I

Ra

b

c

a

ARRAA

KSEDAN

1

rbefhftatrtl(aftTlf

0d

Journal of Hazardous Materials 161 (2009) 1450–1453

Contents lists available at ScienceDirect

Journal of Hazardous Materials

journa l homepage: www.e lsev ier .com/ locate / jhazmat

onic transport behavior of BaO containing sodium borosilicate glasses

.K. Mishraa, R. Mishrab, C.P. Kaushika,∗, A.K. Tyagib, B.S. Tomarc, D. Dasb, Kanwar Raja

Waste Management Division, Bhabha Atomic Research Centre, Trombay, Mumbai 400085, IndiaChemistry Division, Bhabha Atomic Research Centre, Trombay, Mumbai 400085, IndiaRadio Chemistry Division, Bhabha Atomic Research Centre, Trombay, Mumbai 400085, India

r t i c l e i n f o

rticle history:eceived 4 December 2007eceived in revised form 29 April 2008ccepted 29 April 2008

a b s t r a c t

The present manuscript describes the preparation, characterization and electrical behavior of sodiumborosilicate glasses with varying concentrations of BaO, while maintaining the SiO2:B2O3:Na2O ratiosconstant for all the samples. The effect of BaO substitution on the ionic conductivity of glasses was stud-ied by ac impedance analyzer, below the glass transition temperature. The diffusion coefficient (D) of each

vailable online 4 May 2008

eywords:odium borosilicate glasslectrical conductivityiffusion coefficient

sample has been determined from the values of electrical conductivity and density. The activation energyfor sodium ion transport has been calculated from the values of diffusion coefficients at different temper-atures. The electrical properties of the modified glass have been explained on the basis of the structuralfactor.

© 2008 Elsevier B.V. All rights reserved.

lcttamgcntch

cwmii

ctivation energyuclear waste immobilization

. Introduction

During the reprocessing of the spent nuclear fuel, high leveladioactive liquid waste (HLW) is generated, which is usually immo-ilized in the vitrified matrix [1,2]. The vitrification process in Indiassentially consists of metering of pre-concentrated HLW and glassorming additives in the form of slurry into the process vessel ofigh Ni–Cr alloy (Inconel 690) located in a multi-zone induction

urnace [3]. In this process, the metallic melter used for vitrifica-ion often undergoes temperature-based deformations/corrosionnd requires replacement, which is a costly affair. Further, due tohe restriction of process temperature, waste loading in the vit-ified waste product is limited. On account of these limitations,he steps are being taken to switch over to advanced technologiesike Joule heated ceramic melter (JHCM) and cold crucible melterCCM) which will facilitate higher waste loading at higher temper-ture unlike conventional induction pot melter [4]. One of the glassormulation parameters to be used with these melters is glass resis-ivity, which is crucial in governing performance of the joule heater.

his molten glass resistivity dictates the power supply parametersike voltage and current in case of JHCM and additional frequencyor CCM.

∗ Corresponding author. Tel.: +91 22 2559 5528; fax: +91 22 2550 5147.E-mail address: cpk@barc.gov.in (C.P. Kaushik).

sUpaa

brc

304-3894/$ – see front matter © 2008 Elsevier B.V. All rights reserved.oi:10.1016/j.jhazmat.2008.04.122

The decay of radio nuclides, immobilized in the glass matrix mayead to increase in the temperature of the glass product and mayreate the thermal gradient. This temperature gradient can causehe diffusion of radio nuclides in the glass matrix, which is one ofhe most important aspects of the waste immobilization processffecting the leaching behavior of the radio nuclides from the glassatrix. Modeling of leaching behavior of radio nuclides from the

lass matrix requires data on their diffusion. To assess the long termhemical durability of the glass matrix, study of diffusion of radiouclides present in the glass matrices will be an important parame-er. Leach rate measurements [5] also lead to evaluation of diffusionoefficient, however, they have been found to be 500–2000 timesigher than bulk diffusion data.

The knowledge of conductivity of the glass melt as a function ofomposition is required to optimize the composition of the glass,hich will be ideal for the requirements of ceramic melter andelters based on the cold crucible induction technique. However,

n the absence of the conductivity data of the glass melt, the trend inonic conductivity below melting point, as function of glass compo-ition, will provide an insight for the development of suitable glass.nderstanding the ion transport mechanism in the glass matrix isossible from the conductivity measurement at different temper-tures. In addition, the electrical conductivity measurement will

llow determining the diffusion coefficient of alkali ions in the glass.

In this manuscript, we report the ionic conductivity of sodiumorosilicate glasses by ac impedance technique and the effect ofepresentative alkaline earth glass modifier such as BaO on the ioniconductivity of the glass. Based on the electrical conductivity and

R.K. Mishra et al. / Journal of Hazardous Materials 161 (2009) 1450–1453 1451

Table 1Composition and properties of glasses

Glass composition (mol%) SBBa-0 SBBa-8 SBBa-16 SBBa-24

SiO2 53.53 51.68 49.63 47.36B2O3 30.30 29.24 28.10 26.80Na2O 16.17 15.61 14.98 14.31BaO 0.00 3.47 7.29 11.53Density (g cm−3) 2.40 2.57 2.72 2.96CT(

dbtihi

2

ps(scSs

ttwfmsopatn

bttsatoAsl±

pbrdotastww

F3

3

ppcwcciimip1apw[rht1

pehtliB

Na (× 1021 cm−3) 7.38 7.28 7.03 6.93g (K) 820 827 825 824Ea ± 0.01) (eV) 1.00 1.05 1.09 1.14

ensity of the glasses, diffusion coefficients of all the samples haveeen determined. The aim of the present study was to determinehe activation energy of ionic diffusion in barium bearing borosil-cate glasses through ac conductivity measurements, which willelp to understand the effect of barium in the diffusion of sodium

n these glasses.

. Experimental

Glass samples were prepared by thoroughly mixing appro-riate amounts of the components, namely silica, boric acid,odium nitrate and barium nitrate of analytical reagent grade>99.5% purity) for 100 g batch size scale. The compositions ofodium borosilicate glasses (Table 1) were selected having BaOoncentration ranges from 0 to 11.53 mol% while maintaining theiO2/Na2O, SiO2/B2O3, B2O3/Na2O ratios constant for all the glassamples.

The powder mixtures were well ground in agate mortar and thenransferred into a platinum crucible and heated at 700 ◦C for 2 h forhe completion of calcinations and then melted at 1000 ◦C. The meltas maintained at 1000 ◦C for 4 h to ensure homogenization. The

ree flowing melt was poured into preheated cylindrical graphiteould of internal diameter 11 mm and length 20 mm. The glass

amples in the form of rods were cut into uniform circular pelletsf diameter 11 mm and thickness about 1 mm. The glass samplesrepared by soaking the melt at 1000 ◦C for 4 h, were chemicallynalyzed. The compositions were found to be in agreement withhe expected value within an experimental error of ±2% indicatingo significant loss of glass constituents.

The amorphous nature of these glass samples was confirmedy X-ray diffraction (XRD) using a Philips X’Pert Pro diffractome-er with monochromatized Cu K� radiation. The glass transitionemperature (Tg) of the samples was determined by differentialcanning calorimetry (DSC) using Schimadzu instrument in Argontmosphere and at a heating rate of 10 K/min. Structural elucida-ion of borosilicate network of these glass samples was carriedut by recording 29Si and 11B MAS NMR patterns using a Brukerdvance DPX 300 machine. Density of the glass samples was mea-ured by Archimedes’ principle, with distilled water as immersioniquid, on a single pan electronic balance with an accuracy of0.02 g/cc.

The electrical conductivity of the glass samples in the form ofellets (11 mm diameter and 0.5–1 mm thickness) was measuredy a Solatron Impedance Analyzer (Model SI 1260) in the frequencyange of 1 MHz to 1 Hz and with ac voltage amplitude 100 mV andc bias 0 mV. The pellets were uniformly coated with a thin layerf silver paste to have proper electrical contact with platinum elec-rode. The impedance measurement was carried out under static

ir in the temperature range 700–820 K at the interval of 10 K. Theamples were equilibrated for 30 min at the set temperatures beforehe impedance measurements. The temperatures were maintainedithin ±1 K using a microprocessor. The resistances of the pelletsere obtained from the real part of the Cole–Cole plot.

3

t

ig. 1. DSC patterns for sodium borosilicate glasses containing (a) 0 mol% (b).47 mol% (c) 7.29 mol% and (d) 11.53 mol% BaO.

. Results and discussion

Glassy nature of the samples was confirmed by the broad XRDeak centered around 2� = 13◦, which is characteristic of the amor-hous borosilicate network. The Tg of sodium borosilicate glassesontaining different amounts of BaO recorded by DSC techniqueas found to be around 825 K. DSC patterns of glass samples

ontaining different mol% of BaO are shown in Fig. 1. No signifi-ant change in the glass transition temperature was noticed withncrease in BaO content. 29Si MAS NMR patterns of BaO contain-ng glasses showed the characteristic broad asymmetric peak with

axima systematically shifting towards more positive values withncrease in BaO concentration in the glass. 29Si and 11B MAS NMRatterns for sodium borosilicate glasses containing 0, 3.47, 7.29 and1.53 mol% of BaO are shown in Fig. 2A and B, respectively. The broadsymmetric peak 29Si MAS NMR could be deconvoluted into twoeaks with chemical shift values around −99 and −86 ppm whichere attributed to Q3 and Q2 structural units of silicon, respectively

6,7]. With increase in BaO concentration in the glass, initially theelative concentration of Q3 and Q2 structural units remained same,owever, after addition of more than 3.47 mol% BaO, the Q2 concen-ration marginally increases at the expense of Q3 structural units.1B MAS NMR patterns of all the samples are characterized by sharpeak superimposed over broad shoulders placed symmetrically onither side of the sharp peak. Sharp peak is attributed to the tetra-edrally coordinated boron (BO4 structural units) and broad peako the trigonally coordinated boron (BO3 structural units) [6,7]. Theine shape and peak position are identical for all the glass samples,ndicating that the boron network is unaffected by the doping ofaO into the sodium borosilicate glass.

.1. Electrical conductivity

The electrical conductivity �(ω) derived from the real part ofhe Cole–Cole plot versus the ac frequency for glasses containing

1452 R.K. Mishra et al. / Journal of Hazardous Materials 161 (2009) 1450–1453

F0p7

07cfTitb

Fig. 3. Conductivity as function of frequency for sodium borosilicate glass samplesat 773 K containing different mol% of BaO.

Fa

abbfintpB(Fig. 2A).

The Cole–Cole diagram (−Z′′ versus Z′) for glass containing3.47 mol% of BaO recorded at 724, 746,773 and 796 K are shownin Fig. 4. In this diagram, a circular arc-centered slightly above thereal axis is obtained, which varies with the temperature. The inter-

Table 2Ionic conductivity for different compositions at representative temperatures

Temperature (K) Ionic conductivity (ohm−1 cm−1)

0 mol% BaO 3.47 mol% BaO 7.29 mol% BaO 11.53 mol% BaO

ig. 2. (A) 29Si MAS NMR patterns for sodium borosilicate glasses containing (a)mol% (b) 3.47 mol% (c) 7.29 mol% and (d) 11.53 mol% of BaO. (B) 11B MAS NMRatterns for sodium borosilicate glasses containing (a) 0 mol%, (b) 3.47 mol%, (c).29 mol% and (d) 11.53 mol% of BaO. The peaks marked “*” are spinning side bands.

, 3.47, 7.29 and 11.53 mol% of BaO at the selected temperature of73 K is shown in Fig. 3. It can be observed from the figure that theonductivity of the glass samples remains almost constant in the

requency range 1 Hz to 1 MHz, but increases sharply above 1 MHz.he plateau value corresponds to static conductivity for long rangeonic displacement, while the increase at high frequency is dueo relaxation caused by local motion of Na+ cations [8] envisagedy single ionic jump diffusion mechanism as proposed by several

67

77

ig. 4. Cole–Cole plot of NBS glass sample containing 3.47 mol% of BaO, measuredt 724, 746, 773 and 796 K in the frequency range 1 Hz to 10 MHz.

uthors [9,10]. It can also be seen from Fig. 3 that in the case of BaOearing glasses conductivity decreases drastically (by about 50%)y doping 3.47 mol% of BaO but doesn’t significantly changes onurther doping BaO. Decrease in conductivity of glass samples withncrease in BaO concentration can be attributed to the fact that theet Na+ ion concentration decreases with increase in BaO concen-ration. However, the non-linearity in the decrease of conductivityerhaps arises due to breaking of Si network by the presence ofaO beyond 3.47 mol% as also corroborated by the MAS NMR results

65 0.09 0.04 0.01 0.00320 0.28 0.12 0.03 0.01

740 0.41 0.17 0.06 0.0250 0.50 0.21 0.08 0.0375 0.89 0.33 0.11 0.05

R.K. Mishra et al. / Journal of Hazardous

Fg

scwTadssct

3

oc

�

wbr

C

waMatstw(

3

ao

iTaa1aBtzdlwtte

4

mtgtctf31et

A

t

R

[

[

ig. 5. Plots of ln (D) versus 1/T for 0, 3.47, 7.29 and 11.53 mol% of BaO containinglasses below the glass transition temperature.

ection of the curve with the real axis gives the resistance Rdc whichan be related to static conductivity (�dc) given by �dc = (1/Rdc) × l/S,here l is the thickness and S is the cross-sectional area of the pellet.

he conductivity of the sample was found to increase with temper-ture, which can be explained on the basis of thermally stimulatedislocation of alkali cation from its equilibrium position to an inter-titial position followed by random diffusion through interstitialites as proposed by Kelly and Tomozawa [11]. The values of ioniconductivity for different compositions at representative tempera-ures are given in Table 2.

.2. Calculation of diffusion coefficient

The diffusion coefficient of the mobile sodium ion can bebtained by relating the electrical conductivity with the diffusionoefficient D using Nernst–Einstein equation [12].

=(

q2

kT

)× CNa × D� (1)

here q represents the charge of the mobile ion, CNa is the num-er of sodium atom per unit volume calculated using the followingelation

Na = �ZNaN

M(2)

here � represents the sample density, ZNa is the number of sodiumtoms in the chemical formula, N is the Avogadro’s number, and

is the molar mass of the sample. The values of density and CNare given in Table 1. The diffusion coefficient for the transport ofhe charge carrier (Na+ ion) as a function of temperature for glassamples containing different mol% of BaO was calculated by usinghe above relation. The diffusion coefficient was found to decreaseith the increase of BaO concentration at any given temperature

Fig. 5).

.3. Activation energy

The plots of ln (D) versus 1/T below the glass transition temper-ture for barium containing glasses are shown in Fig. 5. It can bebserved from these plots that all glasses have a linear plot, which

[

[

Materials 161 (2009) 1450–1453 1453

s a characteristic of a thermally activated transport phenomenon.he activation energies for the glass samples containing 0, 3.47, 7.29nd 11.53 mol% of BaO calculated from the slope of theses curvesre found to be 1.00 ± 0.01 eV, 1.05 ± 0.01 eV, 1.09 ± 0.01 eV and.14 ± 0.01 eV, respectively. The activation energy for the thermallyctivated transport of Na+ ion is found to increase with increase inaO concentration in the glass matrix, which can be explained onhe basis of the structural factor. The mechanism is based on Diet-el’s structural field strength theory according to which the largeifference in the field strength (�F) values between alkali and alka-

ine earth ions result in more tightly bonded glass network thanhat one gets in the presence of only alkali metal ions [13]. Thus,

he structural factor favor the decreased mobility of Na+ ions inhe barium borosilicate glass explaining the increase in activationnergy.

. Conclusions

The ionic conductivity of sodium borosilicate glasses has beeneasured by ac impedance technique below the glass transi-

ion temperature. The conductivity of the sodium borosilicatelasses decreases significantly with increase in BaO content upo 3.47 mol% and remains invariant thereafter. The diffusionoefficients for migration of Na+ ion have been calculated fromhe electrical conductivity and density data. The activation energyor ion transport calculated from the ln (D) versus 1/T plot for 0,.47, 7.29 and 11.53 mol% of BaO was found to be 1.00 ± 0.01 eV,.05 ± 0.01 eV, 1.09 ± 0.01 eV and 1.14 ± 0.01 eV, respectively. Thelectrical properties of the modified glass have been explained onhe basis of the structural factor.

cknowledgement

Authors thank Dr. V. Sudersan for carrying out NMR studies ofhe glass samples.

eferences

[1] C.W. Forsber, E.C. Beahm, Recovery of fissile materials from wastes and conver-sion of the residual wastes to glass, Nucl. Technol. 123 (1998) 341.

[2] A.J. Freeman, G.H. Lander (Eds.), Handbook on the Physics and Chemistry of theActinides, Elsevier, 1987, pp. 271–312.

[3] K. Raj, K.K. Prasad, N.K. Bansal, Radioactive waste management practices inIndia, Nucl. Eng. Design 236 (2006) 914–930.

[4] S.D. Misra, Proceedings National Symposium on Science & Technology ofGlass/Glass-ceramic (NGSC-06), 2006, pp. 16–21.

[5] B.P. McGrail, D.E. Day, A. Kumar, Sodium diffusion and leaching of simulatednuclear waste glass, J. Am. Ceram. Soc. 67 (1984) 463–467.

[6] R.K. Mishra, V. Sudarsan, C.P. Kaushik, Kanwar Raj, S.K. Kulshreshtha, A.K. Tyagi,Structural aspects of barium borosilicate glasses containing thorium and ura-nium oxides, J. Nucl. Mater. 359 (2006) 132–138.

[7] R.K. Mishra, V. Sudarsan, C.P. Kaushik, K. Raj, S.K. Kulshreshtha, A.K. Tyagi, Effectof BaO addition on the structural aspects and thermo physical properties ofsodium borosilicate glasses, J. Non-Cryst. Solids 353 (2007) 1612–1617.

[8] K. Funke, Jump relaxation in solid electrolytes, Prog. Solid State Chem. 22 (1993)111–195.

[9] J.L. Souquet, M. Duclot, M. Levy, Ionic transport mechanism in oxide-basedglasses in the supercooled and glassy states, Solid State Ionics 105 (1998)237–242.

10] Arpad W. Imre, Stephan Voss, Helmut Mehrer, Ionic conduction, diffusion andglass transition in 0.2 (xNa2O (1−x)Rb2O)0.8 B2O3, J. Non-Cryst. Solids 333(2004) 231.

11] J.E. Kelly, M. Tomozawa, Sodium ion transport in high purity sodium borosilicate

glasses, J. Non-Cryst. Solids 51 (1982) 345–355.

12] A. Grandjean, M. Malki, C. Simonnet, Effect of composition on ionic transportin SiO2-B2O3–Na2O glasses, J. Non-Cryst. Solids 352 (2006) 2731–2736.

13] M. Todorovic, L. Radonjic, Study of the mixed alkali effects in glasses and itsrelation to glass structure and alkali earth ion content, Ceram. Int. 15 (1989)383–388.