Chemical Thermodynamics of Uranium in the Soil Environment Chemical Thermodynamics of Uranium in the Soil Environment

Uranium is present in the soil environment because of human activity, including the usage of U-bearing phosphorus fertilizers. In oxic and many suboxic soil environ-ments, U(VI) is the dominant uranium valence species. With pH, pe (Eh), the partial pressure of CO 2 , the mineralogy of the adsorbing surfaces and the uranium concentra- tion as the key master variables, U(VI) will predictably participate in hydrolysis, ion-pairing, complexation, ion-exchange, mineral precipitation and adsorption reactions. An extensive listing of thermochemical data is currently available for detailed simulations to assist with model setup, data interpretation and system understanding. In this chapter, simulations of U(VI) hydrolysis with variable pCO 2 activities, U(IV) and U(VI) precipitation, U(VI) reduction and U(VI) complexation with carbonate and phosphate assemblages illustrate the usefulness and applicability of simulations in data analysis and experimental design.


Properties, sources, characteristics of soil uranium
Uranium is the third element in the actinide series having an atomic number of 92 and an electronic configuration of [Rn] 5f 3 6d 1 7s 2 . The 5f orbitals are less effective in penetrating the inner core electrons than the 4f orbitals (lanthanide series), thus permitting more favored covalent bonding character [1]. Uranium(IV) and uranium(VI) have ionic radii of 89 and 73 pm, respectively. The two, more abundant, long-lived isotopes of uranium are 235 U 92 and 238 U 92 . The naturally occurring mass abundances of uranium isotopes are 234 U (0.0057%), 235 U (0.71%) and 238 U (99.284%). 235 U is fissile, whereas 238 U in a breeder reactor will yield fissile 239 Pu [2].

Introduction to soil uranium
The Earth's crustal uranium abundance is centered near 2.3 mg U/kg [1]. Soil parent materials vary substantially in their uranium concentrations, with granites (4.4 mg U/kg) and shales (3.8 mg U/kg) having greater abundances than basalts (0.8 mg U/kg) and K-feldspars (1.5 mg U/kg) [3]. The phyllosilicates, muscovite and biotite have U concentrations centering near 20 mg U/kg, and some zircon minerals may have up to 2500 mg U/kg [3]. Aide et al. [4] documented total uranium concentrations by soil horizon depth in numerous southeastern Missouri soils, noting that the uranium concentrations varied from 0.58 to 2.89 mg U/kg, with course-textured soils generally having smaller U concentrations. In their study, uranium in individual soil pedons was well correlated with Fe-oxyhydroxide concentrations. Birke et al. [5] reported that the amount of uranium in river waters in Germany varied from 0.007 to 43.7 μg U/L, with a median of 0.33 μg U/L. Mendez-Garcia et al. [6] observed that high uranium concentrations in sediment in the Rio Grande Basin in Mexico were of natural occurrence. . Uranium(V) species and associated minerals are comparatively rare because of disproportionation into U(IV) and U(VI) species [1].
Soils may become uranium impacted because of nuclear fuel production, nuclear weapons production, depleted uranium in munitions, coal combustion and most importantly by phosphorus fertilizer applications [7][8][9][10][11][12][13][14][15][16][17][18][19]. Stojanovic et al. [17] observed that maize and sunflower plants may be very useful for uranium phytoremediation, with the root mass acquiring much greater uranium accumulations than culms, leaves and grain. Stojanovic et al. [18] documented previous research showing that the use of phosphorus fertilizers may contribute 73% of the total anthropogenic uranium to the global soil resource. Echevarria et al. [20] observed that low pH levels favored increased uranium plant availability. Laroche et al. [21] in a hydroponic study observed that phosphorus reduced uranyl activity, especially at higher pH intervals.
Phyllosilicates (clay minerals) typically manifest a net negative charge density because of isomorphic substitution and unsatisfied edge charges [36][37][38]. Al-, Mn-and Fe-oxyhydroxides have variable charged surfaces (amphoteric) that acquire a positive charge density when the pH is more acidic than the mineral's point of zero net charge density [39,40]. Uranyl ions, along with its hydroxyl monomers and hydroxyl polymers, will participate in adsorption reactions with phyllosilicates and Mn-and Fe-oxyhydroxides [41][42][43][44][45][46][47][48][49]. The transport of U-bearing colloids by wind and water erosion is an important source of U transport from impacted sites.
There lies great interest in understanding the U transport in natural systems such as soil profiles, sediments and aquifers [4,9,10,19,40,[50][51][52]. Johnson et al. [51] investigated depleted uranium soil sites in Nevada (USA), observing that uranium retention is a function of (1) soil type, (2) soil binding site concentrations, (3) the presence of phyllosilicates and their associated Fe-oxyhydroxides, (4) the contaminant concentration, (5) the presence of competing ions and (6) the contaminant speciation based on pH and Eh. They noted that the estimated distribution coefficients (Kd = concentration of the sorbed contaminant/the contaminant in the aqueous phase) increased with soil reaction from pH 7 to pH 11. Roh et al. [16] investigated two U-impacted sites at Oak Ridge, TN using sequential leaching and demonstrated that soil U was associated substantially with carbonates (45%) and Fe-oxyhydroxides (40%).

Uranium hydrolysis
Hydrolysis constants for U(IV) are presented in Table 1.
The solubility of U(IV) may be estimated from thermochemical data, with the assumption that UO 2 is the crystalline phase, as: UO 2 + 2H 2 O + OH − = U(OH) 5 − .

2+
) is an oxycation, given that the high charge polarization of U 6+ prevents this aqueous species from being stable. Hydrolysis constants for U(VI) are presented in Table 2.
In low ionic strength media, the U(VI) polymers are not thermodynamically favored, with the exception of (UO 2 ) 3

Simulation of uranium hydrolysis
Using the MinteqA2 software [59], U(VI) speciation may be estimated from thermochemical data for pH intervals from pH 4 to pH 8. Setting the total U(VI) concentration at 10 −8 mole/liter, the pCO 2 pressure at 0 and then again at 0.02 bar (2 kPa) were the primary model vari able inputs. Establishing a constant ionic strength with 0.01 mole NaNO 3 /liter, activity coefficients were estimated using the Debye-Huckel equation. In the CO 2 closed system, UO 2 2+ is the dominant species in very acidic media, whereas UO 2 (OH) + is the dominant species from pH 6 to pH 8 ( Table 3). The ion pair UO 2 NO 3 + is an important secondary species, particularly in acidic media. In the CO 2 open system, UO 2 2+ is the dominant species in very acidic media; however, the UO 2 CO 3 , UO 2 (CO 3 ) 2 2− and UO 2 (CO 3 ) 3 4− are U(VI) species increasingly dominant upon transition from acidic media to neutral and then to alkaline media ( Table 3). Importantly, the uranyl carbonate complexes are stable at Eh conditions that would promote U(VI) reduction in CO 2 closed systems. This MinteqA2 simulation of dilute U(VI) speciation closely corresponds with the analytical data and its MinteqA2 simulation as presented by Langmuir [38] and also the data analysis from Waite et al. [58].

Uranium oxidation and reduction
The reduction of U(VI) to U(IV) may be presented as [53]: The presence of CO 2 (g) at 2 × 10 −2 bar (2 kPa) and an ionic strength standardized by 0.01 M NaNO 3 . Calcium concentrations were 0.001 mol/L. Within a pH column, ( ) indicates the percentage of the U species. Total U concentration was 10 −3 mole/L, which was allowed to equilibrate and allow precipitation.
Calcium concentrations were standardized at 10 −3 mole/L. The presence of CO 2 (g) at 2 × 10 −2 bar (2 × 10 3 pascal) and an ionic strength standardized by 0.01 M NaNO 3 . Within a pH column, ( ) indicates the percentage of the U species. Activity coefficients were determined by the Debye-Huckel equation. Rutherfordine (UO 2 CO 3 ) was predicted to precipitate from pH 4 to pH 7, whereas calcite was predicted to precipitate at pH 8. At Eh values less than 0.2 volts, U(VI) reduction to uraninite (UO 2 ) is favored. Stewart et al. [60] observed that U(VI) reduction to U(IV) is inhibited in the presence of ferrihydrite. Yajima et al. [61] also observed that U(VI) reduction to U(IV) limited mobility. Goldhaber et al. [62] observed that coffinite formed via reduction processes in sedimentary rocks. Fendorf et al. [63] reviewed the biotic and abiotic pathways for U(VI) reduction in anaerobic soils, and they noted that U(IV) has more limited mobility and binds more preferentially to substrates than U(VI). Uranyl reduction is facilitated by bacterially mediated reactions [64]; however, noncrystalline ferric oxides and nitrate may be effective terminal electron acceptors. Similarly, Burgos et al. [63] observed that soil humic acid partially inhibits U(VI) reduction.

Simulation of uranium reduction
At a pe of 5 (296 mv), indicative of suboxic soil redox conditions, and at a total U concentration of 10 −8 mol/L, the MinteqA2 simulation of U(VI) reduction to U(IV) indicates that U(IV) would Total U concentration was 10 −8 mole/L, which was allowed to equilibrate and allow reduction at a pe of 5 (296 mv). Activity coefficients were determined by the Debye-Huckel equation.

Uranium complexation with an emphasis on phosphorus
Uranium complexation pairs a central cation (coordination center) with a surrounding array of molecules and ions. Phosphorus interactions with U(VI) have been studied to assess whether phosphorus may reduce the availability and mobility of U(VI) [12, 65−67]. Stojanovic et al. [18] reported that phosphorus may readily form uranyl phosphates and subsequently precipitate autunite. They noted that at pH levels greater than 6.0, the dominant U(VI)-phosphorus species was the plant-available UO 2 PO 4 species, whereas at more acidic soil reactions, UO 2 HPO 4 and UO 2 H 2 PO 4 + were more abundant and are not considered as plant-available U-phosphate species. Grabias et al. [65] studied uranyl acetate immobilization in ferruginous soils amended with phosphates. In acidic pH ranges, a strong U(VI) sorption was observed in the presence of phosphate, supporting their premise that adsorption was promoted by the formation of UO 2

21.8
Total U concentration was 10 −3 mole/L, which was allowed to equilibrate and allow reduction at a pe of −3 (−177 mv). Activity coefficients were determined by the Debye-Huckel equation. Calcium and sulfate were present initially at 0.001 mole/L. Uraninite was precipitated and established the U equilibria (saturation index 0.00). No carbonate, sulfate or sulfide minerals were documented to precipitate. Calcium concentrations were standardized at 10 −3 mole/L. The presence of CO 2 (g) at 2 × 10 −2 bar (2 × 10 3 pascal) and an ionic strength standardized by 0.01 M NaNO 3 . Within a column, ( ) indicates the percentage of the U species. Mehta et al. [67] demonstrated that U(VI) flux in soil columns was substantially reduced when phosphate was present. Sequential extractions demonstrated that the U(VI) could be readily extracted by ion-exchange and dilute acid treatments. Laser-induced florescence spectroscopy inferred adsorption to be the dominant retention mechanism.
Sandino and Bruno [68] determined the solubility of (UO 2 ) 3 (PO 4 ) 2 4H 2 O (s) and the formation of U(VI) phosphate complexes over the pH range of pH 4-9. In their study, UO 2 HPO 4 and UO 2 PO 4 − were the dominant U species. Minimum U(VI) solubility for the (UO 2 ) 3 (PO 4 ) 2 4H 2 O (s) system occurred near pH 6, whereas the minimum U(VI) solubility for amorphous (noncrystalline) and crystalline schoepite occurred near the pH levels of pH 7.4 and 8.4, respectively. Thermodynamic data for U(VI) with respect to phosphate and carbonate from the literature are well-documented by Sandino and Bruno [68].
Humic and fulvic acids were demonstrated to strongly bind to U(VI), with humic acid forming a slightly stronger binding complex. The U(VI)-humic acid and U(VI)-fulvic acid complexes were determined to be non-integral (1 U(VI) with between 1 and 2 humic or fulvic acids), suggesting that a 1:1 stoichiometry involving a limited number of high-affinity sites.
Ivanov et al. [70] observed uranyl sorption on bentonite in the presence of humic acid with trace levels of uranium(VI). Uranyl sorption on bentonite was shown to be strongly pH dependent. In the absence of humic acid, U(VI) sorption exhibited a sorption edge between pH 3.2 and pH 4.2. In the presence of humic acid, U(VI) sorption slightly increased at low pH and curtails at moderate pH. Soluble uranyl carbonate species inhibited U(VI) sorption at alkaline pH levels. At pH intervals from pH 3 to pH 4, UO 2 HA was predicted ([U] = 8.4 × 10 −11 and pCO 2 = 10 -3.5 atm, HA = humic acid). From pH 5 to pH 7, UO 2 (OH)HA was predicted to be the dominant species. Tinnacher et al. [71] studied the reaction kinetics of tritium-labeled fulvic acid on uranium(VI) sorption onto silica, demonstrating that metal sorption rates are a complex function of metal and organic ligand concentrations and the nature and abundance of mineral surface sites. 2UO

Uranium solubility and precipitation
The solubility of U(VI) may be estimated from thermochemical data with the assumption that UO 2 (OH) 2 is the crystalline phase [53] as: Hsi and Langmuir [56] investigated the adsorption of U(VI) onto noncrystalline Fe(OH) 3 and goethite (α-FeOOH) in batch 0.1 mole NaNO 3 /liter suspensions prepared with different total carbonate concentrations and pH intervals. Hsi and Langmuir documented that the optimum adsorption pH was near pH 6.3-6.5 for noncrystalline Fe(OH) 3 and in alkaline media, U(VI)-carbonate complexes effectively reduced U(VI) adsorption. The effect of carbonate in the goethite suspensions broadened the pH of maximum U(VI) adsorption from pH 5.7 to pH 8.0, a feature attributed to the lack of U(VI)-carbonate complex desorption. Waite et al. [58] investigated U(VI) adsorption onto hydrous ferric oxides, noting that the maximum U(VI) adsorption occurred from pH 5 to pH 9; however, in the presence of carbonate, the U(VI) adsorption in the pH interval from pH 8 to pH 9 was limited. In general, U(VI) adsorption into Fe-oxyhydroxides is greater than phyllosilicate minerals. Total U concentration was 10 −3 mole/L, which was allowed to equilibrate and allow precipitation of rutherfordine and (UO 2 ) 3 (PO 4 ) 2 .
Activity coefficients were determined by the Debye-Huckel equation.
Calcium and H 3 PO 4 concentrations were initially standardized at 10 −3 mole/L. The presence of CO 2 (g) at 2 × 10 −2 bar (2 kPa) and an ionic strength standardized by 0.01 M NaNO 3 . Within a pH column, ( ) indicates the percentage of the U species.

Uranium -Safety, Resources, Separation and Thermodynamic Calculation
Typically, the pH range of minimal U(VI) mineral solubility coincides with the pH range for optimal U(VI) adsorption. U(IV) complexes are frequently less soluble and less mobile than U(VI) complexes [73]. Duquene et al. [23] noted that U(VI) reduction to less soluble U(IV), by either biotic or abiotic processes, influenced uranium mobility. Stojanovic et al. [17] confirmed that soil temperature, pH, oxidation-reduction potentials and the presence of complexing agents were important factors influencing uranium bioavailability and plant uptake. Shahandeh

Uranium adsorption
In a review, Langmuir [38] reported solution U(VI) speciation data from pH 7 groundwater at Yucca Mountain (Nevada, USA) with a total U(VI) concentration of 10 −8 mol/L. The U(VI) percentage speciation was: (1) UO 2 CO 3 at 7.9%, (2) UO 2 (CO 3 ) 2 at 83.1%, (3) UO 2 (CO 3 ) 3 at 7.8%, Uranium(VI) may be adsorbed onto Fe-oxyhydroxides which may subsequently pursue distinctive pathways: (1) U(VI) undergoes reduction to U(IV) by mobile Fe 2+ or H 2 S or (2) desorbed, especially in alkaline solutions at elevated pH levels. Surface properties of soil mineral phases have altered chemical's reactivity because of the presence of small quantities of noncrystalline Fe-and Al-oxyhydroxides. Thus, these alterations of chemical affinity may be attributed to differences in surface area, abundance and composition of Al-OH, Fe-OH and Si-OH groups, and other features that impact the structure of adsorption surfaces (  Waite et al. [58] investigated U(VI) adsorption onto ferrihydrite as a function of U(VI) concentration and the partial pressure of CO 2 . Using the diffuse double layer model with two site surface complexes (weak and strong ≡FeOH), they hypothesized that UO 2 and at higher pH levels, UO 2 (CO 3 ) formed inner sphere mononuclear, bidentate complexes involving the Fe octahedron edge and the uranyl ion. The U-interacting surface reactions without CO 2 participation were [≡Fe(OH) 2 ] + UO 2 2+ = [FeO 2 ]UO 2 + 2H + with log K = −2.57 for the strong site and log K = −6.28 for the weak site. The U-interacting surface reactions with CO 2 participation were [≡Fe(OH) 2 ] + UO 2 2+ + CO 2 = [FeO 2 ]UO 2 CO 3 2− + 2H + with log K = 3.67 for the strong site and log K = −0.42 for the weak site.
Gao et al. [75] investigated U(VI) sorption on kaolinite using batch experiments to observe the effects of pH, U(VI) concentration and the presence of oxyanions. The sorption of U(VI) on kaolinite increased with pH increases from pH 4.0 to pH 6.5, thereafter, a sorption plateau was indicated. The presence of phosphate increased U(VI) sorption, especially in the pH range from pH 3.0 to pH 6.0, whereas sulfate had no measurable influence. UO 2 HPO 4 is predicted as the major U(VI)-phosphate species from pH 4.0 to pH 6.0, thus, the sorption promotion effect of phosphate was attributed to [≡SOH + UO 2 2+ + HPO 4 2− = ≡SOUO 2 HPO 4 − + H + ].
Barnett et al. [41] observed that U(VI) adsorption on naturally occurring media of mixed mineralogy was nonlinear, suggesting that preferential and finite binding sites exist. Adsorption increased strongly with pH transition from pH 4.5 to pH 5.5 and decreased sharply from pH 7.5 to pH 8.5. The reduced adsorption was associated with carbonate-U(VI) complexes. Hummel et al. [76] provided an excellent companion thermochemical database. The MinteqA2 is able to perform adsorption simulations using: (1) Langmuir, (2), ion-exchange, (3) triple layer, (4) Freundlich, (5) constant capacitance and (6) diffuse layer [59].