VZJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Mansfeldt, T.
Right arrow Articles by Kretzschmar, R.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Mansfeldt, T.
Right arrow Articles by Kretzschmar, R.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Mansfeldt, T.
Right arrow Articles by Kretzschmar, R.
Related Collections
Right arrow Laboratory Column Studies
Right arrow Other Contaminants
Right arrow Soil Pollution
Published in Vadose Zone Journal 3:471-479 (2004)
© 2004 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA

SPECIAL SECTION: COLLOIDS AND COLLOID-FACILITATED TRANSPORT OF CONTAMINANTS IN SOILS

Cyanide Leaching from Soil Developed from Coking Plant Purifier Waste as Influenced by Citrate

Tim Mansfeldta, Heike Leyera,b, Kurt Barmettlerb and Ruben Kretzschmar*,b

a Soil Science and Soil Ecology Group, Faculty of Geosciences, Ruhr-University Bochum, D-44780 Bochum, Germany
b Inst. of Terrestrial Ecology, Swiss Federal Inst. of Technology, Grabenstrasse 3, CH-8952 Schlieren, Switzerland
* Corresponding author (kretzschmar{at}env.ethz.ch).

Received 30 June 2003.



   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
Soils in the vicinity of manufactured gas plants and coal coking plants are often highly contaminated with cyanides in the form of the compound Prussian blue, FeIII43. The objective of this study was to investigate the influence of citrate on the leaching of iron–cyanide complexes from an extremely acidic soil (pH 2.3) developed from gas purifier waste near a former coking plant. The soil contained 63 g kg–1 CN, 148 g kg–1 Fe, 123 g kg–1 S, and 222 g kg–1 total C. Analysis of the soil by X-ray diffraction (XRD) and Fourier transform infrared (FT-IR) spectroscopy revealed the presence of Prussian blue, gypsum, elemental sulfur, jarosite, and hematite. For column leaching experiments, air-dried soil was mixed with purified cristabolite sand at a ratio of 1:3 and packed into chromatography columns. The soil was leached with dilute (0.1 or 1 mM) CaCl2 solutions and the effluent was collected and analyzed for total and dissolved CN, Ca, Fe, SO4, pH, and pe. In the absence of citrate, the total dissolved CN concentration in the effluent was always below current drinking water limits (<1.92 µM), indicating low leaching potential. Adding citrate at a concentration of 1 mM had little effect on the CN concentrations in the column effluent. Addition of 10 or 100 mM citrate to the influent solution resulted in strong increases in dissolved and colloidal CN concentrations in the effluent, which was due to ligand-controlled dissolution of Prussian blue, desorption of FeII4–6 or FeIII3–6 by sorption competition with citrate, and mobilization of colloidal particles by citrate. However, our results indicate that relatively high concentrations of citrate are necessary to significantly increase CN leaching from the strongly acidic soil.

Abbreviations: DOC, dissolved organic carbon • FT-IR, Fourier transform infrared • HPLC, high-pressure liquid chromatography • HFO, hydrous ferric oxides • PDF, powder diffraction files • XRD, X-ray diffraction • XRF, X-ray fluorescence


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
CONTAMINATION OF SOILS with cyanides is a widespread problem on former manufactured gas plant and coal coking plant sites. The first manufactured gas plant worldwide was established in 1812 in London, England. Since then, a large number of manufactured gas plants have been established, producing gas for municipal lighting, heating, and private household use. The number of former manufactured gas plant sites in the USA was estimated to be about 1100 to 3000 (Shifrin et al., 1996). For Europe, recent estimates amount to 3000 ± 1000 sites in England (ERL, 1987), 234 sites in the Netherlands (Meeussen et al., 1994), and more than 1060 sites in Germany (Mansfeldt and Rennert, 2003). Coal coking plants produced coke for the iron and steel industries. Their gas was a by-product and was often sold to the chemical industry or to cities for use as municipal gas. In Germany, about 250 former coking plants are known. Most of them are located in the Ruhr area, which was one of the largest coal and steel producing regions worldwide (Mansfeldt and Rennert, 2003).

During coal gasification, hydrogen cyanide, HCN, was produced in the coke ovens (Grosskinsky, 1958). The main reaction for the formation of HCN in the coke oven is

[1]

In addition, hydrogen sulfide (H2S) was formed from sulfur compounds in the coal. Because both HCN and H2S are toxic and corrosive, the raw gas had to be purified before its distribution (Riesenfeld and Kohl, 1974). At most sites, a dry purification technique was used as the last step in gas purification. The raw gas flowed through the so-called spent oxides or purifiers, which consisted of wood shavings and iron oxides originating from bog iron and iron ores. Spent oxides were filled into boxes, sometimes into large purifier towers. By reactions of HCN with the Fe-rich purifier material, iron–cyanide complexes, [Fe(CN)6], were formed, mostly as the crystalline compound Prussian blue, FeIII43, which is a strong blue pigment. The removal of H2S from the raw gas was also very efficient when using iron oxides. Hydrogen sulfide was transformed into iron sulfides, thereby lowering the content of iron oxides in the purifier. When the oxide content of the purifier became too low, the purifier had to be regenerated. Regeneration was simply performed by aerial oxidation producing reactivated iron oxide and sulfuric acid. For this, the purifier material was stored for several months on the soils around the gas plant. After several cycles of regeneration and reuse of the purifier material, the content of Prussian blue became too large, rendering the purifier less effective. The material was then often disposed on site. Manufactured gas plant purifier wastes from New York State were reported to contain 14.3 and 23.6 g kg–1 CN (Young and Theis, 1991). Some highly contaminated purifier wastes from England contained CN concentrations in the range 30 to 60 g kg–1 (ERL, 1987).

In soils, dissolved CN occurs mostly as iron–cyanide complexes . These CN species are not acutely toxic; however, under the influence of light they may be rapidly photodegraded to form the extremely toxic HCN (Meeussen et al., 1992a; Rader et al., 1993). Therefore, the drinking water limits for total CN are rather low (e.g., Germany: 50 µg L–1 total CN; USEPA: 200 µg L–1 free CN). Some countries also have regulatory threshold values for cyanide contamination of soils (e.g., Germany: 50 to 100 mg kg–1 total CN, depending on land use).

The actual risk of groundwater contamination from highly cyanide contaminated soils depends on the solubility and mobility of the iron–cyanide complexes. The solubility of iron–cyanide complexes in soils is influenced by a variety of chemical processes including oxidation–reduction, precipitation–dissolution, sorption–desorption, complexation with inorganic ions, and chemical or microbial decomposition (Cheng and Huang, 1996; Dursun et al., 1999; Fuller, 1985; Ghosh et al., 1999; Hipps et al., 1988; Meeussen et al., 1992b, 1994; Ohno, 1990; Rennert and Mansfeldt, 2001a, 2001b, 2002; Theis et al., 1988). Iron–cyanide complexes are becoming more soluble with increasing soil pH and increasing redox potential (Meeussen et al., 1994, 1995).

Another factor that could influence the solubility of iron–cyanides in soils may be the type and concentration of organic ligands in solution, for example, low molecular weight organic acids exudated by microorganisms and plant roots. The influence of organic ligands on cyanide leaching may also be relevant for developing in-situ soil remediation treatments for contaminated areas at manufactured gas and coking plant sites. Organic acids such as oxalate, citrate, malonate, and others can adsorb to the surfaces of Fe oxides and Prussian blue and promote mineral dissolution by forming soluble complexes with ferric Fe. Thereby, the iron–cyanide complex may be released either by sorption competition with organic acids on Fe oxides or by ligand-promoted dissolution of Prussian blue. However, the effects of organic ligands on the mobilization and leaching of cyanides from contaminated soils has not yet been studied. Therefore, our objectives were (i) to characterize the chemical and mineralogical composition of a soil developed from purifier waste material on a former coking plant site, and (ii) to assess the potential for cyanide leaching from this soil in the presence and absence of citrate. In this study, citrate was used as an example for low molecular weight carboxylic acids in soils.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 REFERENCES
 
Study Site and Sample Collection
Soil samples were collected at a former coal coking plant site in Dortmund, North-Rhine Westphalia (Germany) in December 1998. The coking plant has been in operation between 1928 and 1992. During this time period, cyanide-enriched gas purifier material has repeatedly been deposited on-site for regeneration, covering an area of roughly 2000 m2. Today, this area is vegetated with birch trees (Betula alba L.) and a weakly developed soil has developed from the gas purifier waste, which is classified as Urbic Anthrosol (FAO Unesco, 1994). For sampling, a 40-cm deep pit was excavated in the contaminated area, which we identified during a preliminary field survey. The high cyanide content was readily recognized by the blue color of the soil, which is due to the compound Prussian blue. Approximately 15 kg of soil was collected from a depth of 10 to 35 cm, dried at 40°C, and sieved to collect the <2 mm fraction for column leaching experiments and chemical analyses.

Soil Characterization
The soil was analyzed for total elemental composition, mineralogy, soil pH, and concentrations of organic C and cyanides. For total elemental analysis, a subsample was ground in a vibratory disc mill (RS 1, Retsch, Haan, Germany) with a tungsten carbide grinding set. Total concentrations of elements with atomic number Z > 11 (Na) were measured by energy-dispersive X-ray fluorescence (XRF) analysis on a Spectro X-Lab 2000 spectrometer equipped with a sequence of secondary targets (Mo, Al2O3, B4C/Pd, Co, and HOPG) producing polarized X-rays. The detection limit was approximately 0.5 mg kg–1 for most elements reported. The total contents of C and N were determined for ground samples using a CHNS Analyzer (CHNS-932, LECO Instrumente GmbH, Moenchengladbach, Germany).

Total CN was determined with an alkaline extraction method (Mansfeldt and Biernath, 2001). Ten grams of dry soil were dispersed in 250 mL of 1 M NaOH, equilibrated for 16 h on an end-over shaker, and centrifuged to collect the supernatant solution. The supernatants from three subsequent extractions were combined and analyzed for total CN content. The samples were digested under acid conditions and boiled with a micro-distillation technique (Mansfeldt and Biernath, 2000). The evolved HCN gas was absorbed in an alkaline solution and CN was determined spectrophotometrically at 600 nm using a barbituric acid–pyridine solution. Soil pH was measured with a combination pH electrode (Type 6.0204.100, Metrohm, Switzerland) after equilibrating 10 g soil with 25 mL 0.01 M CaCl2 solution for 30 min.

For X-ray diffraction (XRD) analysis, the dry soil was carefully ground by hand using an achate mortar and pestle and a randomly oriented powder mount was prepared. The sample was analyzed on a Bruker D4 Endeavor X-ray diffractometer (Bruker AXS GmbH, Karlsruhe, Germany) using Cu–K{alpha} radiation, variable slits, and a secondary monochromator. To obtain a satisfactory signal/noise ratio, a slow scan with a step size of 0.01° 2{theta} and a counting time of 10 s per step was recorded. Identification of crystalline phases was performed using the ICDD powder diffraction files (PDF-2, International Center for Diffraction Data, Newtown Square, PA).

Column Leaching Experiments
The leaching of iron–cyanide complexes from the soil was investigated by conducting laboratory column experiments at controlled flow rates. Air-dried soil was mixed with pre-cleaned cristobalite (SiO2) sand (200–400 µm, Siegfried AG, Zofingen, CH) at a weight ratio of 1:3. The soil–sand mixture was then uniformly packed into chromatography glass columns (Omnifit, Cambridge, England). We added the cristobalite sand to the soil to prevent clogging of pores during leaching experiments. The resulting soil columns were between 30.4 and 39.9 cm long, had an inner diameter of 1 cm, and a bulk density of 1.43 ± 0.06 g cm–3. The resulting pore volume of the columns was between 11.7 and 13.6 mL. The dry columns were purged for at least 10 min with CO2 gas to remove all air from the pore space. Due to the higher water solubility of CO2 compared with N2 and O2 in air, this procedure ensures complete water saturation of the soil shortly after introducing the aqueous solution. The column inlet was then connected to a high-pressure liquid chromatography (HPLC) pump (Jasco PU-980, Jasco, Tokio, Japan) delivering a degassed CaCl2 solution (0.1 or 1 mM) at a constant flow rate between 0.2 and 1.0 mL min–1. The column outlet was connected to an automated fraction collector (Foxy 200, Isco, Lincoln, NE) to sample the column effluent in regular time intervals. To prevent light-induced degradation of iron–cyanide complexes complexes to HCN, the experiment was set up in the dark in a room maintained at 25 ± 1°C. The collected effluent samples were tightly capped and stored in the dark at 4°C until analysis.

Two leaching experiments were conducted without adding citrate to the influent solution. The soil columns were leached with a 1 mM CaCl2 solution at a constant flow rate of 0.3 and 1.0 mL min–1, respectively. Three column experiments were conducted with a sequence of influent solutions. In the first step, we leached the columns with 100 mM CaCl2 solution without citrate. Then the influent was switched to a 0.1 mM CaCl2 solution until the effluent composition was stable and excess CaCl2 had leached from the columns. In the third step, 1, 10, or 100 mM citric acid was added to the 0.1 mM CaCl2 influent solution. To assess the importance of slow kinetics in cyanide release, we conducted several stopped-flow events in which the flow was interrupted for 24 h and then resumed using the same influent solution.

The column effluents were analyzed for CN, SO4, Ca, Fe, pH, and redox potential, Eh. Total and dissolved cyanides (expressed as CN) were analyzed spectrophotometrically after digestion and micro-distillation as described above. The detection limit was approximately 0.4 µM CN. The total CN concentrations were determined by analyzing untreated effluent samples, while dissolved CN concentrations were measured after centrifugation of the samples on an ultracentrifuge (Kontron Centrikon T-1170, Kontron Instruments, Watford, UK) equipped with fixed angle rotor (FFT 70.13) for 2 h at 200000 x g (40000 rpm). This centrifugation step was designed to remove all particles larger than approximately 10 nm in Stoke's spherical equivalent diameter assuming an average specific density of 2.65 kg m–3. The difference between total and dissolved concentrations was interpreted as colloidal CN.

Sulfate was measured using an ion chromatography system (DX 300, Dionex, Idstein, Germany) equipped with an AG 12 A (Dionex) pre-column, a AS 12 A (Dionex) column, and an electrical conductivity detector (CD 20, Dionex). The eluent contained 2.7 mM Na2CO3 and 0.3 mM NaHCO3, and the flow rate was 1.5 mL min–1.

Concentrations of Ca and Fe in the effluent samples were measured by atomic absorption spectrometry (SpectrAA220FS, Varian, Australia) at a wavelength of 422.7 nm for Ca and 248.3 nm for Fe, respectively.

The effluent pH values were measured with a combination pH electrode (Type 6.0204.100, Metrohm, Switzerland). A platinum electrode (Type 6.0434.100, Metrohm, Switzerland) was used to measure the redox potential of the effluent solutions, reported here as pe value (Eh = 0.059 pe, where Eh is the redox potential in V).

Equilibrium calculations were performed for comparison of effluent concentrations of Ca, SO4, Fe, and CN using the computer program ECOSAT (Keizer et al., 1993). The stability constants for cyanide species were adopted from Meeussen et al. (1992b), for other relevant species we used the constants of Lindsay (1979).


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
REFERENCES
 
Soil Chemical and Mineralogical Composition
The results of the soil characterization are summarized in Table 1. Total elemental analysis showed that, in addition to O, Fe (148 g kg–1), C (222 g kg–1), S (123 g kg–1), and N (50 g kg–1) were the most abundant elements in the soil. The total CN content was 63 g kg–1, which accounts for 13% of the total C and 68% of the total N content. The remaining 87% of the total C were present as elemental or organic C.


View this table:
[in this window]
[in a new window]
  		Table 1. Elemental composition and pH of the soil developed from coking plant purifier waste. Elements with contents below 1 g kg–1 are not reported. CN = total cyanide.

 
Crystalline components in the soil were identified by powder XRD analysis. The XRD pattern of the soil (Fig. 1) exhibited diffraction peaks corresponding to d-spacings of 0.509, 0.360, 0.255, and 0.228 nm, which can be assigned to Prussian blue (Fe4[Fe(CN)6]3). Based on the total CN content, the amount of Prussian blue was estimated to be 116 g kg–1, assuming that most of the CN is bound in this solid form. This amount accounts for 35% of the total Fe, the remaining 65% of the Fe is most likely present as hematite and poorly crystalline hydrous ferric oxides (HFO). Rhombic elemental sulfur (S) and gypsum (CaSO4·2H2O) were also clearly identified by XRD (Fig. 1). In addition, the diffraction pattern indicated the presence of small amounts of jarosite (KFe3(SO4)2(OH)6). Given the relatively low Ca content of the soil (25 g kg–1), the amount of gypsum cannot exceed 105 g kg–1 CaSO4·2H2O, corresponding to 16% of the total S. Neglecting the small amounts of jarosite and adsorbed sulfate, the amount of elemental S can therefore be estimated to be approximately 104 g kg–1, accounting for 84% of the total S content.



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 1. Powder X-ray diffraction pattern of the soil developed from purifier waste near a former coal coking plant. Major peaks are labeled with the d-spacing (in nm) and corresponding mineral phases: PB = Prussian blue, GY = gypsum, JR = jarosite, S = elemental sulfur, and H = hematite.

 
The FT-IR absorption spectra of the soil and several reference compounds are shown in Fig. 2 . The soil exhibited infrared absorption bands at wavenumbers 3403, 2087, 1621, 1418, 1194, 1085, 1006, and 630 cm–1, respectively. The bands at 2087 and 1418 cm–1 can be assigned to CN stretching and NH bending vibrations of Prussian blue, respectively, which is shown as a reference spectrum. The band at 1621 cm–1 with a small shoulder at 1687 cm–1 is due to gypsum. Gypsum may also contribute to the absorption bands near 3403 cm–1 and in the region of 1000 to 1200 cm–1, but the band assignment in this region is less clear, since also organic compounds and other minerals may contribute to IR absorption bands at these wavenumbers. Reference spectra of cellulose and lignin are presented as two examples for organic compounds occurring in soils (Fig. 2).



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 2. Transmission FT-IR spectra (KBr pellets) of the cyanide contaminated soil developed from coking plant purifier waste. Spectra of Prussian blue, gypsum, lignin, and cellulose are shown for comparison.

 
The soil suspended in 0.01 M CaCl2 had a pH value of 2.3, which is classified as extremely acidic. The rather unusual composition presented above and the extremely low pH will now be discussed. The high content of Fe originates from the use of iron oxides or spent oxides as gas purifier in the coking plant. The high C content of the soil may have several reasons. First, during coal gasification, thousands of organic compounds were formed, including tar oils and polycyclic aromatic hydrocarbons. Most of these organic compounds were removed from the gas stream by previous cleaning steps, but some compounds were retained in the iron oxide–rich purifier, which represented the last purification step (Grosskinsky, 1958). Second, gas purifier material often contained wood shavings, which may also contribute significantly to the total C content (Riesenfeld and Kohl, 1974). Finally, the soil is now vegetated with birch trees and some of the organic C stems from decaying plant residues and soil organisms. The high S content of the soil can also be explained by its formation from gas purifier material. Hydrogen sulfide is created during coal gasification and is completely retained in the purifier by conversion of iron oxides to iron mono-sulfides (FeS). When purifiers were frequently regenerated, S could be enriched to levels of up to 600 g kg–1 (ERL, 1987). When purifiers are aerated for regeneration, iron sulfides will oxidize in a first step to elemental S. In a second step, some of this elemental S can be oxidized to sulfuric acid. Considering the large amounts of S in the soil, this reaction has probably resulted in the extremely low soil pH and the formation of gypsum.

Column Leaching Experiments
Influence of Flow Rate and Salt Concentration
In a first set of column leaching experiments, we studied the release of iron–cyanide complexes from the soil in the absence of organic acids in the influent. Dissolved cyanide in soils is mostly present as iron–cyanide complexes, because the decomposition of these complexes to free cyanide CN– is very slow in the absence of light (Meeussen et al., 1992a). Here, we report all cyanide concentrations as dissolved, colloidal, or total CN, regardless of the cyanide speciation. Figure 3 shows the results of two experiments in which the soil–sand mixture was leached with 1 mM CaCl2 solution at a flow rate of 1.0 and 0.3 mL min–1, respectively. The dissolved CN concentrations in the effluent never exceeded 1.92 µM, which equals the German drinking water limit of 50 µg L–1 CN. Higher total concentrations were only observed for total CN during the first two pore volumes, that is, directly after rewetting the dry soil with CaCl2 solution. However, the total CN concentrations also dropped to values below 1.92 µM with two pore volumes and remained low during the entire leaching period. The difference between the total and the dissolved CN concentration can be interpreted as colloidal (or particle-bound) CN. The colloidal particles leached during the first two pore volumes may be a result of drying, sieving, packing, and/or rewetting of the soil. The drastic decrease in Ca concentration after the complete dissolution of gypsum did not result in increased colloidal CN concentrations, suggesting that the chemical conditions are unfavorable for colloid release and transport. In general, colloid release in soils can be expected to be negligible at low pH, high ionic strength, and/or high Ca saturation of cation exchange complex (Kretzschmar et al., 1999). Therefore, we interpret the initial release of colloidal CN as an experimental artifact with no significance for CN leaching under field conditions, unless the soil is repeatedly disturbed by soil tillage or other physical impacts.



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 3. Total and dissolved CN, SO4, Ca, pH, and pe in the effluent from packed soil columns leached with a 1 mM CaCl2 solution (pH 6) at a constant flow rate of 0.3 and 1.0 mL min–1, respectively.

 
Figure 3b shows the concentrations of Ca and SO4 in the column effluents. The concentration of Ca exhibited a plateau near 14 mM and remained constant for almost 33 pore volumes, after which it dropped to about 1 mM. The concentration of SO4 also showed a clear plateau near 14 mM over the same leaching period. The leaching behavior of Ca and SO4 can be explained by dissolution of gypsum , which was identified in the soil by XRD and FT-IR analysis (Fig. 2 and 3). The solid and dashed lines in Fig. 3b represent the calculated effluent concentrations of Ca and SO4, assuming that the influent solution (1 mM CaCl2) is equilibrated with gypsum at pH 3 and 25°C. This comparison and the independence of the Ca and SO4 concentrations on flow rate suggest that the effluent solutions were close to equilibrium with respect to gypsum. After about 30 pore volumes, the amount of gypsum left in the column started to limit the dissolution rate and the concentrations of Ca and SO4 dropped. The broad front observed for Ca and SO4 at the higher flow velocity may also be explained by kinetic limitations when only small amounts of gypsum are left in the soil column. If we assume that the gypsum present is completely dissolved, the total amount of gypsum can be estimated from the leaching curves as about 95 g kg–1, which is in reasonable agreement with our previous estimate based on total elemental analysis (105 g kg–1).

Figure 3c and 3d show the pe and pH values of the effluent solutions, respectively. The effluent pH was below pH 2 at the beginning of the leaching experiments and increased to pH 3.3 during the first 15 pore volumes. Afterwards, the effluent pH remained constant for the remaining leaching period of 50 pore volumes. The effluent pH value did not depend on flow velocity. The pe values were near 13 at the beginning and decreased to values near pe 11.5 within the first 15 pore volumes. These pe values are indicative of oxic conditions, which was expected because we made no attempt to exclude atmospheric O2 from the column influent.

Predicting the dissolved CN concentrations in the column effluent by thermodynamic calculations is extremely unreliable, due to several reasons: (i) The thermodynamic solubility constant of the Prussian blue phase present in the soil is not exactly known. For compounds with the same unit formula Fe4(Fe(CN)6)3, log K values of –84.5 (Meeussen et al., 1992b) and –138.11 (Ghosh et al., 1999) have been reported in the literature; (ii) Regardless of the correct log K value of Prussian blue, its solubility strongly depends on solution pH and redox potential. The effluent pH value can be measured with sufficient accuracy, but pe measurements with a platinum electrode are at best semi-quantitative estimates. This can also lead to large errors; (iii) The concentration of CN in equilibrium with Prussian blue strongly depends on the activity of free Fe3+ in solution, which is coupled to the dissolution of Fe minerals in the soil. However, the crystallinity and solubility of iron oxide minerals present in the soil are not known and difficult to quantify; (iv) The concentration of CN in the column effluent may also be influenced by mobilization of adsorbed Fe64– or Fe63– anions, but sorption equilibrium constants for iron–cyanide complexes in strongly acidic soils are lacking; (v) The dissolution kinetics of Prussian blue may be slow and therefore far from equilibrium during a flow-through column experiment. The kinetic rates of Prussian blue dissolution in the presence and absence of citrate are not known. For all these reasons, we were not able to predict the CN concentrations in the column effluents. Later we will show some simple calculations for comparison, which may help to rationalize the possible release mechanisms of CN in the presence of citrate.

Influence of Citrate on Cyanide Leaching
Figure 4 shows the results of a leaching experiment with 1 mM citrate in the influent solution. During the first 71 pore volumes, the soil–sand mixture was leached with a 1 mM CaCl2 solution without citrate at a flow rate of 1 mL min–1. The effluent concentrations of Ca, SO4, and CN are comparable to the previous experiment presented in Fig. 3. After 71 pore volumes, the CaCl2 concentration in the influent was decreased to 0.1 mM and the flow rate to 0.2 mL min–1, which had no effect on the CN concentration in the effluent. After 73 pore volumes, we added 1 mM citrate to the influent solution. This citrate addition also had no significant effect on total or dissolved CN concentration in the effluent (Fig. 4a). After about 100 pore volumes, the dissolved Fe concentration increased and pe values decreased. This can be explained by the retarded breakthrough of citrate, probably as FeIII–citrate complexes formed during ligand controlled dissolution of HFO and/or Prussian blue. After 161 pore volumes, the flow was interrupted for 24 h and then resumed using the same influent solution. Also, this stopped-flow did not result a detectable increase in the effluent concentration of total or dissolved CN.



View larger version (21K):
[in this window]
[in a new window]
  		Fig. 4. Total and dissolved CN, SO4, Ca, pH, and pe in the effluent from a soil column leached with a sequence of influent solutions: (i) 100 mM CaCl2 solution at 1 mL min–1, (ii) 0.1 mM CaCl2 solution at 0.2 mL min–1, and (iii) 0.1 mM CaCl2 solution containing 1 mM citric acid at 0.2 mL min–1. After 161 pore volumes, the flow was interrupted for 24 h and then resumed using the same influent and flow rate.

 
Although concentrations of citrate larger than 1 mM rarely occur in soil solutions (Strobel, 2001), we increased the citrate concentrations to 10 or 100 mM to explore possible effects under extreme conditions. High concentrations of citrate may for example be used for in-situ soil remediation treatments. Figure 5 shows the results of a leaching experiment in which we added 10 mM citrate to the influent solution. The first 92 pore volumes of the experiment were again comparable to the experiments without citrate, except that the column was leached with 100 mM CaCl2 solution, leading to a lower SO4 concentration in equilibrium with gypsum. The addition of 10 mM citrate to the influent at 92 pore volumes resulted in a large peak in the concentrations of total and dissolved CN. Dissolved CN exhibited a maximum concentration of 30 µM, while total CN reached 77 µM. The difference, interpreted as colloidal CN, accounted for 47 µM at the peak maximum. Total and dissolved concentrations of CN dropped to values around 5 µM within about 40 pore volumes after the peak maximum. Note that the increase in CN concentrations was retarded by a factor of 7, since it appeared in the effluent after 99 pore volumes. Coinciding with the CN peak, the total Fe concentration increased to about 3 mM and then decreased to reach a plateau near 1 mM. At 187 and 236 pore volumes, respectively, the flow was interrupted for 24 h and then resumed using the same influent solution. These stopped-flow events resulted in small peaks in CN and SO4 concentrations, but the effects were not very pronounced. Unfortunately, we could not measure Fe, pH, and pe in all effluent samples due to insufficient sample volumes remaining after CN analyses.



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 5. Total and dissolved CN, SO4, Ca, pH, and pe in the effluent from a soil column leached with a sequence of influent solutions: (i) 100 mM CaCl2 solution at 1 mL min–1, (ii) 0.1 mM CaCl2 solution at 0.2 mL min–1, and (iii) 0.1 mM CaCl2 solution containing 10 mM citric acid at 0.2 mL min–1. After 187 and 236 pore volumes, the flow was interrupted for 24 h and then resumed using the same influent and flow rate.

 
The results of a leaching experiment with 100 mM citrate in the influent solution are presented in Fig. 6 . Again, the initial phase corresponded well with results of the previous experiment. At 78 pore volumes, 100 mM of citrate were added to the influent solution. This resulted in large peaks in the concentrations of total CN, dissolved CN, SO4, and Fe. Following these sharp peaks, the concentrations of dissolved CN and Fe reached stable plateau values near 16 µM CN and 0.6 mM Fe, respectively. The measured concentrations of colloidal CN fluctuated somewhat following the peaks, but they stabilized at a low value after 130 pore volumes. Coinciding with the sharp peak, the pe value dropped to about 6, indicating more reducing conditions due to the presence of citrate. Furthermore, in this experiment we interrupted the flow for 24 h after 182, 236, and 255 pore volumes, respectively. Each stopped-flow was followed by a sharp peak in the concentrations of colloidal and dissolved CN, SO4, and Fe. Following these peaks, which lasted only over 1 to 2 pore volumes, the effluent concentrations dropped to the respective plateau values before the flow interruptions. This clearly indicated that slow dissolution reactions led to an increase in dissolved CN and Fe concentrations in the presence of citrate. Kinetic effects were also observed by stopped-flow treatments in the previous two column experiments, but the magnitude of the resulting concentration peaks increased strongly with increasing citrate concentration. This indicates that the dissolution rate of Prussian blue at low citrate concentration may be too slow to produce pronounced peaks within the 24 h flow interruption.



View larger version (23K):
[in this window]
[in a new window]
  		Fig. 6. Total and dissolved CN, SO4, Ca, pH, and pe in the effluent from a soil column leached with a sequence of influent solutions: (i) 100 mM CaCl2 solution at 1 mL min–1, (ii) 0.1 mM CaCl2 solution at 0.2 mL min–1, and (iii) 0.1 mM CaCl2 solution containing 100 mM citric acid at 0.2 mL min–1. After 182, 236, and 255 pore volumes, the flow was interrupted for 24 h and then resumed using the same influent and flow rate.

 
Let us now consider possible mechanisms leading to CN mobilization by citrate and their possible relevance in contaminated soils. Thermodynamic predictions of dissolved CN concentrations are highly uncertain, for several reasons discussed above. However, simple equilibrium calculations may help to rationalize and constrain the possible processes leading to CN mobilization. The results of two calculations, based on the solubility constant for Prussian blue determined by Meeussen et al. (1992b), are summarized in Table 2. In the first calculation, we equilibrated a 10–4 M CaCl2 solution with Prussian blue in the presence of 0, 1, 10, or 100 mM citrate at pH 3.3. Dissolved oxygen was not considered in the calculations. The results reveal that the CN concentrations observed in the experiments without citrate (Fig. 3a) are about one order of magnitude larger than those predicted by the equilibrium calculation with Prussian blue. This discrepancy may be due to slow dissolution kinetics of Prussian blue, namely, the equilibrium concentration of CN may not be reached during a flow-through column experiment, and the variation in flow rate applied not sufficient to produce a significant difference in effluent concentration at levels close to the lower detection limit. Adding citrate to Prussian blue leads to increased equilibrium concentrations of CN and Fe, which is due to complexation of FeIII by citrate. However, comparison with the experiments presented in Fig. 4 to 6 shows that the predicted increase in CN concentration is much larger, while the increase in Fe concentration is smaller than observed in the experiments. In the second calculation, we added HFO (log K = –38.46, for Fe(OH)3(s) -> Fe3+ + 3OH–) to the system and again compared the results of equilibrium calculations (Table 2). Since HFO maintains a relatively high activity of free Fe3+ in solution at pH 3.3, the dissolved concentration of CN is lower than in the absence of HFO, but still higher than observed in the experiments (Fig. 3a). Adding citrate to the system now leads to drastic increases in dissolved Fe concentration, but only moderate increases in dissolved CN concentration. The observed increases in dissolved Fe were much smaller, and the increases in dissolved CN were much larger than predicted by the calculation. This example demonstrates the sensitivity of predicted CN concentrations on the solubility of Fe in the system considered. In soils, the solubility of the Fe minerals may be lower than that of HFO, leading to an intermediate behavior between the two systems calculated. However, in addition to dissolution, adsorption of Fe63– and Fe64– anions to surfaces of soil minerals may also play an important role (Rennert and Mansfeldt, 2001a, 2001b; Theis et al., 1988). Organic acids such as citrate may effectively compete for binding sites and thereby mobilize adsorbed cyanide complexes. Adsorption competition between cyanides and citrate may be responsible for the pronounced peaks in the CN leaching experiments. Dissolution of a mineral phase would be expected to yield a plateau concentration, which was nicely observed for gypsum dissolution. Adsorption of Fe63– and Fe64– in such an extremely acidic and contaminated substrate has not yet been studied, and should be further investigated to improve predictions of CN leaching from soils developed from coking plant purifier waste.


View this table:
[in this window]
[in a new window]
  		Table 2. Calculated pe values and concentrations of total dissolved Fe and iron–cyanide complexes (reported as CN) in a 10–4 M CaCl2 solution at pH 3.3 in equilibrium with either Prussian blue or Prussian blue and hydrous ferric oxide (HFO). The solubility constants were log K = –84.5 for Prussian blue (Meeussen et al., 1992b) and log K = –38.46 for HFO (Lindsay, 1979), respectively.

 
Finally, we will briefly discuss the environmental relevance of low molecular weight carboxylic acids for CN leaching under field conditions. Pore waters in soils often contain between 5 and 100 mM dissolved organic carbon (DOC), which is a complex mixture of biogenic compounds, fulvic acids, and other products of organic matter decomposition. Up to 10% of the DOC in soil solutions may be comprised of aliphatic low molecular weight carboxylic acids (Strobel, 2001). Aliphatic monocarboxylic acids (formic, acetic, propionic, butyric, valeric, lactic) are usually most abundant, while di- and tricarboxylic acids (oxalic, malonic, malic, succinic, tartaric, citric) occur in smaller concentrations (Strobel, 2001). Soils under forest and permanent pasture usually have higher concentrations of low molecular weight carboxylic acids in the soil solution compared with cultivated agricultural soils. Here, we used citrate as an example for low molecular weight carboxylic acids, which are released by plant roots and microorganisms. For citric acid, bulk soil solution concentrations are typically below 1 mM, but local concentrations such as in the rhizosphere of plants may be much higher. The citric acid concentrations used in this study were 1, 10, and 100 mM, that is, much higher than expected in soils under field conditions. Therefore, we conclude that cyanide mobility in the soil studied is extremely low even in the presence of plants exuding low molecular weight organic acids. The soil pH remained extremely acidic for more than 250 pore volumes, indicating that the soil is well buffered at around pH 3.3. Moreover, the soil contains large amounts of elemental S, which will slowly oxidize to form H2SO4 under well-drained conditions, thus keeping the pH extremely low. Liming of such soils should be avoided, because a pH increase would result in much more drastic increases in the solubility of Prussian blue and therefore increased leaching of cyanides into groundwater.


   ACKNOWLEDGMENTS
 
We are grateful to Heidi Biernath and Gerlind Wilde (Ruhr-Universität Bochum) for technical assistance in the laboratory. Financial support was provided by Deutsche Steinkohle AG, Herne, Germany.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 




This article has been cited by other articles:


Home page
 Vadose Zone JHome page
J. Simunek, C. He, L. Pang, and S. A. Bradford
Colloid-Facilitated Solute Transport in Variably Saturated Porous Media: Numerical Model and Experimental Verification
Vadose Zone J., August 24, 2006; 5(3): 1035 - 1047.
[Abstract] [Full Text] [PDF]

Home page
 Vadose Zone JHome page
L. W. de Jonge, C. Kjaergaard, and P. Moldrup
Colloids and Colloid-Facilitated Transport of Contaminants in Soils: An Introduction
Vadose Zone J., May 1, 2004; 3(2): 321 - 325.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Mansfeldt, T.
Right arrow Articles by Kretzschmar, R.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Mansfeldt, T.
Right arrow Articles by Kretzschmar, R.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Mansfeldt, T.
Right arrow Articles by Kretzschmar, R.
Related Collections
Right arrow Laboratory Column Studies
Right arrow Other Contaminants
Right arrow Soil Pollution

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome