Article pubs.acs.org/est
Subsurface Interactions of Fe(II) with Humic Acid or Landfill Leachate Do Not Control Subsequent Iron(III) (Hydr)oxide Production at the Surface Alison Jackson, John W. Gaffney, and Stephen Boult* School of Earth, Atmospheric and Environmental Sciences, University of Manchester, Manchester M13 9PL, United Kingdom S Supporting Information *
ABSTRACT: At least 93% of Fe(II) remained free, as defined by ferrozine assay under anoxic conditions in the presence of humic acid (HA) and two simulated landfill leachates of different maturities. However, tangential flow ultrafiltration showed a weaker but more extensive interaction of Fe with organic carbon (OC); 90% of Fe associated with the less mature leachate. Despite the existence of this weak interaction under anoxic conditions, there was no difference in iron(III) (hydr)oxide production whether HA was added prior to or coincident with the oxidation of Fe(II) on exposure to oxic conditions. Under oxic conditions ferrozine showed that more Fe(II) bound to OC, up to 50% to HA. However, this occurs via oxidation of Fe(II) to Fe(III), which is bound and then thermally reduced. This affinity for Fe(III) and the ability to carry out thermal reduction both increase with the maturity of the OC. The rate at which ferrozine-defined free Fe(II) was lost on exposure to dissolved oxygen was also enhanced by the more mature OC, while it was slowed by acetogenic leachate. The slowing must be a consequence of the filtration-defined Fe(II)/OC interaction.
■
studied by Liang et al.,1 who showed that variation in the relative concentrations of O2, Fe(II), and NOM caused changes in Fe oxidation. However, Liang et al.1 did not investigate the interaction of Fe(II) and NOM in the complete absence of oxygen nor the subsequent exposure of the initially anoxic systems to O2-saturated waters, both of which are likely occurrences in natural systems. This work will, therefore, investigate iron(III) (hydr)oxide production after Fe(II) first interacts with NOM in the absence of O2 and on its subsequent exposure to O2-saturated waters, while also extending the range of NOM type and concentration included in the simulations. Because of the variety of types of NOM, the simulation of aquatic systems requires the selection of representative classes; in previous work typical aquagenic and pedogenic NOM was chosen.4 Given the intention in this work to explicitly simulate Fe interaction in the subsurface, aquagenic NOM was not included. However, in addition to humic acid (HA) as a representative of ubiquitous subsurface NOM, the range was extended by including anthropogenic organic matter (AOM) in the form of a landfill leachate (LL). Because of the widespread use of landfill for putrescible waste disposal, LL is probably the most common type of AOM that
INTRODUCTION Iron(III) (hydr)oxides in natural waters sorb and coprecipitate ions that would otherwise be soluble. Therefore, the prediction of the solubility, mobility, and bioavailability of several toxic metals relies on either measurement or prediction of the concentration of iron(III) (hydr)oxide. Small particles of iron(III) (hydr)oxide with high specific surface area (SSA) will be most important in sorption of toxic metals; the most important source of such material is likely to be from precipitation of Fe(II) rather than from mechanical erosion. There is, therefore, a requirement to understand how the transformation from iron(II) to iron(III) (hydr)oxide varies in natural waters. Natural organic matter (NOM) has been shown to exert a significant control on the oxidation rate of Fe(II)1,2 and observed to be important in maintaining iron(III) (hydr)oxides as colloidal suspensions in natural waters;3 there is, therefore, a requirement to measure the effects of further variation in both the type and concentration of NOM. In natural waters the relevant variables include not only the type and concentration of NOM but also the order in which it mixes with Fe(II) and O2. In most simulations of Fe(II) oxidation in natural waters free Fe(II) ions have been introduced into oxic waters containing NOM of differing compositions. However, it is possible and perhaps more likely that Fe(II) will interact with NOM in subsurface waters prior to exposure to fully oxic waters. Interaction at low dissolved O2 (DO) saturation was © 2012 American Chemical Society
Received: Revised: Accepted: Published: 7543
March 20, 2012 June 12, 2012 June 19, 2012 June 19, 2012 dx.doi.org/10.1021/es301084c | Environ. Sci. Technol. 2012, 46, 7543−7550
Environmental Science & Technology
Article
changing. At this point, while free Fe(II) was determined by color development 30 s after 5 mL of sample was mixed with 5 mL of deoxygenated ferrozine, bound Fe(II) was also determined as the difference in readings between that after 30 s and that after storage of the mixture under anoxic conditions for 24 h (see ref 4 for details). Oxic Interaction of Fe and OC. Subsequent to the anoxic phase, the reaction was monitored as each mixture was exposed to air to simulate passage across the oxycline. Air, passed through a 2 M NaOH solution (Fisher Chemicals) to remove CO2, was continually bubbled through the mixtures at a flow rate of 0.1 L/min for the first 3 h of the reaction. Subsequent gas exchange with atmospheric O2 was via a 10 mm opening containing a molecular filter (Fisher Chemicals) which removed CO2; the DO probe showed 100% saturation throughout this period. During the reaction the pH was maintained at 6.2 by autotitration (TIM 857 radiometer) with 0.01 M NaOH (Fisher Chemicals) and the base additions were recorded. The exclusion of CO2 allowed acid production to be used as a proxy for iron oxy(hydr)oxide formation, this being the only source of H+ production within the simulations.9 The Fe(II) concentration was determined by ferrozine assay at regular intervals throughout the reaction until the reaction was complete. The end point was taken as the point when the Fe(II) concentration had fallen to zero and/or acid production had ceased. At this point the concentrations of both free and bound Fe(II) were determined (see ref 4 for further details of the method). Particle Size Distribution (PSD) of the Leachate. Only the PSD of the leachate that had developed over the shorter period (AL) was analyzed before and after oxidation; it was not possible to extract sufficient NAL for analysis, and the PSD of an Fe/HA mixture had been analyzed previously4 and was not repeated. Two 120 mL samples of leachate were extracted by syringe and immediately filtered through a 5 μm cellulose nitrate filter membrane (Whatman, United Kingdom). Both were sequentially filtered through pairs of 0.2 μm, 50 and 10 kDa tangential flow ultrafiltration (TFU) membranes (Vivaflow 50, Sartorius, Germany). For one sample the procedure was done in an anoxic cabinet, and for the other filtration the procedure was done after the sample was allowed to oxidize in the vessel in the manner described above. The protocol used for size separation in this work followed that of ref 4. Once separation was complete, and within 48 h, the OC concentration was determined in each size fraction using a total organic carbon analyzer (Shimadzu 5050A, Japan). Butyric, propionic, and acetic acid concentrations were also determined on each size fraction below 0.2 μm using ion chromatography (Dionex, United Kingdom). All samples were then acidified using concentrated HNO3 to pH < 2, and the Fe concentration was determined within 10 days by inductively coupled plasma optical emission spectroscopy (ICP-OES) (Optima 5300, Perkin-Elmer, United Kingdom). A set of known standards were run every five samples with drift assumed to be linear and, therefore, corrected by interpolation. The pre-drift-corrected error on all standards was below 5%.
contaminates the subsurface; furthermore, landfill is also often a source of Fe and toxic metals. 5−7 Consequently, the determination of the rate of development of iron(III) (hydr)oxide in the zone within and immediately surrounding a waste emplacement, or “near field”, is of direct relevance to many sites where the dispersal of toxic metals, including radionuclides, is modeled. While HA is known to consist of varying collections of compounds that differ in structural detail, the utility of using a particular type in investigations of generic behavior has been proven;4 this approach was also adopted here. However, there are no such suitable types of LL organic carbon (OC); consequently, a representative synthetic LL was produced and characterized. Subsequent objectives were to determine the degree to which HA and LL control the production of iron(III) (hydr)oxide by monitoring the consecutive processes of their anaerobic interaction with Fe(II) and oxidation of Fe(II).
■
EXPERIMENTAL SECTION Production of Simulated Landfill Leachate. Two 20 L polycarbonate carboys were filled with synthetic groundwater, organic waste, and Fe turnings and sampled over 1800 days. The headspace gas and leachate were analyzed, the latter by ion chromatography (IC) (Dionex, United Kingdom) and excitation emission matrix spectroscopy (EEM), a “fingerprinting” technique used previously to allow comparisons of these complicated mixtures.8 Details of the production, sampling, and analysis are given in the Supporting Information. At the time of their use, one leachate contained acetate (AL) while the other did not (NAL); further explanation is given in the Results and Discussion. Anoxic Interaction of Fe and OC. HA (Acros, United Kingdom) was added to synthetic river water (SRW) (see ref 4 for the composition and preparation) to make a 500 mg L−1 stock solution; this was passed through a glass fiber filter with a 1 μm cutoff (GFA, Whatman, United Kingdom) to remove any large particles and its OC concentration measured (see below). SRWwith and without addition of HA from the filtered stock solutionwas placed inside a custom-fabricated gastight container with several sealable ports. The solution/mixture was deoxygenated by bubbling with N2 at a flow rate of 0.5 L/ minute while being continually stirred. A DO electrode installed in the container allowed anoxia to be detected, and a pH electrode indicated, when the pH stabilized, that all CO2 had been purged. Once anoxic, the N2 flow rate was reduced to 0.1 L/min, causing an internal pressure of 10 kDa) in both simulated leachates is much higher than that normally found in MWLLs,7 and furthermore, this high proportion is indicative of maturity rather than the immaturity12 suggested by the fluorescence measurements (the PSD of Fe is discussed in a later section). These differences in PSD between MWLL and simulated leachate are likely to result from differences in the methods of sample extraction. Samples from landfills were only taken after discarding fluid pumped slowly from a borehole for a long period until a constant minimum turbidity was achieved, this being done to ensure no large particles were mobilized;5,7 the material sampled was only that mobilized by a very low flow and able to pass through the pore spaces of the landfill. However, samples from the simulations were withdrawn quickly, their size being limited only by the 3 mm i.d. of the sample tube; consequently, the proportion of the dissolved OC ( 0) phases of laboratory simulations. Also shown is H+ production during the oxic phase (t > 0) for each system: OC absent, HA, NAL, AL. Gray triangles in the “OC absent” graph represent dissolved oxygen plotted as percent saturation.
shows the same but identifies the proportion of the bound Fe that is Fe(II). Figure 2 shows the changing proportions throughout the anoxic and subsequent oxic phase of free Fe(II), iron(III) (hydr)oxides, and, by subtraction, Fe bound to OC. In Figure 2 the acid production data have been converted to represent the proportion of Fe present as iron(III) (hydr)oxides, and the dotted line on each graph shows their sum with free Fe(II), i.e., (Fe(II) + FeOx). When OC was absent, Fe(II) + FeOx was equal to 100% throughout the oxidation; the loss of free Fe(II) was matched almost exactly throughout the 75 h of reaction by production of iron(III) (hydr)oxides, and no other form of Fe was present. In the presence of OC whenever Fe(II) + FeOx is not equal to 100% some Fe must be present in another form such as bound to OC. Anoxic Interaction of Fe(II) and OC. In anoxic water, when Fe was introduced as Fe(II) at the Fe/OC ratio of 0.14 used here, only 7% of Fe became bound to HA, only 3% to NAL, and none to AL (Figure 1). This shows both that the tendency of leachates to become more like HA as they mature is reflected in their abilities to bind Fe(II) and, furthermore, all types of OC bind only a small proportion of Fe(II) under anoxic conditions. The latter point is of fundamental significance, suggesting that, at concentrations of Fe and OC
similar to those studied here, OC can only have a minor influence on the behavior of Fe when it is in the subsurface. Predicting Iron(III) (Hydr)oxide Production. The aim of this work was, however, to determine the role of OC in controlling production of iron(III) (hydr)oxide from Fe(II), a transformation confined to oxic waters. In attempting to predict the proportions of Fe that either become bound to OC or form iron(III) (hydr)oxides when subsurface waters emerge into oxic waters, it may be possible to disregard the subsurface history of the water, as the amount of subsurface interaction of Fe and OC is only small. That this is the case is supported by the overwhelmingly greater amounts of Fe that become bound to OC under oxic rather than anoxic conditions (Figure 1b): binding to HA increases from 7% to 50%, that to NAL from 3% to 21%, and that to AL from 0% to 36%. Furthermore, in our previous studies,4 interaction of Fe(II) and HA initiated and maintained in oxic conditions, rather than anoxic followed by oxic, resulted in partitioning of Fe that while not identical because it was carried out at pH 6.5 and not 6.2, is little different from that of this study. Consequently, knowledge of the Fe/OC ratio throughout what may be an untraceable and inaccessible subsurface pathway is relatively unimportant compared to this knowledge in the accessible waters at the oxycline. Therefore, in surface water, the relative proportions of 7546
dx.doi.org/10.1021/es301084c | Environ. Sci. Technol. 2012, 46, 7543−7550
Environmental Science & Technology
Article
iron(III) (hydr)oxides and Fe bound to OC should be predictable from the Fe/OC ratio in these relatively easy to sample waters. This was observed to be the case in previous work in HA-rich natural waters4 but could then only be explained as a consequence of the possible similarity of subsurface and streamwater Fe/OC ratios. Although the lack of affinity of Fe(II) for these types of OC observed in these experiments means prediction of Fe speciation in oxic waters may not be complicated by prior subsurface interaction, it also raises questions about the mechanism by which OC controls the behavior of Fe both here and in much previous work. Such control has previously been demonstrated by changes in the reaction rate1,2,17−19 and also more recently, in our work, by the speciation of Fe at the end point of reaction.4 These questions concerning the mechanism of reaction are dealt with next before consideration of the influence of OC on reaction rates under the experimental conditions of this study. Initial Mechanism of Fe(II)/OC Interaction. It is generally agreed that the mechanism of oxidation of Fe(II) to Fe(III) is a combination of its reaction with either of two oxidants, DO or H2O2,17 oxidation by the former releasing superoxide and the latter a hydroxyl radical capable of oxidation of further Fe(II). If OC is present, it may form complexes with Fe(II) and thereby a competing pathway for oxidation of Fe(II), via oxidation of the complexed instead of the free Fe(II). The overall rate has been shown to be both increased1,2,17 and decreased18,19 by OC, indicating that the Fe(II) complex can oxidize faster or slower than free Fe(II). In our experiments, however, the apparent lack of interaction between Fe(II) and OC calls the existence of such a complex into question. This contradiction may be resolved on consideration that whether Fe(II) is free or complexed is only operationally defined by ferrozine assay. Ferrozine is the most common chelating agent used for Fe(II) determination in natural waters20 whether in laboratory, mesocosm, or field samples; it is preferred because OC is less likely to interfere with the analysis than when using other methods.21 It is a simple test used on extracted aliquots but also the analytical method at the heart of sophisticated flow injection analysis.21 However, it has been noted20 that ferrozine may itself be responsible for freeing Fe(II) from weak interactions with OC, and in this study, comparison of PSD analysis with the ferrozine measurements shows that a weak interaction of such type does occur. While ferrozine assay revealed little interaction between Fe(II) and OC under anoxic conditions, TFU of the anoxic waters containing AL (the only experimental system analyzed by TFU) recovered only 10% of the notionally free Fe(II) in the “dissolved”, 5 μm within 1 h. However, in the presence of AL in anoxic conditions, as has already been commented on above, TFU reveals an Fe(II)/OC interaction that occurs to the extent that 90% of the Fe was not dissolved. Furthermore, once oxidized, only 33% of Fe became >5 μm, 45% remaining in 95% of Fe formed iron(III) (hydr)oxides, all 7548
dx.doi.org/10.1021/es301084c | Environ. Sci. Technol. 2012, 46, 7543−7550
Environmental Science & Technology
Article
Notes
While there are similarities in the interaction of HA and NAL with Fe, AL interaction is quite different (Figure 2). The loss of ferrozine-defined free Fe(II) is slower than in the absence of OC, the free Fe(II) concentration remaining above 5% for 150 h rather than 50 h. In these and previous studies the acceleration of Fe(II) oxidation in the presence of OC can be understood to be a result of the interaction between the free Fe(II) and OC. However, the role of OC in slowing the loss of free Fe(II) is apparently paradoxical as it must cause Fe(II) to remain free, something which cannot result from complexation with OC. Given that, as demonstrated in these experiments, the OC is not itself exerting an oxygen demand, it is the extensive filtration-defined interaction of Fe(II) with OC that explains the paradox. The filtration-defined interaction of Fe(II) with AL must be able to protect Fe(II) from oxidation. The loss appears to be first order and for the first 50 h is almost balanced by iron(III) (hydr)oxide formation, showing that while the interaction with AL prevents oxidation of Fe(II), it cannot prevent Fe(III) from hydrolyzing during this period. However, unlike when HA and NAL were present, the rate of binding increases over time as the rate of loss of free Fe(II) decreases, resulting in 36% being bound at the end point largely as Fe(III) with only 6% shown to be Fe(II) (Figure 1). As stated above, for HA, comparison to previous work is possible to show that the progress of the reaction with HA and its end point are little altered by the inclusion of an initial anoxic mixing period, but no such work is available using leachates. Environmental Significance. In anoxic conditions the proportion of Fe(II) that interacts with OC is very small when determined by ferrozine, although a filtration-defined interaction can be much more extensive. Ferrozine shows much more Fe(II) binds to OC under oxic conditions, but this occurs via oxidation of Fe(II) to Fe(III), which is bound and then thermally reduced. This affinity for Fe(III) and the ability to carry out thermal reduction both increase with the maturity of the OC. The rate at which ferrozine-defined free Fe(II) is lost on exposure to DO is also enhanced by the more mature OC, while it is slowed by AL. All three types of OC could, therefore, have significant effects on iron(III) (hydr)oxide production and mobility in natural waters, although through different processes. In the case of Fe interaction with mature OC, that which occurs in the subsurfacethe interactions with Fe(II)is likely to be insignificant in the eventual production of iron(III) (hydr)oxide. However, for an immature OC such as AL this remains to be confirmed.
■
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was funded by an Engineering and Physical Sciences Research Council studentship to A.J. with additional support from Intelisys Ltd.
■
(1) Liang, L.; McNabb, J. A.; Paulk, J. M.; Gu, B.; McCarthy, J. F. Kinetics of Fe(II) oxygenation at low partial pressure of oxygen in the presence of natural organic matter. Environ. Sci. Technol. 1993, 27 (9), 1864−1870. (2) Emmenegger, L.; King, D. W.; Sigg, L.; Sulzberger, B. Oxidation kinetics of Fe(II) in a eutrophic Swiss lake. Environ. Sci. Technol. 1998, 32 (19), 2990−2996. (3) Pédrot, M.; Le Boudec, A.; Davranche, M.; Dia, A.; Henin, O. How does organic matter constrain the nature, size and availability of Fe nanoparticles for biological reduction? J. Colloid Interface Sci. 2011, 359 (1), 75−85. (4) Gaffney, J. W.; White, K. N.; Boult, S. Oxidation state and size of Fe controlled by organic matter in natural waters. Environ. Sci. Technol. 2008, 42 (10), 3575−3581. (5) Gounaris, V.; Anderson, P. R.; Holsen, T. M. Characteristics and environmental significance of colloids in landfill leachate. Environ. Sci. Technol. 1993, 27 (7), 1381−1387. (6) Vilks, P.; Caron, F.; Haas, M. K. Potential for the formation and migration of colloidal material from a near-surface waste disposal site. Appl. Geochem. 1998, 13 (1), 31−42. (7) Jensen, D. L.; Christensen, T. H. Colloidal and dissolved metals in leachates from four Danish landfills. Water Res. 1999, 33 (9), 2139− 2147. (8) Baker, A.; Curry, M. Fluorescence of leachates from three contrasting landfills. Water Res. 2004, 38 (10), 2605−2613. (9) Carlson, L.; Schwertmann, U. The effect of CO2 and oxidation rate on the formation of goethite versus lepidocrocite from an Fe(II) system at pH-6 and pH-7. Clay Miner. 1990, 25 (1), 65−71. (10) Kang, K. H.; Shin, H. S.; Park, H. Characterization of humic substances present in landfill leachates with different landfill ages and its implications. Water Res. 2002, 36 (16), 4023−4032. (11) Gourdon, R.; Comel, C.; Vermande, P.; Veron, J. Fractionation of the organic matter of a landfill leachate before and after aerobic or anaerobic biological treatment. Water Res. 1989, 23 (2), 167−173. (12) Li, R.; Yue, D.; Liu, J.; Nie, Y. Size fractionation of organic matter and heavy metals in raw and treated leachate. Waste Manage. 2009, 29 (9), 2527−2533. (13) Davison, W.; Seed, G. The kinetics of the oxidation of ferrous iron in synthetic and natural waters. Geochim. Cosmochim. Acta 1983, 47 (1), 67−79. (14) Pham, A. N.; Waite, T. D. Oxygenation of Fe(II) in natural waters revisited: Kinetic modeling approaches, rate constant estimation and the importance of various reaction pathways. Geochim. Cosmochim. Acta 2008, 72 (15), 3616−3630. (15) Santana-Casiano, J. M.; González-Dávila, M.; Rodríguez, Ma. J.; Millero, F. J. The effect of organic compounds in the oxidation kinetics of Fe(II). Mar. Chem. 2000, 70 (1−3), 211−222. (16) Millero, F. J.; Sotolongo, S.; Izaguirre, M. The oxidation kinetics of Fe(II) in seawater. Geochim. Cosmochim. Acta 1987, 51 (4), 793− 801. (17) Pullin, M. J.; Cabaniss, S. E. The effects of pH, ionic strength, and iron-fulvic acid interactions on the kinetics of nonphotochemical iron transformations. I. Iron(II) oxidation and iron(III) colloid formation. Geochim. Cosmochim. Acta 2003, 67 (21), 4067−4077. (18) Miles, C. J.; Brezonik, P. L. Oxygen consumption in humiccolored waters by a photochemical ferrous−ferric catalytic cycle. Environ. Sci. Technol. 1981, 15 (9), 1089−1095.
ASSOCIATED CONTENT
S Supporting Information *
Description of the production and sampling of LL and chemical characterization of the leachate and figure showing the change in free Fe(II) during the oxic phase of laboratory simulations in the presence of HA, ML, and IL. This material is available free of charge via the Internet at http://pubs.acs.org/ .
■
REFERENCES
AUTHOR INFORMATION
Corresponding Author
*Phone: (+44) 161-275-3867; fax: (+44) 161-275-3947; email:
[email protected]. 7549
dx.doi.org/10.1021/es301084c | Environ. Sci. Technol. 2012, 46, 7543−7550
Environmental Science & Technology
Article
(19) Theis, T. L.; Singer, P. C. Complexation of iron(II) by organic matter and its effect on iron(II) oxygenation. Environ. Sci. Technol. 1974, 8 (6), 569−573. (20) Pehkonen, S. Determination of the oxidation states of iron in natural waters. A review. Analyst 1995, 120 (11), 2655−2663. (21) Pullin, M. J.; Cabaniss, S. E. Colorimetric flow-injection analysis of dissolved iron in high DOC waters. Water Res. 2001, 35 (2), 363− 372. (22) Pullin, M. J.; Cabaniss, S. E. The effects of pH, ionic strength, and iron-fulvic acid interactions on the kinetics of nonphotochemical iron transformations. II. The kinetics of thermal reduction. Geochim. Cosmochim. Acta 2003, 67 (21), 4079−4089. (23) Munter, R.; Overbeck, P.; Sutt, J. Which is the best oxidant for complexed iron removal from groundwater: The Kogalym case. Ozone: Sci. Eng. 2008, 30 (1), 73−80.
7550
dx.doi.org/10.1021/es301084c | Environ. Sci. Technol. 2012, 46, 7543−7550
