration. In control B (without trees), no change in concentration was detectable and less than 5% of the total iron cyanide was found as ferrocyanide (not shown). When trees were grown in solutions 2 and 3, the total iron cyanide in solution decreased with an average of 44 and 48%, respectively. For both solutions, the speciation of CN changed from being dominated by ferricyanide to be 1958
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 6, 2006
dominated by ferrocyanide. Also here, no loss of CN was detected in the controls, and the speciation remained stable with about 5% found as ferrocyanide. In solution 4, the decrease in concentration was slow in the beginning, but increased toward the end of the experiment where the concentration was decreased by 45%. The speciation remained dominated by ferrocyanide with