Journal of Hazardous Materials 283(2015)432–438
Contents lists available at ScienceDirect
Journal of Hazardous
Materials
j o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /j h a z m a
t
Strontium lectivity in sodium nonatitanate Na 4Ti 9O 20·x H 2O
Arnaud Villard ∗,Bertrand Siboulet,Guillaume Toquer,Aurélie Merceille,
Agnès Grandjean,Jean-Franc
¸ois Dufrêche ∗∗Institut de Chimie Séparative de Marcoule,UMR 5257,CEA-UM2-CNRS-ENSCM,Site de Marcoule,BP 17171,F-30207Bagnols-sur-Cèze,France
h i g h l i g h t s
•Equilibrium constant exchange is cal-culated from the sodium-strontium
exchange.
•Two free parameters models for the solid activity have been checked.•Short-range forces on solid govern the extraction.
•The solid impos the exchange.
g r a p h i c a l
a b s t r a c
t
a r t i c l e
i n f o
Article history:
闻君有两意Received 16June 2014
Received in revid form 4September 2014Accepted 6September 2014Available online 2October 2014
Keywords:
Sodium nonatitanate Strontium adsorption Decontamination
Free energy calculation
a b s t r a c t
We study the extraction of strontium by sodium nonatitanate powder from nitrate strontium and acetate sodium mixture.Experiments show that adsorption is quantitative.The excess Gibbs free energy has been modeled by various models (ideal,2D Coulomb,regular solution model)for the solid pha.We find that the free energy of the solid pha is controlled by short-range interactions rather than long-ranged Coulombic forces.The lectivity is the conquence of a competition between the liquid and solid phas:both phas prefer strontium rather than sodium but the solid contribution is predominant.
©2014Elvier B.V.All rights rerved.
1.Introduction
Nuclear industry produces a wide range of radioactive solutions requiring effective and lective process to decontaminate them.The two main methods ud in nuclear industry are the evaporation and the chemical treatments.The evaporation method allows processing a huge amount of solution which is not suitable for highly concentrated solutions.For this ca,a co-precipitation process is bad on in situ precipitation of solid particles to extract lectively pollutant.The strontiu
m,90Sr,is one of the
∗Corresponding author.Tel.:+33466796961.∗∗Corresponding author.Tel.:+33466339206.
E-mail address:arnaud.villard@univ-montp2.fr (A.Villard),jean-francois.dufreche@univ-montp2.fr (J.-F.Dufrêche).
most abundant radioactive contaminant and most radio-toxic for the environment and human health found as nuclear fission product.Indeed,this element has a specific affinity for bones since strontium and calcium are alkaline earth elements with the same charge and almost the same size.
There are various methods for removing strontium from liq-uid waste using co-precipitation or adsorption process.For example,decontamination process using organic extractants [1–4],adsorption onto resin [5–8]have been propod but they have strong drawbacks due to their low thermal and radiation stability.The current industrial process in ‘La Hague’(French full reprocessing plant)consists in the co-precipitation of radio-strontium by sulphate barium.However,this process gen-erates a large quantity of sludge to be confined due to their radioactivity.So alternative process bad on adsorption pro-cess are investigated.For example,clays [9–12],hybrid materials
/10.1016/j.jhazmat.2014.09.0390304-3894/©2014Elvier B.V.All rights rerved.
A.Villard et al./Journal of Hazardous Materials283(2015)432–438433
[13–16],hydrous metal oxides[17–20],zeolites[21–24]or titanate and silicotitanate[25–31]are widely studied.Among them,sodium nonatitanate prents a lectivity towards strontium with also a suitable radiation resistance.The goal of the prent study is to analyze and model some experimental data concerning the sodium nonatitanate extractive properties in aqueous solution containing strontium nitrate and sodium acetate.This latter has been chon in order to be clo to the industrial solution.
Sodium nonatitanate powder materials have recently been syn-thesized using hydrothermal process at different temperature,and then characterized by various techniques(XRD,TGA and chemical analysis),by hydrothermal process at different temperatures[32]. Depending on the synthesis temperature,two structures appear. Below200◦C,the ionic exchange is rather complete and upper 200◦C,the exchange rate is limited at62%.At low temperature, they are two diffusion limiting steps,diffusion and intra-particle diffusion,whereas at high temperature there exists only an intra-particle diffusion step.
The temperature impacts therefore directly the extraction prop-erties[33].Even if the sodium nonatitanate synthesized at100◦C has the best decontamination performances,the strontium quan-tity adsorbed is cloly related to the sodium concentrations. Moreover,this exchange is specifically due to an ionic exchange (two sodium atoms for one strontium atom).The pH has also an influence on the adsorption rate of the strontium.Indeed when the pH is low,the adsorption rate is minimal whereas this rate is maximum for a pH above eight[33].
The experiments with radioactive species are nsitive due to the radiation and to reduce the number of experiments,it is necessary to develop a predictive model.Predictive model should be bad on a knowledge of the microscopic and macro-scopic phenomena,the former ones still needing much efforts. We propo,as afirst attempt,an investigation on the adsorption mechanisms through some simple free parameter models.The models aim at improving activity coefficients determination,both for the solution and the solid(sodium nonatitanate).They also include the solid characterization[32]via various XRD,TGA,elementary chemical analysis)and take into account the description of the initial conditions of the solution.The equilib-rium constant of the exchange reaction has been calculated from the models.The different models ud allow to discriminate and understand which phenomena drive the extraction.
2.Methods
2.1.Experimental
2.1.1.Synthesis
Sodium nonatitanate powder samples have been synthesized in three steps according to hydrothermal methods[26,32,34,35]: 2.5g of titanium iso-propoxide(Sigma–Aldrich,purity97%)was first added to2g of deionized water and2.73g of50%wt of sodium hydroxide(Sigma–Aldrich,purity>98%).This mixture was then inrted in a Teflon pot at100◦C during24h for an hydrothermal treatment.After this thermal treatment,the resulting gel was next washed three times with deionized water to remove some excess NaOH and centrifuged at4500rpm for5min.The gel was then dried at80◦C during one day.
2.1.2.Sample characterization
The sodium nonatitanate powder was characterized by the fol-lowing techniques:
–In order to characterize and check the purity powder,X-Ray Diffraction(XRD)has been performed at room temperature with
a Bruker®D8advance diffractometer in Bragg-Brentano geome-try with Ni-filtered and Cu-K˛radiation,between2Â=5◦and80◦, step size0.01◦and one cond per step.
–The water amount inside the solid was estimated from Thermo-Gravimetric Analysis(TGA)by a Setaram instrumentation with SetSys Evolution instrument on approximately20–50mg of solid samples at heating rate of10◦C min−1.The analysis was per-formed from room temperature to1000◦C.
–The elementary analysis of powder was performed by inductively coupled plasma atomic emission spectrometry(Thermo Scien-tific).Before the analysis,the powder samples were dissolved in acid media.
–The quantity of strontium adsorbed was determined by mea-suring the strontium concentration in solution before and after contact with sodium nonatitanate powder.This analysis has been performed with a Dionex capillary ionic chromatography with methylsufonic acid as an eluent.We have ud the IonPac CS16 column with a diameter of0.4mm and a length of250mm.
2.1.
3.Sorption experiments
In this paper,we will focus on two specific experiments per-formed from sodium nonatitanate already reported in[32,33].
Thefirst t of experiments,is performed in order to obtain the equilibrium constant.In the experiments,the solid activity coefficients have been neglected by using a solution containing strontium at a very low concentration(trace level).The kind of concentrations have been measured with radioactive strontium (90Sr)by scintillation(Packard,Tri-carb2750TR/LL).A solution of66660±6600Bq L−1,namely13.0±1.3ng L−1,was prepared by dissolving and diluting up to obtain the suitable concentration with a pure radioactive strontium nitrate salt.This solution was added to various sodium acetate concentration solutions,ranging from 0.1to1mol L−1.A cond t of experiments has been performed in order to obtain the adsorption isotherm.The initial solution has been obtained by dissolving strontium nitrate and sodium acetate dissolved into deionized water.Sodium acetate has been ud both to have an alkaline solution and to buffer the solution at pH≈8 [33].The different solutions were prepared by keeping the cationic charge numberfixed at10−2mol L−1(Eq.(1)):
[Na+]+2[Sr2+]=10−2mol L−1(1) Each sample of this t of experiments has been prepared with 10mg of powder added to20mL of initial solution in a vial which was then shacked during one day.
2.2.Modeling
2.2.1.Chemical equilibrium
Initially,the strontium is exclusively in solution,whereas sodium is in the two phas(solid and solution phas).The exchange between two sodium from the solid and one strontium from the solution reads:
2Na++Sr2+ 2Na++Sr2+(2) This also can be written via the mass action law:
K0eq=
a2
Na+
a
Sr2+
a
Sr2+
a Na+2
=[Na
+]2[Sr]
[Sr2+][Na+]
2
2
Na+
Sr2+
Sr2+
Na+2
(3)
where K0eq is the equilibrium constant,a X and a X are respectively cations activities in solution and in solid. X and X are the corre-sponding activity coefficients.
2.2.2.Aqueous pha
The decoupling between the standard term and the activity term depends on the frame of the reference for the activity scales[36].For
434 A.Villard et al./Journal of Hazardous Materials283(2015)
432–438
Fig.1.Diffractogram of the sodium nonatitanate synthesized at100◦C in hydro-thermal conditions.Thefirst peak corresponds to the inter-layer distance.
any solute,the considered standard state is a solution concentration of1mol L−1and the particles are non-interacting(infinite dilution).
When the solute concentration in water is relatively low(typi-cally10−2mol L−1),the Debye–Hückel approximation(DH)is cho-n[37].Thus,the activity coefficients are given by this equation:
ln i=−
z2
i
e2
8 ε0εr k B T
Ä(4)
with z i the charge of ion i,e the elementary charge,ε0εr the water permittivity,k B the Boltzmann constant,T the absolute temperature andÄthe inver Debye length:
Ä2=4 L B
j
C j z2j(5)
where C j is the concentration of ion j in particle m−3and L B is the Bjerrum length:
L B=
e2
4 ε0εr k B T
(6)
word修改模式
When the solute concentration is high(>10−2mol L−1), Mean Spherical Approximation(MSA)model is ud instead of Debye–Hückel model[38–42].This solution model will be devel-oped more precily i
nto a next paper.
2.2.
3.Solid pha
智慧的近义词
For the solid pha,the standard state lected for the cation corresponds to a pure sodium nonatitanate(Na4Ti9O20).If other cations(such as H+,Sr2+)are prent inside the titanate struc-ture,they are considered as infinitely dilute toward the standard concentration of1mol g−1of pure sodium nonatitanate.
The solid has been reported to be a multi-layer material [34,43–45].From Fig.1the inter-layer distance is estimated at9.5˚A. Considering the thick of the titanium oxide octahedral layer(≈6˚A), the space for aqueous part is typically one water molecule diame-ter.Conquently,sodium nonatitanate solid has been modeled as a lamellar system with a mono-layer of hydrated sodium interleaved between two layers of titanium oxide.
Three models have been studied:ideal model,2D Coulomb gas model and regular solution model.
In thefirst model,the ideal one,all the solid activity coefficients are suppod to be equal to one.
For the cond model,the2D Coulomb gas one,is rigorously valid if the main physical interaction between the ions is the elec-trostatic force,similar to the original DH approach in the ca of electrolyte solutions.Here,ions in the Na+for pure Na4Ti9O20)are assumed to be charged point particles(on the TiO2 layers)disperd in a non-polarizable background.This continuous model has already been extensively studied[46–48]thanks to inte-gral equations on DH approximation and Monte Carlo simulation.
The simulations performed by[47]give a very good approxima-tion,the excess Helmholtz energy(A ex),namely the difference from the standard state,of the2D Coulomb gas in thefluid domain can be approximated as
A ex
Nk B T
=˛ (7)
where N is the number of particles(here ions within the solid),the parameter˛has been reported to be equal to−1.04704from[47] and the screening parameter, ,is
= 2L B
N S(8) where is the charge(in unit of e)of the Coulomb gas,N S is the sur-facic concentration of particles(particles m−2)and L B is the Bjerrum
length in the solid(Eq.(6))withεr
solid
the dielectric constant of the water molecules in this confined space between the TiO2layers. For the hydrate ions confined between two TiO2layers,theεr
solid
is lower than theεr in the bulk solution and it has been assumed to be15for the hydrate sodium in the sodium nonatitanate[49].In the ca of mixtures, can be simply generalid by:
=
√
L B
i
N S
i
2
i
i
N S
i
(9)
Thus,the activity coefficients of the i species with a surfacic concentration N S
i
and charge i are obtained by differentiating Eq.
(7)with respect to N.We obtain then:
ln i=˛L B
√
⎛
⎝
j
N j
S
k
x k z2
k
2
+z2
i
−
1
2
+z2
i
N◦
辽宁的简称
1
S
⎞
⎠
(10) where x k=N k/
i
N i is the particle fraction and N◦
1
三八福利is the initial num-ber of cation(here Na+)within the solid.
The third model,the regular solution one,allows to evaluate the short range interactions.The regular solution model can also lead to the expression of the activity coefficient.For a binary mixture with i and j species[50]:
ln i= x2j(11)
where x j reprents the molar fraction and the parameter the interaction coefficient between two ions,which has beenfitted from the experimental data.
This model is a simple model forfirst short range interactions. Contrary to the2D Coulomb gas model,it is valid when the inter-actions between the ions are predominantly short ranged and modeled by a very simple meanfield network model.
Thus the two last models are available to explain the mean inter-actions inside the solid.
3.Results and discussion
3.1.Characterization of powder samples
From the elementary and the TGA analysis,the chemical com-position of the powder has been measured as:Na4.2Ti9O20.1,12.4 H2O.The chemical composition is clo to the theoretical compo-sition of a pure sodium nonatitanate reported by[27].From this result,the Cation Exchange Capacity(CEC)has been calculated to be3.9×10−3mol g−1.
XRD measurement from Bragg’s law gives an inter-layer distance equal to9.5˚A(Fig.1)which is in agreement with [26,34,35,51].Fig.2shows a transmission electronic microscope
A.Villard et al./Journal of Hazardous Materials 283(2015)432–438
435
Fig.2.Transmission Electronic Microscopy picture of sodium nonatitanate.The
alternations of the dark and light fringe are characteristic of a multi-layer
material.
Fig.3.Scanning Electronic Microscopy picture of sodium nonatitanate.
(TEM)picture of the material.An alternation of dark and bright fringes indicates a multi-layer material.The inter-layer distance obrved is in agreement with the XRD results and also with the structure suggested by [27].Fig.3shows the morphology of the powder that is constituted of grains.The average size of grains is clo to one micrometer.
3.2.Equilibrium constant
In the below mentioned study [32],experiments were con-ducted in order to determine the equilibrium constant,K eq .This latter was determined from the trace concentrations,namely at very low strontium concentration.This first t of experiments has been conducted with a strontium concentration below 1.410−10mol L −1,obtained with radioactive strontium (90Sr).The
Table 1
Ionic diameter ud for the liquid activity coefficients calculated with MSA [38,52].
Ion
Diameter (˚A)
Na +
3.05CH 3COO −
4.8Sr 2+
5.26NO −
3
3.78
maximum 90Sr adsorbed on solid is very low compared to the total CEC (3.9×10−3mol g −1).The reference of the solid is chon to be the pure sodium nonatitanate,as mentioned above.We consider the pure sodium nonatitante as a reference for the solid activity coefficients and we assume that the solid state is not altered at this very low strontium concentration.In this way, Sr 2+= Na +=1.From this approximation,the equilibrium constant (Eq.(3))is writ-ten as:
K eq
=[Na +]2[Sr 2+][Sr 2+][Na +]2 2Na + Sr 2+
=K D
[Na +]2
2Na
+[Na +]
2(12)
where K D is the distribution coefficient defined as K D =
a Sr 2+a Sr 2+
=
[Sr 2+][Sr 2+] Sr 2+
(13)
For the liquid state,different experiments have been performed by varying the sodium concentration i
n solution [32].In this latter paper,we have ud the simplest model by considering that the activity coefficients of sodium and strontium in the liquid state are equal to one.
From Eqs.(3)and (13),the slope can be expresd as:
ln 10[K D ]=ln 10K eq +2ln 10[CEC]−2ln 10[a Na +]
(14)
From the decadic logarithm slope of the distribution coefficient as a function of the decadic logarithm of the sodium concentration,the equilibrium constant has been estimated.The linear coeffi-cient has been t to −2becau it corresponds to the number of the sodium exchanged for strontium,which replaces it in order to maintain electro-neutrality.
However,the experiments were conducted up to one molar in sodium acetate and at this high concentration,the activity coefficients are not equal to one.We have therefore to con-sider veral models for the evolution of activity coefficients.The Debye–Hückel theory is valid up to 10−2mol L −1while the Mean Spherical Approximate (MSA)theory is valid up to 1mol L −1.In order to calculate the solution activity coefficients of the first t of experiments,low concentration in strontium and high
concentra-tion in sodium,the MSA theory has been ud.This theory requires the hydrate ionic diameters of the solution which are available in the literature [52,38]and are summarized in Table 1.When the strontium concentration is very low,the strontium activity has been impod by the high sodium concentration.Therefore the strontium solution activity coefficient have not to be neglected.Fig.4reprents the decadic logarithm of the distribution coefficient as a function of the decadic logarithm of the sodium concentration in solution.The red diamonds reprent the exper-imental values with the ideal model,activity coefficients in liquid pha are equal to 1,and the blue diamonds are the experimental values with the activity coefficients in the liquid pha calculated through the MSA theory.The solid line reprents the fit of the experimental values with a slope coefficient t at −2,as mentioned above.
In the ideal ca (red diamond),the equilibrium constant has been estimated at K eq =2.28×107,corresponding approximately to an exchange free energy equal to −44kJ mol −1.
436
A.Villard et al./Journal of Hazardous Materials 283(2015)432–438
Fig.4.Distribution coefficient,K D ,of experimental points measured with radioac-tive strontium as a function of molarity with activity coefficients by MSA theory,,or equal to one,[33].The K d reprents the ration of solid strontium concentration (in mol per gram of solid)to aqueous strontium concentrati
on (in mol per liter).(For interpretation of the references to color in this figure legend,the reader is referred to the web version of this article.)
小学家委会From the same argument,the equilibrium constant of the real solution (blue diamond),modeled with MSA,has been estimated to be K eq =2.24×108,which gives the exchange free energy approxi-mately equal to −47kJ mol −1.The equilibrium constant calculated with the MSA correction is the same order of size magnitude but the reaction is slightly more energetic.This high value of the constant confirms that the solid is more lective for the strontium than for the sodium [33,25].
3.3.Solid activity coefficients
In order to model the strontium adsorption as a function of the initial strontium concentration,we have ud the cond t of experiments,namely at constant cationic charge number.The solid activity model has been fitted with respect to the experimental data.From the experimental data,an apparent equilibrium con-stant is calculated by taking into account the real liquid pha (Eq.(15)).This latter has been calculated as following:
K App
=[Na +]2
[Sr 2+还原论
][Sr 2+][Na +]
2 2
Na + Sr 2+
(15)
The liquid activity coefficients have been obtained by the
Debye–Hückel theory.The excess exchange energy,between the ideal solid state and the real solid state,can therefore be expresd from Eqs.(3)and (15)into the Eq.(16).
G excess 2Na →Sr
=−k B T ln防误触
Sr 2+ Na +2
K eq
=k B T ln K App
(16)
The Coulomb 2D model requires the solid specific area,developed by the lamellars,in order to calculate the sur-facic concentration.This area cannot be obtained by nitrogen adsorption–desorption measurement.Indeed,the hydrate sodium occupying the space between two sodium nonatitanate layers,avoid the inrtion of the nitrogen molecules.A model of the spe-cific area has been developed from the chemical composition of the solid bad on the hydration number of the sodium in solution.It consists in calculating the surface occupied by the sodium tri-hydrate between the TiO 2layers.The simplest way to place three water molecules around sodium is to form a triangle (Fig.5),
the sodium is placed at the gravity center.From this model,a specific area has been calculated at 1050m 2g −1.
The cond model ud to estimated the solid activity coefficient is the regular solution model which requires one parameter called (reprenting the short range energy).This latter has been fitted from the experimental data.
Fig.5.Reprentation of triangular geometry of the sodium hydrate in porous media,the blue and red spheres are water molecules and sodium ions respectively.This geometry give an specific surface area of approximately 1050m 2g −1.(For inter-pretation of the references to color in this figure legend,the reader is referred to the web version of this article.)
0.0
5.0×10-4
1.0
×10-3 1.5 ×10
-3
2.0 ×10
-3
Q Ads / mol g
-
1
6
8101214ΔG e x c h a n g e / k B T
Experiments 2D Coulomb Regular solution
e x c e s s
Fig.6.The excess exchange Gibbs energy as a function of the adsorption quantity;♦are experimental values.The dashed blue line reprents the 2D Coulomb model and the solid red line is the regular solution model.The uncertainty,from the com-prehensive expression of the Gibbs energy,is especially high at saturation becau of the very small amount of non-ideal part.(For interpretation of the references to color in this figure legend,the reader is referred to the web version of this article.)
Fig.6reprents the excess exchange Gibbs energy of the solid calculated from experimental data as a function of the adsorbed strontium quantity by the solid.The experimental data have been plotted using the Debye–Hückel model for the liquid pha.The uncertainty has been calculated for each experiments and this latter is more important at the solid saturation.The uncertainty,from the co
mprehensive expression of the Gibbs energy,is especially high at saturation becau of the very small amount of non-ideal part.The dashed blue line reprents the predictive data calculated by Coulomb 2D model.The calculated curve decreas slightly linearly while the experimental data increa.The solid red line is the regu-lar solution model with a parameter equals to 2.The solid red line goes quite well through the experimental values which indicates that the short range interactions are dominant.Then,the regu-lar solution model has been chon to calculate the solid activity coefficient.
3.4.Model applications
Fig.7reprents the quantity of strontium adsorbed by the solid as a function of the decadic logarithm of the initial strontium con-centration in solution.The total charge concentration has been t equal to 10−2mol L −1.The uncertainty has been evaluated at 10−4mol g −1.
Up to an initial strontium concentration of 8×10−4mol L −1,the quantity of strontium adsorbed on the solid can be simply plotted by Q Ads =V 0/m C Sr initial ,where V 0reprents the volume of the solution and m is the solid mass.This equation is inde-pendent of the equilibrium constant.From an initial strontium
