SSSAJ Grow Your Career with SSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 2 June 2005
Published in Soil Sci Soc Am J 69:996-1008 (2005)
DOI: 10.2136/sssaj2004.0287
© 2005 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
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 Google Scholar
Google Scholar
Right arrow Articles by Essington, M. E.
Right arrow Articles by Holden, W. L.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Essington, M. E.
Right arrow Articles by Holden, W. L.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Essington, M. E.
Right arrow Articles by Holden, W. L.
Related Collections
Right arrow Soil Mineralogy
Right arrow Geochemical Processes
Right arrow Soil Chemistry

Soil Chemistry

Gibbsite and Goethite Solubility

The Influence of 2-Ketogluconate and Citrate

Michael E. Essington*, Julia B. Nelson and William L. Holden

Biosystems Engineering & Environmental Science Dep., Environmental and Soil Sciences Group, 2506 E.J. Chapman Dr., The Univ. of Tennessee, Knoxville, TN 37996-4531

* Corresponding author (messington{at}utk.edu.)


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
2-Ketogluconate (kG) is a microbial byproduct that has been isolated from the rhizosphere of several plants and has been found to accumulate around bacteria adhering to rock surfaces. It has long been suspected that kG may play a role in enhancing the bioavailability of soil nutrients. However, quantitative information relevant to its behavior in soils, its reactions with mineral surfaces, and its aqueous complexation of metal cations is unavailable. The objectives of this research were to examine the influence of kG on gibbsite and goethite solubility and to describe these effects by considering the aqueous complexation of Al3+ and Fe3+ by kG. Secondary to this was to compare the impact of kG on mineral solubility to that of citrate. The equilibrium solubility of gibbsite and goethite was examined at ambient (20–22°C) or controlled (25°C) temperatures as a function of pH, ionic strength, and kG or citrate concentration. Both ligands were observed to significantly enhance mineral solubility (as determined by the total soluble concentrations of Al and Fe), with kG and citrate having similar impacts on gibbsite solubility, but with kG having a lesser impact than citrate on goethite solubility. The solubility data were employed to characterize the aqueous complexation chemistry of Al– and Fe(III)–kG and citrate. In the 4 < pH < 8.5 range, the mononuclear species AlkG2+(aq), Al(OH)2kG0(aq), and Al(OH)3kG–(aq) describe Al–kG speciation, with the latter two species mechanistically described by the bidentate complexation of AlOH2+(aq) and Al+2 [also written as AlOH(H–1kG)0(aq) and Al(OH)2(H–1kG)–(aq)]. Iron(III)–kG complexation is described by the FekG2+(aq) and Fe(OH)3 kG22– species. For the citrate systems, the chemical models required to describe metal speciation differed from those reported in the literature: Alcit0(aq), Al(OH)2cit2–(aq), and Al(OH)3 cit3–(aq) for Al–citrate speciation and Fecit0(aq), Fe(OH)2 cit2–(aq), and Fe(OH)3cit3–(aq) for Fe(III)–citrate.

Abbreviations: AlT, total soluble molar concentration of Al • cit, citrate • kG, 2-Ketogluconate • LMMOA, low-molecular-mass organic acid • MDL, method detection limit


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
LOW-MOLECULAR-MASS organic acids (LMMOAs) comprise one of the more transitory, but ever-present, components of the soil solution pool of organic C. Their occurrence is primarily localized in the microenvironment surrounding soil microbes and in the rhizosphere, because of the intense biological activity in this volume of soil. A wide variety of LMMOAs are known to exist in the soil solution. They range in character from formic acid to complex, multi-ring phenolic acids. The soil solution is dominated by the more soluble, aliphatic acid anions (e.g., formate, acetate, lactate, oxalate, malonate, malate, succinate, and citrate), which are found in the <1 to 1100 µM concentration range (Jones, 1998; Strobel, 2001; van Hees et al., 2000). Due to their innate ability to complex, and in some instances chelate metal cations, LMMOAs can influence the bioavailability and transport of metals. Organic acids may also form specific complexes with mineral surface functional groups, affecting surface charge characteristics (Yao and Yeh., 1996), compete with other specifically retained substances for surface sites (Geelhoed et al., 1998; Grafe et al., 2002; Kafkafi et al., 1988; Wijnja and Schulthess, 2000), and inhibit mineral crystallization, directly impacting mineral precipitation and dissolution (Huang and Violante, 1986; Jardine and Zelazny, 1995; Lebron and Suarez, 1999). Indeed, many organic acid anions in soils (principally the di- and tricarboxylates) reside almost completely in the adsorbed phase (Jones, 1998; Strobel, 2001).

Both plant roots and soil microbes are known to exude greater concentrations of organic acid anions (primarily citrate and malate) when subjected to Fe, P, and K (and possibly Ca and Zn) deficiencies, and as a mechanism to mitigate Al toxicity effects (Hocking, 2001; Inskeep and Silvertooth, 1988; Jones et al., 2003). Among the LMMOA anions that have been identified in the rhizosphere soil solution and in microbial exudates is kG (C6H9O7) (Duff et al., 1963; Moghimi et al., 1978; Vance et al., 1995). 2-Ketogluconate is a microbial byproduct of glucose oxidation. It has been isolated from the rhizosphere of several plants, including wheat (Triticum aestivum L.), corn (Zea mays L.), and peas (Pisum sativum L.). It has also been suggested that kG may accumulate around bacteria adhering to mineral surfaces, thereby facilitating phosphate solubilization (Chiyonobu et al., 1973; Duff et al., 1963; Erlich, 1981; Halder and Chakrabartty, 1993; Klasen et al., 1992; Kucey et al., 1989; Neijssel and Tempest, 1975; Sokatch, 1969; Webley and Duff, 1965). Neijssel and Tempest (1975), using Klebsiella aerogenes, noted a rapid rate of glucose uptake and a lower mass conversion of glucose to bacterial mass when P was limiting. The accelerated rate of glucose metabolism resulted in the increased production of kG and thereby an increased solubilization of P. Indeed, it has long been suspected that kG may play an important role in enhancing P bioavailability; however, it is the generally high concentrations (10 mM) of kG that were required to extract even small amounts of P from alkaline soils (Holden, 1996; Moghimi et al., 1978) that led researchers to disregard the potential importance of this compound in the soil environment (Jones, 1998).

