Evaluation of polymer inclusion membranes containing calix[4]-bis-2,3-naptho-crown-6 for Cs recovery from acidic feeds: Transport behavior, morphology and modeling studies

Journal of Membrane Science 407–408 (2012) 17–26

Contents lists available at SciVerse ScienceDirect

Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci

Evaluation of polymer inclusion membranes containing calix[4]-bis-2,3-naptho-crown-6 for Cs recovery from acidic feeds: Transport behavior, morphology and modeling studies D.R. Raut a , P. Kandwal a , G. Rebello b , P.K. Mohapatra a,∗ a b

Radiochemistry Division, Bhabha Atomic Research Centre, Mumbai 400 085, India Perkin Elmer Laboratory, Andheri, Mumbai, India

a r t i c l e

i n f o

Article history: Received 4 October 2011 Received in revised form 23 February 2012 Accepted 26 February 2012 Available online 6 March 2012 Keywords: Calix-crown Cesium Polymer inclusion membrane Morphology Radioactive waste management

a b s t r a c t Transport of cesium across polymer inclusion membrane (PIM) containing calix[4]-bis-2,3-napthocrown-6 (CNC) was investigated using acidic feed conditions. The PIMs were prepared using cellulose triacetate (CTA) as the polymer, 2-nitrophenyloctyl ether (NPOE) as the plasticizer and CNC as the carrier extractant. The studies included effects of nature of plasticizer, plasticizer concentration, CNC concentration, CTA fraction, feed acidity, and cesium concentration. Optimized membrane composition was found to be 33% CTA, 5% CNC and 62% NPOE which could result in ca. 85% Cs transport in 24 h. While significant Cs transport was seen with NPOE and TBP as the plasticizer, nearly no transport was noticed with TEHP (tris-2-ethylhexyl phosphate) and di-n-octyl phthalate (DOP). The transport behavior of Cs(I) was correlated with membrane morphology using Transmission Infrared Mapping Microspectroscopy (TIMM) plots and AFM profiles. Selectivity studies were also carried out which were found to be quite encouraging. A mathematical model was developed to simulate the transport process of Cs(I) ion from feed side to the strip side. © 2012 Elsevier B.V. All rights reserved.

1. Introduction 137 Cs

(half-life: 30.1 y) is one of the important fission products present in the nuclear waste and responsible for the MANREM problems associated with radioactive waste management due to the 662 keV gamma ray emission. There is, therefore, a strategy evolving for the separation of radio-cesium prior to the vitrification of the high level waste (HLW). Moreover, the separated radio cesium can be used as a potential gamma ray source in place of 60 Co for different purposes like sterilization of foods, sterilization of medical accessories as well as sewage sludge treatment [1]. The separation of radio cesium from acidic radioactive waste solutions is achieved using three classes of reagents, viz., ion-exchangers such as AMP coated resins, cobalt dicarbollides, and macrocyclic compounds such as crown ethers and calix-crown derivatives [2–7]. Out of these, AMP based separations are really hard to be reversible while solvent extraction methods using cobalt dicarbollides use corrosive or hazardous diluents like nitrobenzene and FS-13 [8–10]. Crown ether based separations are attractive from the point of view of facile back extraction of Cs(I) using distilled water as the strippant [11]. The selectivity of a crown ether is derived by the cavity fit-

∗ Corresponding author. Tel.: +91 22 25594576; fax: +91 22 25505151. E-mail address: [email protected] (P.K. Mohapatra). 0376-7388/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.memsci.2012.02.050

ting parameter and cation–pi electron cloud interactions and hence can be tuned by synthesizing molecules of the required cavity size. Calix-crowns are reported to be more efficient extractants due to their high complexation ability with Cs+ ion and display very high selectivity in the presence of large concentrations of other alkali metal ions [12]. Separation of Cs(I) from alkaline nuclear waste solutions has been extensively studied using calix[4]arenebis(tert-octylbenzo-crown-6), better known as BOBCalixC6 [13,14]. However, degradation of this reagent in acidic conditions limits its use for Cs(I) recovery from the acidic wastes such as the high level waste (HLW) [15]. We have studied several commercial calix-bis-crown-6 ligands for Cs(I) recovery, out of which calix[4]-bis-2,3-naptho-crown-6 (CNC, Fig. 1) was found to be the most efficient [16]. Crystal structure and molecular dynamics calculations have revealed its excellent selectivity (1:45,000) over Na(I) [17]. The favorable cation–pi interactions between the cation and phenyl rings of ligand led to selective binding of the metal ion [18]. Using solvent extraction techniques we have demonstrated the selective extraction of Cs(I) employing the CNC as the extractant from acidic waste solutions with nitrobenzene–toluene mixture as the diluent [19]. Its radiation and chemical stability were found to be satisfactory suggesting its potential for long term applications [20]. Due to the concern for the environment and possibility of using very low inventory of the extractant, liquid membrane based

18

D.R. Raut et al. / Journal of Membrane Science 407–408 (2012) 17–26

room temperature. The resulting PIM was peeled out by spreading a few mL of water on it and was subsequently used for the transport studies. Thickness of the membranes was measured using digimatic micrometer (Mitutoyo Corporation, Japan). 2.3. Transport studies of Cs(I)

Fig. 1. Structural formula of calix[4]-bis-2,3-naphtho-crown-6 (CNC).

separation methods are considered viable alternatives to the solvent extraction methods [21–23]. We have carried out systematic investigations on the Cs transport using CNC as the carrier in flat sheet supported liquid membranes [24,25]. The flux, however, was low resulting in slow mass transfer rates. It was required, therefore, to evaluate alternative membrane separation methods such as polymer inclusion membranes which display high stability [26]. Schow et al. [27] reported higher fluxes for PIMs for K+ ion using a crown ether as carrier than those for flat sheet supported liquid membranes. Miguel et al. [28] reported quantitative recovery of Au(III) using polymer inclusion membranes from concentrated hydrochloric media using Kelex-100 (7-(4-ethyl-1-methyloctyl)8-hydroxyquinoline) as carrier. The transport mechanism in PIMs depends on factors like membrane composition, homogeneity of the membrane and surface morphology. Therefore, it is required to investigate the effects of membrane morphology on the transport rates. In the present study, we have prepared several PIM compositions containing cellulose triacetate (CTA) as the polymeric material and calix[4]-bis-2,3-naptho-crown-6 as the carrier extractant. Several plasticizers such as TBP, NPOE, DOP and TEHP were evaluated and the Cs(I) transport rates were measured. Other transport studies included carrier concentration variation, CTA concentration variation, and effect of feed acidity. Selectivity and long term stability studies were also carried out. Role of membrane homogeneity and morphology on the Cs(I) transport rates were also investigated by carrying out TIMM (Transmission Infrared Mapping Microspectroscopy and AFM (Atomic Force Microscopy) measurements. Mathematical modeling of the transport data was also done and compared with the experimentally obtained data points. 2. Experimental 2.1. Materials Calix[4]-bis-2,3-naptho-crown-6 (CNC) was purchased from Acros Organics, Belgium. 137 Cs tracer was procured from Board of Radiation and Isotope Technology (BRIT), Mumbai, and its radiochemical purity was ascertained by gamma ray spectrometry employing high resolution HPGe detector. 2-Nitrophenyloctyl ether (Fluka), tris-2-ethylhexyl phosphate (Fluka), tri-n-butyl phosphate (Aldrich), di-n-octyl phthalate (Aldrich) and cellulose triacetate (CTA) were used as obtained. All the other reagents used were of analytical reagent grade. 2.2. Preparation of PIM The PIMs were prepared using previously reported literature methods [27,29]. A mixture of cellulose triacetate (CTA), extractant (CNC) and a plasticizer (TBP/NPOE/DOP/TEHP) was dissolved in dichloromethane and homogenized by sonication. The solution was poured into a flat petri dish and was allowed to evaporate at

The Cs(I) transport experiments were carried out by using a twocompartment Pyrex glass cell as reported earlier [29]. Volumes of the feed and the receiver compartments were 20 mL each and the exposed membrane area was 4.94 cm2 . The feed solution for these studies contained 1 M HNO3 while distilled water was used as the strippant. Equal volumes of the feed (spiked with 137 Cs tracer) and receiver phase were transferred into the respective compartments and were stirred at 200 rpm by synchronous motors. Assay of 137 Cs in the feed as well as in the receiver phase was carried out at different time intervals to calculate the transport parameters. Counting was done by the NaI(Tl) detector and High Purity Germanium (HPGe) detector both coupled to multi-channel analyzers. All the transport studies were carried out at ambient temperature (25 ± 1 ◦ C). The material balance in these studies was found to be within ±5%. 2.4. Membrane characterization 2.4.1. TIMM measurements Transmission Infrared Mapping Microspectroscopy (TIMM) plot measurements were performed by using a PERKIN ELMER auto image microscope coupled with FTIR. Scanning was performed using 25 ␮m × 25 ␮m aperture areas with 16 cm−1 resolution having intervals of 8 cm−1 using 2 scans per point in the region, 4000–720 cm−1 . Membranes made of CTA served as blanks and all measurements were baseline corrected with the 4000 cm−1 line as the reference. The distribution of the components was measured using the following bands: NPOE: 1528 cm−1 for (asym. NO2 stretching), TEHP: 1943 cm−1 (P O stretching), DOP: 1730 cm−1 (C O stretching), TBP: 1899 cm−1 (P O stretching), CNC: 2356, 1198 cm−1 (C H bending, C H stretching, respectively), unless mentioned otherwise. 2.4.2. Atomic force microscopy Surface morphology of the polymer inclusion membranes was characterized by atomic force microscopy (AFM). AFM measurements were carried out in contact mode using a scanning probe microscope (SPM-Solver P47, NT-MDT, Russia). Rectangular cantilevers of silicon nitride having force constant of 3 N/m was employed for measurement. The membrane surfaces were compared by means of roughness parameters, such as maximum height (Rmax ), mean height (Rmean ), mean roughness (Ra ) and root mean square (rms) of the z data (Rq ) [30]. 2.5. Mathematical modeling of Cs transport data A mathematical model was developed to simulate the transport process of Cs(I) in a manner similar to that reported in a recent publication [25]. The model was developed based on carrier (ligand) mass balance in the membrane phase. Distribution ratio values were calculated by equilibrating a known volume and weight of piece of PIM with certain volume of nitric acid (1 M) containing 137 Cs. The membrane phase diffusivity (o ) was calculated by time-lag experiment and used to predict the transport data. It is assumed that the facilitated transport of the Cs(I) takes place through the ion-pair mechanism similar to that of supported liquid membranes as given below, Kex

Cs+ (a) + NO3 − (a) + L(o) (Cs · L)+ · NO3 − (o)

(1)

D.R. Raut et al. / Journal of Membrane Science 407–408 (2012) 17–26

19

Considering strip reaction to be instantaneous C¯ is∗ can be assumed to be 0 and the carrier mass balance inside the membrane phase leads to the following equation [25]: ko−1 DCBf,t Kd ka−1 + ko−1

+

Kd − ET = 0 [(Xfo − CBf,0 ) + CBf,t ]Kex

(7)

where ko and ka are the reciprocals of o (Do /do ) and a (Da /da ), respectively. Do is the diffusion coefficient of the metal–carrier complex inside the membrane while Da is the diffusion coefficient of the metal ion in the aqueous feed phase. The term da is the thickness of the aqueous stagnant film, while do is the membrane thickness, o is the membrane phase diffusivity. Xfo is the initial nitrate ion concentration, A is the effective membrane area, V is the volume of the feed solution, Kd is the distribution ratio at feed membrane side (though D is the conventional notation for distribution ratio, Kd is used to avoid confusion between diffusion coefficient and the distribution ratio). The equation giving the rate of change of metal ion concentration with respect to time can be given similar to the flat sheet SLM modeling equation given as [25], Fig. 2. Schematic presentation of the PIM transport process.

dCBf,t

where the species with the subscripts (a) and (o) refer to those in the aqueous and the organic phases, respectively while Kex refers to the two-phase extraction equilibrium constant defined as, Kex =

[(Cs · L)+ · NO3 − ](o) [Cs+ ](a) [NO3 − ](a) [L](o)

(2)

The membrane flux can be derived by applying Fick’s diffusion law with few simple assumptions, viz. (a) the composition of the strip solution is such that the complex of Cs(I) and CNC is instantaneously dissociated at the membrane–aqueous strip interphase, (b) linear concentration gradient in the boundary layer and membrane phase, (c) there is no extraction of Cs(I) by the pure CTA/NPOE and mixture of both. Fig. 2 presents the PIM transport profile. Assuming that the transport of Cs(I) occurs under steady state conditions, the concentration gradients are linear, and feed concentration of NO3 − is constant, the equations describing the aqueous flux of metal ion is given by Eq. (3): Ja = ka (CBf − Cif )

(3)

where Ja is the feed phase flux, ka is the aqueous mass transfer coefficient, C¯ if is the metal ion concentration in the membrane phase at feed–membrane interface. The membrane flux of metal–carrier complex is given by Eq. (4): Jo = ko (C¯ if − C¯ is∗ )

(4)

where Jo is the membrane flux, ko is the membrane mass transfer coefficient, while C¯ is∗ is the metal ion concentration in the membrane phase at strip–membrane side interface. If volume of the feed does not change significantly during the experiment, then one gets the expression



ln

CBf,t



CBf,0



=−

A Vf



Pt

(5)

A is the geometrical surface area. CBf,t and CBf,0 refer to the concentration of the metal ion in bulk feed at a given time ‘t’ and at the start of the experiment. The permeability coefficient (P) values were calculated using Eq. (5). The cumulative percent transport (%T) at a given time is determined by the following equation, %T = 100 ×

CBf,0 − CBf,t Cf,0

(6)

The term (CBf,0 − CBf,t ) was found to be nearly equal to the concentration of the metal ion in the receiver phase for the given data point.

dt

=−

A V



Kd CBf,t



Kd ka−1 + ko−1

(8)

The above two equations were solved using the known transport parameters (Kex and D), the prediction of the Cs(I) transport rate has been plotted at three different CNC concentrations. The Kex value calculated was found to be 1428 by carrying out Cs(I) extraction studies from 1 M HNO3 feed solution using a 1 cm2 PIM containing the carrier extractant and measuring the Kd value after attainment of equilibrium. The membrane phase diffusivity (Do ), calculated by time-lag experiment using the relation Do = (dm )2 /6tlag , where dm = membrane thickness (in cm), tlag = time lag (in s), were used to predict the transport data [31]. As the transport occurs by the ion pair mechanism, the aqueous diffusion layer thickness was taken from our previous work for Cs(I) transport using CNC as the carrier ligand [25]. 3. Results and discussion 3.1. Effect of plasticizer type and concentrations Plasticizer plays a vital role in the transport of metal ions as well as for membrane softness and flexibility in the PIM. The plasticizer provides the elasticity and constitutes a fluid phase in which the metal ion/complexed species can diffuse. It penetrates between the polymer molecules and neutralizes the polar groups of polymer by interaction with its own polar groups or merely increases the distances between the polymer molecules and reduces the strength of intermolecular forces [32]. In the present study, DOP, TEHP, TBP and NPOE were evaluated as plasticizers. DOP, TEHP, TBP have long alkyl chains but with less polar groups while NPOE has long alkyl chains as well as has high dielectric constant increasing the polar nature. In present study, PIMs were prepared with these plasticizers in varying amounts (50–150 mg), with CNC (4 mg) as a carrier ligand and CTA (80 mg) as the polymeric material. PIMs generated with the plasticizers were transparent with uniform thickness. Table 1 presents the transport data obtained with the PIMs made from different plasticizers. With TEHP as the plasticizer, poor Cs(I) transport (2.27% after 24 h) was observed which did not change with the TEHP content (up to 200 mg). On the other hand, acid co-transport was significant (30% after 24 h) which may be one of the reasons for the poor Cs(I) transport rate. In the case of DOP, both Cs(I) (3.86% after 24 h) and acid transport (2.5% after 24 h) were insignificant while with TBP as the plasticizer, significant amounts of Cs(I) (33.5% after 24 h) and acid transport (50% after 24 h) were observed. On the other hand, NPOE was found to be the

20

D.R. Raut et al. / Journal of Membrane Science 407–408 (2012) 17–26

Table 1 Correlation of physical parameters of different plasticizers with Cs(I) and acid transport in PIMs, CNC = 4 mg, CTA = 80 mg, plasticizer = 150 mg, feed: 1 M HNO3 , receiver: distilled water. Plasticizer

% Cs(I) transport (24 h)

% acid transport (24 h)

Viscosity (mPa s)

Dielectric constant

TEHP DOP TBP NPOE

2.27 2.26 33.49 36.80

30.0 2.5 33.5 1.0

12 41.5 3.39 13

4.8 5.0 8.0 24.8

most suitable plasticizer with 42% of Cs(I) transport and only ∼1% acid transport after 24 h. Experiments carried out with PIMs containing only the plasticizers (no CNC) revealed insignificant Cs(I) transport (<0.1%) while acid transport was comparable to PIMs containing CNC suggesting acid transport is due to the dual role played by the plasticizer (as plasticizer as well as carrier). Moreover, PIM with the carrier ligand (4 mg CNC) but without plasticizer resulted in no Cs(I) transport while acid transport was also very low (∼5%). The reason behind this different Cs(I) transport behavior is related to the viscosity and the dielectric constant of the plasticizer. The diffusion of metal–carrier complex inside the membrane is the rate determining step in the overall transport process and is inversely proportional to the viscosity of the medium as per the Stokes–Einstein equation: kT 6˘r

3.3. Effect of CNC concentration on Cs(I) transport Fig. 5 presents the transport data obtained using different amounts of CNC where 150 mg of NPOE was used along with 80 mg of CTA for casting of the membranes. As shown in the figure, the rate of transport increased with increasing CNC concentration up -5

7.0x10

-5

6.0x10

(9)

where Do is the diffusion coefficient of the metal–carrier complex in the solvent (viz., NPOE) at low solution concentration, k is Boltzmann constant,  is the viscosity (mPa s) and r is the radius of solute. Less permeation of Cs(I) observed with DOP, TEHP can be explained on this basis. Table 1 also lists the physical parameters of the plasticizers used in PIM with the % Cs(I) transport obtained. It is clear that, though NPOE has similar viscosity to that of TEHP, promising transport of Cs(I) with the former plasticizer was observed because of its favorable dielectric constant. In high dielectric constant media, the formation of the extractable ion-pair species will be more favorable [33]. Thus, NPOE with higher dielectric constant than TEHP favors Cs(I) extraction resulting in better Cs(I) transport. This was consistent with our earlier studies [29] and the work carried out by Sugiura et al. [34]. Further, variation in the amount of NPOE gives higher permeability up to 150 mg of NPOE which might be due to the favorable plasticization effect (Fig. 3) beyond which a decrease in transport was seen which was attributed to intermolecular interactions resulting in lower mass transfer.

-5

5.0x10

P (cm/s)

Do =

transport rates. Thus, polymer inclusion membranes prepared with 80 mg of CTA (34.2% CTA) were optimum for transport studies and were used in major part of the present study.

-5

3.0x10

-5

2.0x10

-5

1.0x10

35

40

45

50

55

60

65

70

75

NPOE content (%) Fig. 3. Effect of NPOE content on the permeability of Cs. PIM composition: CNC = 4 mg, CTA = 80 mg, feed: 1 M HNO3 , receiver: distilled water.

6.0x10-5

3.2. Effect of polymer concentration on Cs(I) transport

5.0x10-5

-1

P (cm.s )

The polymer plays a crucial role on membrane performance by providing mechanical strength [34,35]. Studies were carried out using PIMs made with varying amounts of CTA and fixed concentrations of CNC (4 mg) and NPOE (150 mg). Four PIMs were prepared containing 20.6%, 34.2%, 43.8% and 51.0% of CTA and transparent membranes were formed with 0.026, 0.032, 0.048 and 0.058 cm as the mean thicknesses, respectively. Cs(I) transport studies were carried out and permeability coefficient values were determined (Fig. 4). As shown in the figure, the P values showed a decreasing trend with increasing CTA fraction which is a direct consequence of the increasing membrane thickness. Similar observations have been made while studying Cs transport from acidic feeds using CTAbased PIMs containing crown ethers as the carrier extractant [36]. Though higher Cs permeability was observed with PIM containing 20.6% CTA, it was not used in the subsequent studies as the membranes were found to have inadequate stability with acidic feeds. On the other hand higher CTA fraction resulted in slow Cs(I) ion

-5

4.0x10

4.0x10-5

3.0x10-5

2.0x10-5 20

25

30

35

40

45

50

55

%CTA in PIM Fig. 4. Effect of CTA content on the permeability of Cs. PIM composition: CNC = 4 mg, NPOE = 150 mg, feed: 1 M HNO3 , receiver: distilled water.

D.R. Raut et al. / Journal of Membrane Science 407–408 (2012) 17–26

% Cs Transported

60

Table 2 Effect of H+ ion concentration at fixed nitrate ion concentration on the transport behavior of Cs(I), PIM composition: CNC = 4 mg, CTA = 80 mg, NPOE = 150 mg, receiver: distilled water.

1 mg CNC 4 mg CNC 8 mg CNC 10 mg CNC 12 mg CNC 14 mg CNC

40

21

Feed solution

% Cs transport (24 h)

P × 104 (cm/s)

1 M HNO3 a 2 M HNO3 + 1 M NaNO3 1 M HNO3 + 2 M NaNO3 3 M NaNO3

36.80 42.52 68.32 10.44

0.520 0.566 0.720 0.137

a

20

± ± ± ±

0.06 0.06 0.05 0.02

1 M HNO3 data are included for comparison purpose.

3.4. Effect of hydrogen ion concentration on Cs(I) transport

0

200

400

Time (minutes) Fig. 5. Transport of Cs(I) as a function of CNC content, PIM composition: CTA = 80 mg, NPOE = 150 mg, feed: 1 M HNO3 , receiver: distilled water.

to 12 mg of CNC beyond which a decrease in the Cs(I) transport was seen. The increase in Cs(I) transport with increasing CNC concentration is straightforward and can be explained on the basis of the extraction equilibrium (Eq. (1)). Though about 45% Cs transport was seen with 12 mg of CNC, it increased to >85% Cs(I) transport after 24 h which is quite promising from their application point of view. The permeability coefficient value for this particular concentration was (1.84 ± 0.05) × 10−4 cm/s which was quite comparable with the permeability coefficient values ((1.88 ± 0.05) × 10−4 cm/s) obtained in SLM though higher CNC concentration was needed in PIM for quantitative transport of Cs(I) in 24 h [24]. (Note: CNC concentration for SLM was 5 × 10−4 M in 80% NPOE in n-dodecane while feed acidity was kept at 3 M HNO3 .) The optimized membrane composition of PIM made with 12 mg CNC was found to be 33% CTA, 5% CNC and 62% NPOE. On the other hand, decrease with increasing CNC concentration may be due to possible interactions between the plasticizer and carrier. Similar decrease in Cs(I) permeability was also reported by us in SLM studies using NPOE as the diluent. Moreover, a decrease in the % transport at higher CNC concentration may be due to the crowding of the carrier extractant (CNC) molecules inside the membrane. In transport studies involving PIMs, a minimum amount of the carrier ligand for transport to occur is referred to as the ‘percolation threshold’. Miguel et al. have observed such ‘percolation threshold’ value for the gold transport through the PIM using Kelex-100 as the carrier ligand which reveals the ‘fixed site jumping’ mechanism for metal ion transport [28]. The presence of a percolation threshold in PIMs has also been reported for Cd(II) transport with Lasalocid A and for Pt(IV)transport with Aliquat 336 when both systems used NPOE as the plasticizer [37,38]. However, in the present case, transport observed with even 0.5 mg of CNC ruled out the possibility of percolation theory and confirms the diffusion of the ion-pair complex given above (Eq. (1)). Interestingly, in all cases, acid transport was ≤1% suggesting the effectiveness of NPOE as the plasticizer as compared to the others. Acid transport observed here, was due to the extraction of H+ ion as the CNC–hydronium ion complex which competes with the Cs–CNC complex for transport [24]. However, the hydrogen ion competes more strongly at higher acidities than at 1 M acidity. In the light of the literature and our earlier work with CNC, it could be concluded that at lower acidity the extracted species was [CsL]+ ·[NO3 ]− . This point was verified by observing membrane fluxes at fixed nitrate ion concentration at variable H+ ion concentration (vide infra).

As evident from Table 2, the permeability coefficient at 1 M HNO3 as the feed was (0.552 ± 0.06) × 10−4 cm/s which does not vary significantly at 2 M HNO3 + 1 M NaNO3 (P = (0.566 ± 0.06) × 10−4 cm/s). On the other hand, at 1 M HNO3 + 2 M NaNO3 as the feed, higher permeability coefficient for Cs(I) transport was obtained while when 3 M NaNO3 was used as the feed, the permeability coefficient values decreased drastically (P: (0.137 ± 0.02) × 10−4 cm/s). Apparently, two effects are seen, viz., higher nitrate ion concentration facilitates Cs+ transport; higher Na+ or H+ concentration facilitates the transport of these ions which affects the Cs+ transport rates. Therefore, an optimum balance of Na+ , H+ and NO3 − concentration is required for Cs+ ion concentration. Similar observations were made in SLM studies by us with CNC in NPOE-n-dodecane as the carrier solvent which confirms the mechanism in PIM was similar to that of SLM [24]. 3.5. Selectivity studies The selective transport of Cs(I) with respect to other fission products was investigated by irradiating a natural uranium target in APSARA reactor at a thermal neutron flux of 1012 n cm−2 s−1 . Subsequently, the irradiated target was dissolved in 1 M HNO3 and the fission product mixture was spiked with 137 Cs tracer. The transport of various fission products was monitored after 24 h in the strip solution of distilled water using a PIM containing 4 mg of CNC. Fig. 6 represents the recorded gamma ray spectra of the fission products. It was observed that fission products such as 143 Ce, 140 La, 140 Ba, 103 Ru, 99 Mo, 99m Tc, 97 Zr and 91 Sr were not transported to any significant extent by the PIM while 137 Cs transport was clearly seen along with a small amount of 131 I suggesting the selective transport of the cesium. Duhart et al. [39] have also reported very high Cs(I) selectivity while using a fixed site membrane containing unsymmetrical calix[4]arenebiscrown-6 bonded to an immobilized polysiloxane framework. The transport of Cs(I) was comparatively low, as only 42% Cs was transported in 24 h. However, as shown above, the

Counts (for 5 minutes)

0

4000

Strip-24 Hrs

2000

137

Cs

0 4000

Feed-0 Hrs

137

Cs

2000 0 0

200

400

600

800

1000

1200

Energy KeV Fig. 6. Selectivity of the PIM, feed: irradiated uranium target dissolved in 1 M HNO3 , receiver phase: distilled water, membrane composition: CNC = 4 mg, CTA = 80 mg, NPOE = 150 mg.

22

D.R. Raut et al. / Journal of Membrane Science 407–408 (2012) 17–26

Table 3 Effect of Cs ion variation on permeability coefficient of Cs(I), composition of PIM: CNC = 12 mg, CTA = 80 mg, NPOE = 150 mg, feed: 1 M HNO3 , receiver: distilled water. Feed solution, Cs (g/L)

%T at 300 min

P × 104 (cm/s)

0.01 0.02 0.03 0.05

42.36 50.40 44.43 35.77

1.97 2.59 2.20 1.52

± ± ± ±

0.06 0.05 0.04 0.01

transport rates can be enhanced by using higher concentration of the carrier. 3.6. Effect of cesium concentration The HLW contains varying amounts of Cs (mixture of isotopes) depending on the nature of the fuel as well as burn up. A typical PHWR fuel with a burn up of 6500 MWd/t has about 0.32 g/L of Cs in the HLW. However, considering the lower transport rates at higher cesium loading conditions, the Cs concentration in feed compartment was varied from 0.01 g/L up to 0.05 g/L. The data are presented in Table 3. An increase in the permeability coefficient ((2.59 ± 0.05) × 10−4 cm/s) was observed up to 0.02 g/L (∼88% Cs(I) transport at 24 h) beyond which a decrease in transport rate was seen (P was (1.52 ± 0.01) × 10−4 cm/s for 0.05 g/L Cs). Though the overall changes in P values are not significant, there is a decrease in the Cs transport rate when the CNC:Cs ratio was lowered. As shown below (Section 3.8), there was significant deviation in the mass transfer modeling when the CNC concentration was low. The exact reason for this is not understood and could be based on a threshold value for CNC:Cs ratio to result in favorable mass transfer. 3.7. Membrane morphology studies Morphology of the PIM has a very important role on its transport properties [26,30]. The plasticizer which acts as a solvent in the PIM helps in getting a homogeneous distribution of the carrier

extractant and also making the membrane softer. It also helps in reducing the repulsive interactions between the polymer matrices and hence facilitates the carrier transport. On the other hand, Hbonding interactions between the plasticizer and the carrier can hinder the movement of the carrier molecules and hence can possibly result in what is known as ‘fixed site jumping’ mechanism of cation transport which often shows a ‘percolation threshold’. In the present studies, the membrane homogeneity and morphology were studied by FTIR TIMM measurements and also by AFM. 3.7.1. TIMM measurements Fig. 8 presents the TIMM plots obtained for PIMs made from different plasticizers with CTA as the polymeric matrix and CNC as the carrier. For CTA + TEHP system, the TEHP distribution shows a flat profile for its characteristic wavelength 1465 cm−1 (Fig. 7(a)) indicating homogeneous distribution of the plasticizer in the PIM. In case of all the other plasticizers systems (CTA + DOP, CTA + TBP, CTA + NPOE), the distribution of plasticizers was non-uniform as revealed by the TIMM plots in Fig. 7(b), (c) and (d). This was due to the lower affinity of the plasticizers for the CTA. Fig. 8 shows the TIMM plots for four different membranes where distribution of the plasticizer and CNC was obtained. (Note: Distribution profiles were obtained at characteristic wavelength which was not interfering with other component wavelength. Further, due to three component system, characteristic wavelength may be different which was confirmed by the merging the FTIR spectrum of individual components.) As shown in Fig. 8(a) and (b), the distribution of TEHP and CNC was relatively uniform in PIMs containing CTA + TEHP + CNC, suggesting affinity of the polymer matrix for CNC as well as TEHP. This could result in restricted transport of the metal ion. Though similar trend was also expected for hydrogen ion transport, significant acid transport was noticed which was attributed to TEHP acting as a carrier (in addition to its role as a plasticizer). For CTA + DOP + CNC system, though DOP distribution was uniform (Fig. 8(c)), a non-uniform CNC distribution (Fig. 8(d)) indicated possibilities of the Cs(I) transport. However, poor transport rates were

Fig. 7. TIMM distribution profiles of PIMs components in the absence of CNC with 150 mg of plasticizer. (a) TEHP; (b) DOP; (c) TBP; (d) NPOE.

D.R. Raut et al. / Journal of Membrane Science 407–408 (2012) 17–26

23

Fig. 8. TIMM distribution profiles of PIM components in the presence of CNC. (a)TEHP; (b) TEHP; (c) DOP; (d) DOP; (e) TBP; (f) TBP; (g) NPOE; (h) NPOE. All PIMs contain 150 mg of the plasticizer and 80 mg of CTA and 4 mg of CNC. For NPOE plasticized PIMs 12 mg CNC was used. (b), (d), (f) and (h) refer to CNC distribution.

attributed to the high viscosity of the medium. In CTA + TBP + CNC system, distribution of both TBP (Fig. 8(e)) and CNC (Fig. 8(f)) was non uniform indicating the non affinity of the polymer matrix for both CNC and TBP. This was reflected in transport of both Cs(I) and acid to a significant degree. However, in view of high concentration of TBP, the acid transport rates were dominating and may be the reason for a relatively lower Cs(I) transport. Though similar non-uniform patterns were obtained for the CTA + NPOE + CNC system with both NPOE (Fig. 8(g)) and CNC (Fig. 8(h)), reasonably good Cs(I) transport was seen with very low acid transport. This is attributed to inefficient acid transport of NPOE as compared to phosphate plasticizers such as TBP and TEHP. The NPOE system also

indicates no percolation threshold which is in conformity with the carrier concentration variation studies mentioned above. 3.7.2. Surface morphology data from AFM Surface morphology of the plasticized PIMs containing CNC was studied using AFM with TBP and NPOE as the plasticizers [30]. The PIMs with DOP and TEHP were not investigated as the Cs(I) transport rates were poor in these membranes. The AFM profiles are shown in Fig. 9 and the roughness parameters are listed in Table 4. The roughness parameters are higher in the TBP plasticized membrane as compared to the NPOE plasticized one. This indicates that in case of NPOE plasticized membrane, the height

24

D.R. Raut et al. / Journal of Membrane Science 407–408 (2012) 17–26

Table 4 AFM data obtained for different PIMs. Composition of PIM: 4 mg CNC + 80 mg of CTA + 150 mg of plasticizer. Plasticizer

Surface area (nm2 )

Rmax (nm)

Rmean (nm)

Roughness Ra (nm)

Root mean-square (of Ra ) Rq (nm)

TBP NPOE

7081.7 × 7094.8 7068.2 × 7072.9

73.379 40.810

37.553 15.845

6.304 4.743

9.041 6.030

Fig. 9. AFM pictures obtained for PIMs with (a) TBP and (b) NPOE as the plasticizers. Composition of PIM: 4 mg CNC + 80 mg of CTA + 150 mg of plasticizer. Mean roughness parameters are indicated below the respective AFM images.

1

0.95

0.9

Ct / C0

difference between the highest and lowest points in the surface (Rmax ) and standard deviation in z-values are less and hence forms relatively smoother surface. This was also indicated in their transport properties. Usually, a rough PIM should facilitate higher transport rates. In the present case, though higher Cs(I) transport rates are expected with TBP plasticized PIMs, competitive acid transport to the receiver compartment leads to inefficient Cs(I) stripping resulting in effectively lower Cs(I) transport rates than expected. On the other hand, NPOE plasticized PIMs appear better as the acid transport rates are significantly lower.

0.85

3.8. Mathematical modeling of transport data 0.8

Based on the developed mathematical model and the known transport parameters, simulation of the transport process was carried out at three CNC concentrations viz., 4 mg, 8 mg and 10 mg and the data are presented in Fig. 10. The data used for the mathematical modeling are listed in Table 5. The lines represent the transport profiles as predicted by the mathematical model described above (Section 2.5) while the symbols represent the experimentally obtained data. The predicted data found to be matching reasonably well (within ±5%) with those obtained from the transport studies for PIMs containing 10 and 8 mg CNC. On the other hand, significant deviation was noted when PIM containing 4 mg CNC was used.

0.75

0.7

0

50

100

150

200

250

300

350

400

Time (min) Fig. 10. Prediction of Cs(I) transport (indicated by the lines in the figure) by the presented model for the variable CNC concentration in the PIM and actual (indicated by the symbols in the figure) transport rate observed. CTA = 80 mg, NPOE = 150 mg. (䊉): 4 mg CNC; (*): 8 mg CNC; (): 10 mg CNC; (—): 4 mg CNC; (. . .. . .): 8 mg CNC; ): 10 mg CNC. (

D.R. Raut et al. / Journal of Membrane Science 407–408 (2012) 17–26 Table 5 Different parameters of PIM used in the present study for the mathematical modeling for Cs(I) transport and calculation of membrane phase diffusivity. Parameters

Description

Effective membrane area (A) Membrane thickness Initial nitrate concentration [N]0 (M) ko (cm/s) ka (cm/s) Carrier (CNC) concentration [E] Kex

4.94 cm2 0.032 cm 1 4.43 × 10−6 2.67 × 10−4 Variable 1428

Higher Cs transport rates can only be explained by possible complexation of two Cs+ ions with one CNC carrier molecule. Predicted line with 12 mg CNC also showed very large deviation from the experimental data points (not shown in the figure). At higher CNC concentrations, may be the crowding of ligands inside the membrane may cause change in membrane parameters like diffusivity of species affecting the transport process. Similar modeling of supported liquid membrane data has shown very good agreement with the predicted data [25] suggesting a change in the transport mechanism in case of PIMs. This change in transport mechanism may be responsible for reduced permeability of Cs(I) ion at higher CNC concentration as reported above (Section 3.3). 4. Conclusions More than 85% Cs(I) transport was achieved using PIM containing 12 mg of CNC as the carrier ligand. NPOE was found to be the optimum plasticizer for quantitative transport to occur with only 1% of acid transport. TBP as a plasticizer shows the high acid transport which affects Cs transport. The characterization of membranes with FTIR microscopy gives useful insights for plasticizer and CTA interactions. Increase in the carrier ligand concentration in PIM compositions led to increase in the Cs(I) transport. However, permeability coefficient value decreased as the function of increased cesium loading in the feed side. At feed side compositions, a constant H+ ion concentration with variable NO3 − increases the Cs(I) transport due to availability of more nitrate ion for ion-pair formation. Selectivity studies carried out with irradiated uranium target dissolved in 1 M HNO3 revealed excellent selectivity for Cs(I) ion over different elements. The membrane shows promises for recovery of Cs(I) from nuclear waste at lower acidity. Acknowledgements The authors thank Dr. A. Goswami, Head, Radiochemistry Division, for his keen interest in this work. They also thank Dr. P.A. Hassan, Chemistry Division, BARC for help in recording AFM profiles.

Nomenclature Abbreviations AFM atomic force microscopy ammonium molybdophosphate AMP BRIT Board of Radiation and Isotope Technology calix[4]-bis-2,3-naphtho-crown-6 CNC CTA cellulose triacetate dioctylphthalate DOP HLW high level waste NPOE 2-nitrophenyloctyl ether polymer inclusion membrane PIM TBP tri-n-butyl phosphate TEHP tris-2-ethylhexyl phosphate

25

TIMM Transmission Infrared Mapping Microspectroscopy MANREM exposure to personnel Symbols CBf,0 initial concentration of metal ion in bulk feed phase CBf,t concentration of metal ion in bulk feed phase at time t concentration of metal ion in the membrane phase C¯ if at feed–membrane interface concentration of metal ion at membrane–strip C¯ is∗ interface (in membrane) D distribution ratio of Cs Da diffusion coefficient of the metal ion in aqueous phase diffusion coefficient of the metal carrier complex Do inside the membrane da thickness of the aqueous stagnant film membrane thickness do [E]T total extractant concentration Jo membrane flux Kex two-phase extraction equilibrium constant membrane mass transfer coefficient ko ka aqueous mass transfer coefficient permeability coefficient P tlag lag time Greek letters aqueous phase resistance a o membrane phase resistance  dynamic viscosity

References [1] IAEA Technical Data Series No. 356, IAEA, 1993. [2] R.L. Miller, A.B. Pinkerton, P.K. Hurlburt, K.D. Abney, Extraction of Cs and Sr in to hydrocarbon solvent using-tetra-cobalt dicarbolide, Solvent Extr. Ion Exch. 13 (1995) 813–827. [3] T.J. Tranter, T.A. Vereshchagina, V. Utgikar, An inorganic microsphere composite for the selective removal of 137 cesium from acidic nuclear waste solutions. 1. Equilibrium capacity and kinetic properties of the sorbent, Solvent Extr. Ion Exch. 27 (2009) 199–218. [4] D. Banerjee, M.A. Rao, J. Gabriel, S.K. Samanta, Recovery of purified radiocesium from acidic solution using ammonium molybdophosphate and resorcinol formaldehyde polycondensate resin, Desalination 232 (2008) 172–180. [5] L.H. Gerow, J.E. Smith Jr., M.W. Davis Jr., Extraction of Cs and Sr from HNO3 medium using macrocyclic polyethers, Sep. Sci. Technol. 16 (1981) 519–548. [6] V.S. Talanov, G.G. Talanov, M.G. Gorbunova, R.A. Bartsch, Novel cesium selective, 1,3-alternate calix[4]arene(bis-crown-6 ethers) with proton ionizable groups for enhanced extraction efficiency, J. Chem. Soc., Perkin Trans. 2 (2002) 209–215. [7] R. Ungaro, A. Casnati, F. Ugozzoli, A. Pochini, J.-F. Dozol, C. Hill, H. Rouquette, 1,3-Dialkoxycalix[4]arenecrowns-6 in 1,3-alternate configuration: cesium-selective ligands that exploit cation–arene interactions, Angew. Chem. Int. Ed. Engl. 33 (1994) 1506–1509. [8] J. Rias, P. Selucky, M. Kyrs, Extraction of alkali metals into nitrobenzene in the presence of univalent polyhedral borate anions, J. Inorg. Nucl. Chem. 38 (1976) 1376. [9] S.D. Reilly, C.P. Mason, P.H. Smith, Cobalt(III)dicarbollide A Potential 137 Cs and 90 Sr Waste Extraction Agent, LA-11695, Los Alamos National Laboratory, NM, 1992. [10] V.A. Babain, A.Y. Shadrin, L.N. Lazarev, R.N. Kiseleva, Extraction of cesium and strontium by solutions of chlorinated cobalt dicarbollide in diethyleneglycol ditetrafluoropropyl ether, Radiokhimiya 35 (1993) 81. [11] A. Kumar, P.K. Mohapatra, V.K. Manchanda, Extraction of cesium-137 from nitric acid medium in the presence of macrocyclic polyethers, J. Radioanal. Nucl. Chem. 229 (1998) 169–172. [12] P. Thuery, M. Nierlich, V. Lamare, J.F. Dozol, Z. Asfari, J. Vicens, Bis-(crown ether) and azobenzocrown derivatives of calix[4]arene. A review of structural information from crystallographic and modeling studies, J. Incl. Phenom. Macrocycl. Chem. 36 (2000) 375–408.

26

D.R. Raut et al. / Journal of Membrane Science 407–408 (2012) 17–26

[13] R.A. Leonard, C. Conner, M.W. Liberatore, J. Sedlet, S.B. Aase, G.F. Vandegrift, L.H. Delmau, P.V. Bonnesen, B.A. Moyer, Development of solvent extraction process for cesium removal from SRS tank waste, Sep. Sci. Technol. 36 (2001) 743–766. [14] P.V. Bonnesen, L.H. Delmau, T.J. Haverlock, B.A. Moyer, ORNL Report, ORNL/TM13704, 1996. [15] P.V. Bonnesen, T.J. Haverlock, N.L. Engle, R.A. Sachleben, B.A. Moyer, in: G.J. Lumeta, R.D. Rogers, A.S. Gopalan (Eds.), ACS Symposium Series 757, 2002, pp. 27–44. [16] P.K. Mohapatra, S.A. Ansari, A. Sarkar, A. Bhattacharyya, V.K. Manchanda, Evaluation of calix-crown ionophores for selective removal of cesium from acidic nuclear waste solution, Anal. Chim. Acta 517 (2006) 308–314. [17] G. Wipff, M. Lauterbach, Complexation of alkali cations by calix[4]crown ionophores: conformation and solvent dependent Na/Cs binding selectivity and extraction: MD simulations in the gas phase, in water and at the chloroform–water interface, Supramol. Chem. 6 (1995) 187–207. [18] R. Ungaro, A. Casnati, F. Ugozzoli, A. Pochini, J.F. Dozol, C. Hill, H. Rouquette, 1,3-Dialkoxycalix[4]arenecrowns-6 in 1,3-alternate conformation: cesiumselective ligands that exploit cation–arene interactions, Angew. Chem. Int. Ed. Engl. 33 (1994) 1506–1509. [19] D.R. Raut, P.K. Mohapatra, S.A. Ansari, V.K. Manchanda, Use of calix[4]-bis2,3-naphthocrown-6 for separation of cesium from pressurized heavy water reactor simulated high level waste solutions (PHWR-SHLW), Sep. Sci. Technol. 44 (2009) 3664–3678. [20] D.R. Raut, P.K. Mohapatra, V.K. Manchanda, Cs(I) extraction/transport studies using irradiated calix [4]-bis 2,3-naphtho-crown-6, Desalination Water Treat. 12 (2009) 52–56. [21] L. Boyadzhiev, Z. Lazarova, in: R.D. Noble, S.A. Stern (Eds.), Membrane Separations Technology: Principles and Applications, Elsevier Science B.V., 1995, pp. 283–300. [22] P.K. Mohapatra, V.K. Manchanda, Liquid membrane based separations of actinides and fission products, Indian J. Chem. 42A (2003) 2925–2939. [23] N.M. Kocherginsky, Q. Yang, L. Seelam, Recent advances in supported liquid membrane technology, Sep. Purif. Technol. 53 (2007) 171–177. [24] D.R. Raut, P.K. Mohapatra, S.A. Ansari, V.K. Manchanda, Evaluation of a calix[4]-bis-crown-6 ionophore-based supported liquid membrane system for selective 137 Cs transport from acidic solutions, J. Membr. Sci. 310 (2008) 229–236. [25] P. Kandwal, S. Dixit, S. Mukhopadhyay, P.K. Mohapatra, V.K. Manchanda, Mathematical modeling of Cs(I) transport through flat sheet supported liquid membrane using calix-[4]-bis(2,3-naptho)-18-crown-6 as the mobile carrier, Desalination 278 (2011) 405–411.

[26] L.D. Nghiem, P. Mornane, I.D. Potter, J.M. Perera, R.W. Cattrall, S.D. Kolev, Extraction and transport of metal ions and small organic compounds using polymer inclusion membranes (PIMs), J. Membr. Sci. 281 (2006) 7–41. [27] A.J. Schow, R.T. Peterson, J.D. Lamb, Polymer inclusion membranes containing macrocyclic carriers for use in cation separations, J. Membr. Sci. 111 (1996) 291–295. [28] E.R.S. Miguel, A.V. Garduno-Garcıa, J.C. Aguilar, J. Gyves, Gold(III) transport through polymer inclusion membranes: efficiency factors and pertraction mechanism using Kelex 100 as carrier, Ind. Eng. Chem. Res. 46 (2007) 2861–2869. [29] P.K. Mohapatra, P.N. Pathak, A. Kelkar, V.K. Manchanda, Novel polymer inclusion membrane containing a macrocyclic ionophore for selective removal of strontium from nuclear waste solution, New J. Chem. 28 (2004) 1004–1009. [30] K.C. Khulbe, B. Kruczek, G. Chowdhury, S.Gagne, T. Matsuura, Surface morphology of homogeneous and asymmetric membranes made from poly(phenylene oxide) by tapping mode atomic force microscope, J. Appl. Polym. Sci. 59 (1996) 1151–1158. [31] J. Crank, The Mathematics of Diffusion, 2nd edn, Oxford University Press Inc., New York, 1975, p. 414. [32] J.K. Sears, J.R. Darby, Technology of Plasticizers, John & Wiley Sons, New York, 1982, p. 1174. [33] M.E. Duffey, D.F. Evans, E.L. Cussler, Simultaneous diffusion of ions and ion pairs across liquid membranes, J. Membr. Sci. 3 (1977) 1–14. [34] M. Sugiura, M. Kikkiwa, Effect of plasticizer on carrier mediated transport of zinc ion through cellulose triacetate membranes, Sep. Sci. Technol. 22 (1987) 2263–2268. [35] J.S. Gardner, J.A. Walker, J.D. Lamb, Permeability and durability effects of cellulose polymer variation in polymer inclusion membranes, J. Membr. Sci. 229 (2004) 87–93. [36] P.K. Mohapatra, D.S. Lakshmi, A. Bhattacharyya, V.K. Manchanda, Evaluation of polymer inclusion membranes containing crown ethers for selective cesium separation from nuclear waste solution, J. Hazard. Mater. 169 (2009) 472–479. [37] R. Tayeb, C. Fontas, M. Dhahbi, S. Tingry, P. Seta, Cd(II) transport across supported liquid membranes (SLM) and polymeric plasticized membranes (PPM) mediated by Lasalocid A, Sep. Purif. Technol. 42 (2005) 189. [38] C. Fontas, R. Tayeb, S. Tingry, M. Hidalgo, P. Seta, Transport of platinum(IV) through supported liquid membrane (SLM) and polymeric plasticized membrane (PPM), J. Membr. Sci. 263 (2005) 961. [39] A. Duhart, J.F. Dozol, H. Rouquette, A. Deratani, Selective removal of cesium from model nuclear waste solutions using a solid membrane composed of an unsymmetrical calix[4]arenebiscrown-6 bonded to an immobilized polysiloxane backbone, J. Membr. Sci. 185 (2001) 144–155.