A thermodynamic model for reactive extraction of macro amounts of zirconium and hafnium with TBP

https://doi.org/10.1016/j.seppur.2019.116491Get rights and content

Highlights

  • Robust thermodynamic building block modeling approach for reactive extraction of nuclear metals.

  • Complex metal extraction system involving macro concentrations is modeled using a hierarchy of sub-systems.

  • Unique feature of predicting equilibrium concentrations of all the species.

  • Models developed applicable at industrial scale and can be extended to develop mixture models for separation of Zr and Hf.

Abstract

Pure component thermodynamic models are developed for solvent extraction involving ionic reaction equilibrium for distribution of macro concentrations of zirconium and hafnium between an aqueous phase containing HNO3 and an organic solvent phase containing diluted TBP. The concentration range considered is 10−03 to 100 M. A framework based on chemical speciation calculation is used for the purpose. Experimental procedures adopted to obtain the required solvent extraction data for modeling and validating these models are detailed. Measured equilibrium concentrations of total zirconium, total HNO3, total nitrate in the aqueous phase and total zirconium in the organic phase are used to obtain the thermodynamic parameters such as equilibrium constants and activity coefficient model parameters for zirconium and hafnium extraction systems.

The novel framework is useful for computing equilibrium concentrations of all the species present in the respective systems. It is also demonstrated that the pure component models developed for extraction of zirconium and hafnium are effective over a wide range of concentrations. These pure component models can be used to simulate and optimize industrial scale zirconium – hafnium separation process.

Introduction

Solvent extraction and more specifically reactive extraction is used in the nuclear industry for purification of metals like Zr, Hf, Pu, Th, U, Np [1], [2], [3], [4], [5], [6]. Typically, impure forms of oxides or hydroxides of these metals are dissolved in nitric acid and the resultant nitrate salts are extracted from aqueous nitric acid solution using the solvent Tri-n-Butyl Phosphate (TBP) in an inert diluent such as n-dodecane or kerosene. Purity achieved depends on the selectivity of solvent towards the metal, which is influenced by the chemical form of metal ion present under the operating conditions of solvent extraction. This inturn depends on the concentrations of the metal ion and nitric acid in the feed solution. These metals in acidic aqueous solutions are ionisable and hydrolysable; therefore, the feed solution of metal in nitric acid contains several ions like metal-hydroxyl, and metal-nitrate species apart from hydrogen ions, nitrate ions and hydroxyl ions. These metals also form polymeric-metal complexes of different stoichiometry, specifically at macro concentration levels (10−03–100 M) [7], [8]. Undissociated nitric acid and water are also present in the aqueous solution. In view of the selectivity of TBP only towards unionized or neutral species, efficiency of extraction process is dependent on the ionic strength of system. Extraction efficiency is also affected by temperature, acid concentration, metal concentration in feed, solvent concentration and solvent to aqueous phase ratio.

Knowledge of the distribution of metals under different process conditions such as metal concentration and acid concentration in the feed, solvent concentration, phase ratios of mixing is desirable to optimize the process conditions. This can be accomplished experimentally but is labour intensive and a theoretical approach is preferred. Among the theoretical approaches, thermodynamic models are rigorous and comprehensive, compared to semi-empirical and empirical models, as they capture different complex interactions in the system [9], [10]. Metal extraction models for neptunium [11], [12], [13], thorium [13], [14], [15], plutonium [13], [16], [17], [18], [19], [20], [21], zirconium [13], [22], [23], [24], [25], etc., are available and in all these works, the metal concentrations considered were at trace level concentration (10−07–10−11M). At trace level metal concentrations in the aqueous phase, these metal ions hydrolyze and form hydroxo-complexes which are predominately monomeric (M-OHy) in nature. Whereas, at higher metal concentrations, these metal ions hydrolyze to form polymeric complexes (Mx-OHy), while monomeric complexes are negligible [7]. Thus, the type of species to be considered in the model developed for the trace level concentrations (10−07–10−11M) and macro concentration (10−03–100 M) are different.

The available approaches (a) do not consider formation of poly-nuclear species at higher concentrations [25], (b) are inaccurate in estimating unreacted or free solvent concentration as they ignore the amount of solvent complexed with metal ion [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], which introduces inaccuracy in predictions at high metal concentrations, and (c) measurements other than distribution coefficient such as free acid concentration in the aqueous phase, total nitrate concentration in the aqueous phase, total metal concentration in aqueous and organic phases are not used [26], [27], [28]. Consequently, the presently available models are not convenient to predict metal extraction involving macro concentrations (10−03–100 M). Besides they cannot accurately capture the dependency of distribution coefficient on metal concentration in aqueous phase, TBP concentration in solvent phase, or the ratio of solvent to aqueous phases over a wide range, and therefore are of limited utility.

Although these limitations can be resolved with chemical speciation calculation algorithms, these algorithms inherently suffer from a numerical convergence problem and are sensitive to the initial guess. As a result the available speciation calculation algorithms are not useful for modeling liquid-liquid extraction systems where the species concentrations of ionic complexes range over several orders of magnitude. A robust algorithm addressing the aforesaid limitations for speciation calculation was developed by Ravi Kanth et al. [29] and was found to be effective to develop thermodynamic models of chemical systems in estimating the unknown parameters [30] and in modeling complex metal extraction systems [32]. In the present work, this framework is adopted for developing more complex solvent extraction models involving macro concentrations of zirconium and hafnium metals from nitric acid solutions using diluted TBP as solvent.

Section snippets

Approach for modeling zirconium and hafnium extraction

The hierarchical framework developed by Ravi Kanth et al. [31] is adopted to model the metal extraction systems involving macro concentrations of metals encountered in the industrial scale process of zirconium (Zr) extraction and its purification from the co-existing nuclear impurity, hafnium (Hf). These models aim at predicting the distribution of pure components of Zr, Hf (where either of these is exclusively present). These pure component models can ultimately be used, to develop models for

Experimental methods and materials

Variables that are measurable to study the distribution of metal salts are the equilibrium concentrations of (i) total metal (Zr, Hf), (ii) stoichiometric concentration of nitric acid, (iii) total nitrate in the aqueous phase and (iv) total metal (Zr, Hf) in the organic phase under different conditions of interest. The experimental procedure adopted to obtain these variables is detailed next.

Results and discussion

Distribution coefficient of metal (DMExp) was determined at various free acid concentrations by mass balance of aqueous phase before and after extraction, using Eq. (10). Though the measured metal concentration in the organic phase was available, it was not used to calculate DMExp, since these measurements are not sufficiently accurate. This error is caused because the organic sample is analyzed by stripping into aqueous phase. The result thus obtained introduces an additional error due to

Conclusions

Thermodynamic models for distribution of macro concentrations of pure zirconium and hafnium between aqueous HNO3 phase and the organic TBP diluted in kerosene are developed for industrially prevailing operating conditions. Required experimental data in the range of industrial applicability was generated for developing and validating these models. Both the Zr and Hf models provided reasonable predictions in the range of 2–6 M HNO3 in the aqueous phase. It is demonstrated that using the

CRediT authorship contribution statement

M.V.S.R. Ravi Kanth: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Data curation, Writing - original draft. S. Pushpavanam: Writing - review & editing, Supervision. Shankar Narasimhan: Writing - review & editing, Supervision. B. Narasimha Murty: Writing - review & editing, Supervision.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

M.V.S.R. Ravi Kanth and B.N. Murty wish to express their heartfelt thanks to Dr. Dinesh Srivastava, Chief Executive, NFC for the motivation and support. Thanks are also due to Shri Sudhir Thakur, Shri Phani Babu, Shri MRS Prasad, and Smt Rama Ramesh, for their encouragement and support.

References (51)

  • H.J. Bart

    Reactive Extraction

    (2001)
  • J. Rydberg et al.

    Solvent extraction equilibria

    Solvent Extr. Princ. Pract.

    (2004)
  • P.L. Brown et al.

    Hydrolysis of Metal Ions

    (2016)
  • P.L. Brown et al.

    Chemical Thermodynamics of Zirconium

    (2008)
  • N.V. Jarvis

    Thermodynamic modeling of solvent extraction systems: Successes and problems

    Sep. Sci. Technol.

    (1991)
  • M.J. Sarsfield et al.

    Solvent extraction and ion exchange modelling Np (IV) solvent extraction between 30% tri-butyl phosphate and nitric acid in the presence of simple hydroxamic acids

    Solvent Extr. Ion Exch.

    (2011)
  • M.J. Sarsfield et al.

    Extraction of neptunium (IV) ions into 30% tri-butyl phosphate from nitric acid

    Solvent Extr. Ion Exch.

    (2009)
  • A.S. Solovkin

    Thermodynamics of extraction of tetravalent plutonium, uranium, thorium and zirconium

    J. Radioanal. Chem.

    (1974)
  • V. Neck et al.

    Ion interaction (SIT) coefficients for the Th4+ ion and trace activity coefficients in NaClO4, NaNO3 and NaCl solution determined by solvent extraction with TBP

    Radiochim. Acta

    (2006)
  • A.S. Solovkin et al.

    Physicochemical approach to the description of the distribution of macroquantities of Pu(IV) in extraction by tributylphosphate from nitrate solutions in the presence of complex formers applicable to the generation of spent nuclear fuel from fast reactors

    Transl. At. Energiya.

    (1984)
  • A.S. Solovkin

    Thermodynamics of extraction of Pu (IV) from nitric acid solutions by Tri-n-Butyl-Phosphate (TBP)

    Transl. from At. Energiya.

    (1971)
  • W.L. Kubic et al.

    A thermodynamic model of plutonium (IV) nitrate solutions

    J. Radioanal. Nucl. Chem.

    (2012)
  • D.J. Chaiko et al.

    Thermodynamic modeling of chemical equilibria in metal extraction

    Separation Science and Technology

    (1988)
  • Y. Enokida et al.

    Distribution coefficient correlations for nitric acid, U(VI) and Pu(IV) in two-phase system with aqueous nitric acid and 30% Tri-n-buthyl phosphate solutions

    J. Nucl. Sci. Technol.

    (1997)
  • P. Tkac et al.

    Modeling of Pu(IV) extraction from acidic nitrate media by Tri-n-butyl Phosphate

    J. Chem. Eng. Data.

    (2009)
  • Cited by (0)

    View full text