In solution, kG may exist in three different structural configurations (assuming a structural analogy to gluconic acid): five-member ring, linear, and six-member ring. In its linear form (Fig. 1) , kG is comprised of a carboxyl group in the first position, a carbonyl group in the second position, and alcohol groups on the remainder of the carbon atoms. The carboxylic acid moiety is a relatively strong weak acid, with a pKa1 of 3.00 ± 0.06 (Nelson and Essington, 2005). A second acidic functional group, most likely the alcohol moiety in the third position (ortho to the carbonyl), has an acidity described by a pKa2 of 11.97 ± 0.41. Therefore, in ‘typical’ soil solutions, the monovalent anion is predicted to predominate, although a dissociable alcohol moiety may be involved in aqueous metal complexation and surface ligand exchange reactions in acidic environments (again, assuming a structural analogy to gluconic acid) (Motekaitis and Martell, 1984).



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 1. The structure of 2-keto-D-gluconate.

 
The solubility-enhancing characteristics of kG were first deduced from studies involving the solubilization of calcium phosphates in alkaline soils. Duff and Webley (1959) suggested that kG could be a strong complexing agent for Ca2+, promoting the dissolution of calcium phosphates. Webley et al. (1963) suggested that microbes might enhance the dissolution of aluminosilicates due to the production of extracellular complexing substances. Subsequently, Duff et al. (1963) showed that kG forms soluble complexes with divalent metal cations (Ca, Mg, Mn, Zn, Sr, and Ni). However, Moghimi and Tate (1978) were unable to validate the formation of a Ca2+–kG complex using a potentiometric acid-base titration method. Berrow et al. (1982) demonstrated that a 0.1-M kG solution was able to extract more Co, Cu, Ni, Zn, Fe, Ti, and V from soil than a 1-M ammonium acetate solution and more Fe, Mo, and Ti than a 0.43-M acetic acid solution. They concluded kG was complexing a portion of the non-exchangeable forms of these metals in soil. Although apparently a more efficient extractant than acetate, a 0.1-M kG solution was two- to ten-times less effective than a 0.05-M EDTA or a 0.05-M DTPA solution in extracting soil transition metals (Berrow et al., 1982).

The elevated dissolution of calcium phosphates observed in the presence of kG was postulated by Moghimi and Tate (1978) to be promoted by proton attack rather than by Ca2+ complexation. They noted kG to be one of the stronger monobasic carboxylic acids, as indicated by the pKa1 value reported above. The Lowry-Brönsted acidity, coupled with the perceived inability of kG to complex Ca2+ led to the conclusions of Moghimi and Tate (1978); although the insensitivity of the methodology employed in their evaluation was not considered. More recent potentiometric titration studies using a Ca2+ ion selective electrode, indicate that kG indeed forms an outer-sphere complex, CakG+(aq), with an ion association constant of log Kf = 1.74 (Nelson and Essington, 2005).

The ligand-enhanced dissolution of soil minerals in the rhizosphere is an important mechanism used by soil biota (plant roots and microbes) to increase the bioavailability of nutrients (e.g., Fe and P) and to mitigate the toxic effects of Al. If kG is an important rhizosphere component, playing a role in soil nutrient cycling, it must be capable of affecting the enhanced dissolution of soil minerals. The objective of the current study was to investigate the influence of kG on the equilibrium solubility of gibbsite [Al(OH)3(s)] and goethite [FeOOH(s)], and to describe these effects by quantifying the aqueous complexation of Al and Fe by kG. Both minerals are scavengers for and control the soil solution P composition in neutral to acidic soils, and they are reservoirs for a nutrient (Fe) and a potential toxin (Al). Presumably, their dissolution would render P more bioavailable. In addition, the enhanced solubility of these accessory minerals in the presence of kG would indicate the formation of stable Al and Fe aqueous complexes with kG; processes that detoxify Al and increase the bioavailability of Fe.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Preparation of Solids
Hydrated alumina (C31) was obtained from the Aluminum Company of America (ALCOA, Pittsburgh, PA). X-ray diffraction indicated C31 to be composed of synthetic gibbsite, without detectable impurities. Samples of C31 gibbsite were prewashed with 0.1 M HCl to remove any poorly crystalline Al(OH)3 coatings (Bloom and Weaver, 1982). Following the 14-d cleansing period, with constant agitation, the samples were centrifuge-washed with the KCl background electrolyte used in the equilibrium solubility experiments. The centrifuge washings were conducted until the suspensions were pH-neutral. A second hydrated alumina (SF-4) was obtained from Alcan Chemicals (Beachwood, OH). This material, also consisting of monoclinic gibbsite without impurities (by x-ray diffraction), was treated with CO2–free, 0.01 M NaOH for 30 min to remove poorly crystalline Al(OH)3 (Sarkar et al., 1999). Following the base treatment, the gibbsite was centrifuge-washed with NaNO3 background electrolyte (used in a second set of solubility studies) until a pH-neutral suspension was obtained. Synthetic goethite (as verified by x-ray diffraction) was obtained from Kraemer Pigments (New York, NY). Poorly crystalline materials were removed by reaction with 0.4 M HCl for 30 min (Schwertmann and Cornell, 2000). The cleansed goethite was repeatedly washed with the NaNO3 background electrolyte to be used in the solubility studies, and then brought to neutral pH with a small volume of 1 M NaOH.

Mineral Solubility
Two sets of gibbsite solubility experiments were conducted. The first set of studies was conducted using a 100 g L–1 solid/solution ratio, ionic strengths of 0.05 or 0.1 M KCl, kG concentrations of approximately 0, 0.005, or 0.01 M, and equilibrium was approached from both undersaturated and supersaturated initial conditions. 2-Keto-D-gluconate was obtained as a hemicalcium dihydrate salt (Ca0.5C6H9O7 · 2H2O, 98% purity) and used without further purification. Prewashed samples of C31 gibbsite (10 g) were placed in 250-mL polypropylene bottles. For the undersaturated initial condition experiments, 100 mL of background electrolyte/kG solution was added to each bottle. For the supersaturated initial condition experiments, 100 mL of the background electrolyte/kG solution containing 0.002 M AlCl3 was added to each bottle. Solution pH (approximately 3, 4, 5, and 6) was adjusted with 0.1 M HCl or 0.1 M KOH. Solution pH was monitored three more times, within a 2-wk period, and adjusted accordingly to ensure that the desired pH existed in solution. The sample bottles were capped and placed in a temperature-controlled water bath maintained at 25 ± 0.1°C. The samples were equilibrated, with periodic shaking, for a 6-wk period. After equilibration, the solid and equilibrating solution phases were separated by centrifugation at 750 x g for 10 min. Aqueous-phase pH was determined with a combination pH electrode and the aqueous extracts were analyzed for Al, Ca, and K by ICP–AES using a Model 61 Thermo-Jarrell Ash (Franklin, MA). Standard solutions were prepared and verified against USEPA standard reference solutions. The prepared standards, over the duration of the study, were within +2 and +4% for Ca, +1 and +5% for Al, and +0.8 and +2% for K of the standard reference solution values. The aqueous extracts were also analyzed for Cl– using a Dionex (Sunnyvale, CA) DX-100 ion chromatograph. A standard calibration curve was established and verified against USEPA standard reference solutions. Over the duration of the study, the prepared standards were within ±1% of the standard reference values. Total C concentrations were determined using a Dohrmann (Mason, OH) DC80 total organic C analyzer. Total C values obtained were assumed to represent kG concentrations in solution, an assumption that was later confirmed by ion chromatography (described below).

A second set of gibbsite solubility studies, and a single set of goethite solubility studies, were conducted using a 100-g L–1 solid-to-solution ratio for gibbsite and 25 g L–1 for goethite, background electrolyte compositions of 0.01, 0.04, or 0.4 M NaNO3, and kG concentrations of approximately 0, 0.003, or 0.035 M or citrate concentrations of 0 and 0.05 M. Solutions prepared from the 2-keto-D-gluconate hemicalcium dihydrate salt were passed through Na-saturated Dowex HCR-W2 cation exchange resin (Supelco, Bellefonte, PA) to generate NakG salt solutions. Sodium citrate solutions were prepared from ACS certified Na3C6H5O7 · 2H2O without further purification. Prewashed samples of either SF-4 gibbsite or synthetic goethite (0.5–0.6 g) were placed in 50-mL polypropylene tubes, to which was added 20 mL (for gibbsite) or 15 mL (for goethite) of the background electrolyte/kG solution. Solution pH (ranging from 4 to 8) was adjusted weekly under N2 gas with 0.1 M HCl or 0.1 M NaOH for the duration of the equilibration period. The samples were equilibrated at ambient temperature (20–22°C), with periodic shaking, until solution pH did not vary from one week to the next. For gibbsite, the equilibration period was 12 wk. For goethite, the equilibration period was 8 wk. Gibbsite and goethite solubility studies in the presence of citrate were similarly performed, but with the samples being equilibrated for 8 wk in the dark.

After equilibration, the solid and equilibrating solution phases were separated by centrifugation at 750 x g for 10 min, followed by pressure filtration through a 0.45-µm nylon membrane filter. Aqueous-phase pH was determined with a combination pH electrode and the aqueous extracts were analyzed for Al and Fe using a Spectro (Fitchburg, MA) CIROS ICP–AES. Nitrate concentrations were determined by ion chromatography (described above). Sample replicates and standard checks were within ±4% for all elements. Ketogluconate and citrate concentrations were verified from total C analyses (Dohrmann DC80 total organic C analyzer). A Dionex DX-100 ion chromatograph equipped with a 25-µL injection loop, an AS4A analytical column, an AG4A guard column, suppressed conductivity detection, a 1.8 mM Na2CO3/1.7 mM NaHCO3 eluent, and a flow rate of 2 mL min–1 was also used to verify kG concentrations. The method detection limit (MDL) for kG was approximately 0.05 µM, and sample replicates and standard checks were within ±10%.

Data Analysis
The data analysis was designed and employed to extract a maximum amount of information from the solubility data (Essington, 1990). Specifically, it was anticipated that kG and citrate would enhance the total dissolution of gibbsite and goethite via a chelation effect. Consider, for example, the equilibrium solubility of gibbsite. The total soluble molar concentration of Al (denoted as AlT) in a conservative background electrolyte containing a complexing ligand, Ll–, is described according to the mass balance relationship:

[1]
where n is the number of OH– ions in the hydrolysis products, a, b, and c are stoichiometric coefficients and q is charge on the complex ion AlaHbLqc (aq) (q = 3 – a + b – lc), the brackets denote molar concentrations, and it is assumed that only the monomeric Al species are of significance. The formation of each of the Al hydrolysis products may be described by the generalized reaction:

[2]
For each value of n one can write an expression for the conditional hydrolysis constant (cK1n):

[3]
Each cK1n value is related to the true thermodynamic equilibrium constant for the hydrolysis reaction (K1n) according to the relationship:

[4]
where K1n is described by Eq. [3], but with parentheses (denoting thermodynamic activity) replacing brackets, and where the {gamma} values are the single ion activity coefficients (a function of ion charge and ionic strength if the Davies equation is employed). Rearranging Eq. [3] leads to an expression that describes the concentration of the hydrolysis product:

[5]
The formation of the complex is described by the reaction:

[6]
with

[7]
and

[8]
where cKf is the conditional formation constant. Substituting Eq. [5] and [8] into Eq. [1] yields:

[9]
Simplifying,

[10]
The dissolution of gibbsite and the associated conditional solubility product constant (cKsp, gibbsite) can be described by

[11]

[12]
which assumes equilibrium and unit activity of liquid water and the solid phase. Rearranging Eq. [12] and substituting for [Al3+] in Eq. [10] yields:

[13]
or

[14]
Equation [14] states that the total concentration of Al in a solution in equilibrium with gibbsite is determined by the innate thermodynamic stability of gibbsite (cKsp, gibbsite, a known value that was validated in this study), the magnitude of the hydrolysis constants (cK1n, known values), the solution pH (controlled or measured property), the ionic strength of the solution (controlled property), the aqueous speciation of the complexing ligand (the proportion of LT that exists as Ll–), and the magnitude of the complex ion formation constants (cKf).

An expression similar to Eq. [14], but for LT, can also be developed. The total molar concentration of L, in a solution containing Al3+, is given by:

[15]
The formation of each HjLj–l(aq) is described by

[16]
where

[17]
Substituting Eq. [8] and [17] into Eq. [15] yields:

[18]
Rearranging and substituting Eq. [12] for [Al3+],

[19]

Equations [14] and [19] are coupled nonlinear algebraic expressions that may be solved for [Ll–] and values of cKf when the concentration of AlT controlled by gibbsite in the presence of a complexing ligand is greater than that observed in the absence of the ligand. To solve for the unknown variable in Eq. [14] and [19], a numerical algorithm is required, as is a chemical model that identifies the nature of the Al3+ and Ll– interactions. A chemical model is a set of chemical reactions that, when employed, will predict a total soluble metal concentration that matches the analytically determined values. The computer code FITEQL (Herbelin and Westall, 1996) was employed to optimize the chemical model dependent cKf values using AlT or FeT, kGT or citT, pH, cKsp, gibbsite or cKsp, goethite, cK1n values for Al3+ or Fe3+ hydrolysis, and cKj1 values for kG– or cit3– protonation. The known equilibrium constants for mineral solubility and ion speciation common to all chemical models employed in this study are given in Table 1. These constants were converted to conditional constants using the Davies equation or the Setchénow equation (for neutral species and a salting-out coefficient of 0.1 mol kg–1), as illustrated in Eq. [4]. For any given system (mineral, ionic environment, ligand type, and concentration), it was anticipated that the chemical model that resulted in the smallest FITEQL-generated numerical value of WSOS/DF (weighted sum of squares divided by the degrees of freedom) would be the one selected to define the complexation chemistry. In general, WSOS/DF values between 0.1 and 20 represent a reasonably good fit of the chemical model to the experimental data. FITEQL also generates a standard deviation for each adjustable parameter (cKf values). These error values are a function of the goodness-of-fit of the chemical model, as well as the standard deviation in the experimental data (defined as the analytical error).


View this table:
[in this window]
[in a new window]
  		Table 1. Chemical reactions and associated equilibrium constants (25°C) used in modeling the influence of 2-ketogluconate (kG) and citrate (cit) on the solubility's of gibbsite and goethite.

 
The conditional formation constants (cKf values) generated by FITEQL for Al3+ and Fe3+ complexation were converted to the zero ionic strength equilibrium constants (Kf values) using the Davies equation for charged species and the Setchénow equation for neutral species. The statistical error (standard deviation) associated with each mean log Kf value arose from two sources: experimental error and the goodness-of-fit of the FITEQL-predicted chemical model. The experimental error was assigned to the standard deviation of the mean log Kf value (averaged over ionic strength and kG concentration); whereas, the ability of a chemical model to describe the experimental data was embodied in a FITEQL-generated standard deviation on the predicted log Kf values. The reported standard deviation on a computed mean log Kf value is the greater of the two error sources. The log Kf values were then input to the GEOCHEM-PC (Parker et al., 1995) thermodynamic database. The chemical characteristics of the equilibrium solutions in contact with either gibbsite or goethite were input to GEOCHEM-PC, and the speciation model was then used to compute the activities of Al3+ and Fe3+ (gibbsite and goethite were not allowed to precipitate during the modeling exercise). Activity diagrams were then created to illustrate log (Al3+) or log (Fe3+) as a function of pH, and to further evaluate the chemical speciation models predicted by FITEQL to characterize aqueous speciation in the gibbsite and goethite systems.


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
The Peculiar Nature of the Citrate Systems
Mineral dissolution in the presence of citrate was investigated with the objective to confirm the solubility product constants of gibbsite and goethite. Using gibbsite as an example, the IAP (a measured property of the equilibrating solution) is computed as:

[20]
where IAP is numerically equal to Ksp, at equilibrium. Although this approach was not strictly necessary for the gibbsite systems, as AlT concentrations in KCl and NaNO3 solutions were well above the MDL and easily determined, FeT concentrations supported by goethite in NaNO3 solutions were less than the MDL. Thus, a mechanism was required to enhance FeT concentrations in the equilibrium systems, while allowing for the determination of Ksp, goethite. The chelation effect of citrate is evidenced by the enhanced solubility of gibbsite and goethite (Fig. 2) . Central to this mechanism of evaluating the equilibrium solubility of these minerals is the presumption that the metal complexation chemistry of cit3–(aq) is well-known. Unfortunately, this is not the case. Indeed, there is very little commonality amongst the various studies that examine Al–citrate and Fe(III)–citrate complexation (e.g., ratios of metal to citrate concentrations and ionic environment), and the chemical models used to describe the complexation chemistry (Table 2) (Königsberger et al., 2000; Pierre and Gautier-Luneau, 2000; Harris et al., 2003). For example, Motekaitis and Martell (1984) employed a chemical model consisting of the Alcit(aq)0, AlHcit+(aq), and AlH–1cit–(aq) species to describe their potentiometric titration data; whereas, Öhman and Sjöberg (1983) used Alcit0(aq), AlHcit+(aq), Alcit3–2, and Al34cit4–3, and Öhman (1988) used Alcit0(aq), AlHcit+(aq), Alcit3–2, Al34cit4–3, AlH–1cit–(aq), AlOH(H–1cit)2–(aq), and Al3(OH)4 7–3. The numerous Fe(III)–citrate complexation models were reviewed by Königsberger et al. (2000), and include the models of Timberlake (1964) [Fecit0(aq) and Fe22–2], Ramamoorthy and Manning (1973) , and their own model [Fecit0(aq), Fe(H–1cit)–(aq), Fecit3–2, Fe(Hcit)(cit)2–(aq), and Fe(H–1cit)(cit)4–(aq)]. Much of the disparity among the various chemical models has been attributed to the slow kinetics of metal–citrate systems (in all but strongly acidic systems) and that numerous chemical models can be employed to explain experimental data when the protonation of citrate is the dominant chemical process (when the citrate concentration greatly exceeds that of the metal) (Öhman, 1988; Harris et al., 2003).



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 2. Influence of citrate (cit) on the solubility of (a) gibbsite and (b) goethite at ambient temperature (20–22°C). The lines represent the equilibrium solubility of gibbsite and goethite in a 0.05 M NaNO3 solution at 25°C generated using Eq. [14] and Table 1 data.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Aluminum–citrate and Fe(III)–citrate binding constants (log Kf at 25°C) used in GEOCHEM-PC to speciate the equilibrium solutions in contact with gibbsite (Fig. 3a) and goethite (Fig. 3b). Selected chemical models and associated binding constants (log cKf) from the literature and obtained from potentiometric titrations are also shown.{dagger}

 


View larger version (21K):
[in this window]
[in a new window]
  		Fig. 3. Activity diagrams illustrating the activities of Al3+(aq) and Fe3+(aq) as a function of pH relative to the stability of (a) gibbsite and (b) goethite at ambient temperature (20–22°C). Metal cation activities were computed using GEOCHEM-PC and the metal–citrate (cit) binding constants listed in Table 2 (closed symbols, T2) or Table 3 (open symbols, T3). The solid lines represent the equilibrium solubility of gibbsite and goethite at 25°C (Table 1).

 
The gibbsite and goethite solubility data were evaluated using GEOCHEM-PC and the Al–citrate and Fe(III)–citrate binding constants listed in Table 2. The computed Al3+(aq) and Fe3+(aq) activities were then used to construct the activity diagrams in Fig. 3 . The diagrams indicate that the solutions were highly undersaturated with respect to gibbsite and undersaturated to slightly supersaturated (at neutral pH values) with respect to goethite. The diagrams also illustrate that the relative saturation of the solutions with respect to the minerals varied with pH. Saturation indices (SI = log IAP–log Ksp) varied from approximately –5 to –2 for gibbsite and –2 to +0.7 for goethite. If all conditions for a thermodynamic evaluation of a system are met (such as, equilibrium conditions, macrocrystallinity and purity of the solid, analytical confidence, appropriate complexation reactions and associated binding constants), then the computed metal ion activities (as a function of pH) must fall on or only deviate to a minor extent from the mineral stability lines (fall within SI = –1 to +0.1) (Essington, 2003). In this case, there exists considerable ambiguity in the scientific literature concerning the detailed speciation Al–citrate and Fe(III)–citrate as a function of pH and other solution properties. Therefore, the most plausible explanation for the disparity between the gibbsite and goethite Ksp and IAP values is that the Al–citrate and Fe(III)–citrate complexation chemistry and associated binding constants (listed in Table 2) are not adequate to describe these systems.


View this table:
[in this window]
[in a new window]
  		Table 3. Al–citrate (cit) and Fe(III)–citrate chemical models and associated binding constants optimized by FITEQL to describe the equilibrium solubility of gibbsite and goethite at ambient temperature (20–22°C).

 
Several published Al–citrate and Fe(III)–citrate binding models (with associated log cKf values) were evaluated using the FITEQL code, the analytical AlT and FeT concentration data, and the imposed condition of solution equilibrium with gibbsite or goethite. Attempts to optimize the cKf values for the chemical binding models reported in the literature (Table 2) were unsuccessful. The chemical models, as a whole or in part, could not be optimized (FITEQL would not converge). Instead, an independent series of models involving Al–citrate and Fe(III)–citrate species were developed and tested, with the assumed complexes having the general form, xycit3x–y–3zz. This general formula conceals the highly complicated metal complexation chemistry of metal citrates. Citrate has four potential metal binding sites, three carboxyls and an alcohol group, which are also involved in proton binding. In addition, the metals can be in various stages of hydrolysis. For example, the Al-citrate species described in Table 2 as AlOHcit–(aq), AlOHcit4–2, Al22cit2–2, and Al34cit4–3 are actually Al(H–1cit)–(aq), Al(H–1cit)(cit)4–(aq), Al22–2, and Al3(H–1cit)3(OH)–4 (where H–1cit4– represents the citrate ligand with all four binding sites dissociated). The chemical models that generated the lowest WSOS/DF values are shown in Table 3 and consist of the monomeric Al–citrate complexes, Alcit0(aq), Al(OH)2cit2–(aq), and Al(OH)3cit3–(aq); and the Fe(III)–citrate complexes, Fecit0(aq), Fe(OH)2cit2–(aq), and Fe(OH)3cit3–(aq). Di- and trimeric species were included in the various chemical models, but FITEQL would not converge in those instances. This result (the absence of di- and trimeric species) is consistent with that of Harris et al. (2003). They observed that monomeric Al–citrate complexes were predominant when AlT concentrations were exceedingly small relative to citrate (10 µM AlT to 1 mM citrate), a condition that is generally present in the current study. Further, the log Kf values generated from the different ionic strength systems are comparable, an indication that the chemical models are valid. As expected, the inclusion of the generated log Kf values in GEOCHEM-PC for the geochemical analysis of the solubility data results in Al3+(aq) and Fe3+(aq) activities that very closely follow the gibbsite and goethite stability lines as a function of pH (–0.28 < SI < 0.36 for gibbsite; –0.09 < SI < 0.36 for goethite) (Fig. 3).

The chemical models developed in the present treatment of the solubility data differ from those generated in previous studies (as obtained through potentiometric titration). Öhman and Sjöberg (1983), Öhman (1988), Motekaitis and Martell (1984), and Harris et al. (2003) consistently invoke the mononuclear species Alcit0(aq), AlH–1cit–(aq), AlHcit+(aq) to describe Al-citrate complexation (Table 2). With the exception of Motekaitis and Martell (1984), these authors also resolve various multinuclear species, most commonly Al3OH4–3 (aq), which is predicted to dominate AlT in the pH 5 to 8 range. In addition, the application of various spectroscopic techniques (1H-, 27Al-, and 13C-NMR, differential UV, and electrospray mass spectrometry) have led to the conclusion that multinuclear Al–citrate complexes, having Al-to-citrate stiochiometries of 1:2, 2:2, 2:3, and 3:3, may exist in solutions (Karlik et al., 1983; Matzapetakis et al., 1999; Bodor et al., 2002; Harris et al., 2003). These direct findings are commonly invoked to confirm titration-based chemical models. However, as indicated by Harris et al. (2003), the ratio of citT to AlT in solution impacts the formation of multinuclear species. Low citT to AlT ratios (1:1 to 4:1) are common to titration and spectroscopic studies that resolve multinuclear species. Using a citT to AlT ratio of 100:1 and differential UV spectroscopy, Harris et al. (2003) could not detect multinuclear species, reasoning that slow reaction kinetics restrict their formation. Although not constant, the citT to AlT ratios controlled by gibbsite dissolution (approximately 5:1 at pH 4 to 50:1 at pH > 6) were apparently not conducive to the resolution of multinuclear complexes in the pH 5 to 8 range.

As surmised by Pierre and Gautier-Luneau (2000) and based on their review of the pertinent literature, the solution chemistry of Fe(III)–citrate systems is not obvious. Like studies of Al–citrate complexation, investigations of Fe(III)–citrate systems are typically constrained to low citT to FeT ratios (<13:1), which are quite different to those controlled by goethite dissolution (approximately 500:3 to 5000:1). The potentiometric titration results of Field et al. (1974), which are considered to be the most reliable for pH < 4 solutions (Martin, 1986; Pierre and Gautier-Luneau, 2000), are described by invoking Fe(III)–citrate complexes having 1:1 Fe-to-citrate stoichiometric ratios [Fecit0(aq), FeHcit+(aq), and FeOHcit–(aq)] (Table 2). In pH 6 to 7 solutions, the Fe(Hcit)23– species has been proposed by analogy to Al–citrate complexation chemistry (Martin, 1986); however, the existence of this species has not been confirmed.

The difference between our log Kf values for the formation of Alcit0(aq) and Fecit0(aq) (6.53 log Kf units) is greater than literature values (approximately 4 log Kf units) (Tables 2 and 3). This log Kf difference is also greater than that between Al3+ and Fe3+ complexes with other carboxylate ligands, which general ranges between 1 and 5 log Kf units (May and Murray, 2000). This is an inconsistency that we can only attribute to the species included in the various chemical models, and supports the argument that species and associated formation constants from disparate chemical models should not be mixed when employed to predict aqueous speciation.

The Complexation Chemistry of 2-Ketogluconate
The principal objective of this study was to investigate the impact of kG on the solubility of gibbsite and goethite. As illustrated in Fig. 4 and 5 , kG indeed has a significant impact on the solubility of both gibbsite and goethite. Further, the solubilities of gibbsite and goethite (i.e., the concentrations of AlT and FeT) increase in a manner that is directly related to kG concentration, and this enhanced solubility effect is consistent over a broad pH range. The influence of kG on gibbsite solubility was similar to that of citrate (also shown in Fig. 4), indicating that kG may be as effective as citrate in forming stable aqueous Al–kG species. Although significant, the influence of kG on goethite solubility enhancement is less than that of citrate, particularly at pH values of 6 and less.



View larger version (23K):
[in this window]
[in a new window]
  		Fig. 4. Influence of 2-ketogluconate (kG) and citrate (cit) on the solubility of gibbsite at ambient temperature (20–22°C; SF-4) or 25°C (C31). The line represents the equilibrium solubility of gibbsite in a 0.05 M NaNO3 solution at 25°C generated using Eq. [14] and Table 1 data.

 


View larger version (26K):
[in this window]
[in a new window]
  		Fig. 5. Influence of 2-ketogluconate (kG) and citrate (cit) on the solubility of goethite at ambient temperature (20–22°C). The line represents the equilibrium solubility of goethite in a 0.05 M NaNO3 solution at 25°C generated using Eq. [14] and Table 1 data.

 
The solubility enhancement data were employed to elucidate the Al–kG and Fe(III)–kG complexation chemistry in a manner similar to that described above for the citrate systems. Chemical models involving the species, xykG3x–y–zz, were developed and evaluated using FITEQL under the assumed conditions that Al3+ and Fe3+ activities are controlled by the Ksp values of gibbsite and goethite. In the gibbsite systems, FITEQL convergence was attained for the SF-4 gibbsite systems when the chemical model considered the formation of Al(OH)2kG0(aq) and Al(OH)3kG–(aq) (Table 4). Additionally, the AlkG2+(aq) species was resolved in the C31 gibbsite systems. As discussed above for citrate, the actual complexation mechanism cannot be ascertained from the solubility data, although the AlkG2+(aq) species is probably outersphere (based on the magnitude of log Kf = 3.31). For example, Al(OH)2kG0 (aq) may be a monodentate species generated via the reaction:

[21]
which involves the complexation of Al+2 by the carboxyl moiety (Fig. 6a) . However, Al(OH)2kG0(aq) may be a bidentate species generated via the reaction:

[22]
which involves the complexation of AlOH2+(aq) by kG carboxyl and dissociated alcohol moieties to form a six-member ring structure (Fig. 6b). Given the strength of the complex {log Kf = –5.00 for Eq. [21] or [22], or log Kf = 12.07 for AlOH2+(aq) + H–1kG2–(aq) = AlOH(H–1kG)0(aq)}, it is very likely that the Al(OH)2kG0(aq) and Al(OH)3kG–(aq) [which also may be written as Al(OH)2(H–1kG)–(aq)] species are predominantly bidentate.


View this table:
[in this window]
[in a new window]
  		Table 4. Al–2-ketogluconate (kG) chemical models and associated binding constants optimized by FITEQL to describe the equilibrium solubility of gibbsite at ambient temperature (20–22°C; SF-4) and 25°C (C31).

 


View larger version (35K):
[in this window]
[in a new window]
  		Fig. 6. Possible structural configurations of Al–2-ketogluconate (kG) and Fe(III)–kG complexes: (a) monodentate and (b) bidentate.

 
The solubility of goethite could be described in all four systems by a chemical model consisting of the FekG2+(aq) and Fe3kG2–2 species, where the latter species is a diketogluconate. As noted for the bidentate Al–kG species, the Fe(III)–kG complexes are also thought to be inner-sphere. For example, the monodentate complex formed via the reaction, Fe3+(aq) + kG–(aq) = FekG2+(aq), is probably inner-sphere owing to the magnitude of the formation constant (log Kf = 9.78). The Fe3kG2–2 species may also be viewed as consisting of two ring structures that chelate FeOH2+(aq) {which may be written as FeOH2–2(aq)} (Fig. 6b).

The FITEQL-generated log Kf values for the various Al–kG and Fe(III)–kG species (Tables 4 and 5) were input to the GEOCHEM-PC thermodynamic datafile. The computer code was then employed to predict the activities of Al3+(aq) and Fe3+(aq) in the gibbsite and goethite equilibrium solutions. Because the solubility data were employed by FITEQL to establish the influence of kG on metal speciation, it is expected that the computed activities of Al3+(aq) and Fe3+(aq) will mirror the stabilities lines for gibbsite and goethite. This is indeed the case for the gibbsite and goethite systems (Fig. 7 and 8) , where SI ranges between –1.50 and 1.25 for gibbsite and –0.94 and 0.45 for goethite.


View this table:
[in this window]
[in a new window]
  		Table 5. Fe(III)–2-ketogluconate (kG) chemical model and associated binding constants optimized by FITEQL to describe the equilibrium solubility of goethite at ambient temperature (20–22°C).

 


View larger version (23K):
[in this window]
[in a new window]
  		Fig. 7. Activity diagram illustrating the activity of Al3+ as a function of pH at ambient temperature (20–22°C; SF-4) or 25°C (C31). The chemical model defined in Table 4 was used to describe Al–2-ketogluconate (kG) complexation. The solid lines represent the equilibrium solubility of gibbsite at 25°C (Table 1).

 


View larger version (25K):
[in this window]
[in a new window]
  		Fig. 8. Activity diagram illustrating the activity of Fe3+ as a function of pH at ambient temperature (20–22°C). The chemical model defined in Table 5 was used to describe Fe(III)– 2-ketogluconate (kG) complexation. The solid lines represent the equilibrium solubility of goethite at 25°C (Table 1).

 
The species distribution diagrams in Fig. 9 illustrate the significance of the Al–kG species, relative to free Al3+(aq) and its hydrolysis products, as a function of pH and the ratio of AlT to kGT. When AlT and kGT are present in equimolar concentrations (AlT = kGT = 10–3 M, Fig. 9a), the Al–kG species are predicted to predominate when solution pH values are greater than 4.5. Indeed, the Al(OH)2(H–1kG)–(aq) species predominates at all pH values above approximately 5. Below pH 4.5, Al3+(aq) is predicted to predominate. However, when the concentration of kGT exceeds that of AlT (AlT = 10–4 M, kGT = 10–3 M, Fig. 9b), only the Al–kG species are predicted to predominate, with the Al(OH)2(H–1kG)–(aq) species again predominating above pH 5. When FeT and kGT are present in equimolar concentrations (FeT = kGT = 10–3 M, Fig. 10a) , the FekG2+(aq) is predicted to predominate in acidic solutions (pH < 5.5). Above pH 5.5, the predicted concentrations of FeOH2–2 are equal to the sum of the Fe3+(aq) hydrolysis products. However, when the concentration of kGT exceeds that of FeT (FeT = 10–4 M, kGT = 10–3 M, Fig. 10b), the Fe(III)–kG species account for approximately 100% of FeT.



View larger version (30K):
[in this window]
[in a new window]
  		Fig. 9. Predicted species distribution for Al in solutions containing (a) 10–3 M AlT and 10–3 M kGT and (b) 10–4 M AlT and 10–3 M kGT in 0.01 M NaNO3.

 


View larger version (23K):
[in this window]
[in a new window]
  		Fig. 10. Predicted species distribution for Fe(III) in solutions containing (a) 10–3 M FeT and 10–3 M kGT, and (b) 10–4 M FeT and 10–3 M kGT in 0.01 M NaNO3.

 
The log Kf for the formation of AlkG2+(aq) is consistent with the log Kf values for Al3+ complexation with monocarboxylates. However, the log Kf for the formation of FekG2+(aq) is more consistent with the log Kf values for Fe3+ complexation with di- and tricarboxylates. As was noted for the citrate systems, the difference between the log Kf values for the formation of AlkG2+(aq) and FekG2+(aq) (6.50 log Kf units) is also greater than literature values for Al3+ and Fe3+ complexes with other carboxylate ligands (between 1 and 5 log Kf units). This large differential may be associated with the chemical model employed to describe Fe(III)–kG speciation (one 1:1 and one 1:2 species), relative to that invoked for Al–kG speciation (three 1:1 species).


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The chemical nature of the soil environmental is exceedingly complex, particularly in the rhizosphere where soil biota degrades organic detritus and exudes a myriad of organic substances. Detailed characterizations have identified numerous organic compounds in soils that may play an important role in the cycling of nutrients and in the detoxification of potentially toxic elements. It has long been suspected that kG may be one such compound, particularly in respect to P cycling where it enhances the solubility of metal phosphates. Our findings indicate the kG can significantly enhance the solubility of two common soil accessory minerals, gibbsite and goethite, which are responsible for a large portion of a soil's capacity to specifically retain metals and ligands. As a result of our equilibrium solubility studies, conducted as a function of pH, ionic strength, and kG concentration, we have developed chemical models that quantitatively describe the influence of kG on the aqueous speciation of Al and Fe. Specifically, the gibbsite solubility data were employed to generate formation constants for the AlkG2+(aq), Al(OH)2kG0(aq), and Al(OH)3kG–(aq) species. With the exception of the AlkG2+(aq) species, the magnitude of the association constants suggest that the complexes are bidentate, involving the inner-sphere complexation of AlOH2+(aq) and Al2+ via the dissociated carboxyl and alcohol moieties of kG. Therefore, the configuration of the bidentate species may be more appropriately expressed as AlOH(H–1)kG0(aq) and Al(OH)2(H–1)kG–(aq). The modeled speciation of Al (when kGT > AlT) indicates that AlkG2+(aq) will predominate in strongly acidic solutions (pH < 4.3), AlOH(H–1)kG0(aq) predominates in a narrow pH range (4.3 < pH < 5.1), and Al(OH)2(H–1)kG–(aq) predominates when solution pH is greater than 5.1.

The goethite solubility data were employed to generate formation constants for the FekG2+(aq) and Fe(OH)3 kG2–2 species. Again, based on the magnitude of the formation constants, both species probably involve the inner-sphere complexation of Fe3+(aq) (monodentate in FekG2+(aq)) or FeOH2+(aq) [bidentate in Fe(OH)3 kG2–2 or FeOH2–2]. The modeled speciation of Fe in solutions with kGT > FeT indicates that Fe–kG species will account for approximately 100% of FeT. The complexation chemistry of kG indicates that it may play an important role in biogeochemical soil processes. Strong complexation of Fe(III) and the resulting enhanced solubility of goethite may enhance bioavailability, and the strong complexation of Al may result in detoxification.

The solubilities of gibbsite and goethite in citrate solutions were used to derived Al–citrate and Fe(III)–citrate complexation models. As a result, formation constants for the Al–citrate species, Alcit0(aq), Al(OH)2cit2–(aq), and Al(OH)3cit3–(aq), and the Fe(III)–citrate species, Fecit0(aq), Fe(OH)2cit2–(aq), and Fe(OH)3cit3–(aq) were generated. The observation that the developed chemical models for the metal–citrate systems differ from those reported in the literature is a further indication that it is difficult to identify clearly the unique speciation of citrate using the available methodologies.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Support for this research by the USDA-NRI Competitive Grant Program (Project No. 2000-00537) is gratefully acknowledged.

Received for publication August 26, 2004.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES