Highly selective recovery of palladium by a new silica-based adsorbent functionalized with macrocyclic ligand

Separation and Purification Technology 106 (2013) 38–46

Contents lists available at SciVerse ScienceDirect

Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur

Highly selective recovery of palladium by a new silica-based adsorbent functionalized with macrocyclic ligand Feifei Bai a,b, Gang Ye a,⇑, Guangjin Chen b, Jichao Wei a, Jianchen Wang a, Jing Chen a a b

Institute of Nuclear and New Energy Technology, Tsinghua University, Beijing, PR China Faculty of Chemical Science and Engineering, China University of Petroleum, Beijing, PR China

a r t i c l e

i n f o

Article history: Received 18 September 2012 Received in revised form 18 December 2012 Accepted 19 December 2012 Available online 27 December 2012 Keywords: Palladium recovery Adsorption Silica Macrocyclic ligand High level liquid waste (HLLW)

a b s t r a c t A new kind of silica-based adsorbent with high selectivity and efficient adsorption for palladium ion (Pd(II)) was synthesized in this study. A macrocyclic polyether isomer was modified for incorporation to the silica matrix as a ligand to realize the complexation between Pd(II) and the adsorbent. The adsorption of Pd(II) in HNO3 media was evaluated by both batch and column operations. Influences including acidity, contact time, initial metal concentration and elution conditions were detailed. From a practical viewpoint, the functionalized adsorbent was employed for the recovery of Pd(II) from the simulated high level liquid waste (HLLW) containing a large amount of interferences. Superior selectivity to Pd(II) as well as a recovery rate higher than 90% was obtained. A mechanism concerning the formation of complex ionpair was proposed for the description of the Pd(II)-binding process. The macrocyclic ligand functionalized silica adsorbent possesses potential for the recovery of the palladium resource in radioactive liquid waste. Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction As an important member of platinum group metals (PGMs), palladium is of great value in modern industry due to its favorable physical and chemical properties, such as high melting point, corrosion resistance and extraordinary catalytic ability [1,2]. In recent years, the rapid development of catalytic systems for petrochemical industry and automotive exhaust has led an increasing consumption of the precious metal. The gap between supply and demand is becoming more and more overt because of the limited palladium source in terrestrial crust with low availability. Meanwhile, from an environmental point of view, the release and accumulation of palladium in environment as a new pollutant may cause adverse health effects to human beings (e.g. eye irritations, asthma, rhinoconjunctivitis, skin problems, etc.) [3–5]. Therefore, the research on the efficient separation and recovery of palladium is in high demand. Nowadays, it has been widely accepted that palladium recovery from industrial waste such as spent catalyst and electronic devices is valuable, and much efforts have been made to reach the goal [6– 9]. However, the palladium resource present in radioactive waste stream still seems to be lack of enough concern [10,11]. According to a previous report, around 11 kg fission-generated palladium is estimated in every metric ton of the spent fuel of fast reactor, which could significantly cover the amount obtained naturally ⇑ Corresponding author. Tel.: +86 10 8979 6063; fax: +86 10 6277 1740. E-mail address: [email protected] (G. Ye). 1383-5866/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.seppur.2012.12.021

[12,13]. Furthermore, as the volume of the accumulated high level liquid waste (HLLW) is constantly increasing worldwide, the recovery of palladium from radioactive waste stream will be considerable in economic benefit. Nevertheless, it must be admitted that, owing to the complexity of the HLLW circumstance and overlapping properties between palladium and other interfering metals, separation of palladium still faces enormous challenge [11]. In general, precious metal recycling techniques are categorized as pyro- and hydrometallurgical processes. The latter gains more popularity for easy controllability over the other one [14]. In coordination with the hydrometallurgical process, many methods including ion exchange, solvent extraction, electro-deposition, membrane separation and adsorption have been developed. Among which, adsorption has been extensively applied due to high enrichment factor, fast kinetics, minimal organic diluents utilization, and less waste accumulation [15]. Besides, the most appealing aspect of adsorption lies in the vast choice of adsorbents with wide possibilities of functionalization, making them capable for palladium ion (Pd(II)) recycling in varieties of wastewater systems. Typical cases are usually developed by using chelating resins [16,17], chemically-modified activated carbon [18,19], nanotubes [20], biomass [21–24] or functionalized silica gel [25–30] as adsorption materials. However, up to now, most of the separation protocols still suffer from a low efficiency and/or a poor selectivity. Adsorbents with selective recognition to Pd(II) have been rarely reported in the literatures [31], let alone those specialized for Pd(II) enrichment in HLLW system which contains a great many of interferences.

F. Bai et al. / Separation and Purification Technology 106 (2013) 38–46

The marriage between adsorption material and supramolecular chemistry has provided a boost to researches on novel adsorbents for selective metal separation lately [32–34]. Macrocyclic ligands such as crown ether, calixarene and cyclodextrin were incorporated into certain substrates (e.g. polymeric resin, silica gel, supported liquid membrane, etc.), and selective recognition towards metal ions was realized based on host–guest interaction [35–41]. Unlike binding of hard alkali metal cations, the complexation between macrocyclic ligands containing hard oxygen donors and the soft palladium ion is atypical. Talanova et al. reported a kind of divinylbenzene copolymer bearing dibenzo-18-crown-6 ligand, which showed highly efficient sorption of Pd(II) from aqueous K2PdCl4 solution [42]. Also, Fontàs et al. prepared a solid phase extraction support with high affinity for Pd(II) by using thiacalix[4]arene derivative [43]. For these cases, the hole-size fitting concept in traditional supramolecular chemistry is unconvincing, while ion-pair complex mechanism has been proposed as a seemingly reasonable explanation [44]. Overall, much more effort still needs to be made for either preparation of new functionalized adsorbents with selective Pd(II)-separation or the probing of the recognition mechanism. In this study, we presented a novel example of macrocyclic ligand functionalized silica adsorbent which had efficient adsorption and excellent selectivity towards Pd(II) in HNO3 media and simulated HLLW. For preparation of the adsorbent, a macrocyclic polyether isomer cis-di(aminocyclohexyl)-18-crown-6 was incorporated into silica matrix via a co-condensation synthetic method. The adsorption of Pd(II) in HNO3 solution was carried out by both batch and column operations. Influences including kinetics, acidity, initial metal concentration, phase ratio, and elution condition were investigated. From a practical point of view, selective recovery of Pd(II) form simulated HLLW of a light water reactor (LWR) was performed. Mechanism about the complexation between the macrocyclic ligand and Pd(II) was discussed, and cycle use of the adsorption material was evaluated. 2. Experimental 2.1. Chemicals Macrocyclic polyether cis-di(aminocyclohexyl)-18-crown-6 was synthesized according to Deetz’s work [45], and molecular structure was confirmed by 1H and 13C NMR spectra. Tetraethoxysilane (TEOS, 98%), di-n-butyltin dilaurate (DBTL, 95%) and 3-chloropropyltriethoxysilane (CPTES, 97%) were purchased form Aldrich. Deionized water with resistivity >18 MX cm was obtained from a Milli-Q water purification system. Stock solution of Pd(II) (92 g/L) were purchased from General Research Institute for Nonferrous Metals. Before use, the stock solutions were further diluted by HNO3 with different concentration. The acidity was calibrated by titration. The analytical grade chemicals, such as potassium iodide, tetrahydrofuran, and other reagents were commercially obtained and used as received without further purification.

39

MAS) technique. FT-IR spectra (4000–400 cm1) were recorded by Nicolet Nexus 470 FT-IR in KBr matrix. X-ray photoelectron spectroscopy (XPS) measurement was performed using a PHI5300 ESCA instrument. Elemental analysis of C, H and N was performed on Elementar Vario EL III. Inductively Coupled PlasmaAtomic Emission Spectrometry (ICP-AES) (RF power supply: 1.15 kW; wavelength of element: Pd(II) 363.470 nm, Thermo Jarrel Ash mod. IRIS Advantage) and Atomic Absorption Spectroscopy (HITACHI Z-2000) were used for concentration measurement of Pd(II). The micro-morphology of the adsorbent particles was recorded by FEI Quanta 200 environmental scanning electron microscopy (ESEM). For column experiment, an YL-110 peristaltic pump and a PTFE (polytetrafluoroethylene) column (50 mm  7.0 mm i.d.) were used. 2.3. Synthesis of functionalized silica adsorbent 2.3.1. 4,40 -Bis((3-(triethoxylsilyl)propyl)amino)dicyclohexyl-18crown-6 [1] Cis-di(aminocyclohexyl)-18-crown-6 (4.02 g, 10 mmol) was dissolved in 250 mL tetrahydrofuran (THF), followed by the addition of 3.45 g potassium carbonate (25 mmol) and 3.32 g potassium iodide (20 mmol). Under nitrogen protection and stirring, 4.80 g 3-chloropropyltriethoxysilane (CPTES) (20 mmol) was dropwise injected. After 5 h reflux, the solid salt was separated by filtration. A brown oily product was obtained after removing the solvent of the filtrate. The product was purified via a procedure similar to that reported by Dubois et al. [37]. The final product was obtained with a yield of 47%. 1H NMR (300 MHz, CDCl3): d 3.79–3.70 (q, 12H), 3.67–3.54 (m, 16H), 3.08 (m, 4H), 2.69 (m, 6H), 1.95–1.86 (m, 4H), 1.60–1.52 (m, 8H), 1.40–1.22 (m, 4H), 1.19 (t, 18H), 0.74 (t, 4H). IR (KBr) (m, cm1): 3328.6, 2985.6, 2922.8, 2857.5, 1452.3, 1354.1, 1291.6, 1220.3. ESI MS: [M + Na]+ 834.20 (calc. value for [C38H78N2O12Si2 + Na]+ 833.50). 2.3.2. Functionalized silica adsorbent Co-condensation method was used to prepare the functionalized silica adsorbent. The silanized monomer [1] (7.5 mmol) and tetraethoxysilane (TEOS) (2.5 mmol) were dissolved in quantitative anhydrous ethanol with total concentration of 1 mol/L. Under mild agitation, a stoichiometric amount of water was slowly added with a syringe, followed by 2 wt.% di-n-butyltin dilaurate (DBTL) as catalyst. Then, the mixture was further stirred for 10 min. Homogeneous solution was obtained and set for 3 days at ambient temperature for gelation. The product was granulated, followed by thermal treatment at 120 °C under vacuum for 8 h. The catalyst and the unreacted monomers were washed off with hot acetone. Finally, the particles were collected and dried in 60 °C vacuum oven for 24 h. Solid state 13C CP-MAS NMR (300 MHz, ppm): 77.5, 71.0, 60.3, 54.1, 47.9, 26.4, 23.0, 17.6, 10.5. Solid state 29Si MAS NMR (300 MHz, ppm): 57.7, 67.5, 102.2, 110.2. IR (KBr) (m, cm1): 3385.5, 2929.7, 2863.2, 1465.1, 1347.4, 1101.6, 789.2, 689.8. 2.4. Pd(II) recovery by batch method

2.2. Characterization Nitrogen adsorption measurement of the functionalized organosilica adsorbent was carried out on a NOVA 3200e Surface Area and Pore Size Analyzer. Samples were pretreated at 140 °C under vacuum for at least 2 h before the nitrogen adsorption experiments. Surface areas were calculated based on the Brunauer-Emmett-Teller (BET) method. 1H NMR spectra were recorded by a JOEL JNM–ECA600 NMR spectrometer. 29Si and 13C solid-state NMR spectra were obtained with a Bruker AV300 Spectrometer with the cross-polarization/magic-angle spinning (CP/

The recovery of Pd(II) in aqueous HNO3 solutions and simulated HLLW was carried out by batch operation. In a typical process, 0.02 g functionalized organosilica adsorbent was mixed with HNO3 solution containing about 5.0  103 mol/L Pd(II). The acidity of the solution varied from 0.1 to 6 mol/L, and phase ratio was chosen as 0.01 g/1 mL. With vigorous agitation in 25 °C thermostat for 1 h, the aqueous phase was separated with 0.45 lm micro-pore filter. Inductively Coupled Plasma-Atomic Emission Spectrometry (ICP-AES) was employed to determine the residual amount of the Pd(II) in the solution. The adsorption capacity qt

40

F. Bai et al. / Separation and Purification Technology 106 (2013) 38–46

Fig. 1. Structure illustration of the macrocyclic ligand functionalized silica (A) and ESEM morphology of the adsorbent particles (B).

and recovery rate R% are defined as shown in the following equations:

qt ¼

ðC 0  C t Þ  V W

R% ¼

ðC 0  C t Þ  100 C0

ð1Þ

ð2Þ

where C0 and Ct are the initial and residual Pd(II) concentration in the aqueous phase, respectively. V is the volume of solution, and W is the weight of adsorbent. The batch experiment for Pd(II)-recovery in simulated HLLW was also performed with a phase ratio of 0.01 g/1 mL. The composition of simulate HLLW used is identical with the real HLLW generated during the spent fuels reprocessing of a typical light water reactor (LWR) nuclear power plant. The spent fuels are designed to have a burn up of 55,000 MWd/T followed with 8 years cooling. The acidity of the simulated HLLW was adjusted to be 1 mol/L, which was determined by acid–base titration. The measurement of the metal elements was performed by ICP-AES, except for the cesium ion which was detected by Atomic Absorption Spectroscopy. 2.5. Column experiment and elution A PTFE column (50 mm effective length and 7.0 mm i.d.) loaded with 0.2 g of functionalized organosilica adsorbent was used for column solid phase extraction (SPE). The column was connected to a peristaltic pump to control the flow rate. Pd(II) solution with the concentration of 5.0  103 mol/L was passed through the column with an appropriated rate. Effluent was collected in a 10 mL tube and the residual Pd(II) amount was measured by ICP-AES. To determine the optimal elution condition, the loaded metal was eluted by 5 mL different eluents including deionized water, 0.1 mol/L CH3COOH, 0.1 mol/L oxalic acid, 0.1 mol/L HCl, 0.1 mol/L HNO3, 1 mol/ L HNO3, 4 mol/L HNO3 and 1% thiourea in 1 mol/L HNO3. The concentration of Pd(II) in the eluent was also determined by ICP-AES. 3. Results and discussion

co-condensation can result in a better control of functionalization degree and a uniform distribution of the functional groups. In this study, the new silica-based adsorbent functionalized with macrocyclic ligand was prepared through the co-condensation approach. At first, synthesis of macrocyclic polyether isomer cis-di(aminocyclohexyl)-18-crown-6 was a critical step, which required multiple reactions including selective nitration, reduction and catalytic hydrogenation without damaging the reactive amino groups. Here, the cis-isomer of the crown ether was elaborately separated and used as functional monomer, because it tended to have a better coordination with metal ion [46]. The cis-isomer was then silanized by 3-chloropropyltriethoxysilane (CPTES) based upon a nucleophilic substitution reaction. Finally, co-condensation synthesis was carried out following a typical sol–gel process with the silanized macrocyclic polyether monomer and tetraethoxysilane (TEOS). By weighing an abundance of the macrocyclic ligand groups for metal binding as well as a favorable morphology of the adsorbent particles, we determined the molar ratio of the macrocycle monomer to TEOS as 3:1. Fig. 1A illustrated the structure of the new macrocyclic ligand functionalized silica adsorbent. The detailed synthetic procedure can be seen in our recent report [47]. To confirm that the macrocyclic ligand has been incorporated in the silica-based matrix, solid state 13C CP-MAS NMR was employed to identify the organic functional groups (Fig. S1). The signal peaks centred at 77.5, 70.8, 54.5, 23.0 ppm were reasonably attributed to the carbon atoms in the polyether ring and cyclohexyl group of macrocyclic ligand. Further evidence was provided by the survey of X-ray photoelectron spectroscopy (XPS) (Fig. S2). The N 1s signal presented at around 400 eV corresponded to the N atom in the amino group of the monomer. Exact amount of the macrocyclic ligand groups could be inferred based on the result of elemental analysis, in which the C, H and N elements, took up 38.75 wt%, 6.80 wt% and 2.77 wt% of the gross weight, respectively. The SEM morphology of the silica-based adsorbent particles is shown in Fig. 1B. Owing to the higher proportion of the organic ingredients, the particles tend to partially aggregate. The average size of particles was estimated to be 3.5 lm. N2 adsorption– desorption measurement shows that the sample has a BET specific surface area of 72.0 m2/g with a pore volume 0.04 cm3/g.

3.1. Synthesis and characterization 3.2. Effect of acidity Silica-based adsorption materials containing special functional groups can be generally prepared through two methods, namely, co-condensation and post-grafting. Compared to the latter one,

The effect of acidity on the adsorption ability of the macrocyclic ligand functionalized silica towards Pd(II) was investigated. Taking

41

F. Bai et al. / Separation and Purification Technology 106 (2013) 38–46

into account the circumstance of radioactive stream, HNO3, rather than HCl which was usually engaged, was chosen to adjust the acidity of the solution. The Pd(II) stock solutions were prepared in 0.1–6 mol/L HNO3 with initial metal concentration of 5.0  103 mol/L. Phase ratio was set as 0.01 g/1 mL. The curve of adsorption capacity versus HNO3 concentration was plotted in Fig. 2. The result showed that acidity of the solution had a evident influence on the Pd(II) adsorption. The capacity declined rapidly with the increase of the HNO3 concentration. The maximum uptake of Pd(II) was obtained at 0.1 mol/L HNO3. It has been reported that nitric acid can compete for oxygen atoms in the macrocyclic polyether, which will impede the coordination between the metal ion and the ligand group [48]. This competitive effect might give a reasonable explanation to our experimental result. 3.3. Adsorption kinetics The influence of contact time on Pd(II) adsorption was shown in Fig. 3. It can be seen that the adsorption capacity has a rapid increase as soon as the contact happens, and the equilibrium with qe around 56.2 mg/g can be reached in only 30 min. Such a fast kinetics will benefit a highly efficient metal adsorption. Further prolonging the contact time does not result in a higher uptake of Pd(II). Hence, the contact time of 1 h was sufficient for subsequent adsorption experiments. Pseudo-first-order [49] and pseudo-second-order models [50] were employed to fit the experimental data. A good correlation

Table 1 Kinetic parameters fitted by pseudo-first-order and pseudo-second-order model for Pd(II) adsorption. Pseudo-first-order model

Pseudo-second-order model

k1 (min1)

qe,1 (mg/g)

R2

k2 (g/mg min)

h (mg/g min)

qe,2 (mg/g)

R2

12.7

67.8

0.912

0.0173

51.5

54.6

0.996

of the kinetic data may help to understand the kinetic mechanism of the adsorption process. Pseudo-first-order model is represented by the following equation:

1 k1 1 ¼ þ qt qe t qe

ð3Þ

where k1 (min1) is the adsorption rate constant of pseudo-first-order model, qt (mg/g) denotes the adsorption capacity at certain time t (min), and qe (mg/g) is the adsorption capacity at equilibrium. Pseudo-second-order can be written in a linear form as the following equation:

t 1 1 ¼ þ t qt k2 q2e qe

ð4Þ

where k2 (g/(mg min)) is the pseudo-second-order adsorption rate constant. The initial adsorption rate (h) can be determined from fitted k2 and qe values based on,

h ¼ k2 q2e

ð5Þ

The fitted kinetic parameters are summarized in Table 1. The results show that pseudo-second-order model has a better correlation coefficient R2 than that of pseudo-first-order model. The calculated value of qe by pseudo-second-order model is much closer to the experimental value, which also confirms the validity of the model. Besides, the fast initial adsorption rate (h = 51.5 mg/ g min) is favorable for the adsorption application. The fitted curve of linear pseudo-second-order model was plotted in Fig. 4. The results suggested that chemical adsorption was the main factor in the Pd (II) adsorption process [8]. 3.4. Adsorption isotherm

Fig. 2. Effect of the HNO3 concentration on the adsorption ability of the silica-based adsorbent towards Pd(II).

For a further understanding of the functionalized silica’s adsorption behavior, the effect of the initial Pd(II) concentration on the adsorption ability was examined. Solutions with metal ion in the range of 0.1–3.0 g/L were used in the batches containing 0.01 g/1 mL adsorbent at room temperature. With the raise of initial metal concentration, the adsorption capacity, typically, showed a sharp increase, followed with a slow down to equilibrium state. The maximum adsorption capacity was obtained as 83.3 mg/g. This value was reached for the initial concentration of Pd(II) of 2.0 g/L. The adsorption equilibrium data were fitted with Langmuir and Freundlich isotherm models for a comparative study of the sorption behavior (Fig. 5). The Langmuir model represents the nonlinear sorption and suggests that metal uptake occurs on a homogeneous surface by monolayer sorption without interaction between adsorbed species [24]. The model can be written as Eq. (6),

qe ¼

Fig. 3. Pd(II) adsorption capacity versus the contact time (5.0  103 mol/L Pd(II) in 0.1 mol/L HNO3).

qmax K L C e 1 þ K LCe

ð6Þ

where qe (mg/g) is the equilibrium Pd(II) concentration on the adsorbant, Ce (mg/L) is the equilibrium Pd(II) concentration in the solution, qmax (mg/g) is the maximum monolayer adsorption capacity, and KL (L/mg) is the Langmuir sorption constant. Yet, the

42

F. Bai et al. / Separation and Purification Technology 106 (2013) 38–46 Table 3 Adsorption capacity of Pd(II) as well as the other metals in simulated HLLW. Element

Original amount (g/L)

Adsorption capacity (mg/g)

Pd Cs Fe Mo Na Nd Ni Ru Sr Zr Rb Ba Cd Cr

0.380 0.485 0.250 0.674 1.136 2.167 0.400 0.780 0.159 0.292 0.298 0.394 0.085 0.051

34.01 1.90E  04 1.90 17.57 3.20E  04 8.40 2.84 2.62 3.28 3.24 0.39 14.54 1.42 1.35

Fig. 4. Fitted kinetic curve based on the pseudo-second-order model with R2 of 0.996.

3.5. Selectivity and Pd(II) recovery in simulated HLLW

Fig. 5. Data fitting by using Langmuir and Freundlich isotherm models for the adsorption of Pd(II) .

Freundlich model describes adsorption on a heterogeneous (multiple layer) surface with uniform energy, which is represented by the following equation:

qe ¼ K F C e1=n

ð7Þ

where KF and n are the Freundlich constants concerning the multilayer adsorption capacity and adsorption intensity, respectively [2]. The fitted parameters of the adsorption isotherm were shown in Table 2. In comparison, the adsorption behavior of Pd(II) was better fitted by the Langmuir model with the coefficient (R2) of 0.983. The maximum sorption capacity was calculated to be 83.0 mg/g, which was very close to the experimental data. The R2 value (0.862) of Freundlich model indicated that the model was not able to adequately describe the adsorption behavior in our experiment. In other words, the Pd(II) uptake on the silica-based adsorbent functionalized with macrocyclic ligand could be more reasonably regarded as monolayer adsorption.

Table 2 Fitted parameters of Langmuir and Freundlich models for Pd (II) adsorption of the macrocyclic ligand functionalized silica. Langmuir model

Freundlich model

qmax (mg/g)

KL (L/mg)

R2

n

KF

R2

83.0

26.6

0.983

4.83

77.2

0.862

Selectivity is an important aspect for adsorbents aiming to recover noble metals from waste streams. For a practical evaluation, the macrocyclic ligand functionalized silica adsorbent was employed for Pd (II) recovery in simulated HLLW. The preparation of simulated HLLW can be seen in our previous work [47]. Batch separation was carried out with the phase ratio of 0.01 g/1 mL and the contact time of 1 h. The original concentration of main elements in simulated HLLW as well as the adsorption capacity was summarized in Table 3. The highest value of 34.01 mg/g was obtained for Pd(II). Considering that the metals had different initial concentration, more intuitive comparison of recovery rate (R%) was shown in Fig. 6. The result revealed that the macrocyclic ligand functionalized silica adsorbent had dominant selectivity towards Pd(II), and more than 90% of the Pd(II) in the multicomponent simulated HLLW could be recovered. Among the miscellaneous interferences, only the Ba(II) showed a certain extent of affinity to the adsorbent (R% = 37.3). Overall, this kind of macrocyclic ligand functionalized silica adsorbent is believed to have the potential for recovery of palladium resource in radioactive streams. 3.6. Column experiment and elution Column experiments were carried out in a PTFE column connected to a peristaltic pump for controlling the flow rate. Feed solution with Pd(II) concentration of 5.0  103 mol/L in 0.1 mol/ L HNO3 was passed through the loaded silica-based adsorbent at a rate of 5 mL/min. The breakthrough curve in Fig. 7 showed a fast adsorption kinetic which was in agreement with the result obtained in batch operations. Elution condition was investigated by selecting various eluents including HCl, oxalic acid, acetic acid, thiourea and HNO3 with different concentration. 5 mL eluent was used for stripping the loaded Pd(II) on the silica adsorbent. The metal concentration in the effluent was determined by ICP-AES for analysis. Fig. 8 shows the elution effect with different eluents. It can be seen that thiourea has excellent property for Pd(II) elution with a recovery rate of 98%, while the others, even HNO3 with high concentration (4 mol/L), cannot afford a satisfactory elution effect. It has been reported that Pd(II) can form a positively charged stable complex with thiourea molecule, which cannot be retained on the column during the elution [15]. Therefore, thiourea could be the optimal choice for the regeneration of macrocyclic ligand functionalized silica adsorbent. Fig. 9 gives visualized evidences of the adsorption of Pd(II) and regeneration of the silica-based adsorbent. The HNO3 solution containing 5.0  103 mol/L Pd(II) showed a yellow color (A), which turned colorless (B) after passing the column loaded with the silica adsorbent. The Pd(II) uptake could be effectively washed out by 1%

F. Bai et al. / Separation and Purification Technology 106 (2013) 38–46

43

Fig. 6. Recovery rate of Pd(II) and the other 13 interfering metals in simulated HLLW by the functionalized silica adsorbent.

seemed to give a more reasonable explanation. In our HNO3 system, with the absence of chlorine ion, Pd(II) ion could stably exist with the form of [Pd(NO3)4]2 when the pH of the solution is lower than 3. Based on this consideration, here, we propose an analogous ion-pair mechanism to explain the interaction between the macrocyclic ligand functionalized silica and Pd(II) ions in our experiment, which may be described as follows: 2 þ 2Hþaq þ ½PdðNO3 Þ4 2 aq þ 2Ls ¼ ½ðH LÞ2 ðPdðNO3 Þ4 Þ s

and/or 2 þ þ 2Hþaq þ ½PdðNO3 Þ4 2 aq þ Ls ¼ ½ðH LÞðPdðNO3 Þ4 Þ ðH Þs

Fig. 7. Breakthrough curve of column adsorption of Pd(II) in 5.0  103 mol/L feed solution with a flow rate of 5 mL/min.

thiourea solution. Accordingly, the color of the effluent became yellow (C). The functionalized silica adsorbent showed a light yellow color initially (D). Adsorption of the Pd(II) made the adsorbent turn black (E), which implied the formation of complex in between. After the elution by thiourea, the regenerated silica regained the same appearance as the original adsorbent (F). After 5 cycles of regeneration, the adsorption capacity of the silica-based adsorbent was still found to be apparently constant. 3.7. Mechanism discussion Basically, the mechanism concerning the macrocyclic ligand for the binding of noble metals is still indeterminate. The hole-size fitting concept in the traditional supramolecular chemistry is not suitable because of the obvious discrepancy between the noble metal ions and the hole-size of the macrocycle. Meanwhile, according to the hard and soft acids and bases theory, it is also atypical that soft electron acceptors, such as Pd(II) and Pt(II), coordinate with the hard oxygen electron donors [43]. In previous reports, some mechanisms have been suggested for the binding of PGMs’ chloro complexes. Among them, ion-pair complex mechanism

where the subscripts aq and s represent aqueous and solid phase, respectively. Ls refers to the silica-based adsorbent. The immobilized complex ion-pairs are ‘‘ion triads’’ (two cations and one anion) [42]. As stated above, the color change in Fig. 9 (E) has offered an evidence of the complex formation between Pd(II) and the macrocyclic ligand functionalized silica. Nevertheless, our further confirmatory tests showed that such kind of complex did not happen between Pd(II) and free macrocyclic ligands in liquid–liquid extraction. We intentionally examined the Pd(II) extraction ability of two macrocycle molecules, namely, di(cyclohexyl)-18-crown-6 and cis-di(aminocyclohexyl)-18crown-6. The structure of the extractants was shown in Fig. S4. The ligand molecules had the same macrocycle as that immobilized on the silica adsorbent. But poor extraction to Pd(II) was found for both ligands in HNO3 solutions, which only gave a distribution coefficient of 0.23 and 0.09. Therefore, further effort still needs to be made for an in-depth study of the mechanism on the complex between macrocyclic ligand and Pd(II). 4. Conclusions In this study, a novel macrocyclic ligand incorporated silica adsorbent was synthesized by co-condensation method. The functionalized adsorbent showed highly selective adsorption towards Pd(II) in HNO3 solution. Influences such as acidity, contact time, initial metal concentration and elution conditions were investigated. The adsorption kinetics and isotherms were evaluated. Pd(II)

44

F. Bai et al. / Separation and Purification Technology 106 (2013) 38–46

Fig. 8. Elution of Pd(II) loaded on the silica-based adsorbent by 5 mL different eluents including deionized water, 0.1 mol/L CH3COOH, 0.1 mol/L oxalic acid, 0.1 mol/L HCl, 0.1 mol/L HNO3, 1 mol/L HNO3, 4 mol/L HNO3 and 1% thiourea in 1 mol/L HNO3.

Fig. 9. Digital images of the Pd(II) adsorption and elution. (A) 5.0  103 mol/L Pd(II) solution in 0.1 mol/L HNO3, yellow solution; (B) stock solution after Pd(II) adsorption, colorless solution; (C) effluent after the stripping of Pd(II) by 1% thiourea, yellow solution; (D) original silica-based adsorbent, light yellow powder; (E) adsorbent after Pd(II) uptake, black powder; (F). regenerated adsorbent, light yellow powder. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

F. Bai et al. / Separation and Purification Technology 106 (2013) 38–46

recovery from the simulated HLLW of a light water reactor (LWR) was realized with a rate of more than 90%. A mechanism concerning the formation of complex between Pd(II) and the adsorbent was proposed. This kind of macrocyclic ligand functionalized silica adsorbent is of potential value for the application of Pd(II) recovery. Acknowledgement The financial support from the NSFC under Project 51103079 is gratefully acknowledged.

Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.seppur.2012.12. 021. References [1] L.M. Zhou, J.P. Xu, X.Z. Liang, Z.R. Liu, Adsorption of platinum(IV) and palladium(II) from aqueous solution by magnetic cross-linking chitosan nanoparticles modified with ethylenediamine, J. Hazard. Mater. 182 (2010) 518–524. [2] K. Fujiwara, A. Ramesh, T. Maki, H. Hasegawa, K. Ueda, Adsorption of platinum(IV), palladium(II) and gold(III) from aqueous solutions onto l-lysine modified crosslinked chitosan resin, J. Hazard. Mater. 146 (2007) 39–50. [3] R. Merget, G. Rosner, Evaluation of the health risk of platinum group metals emitted from automotive catalytic converters, Sci. Total Environ. 270 (2001) 165–173. [4] J. Kielhorn, C. Melber, D. Keller, I. Mangelsdorf, Palladium – a review of exposure and effects to human health, Int. J. Hyg. Envir. Heal. 205 (2002) 417– 432. [5] C.L.S. Wiseman, F. Zereini, Airborne particulate matter, platinum group elements and human health: a review of recent evidence, Sci. Total Environ. 407 (2009) 2493–2500. [6] J. Panchompoo, L. Aldous, R.G. Compton, Irreversible uptake of palladium from aqueous systems using l-cysteine methyl ester physisorbed on carbon black, J. Mater. Chem. 21 (2011) 9513–9522. [7] Z. Hubicki, A. Wolowicz, A comparative study of chelating and cationic ion exchange resins for the removal of palladium(II) complexes from acidic chloride media, J. Hazard. Mater. 164 (2009) 1414–1419. [8] L.M. Zhou, J.H. Liu, Z.R. Liu, Adsorption of platinum(IV) and palladium(II) from aqueous solution by thiourea-modified chitosan microspheres, J. Hazard. Mater. 172 (2009) 439–446. [9] C.R. Adhikari, D. Parajuli, H. Kawakita, K. Inoue, K. Ohto, H. Harada, Dimethylamine-modified waste paper for the recovery of precious metals, Environ. Sci. Technol. 42 (2008) 5486–5491. [10] Z. Kolarik, E.V. Renard, Recovery of value fission platinoids from spent nuclear fuel part I: general considerations and basic chemistry, Platin. Met. Rev. 47 (2003) 74–87. [11] Z. Kolarik, E.V. Renard, Recovery of value fission platinoids from spent nuclear fuel part II: separation processes, Platin. Met. Rev. 47 (2003) 123–131. [12] S. Koyama, T. Suzuki, M. Ozawa, From waste to resource, nuclear rare metals as a dream of modern alchemists, Energy Convers. Manage. 51 (2010) 1799– 1805. [13] D. Parajuli, K. Hirota, N. Seko, Effective separation of palladium from simulated high level radioactive waste, J. Radioanal. Nucl. Ch. 288 (2011) 53–58. [14] A. Parodi, T. Vincent, M. Pilsniak, A.W. Trochimczuk, E. Guibal, Palladium and platinum binding on an imidazol containing resin, Hydrometallurgy 92 (2008) 1–10. [15] R.K. Sharma, A. Pandey, S. Gulati, A. Adholeya, An optimized procedure for preconcentration, determination and on-line recovery of palladium using highly selective diphenyldiketone-monothiosemicarbazone modified silica gel, J. Hazard. Mater. 209–210 (2012) 285–292. [16] E. Birinci, M. Gülfen, A.O. Aydın, Separation and recovery of palladium(II) from base metal ions by melamine–formaldehyde–thiourea (MFT) chelating resin, Hydrometallurgy 95 (2009) 15–21. [17] A. Wołowicz, Z. Hubicki, Applicability of new acrylic, weakly basic anion exchanger purolite a-830 of very high capacity in removal of palladium(II) chloro-complexes, Ind. Eng. Chem. Res. 51 (2012) 7223–7230. [18] Z. Tu, S. Lu, X. Chang, Z. Li, Z. Hu, L. Zhang, H. Tian, Selective solid-phase extraction and separation of trace gold, palladium and platinum using activated carbon modified with ethyl-3-(2-aminoethylamino)-2-chlorobut-2enoate, Microchim. Acta 173 (2011) 231. [19] H. Sharififard, M. Soleimani, F.Z. Ashtiani, Evaluation of activated carbon and bio-polymer modified activated carbon performance for palladium and platinum removal, J. Taiwan Inst. Chem. Eng. 445 (2012) 8.

45

[20] R. Li, Q. He, Z. Hu, S. Zhang, L. Zhang, X. Chang, Highly selective solid-phase extraction of trace Pd(II) by murexide functionalized halloysite nanotubes, Anal. Chim. Acta 713 (2012) 136–144. [21] T. Maruyama, H. Matsushita, Y. Shimada, I. Kamata, M. Hanaki, S. Sonokawa, N. Kamiya, M. Goto, Proteins and protein-rich biomass as environmentally friendly adsorbents selective for precious metal ions, Environ. Sci. Technol. 41 (2007) 1359–1364. [22] P. Ramakul, Y. Yanachawakul, N. Leepipatpiboon, N. Sunsandee, Biosorption of palladium(II) and platinum(IV) from aqueous solution using tannin from indian almond (terminalia catappa l.) leaf biomass: kinetic and equilibrium studies, Chem. Eng. J. 193–194 (2012) 102–111. [23] J. Park, S.W. Won, J. Mao, I.S. Kwak, Y. Yun, Recovery of Pd(II) from hydrochloric solution using polyallylamine hydrochloride-modified escherichia coli biomass, J. Hazard. Mater. 181 (2010) 794–800. [24] A. Sari, D. Mendil, M. Tuzen, M. Soylak, Biosorption of palladium(II) from aqueous solution by moss (racomitrium lanuginosum) biomass: equilibrium, kinetic and thermodynamic studies, J. Hazard. Mater. 162 (2009) 874–879. [25] S.Y. Kim, Y. Xu, T. Ito, Y. Wu, T. Tada, K. Hitomi, E. Kuraoka, K. Ishii, A novel partitioning process for treatment of high level liquid waste using macroporous silica-based adsorbents, J. Radioanal. Nucl. Ch. 1 (2012). [26] E. Mladenova, I. Dakova, I. Karadjova, M. Karadjov, Column solid phase extraction and determination of ultra-trace au, pd and pt in environmental and geological samples, Microchem. J. 101 (2012) 59–64. [27] K. Mehrani, A. Mehrani, M. Amini, O. Sadeghi, N. Tavassoli, Dipyridylaminemodified nanoporous silicas as new sorbents for the separation and preconcentration of palladium, Microchim. Acta 173 (2011) 521. [28] S. Sivrikaya, H. Altundag, M. Zengin, M. Imamoglu, Separation, preconcentration, and recovery of Pd(II) ions using newly modified silica gel with bis(3-aminopropyl)amine, Sep. Sci. Technol. 46 (2011) 2032–2040. [29] H. Ebrahimzadeh, N. Tavassoli, M.M. Amini, Y. Fazaeli, H. Abedi, Determination of very low levels of gold and palladium in wastewater and soil samples by atomic absorption after preconcentration on modified mcm-48 and mcm-41 silica, Talanta 81 (2010) 1183–1188. [30] A. Zhang, E. Kuraoka, M. Kumagai, Removal of Pd(II), Zr(IV), Sr(II), Fe(III), and Mo(Vi) from simulated high level liquid waste by extraction chromatography utilizing the macroporous silica-based polymeric materials, Sep. Purif. Technol. 50 (2006) 35–44. [31] H. Zheng, D.Z. De Zhang, W.Y. Wang, Y.Q. Fan, J. Li, H.P. Han, Highly selective determination of palladium(II) after preconcentration using Pd(II)-imprinted functionalized silica gel sorbent prepared by a surface imprinting technique, Microchim. Acta 157 (2007) 7. [32] K. Ariga, A. Vinu, J.P. Hill, T. Mori, Coordination chemistry and supramolecular chemistry in mesoporous nanospace, Coordin. Chem. Rev. 251 (2007) 2562– 2591. [33] G.W. Gokel, W.M. Leevy, M.E. Weber, Crown ethers: sensors for ions and molecular scaffolds for materials and biological models, Chem. Rev. 104 (2004) 2723–2750. [34] J.S. Bradshaw, R.M. Izatt, Crown ethers: the search for selective ion ligating agents, Accounts Chem. Res. 30 (1997) 338–345. [35] X.J. Pan, Y.L. Wang, H. Jiang, G. Zou, Q.J. Zhang, Benzo-15-crown-5 functionalized polydiacetylene-based colorimetric self-assembled vesicular receptors for lead ion recognition, J. Mater. Chem. 21 (2011) 3604–3610. [36] G. Nasr, H. Bestal, M. Barboiu, B. Bresson, T. Thami, Functionalization of polymethylhydrosiloxane gels with an allyl ureido benzocrown ether derivative: complexation properties, J. Appl. Polym. Sci. 111 (2009) 2785– 2797. [37] G. Dubois, C. Reye, R. Corriu, C. Chuit, Organic–inorganic hybrid materials. Preparation and properties of dibenzo-18-crown-6 ether-bridged polysilsesquioxanes, J. Mater. Chem. 10 (2000) 1091–1098. [38] M.M. Collinson, Sol–gel strategies for the preparation of selective materials for chemical analysis, Crit. Rev. Anal. Chem. 29 (1999) 289–311. [39] K. Kimura, T. Sunagawa, S. Yajima, S. Miyake, M. Yokoyama, Neutral carriertype ion sensors based on sol–gel-derived membranes incorporating a bis(crown ether) derivative by covalent bonding, Anal. Chem. 70 (1998) 4309–4313. [40] I.M. El-Nahhal, N.M. El-Ashgar, A review on polysiloxane-immobilized ligand systems: synthesis, characterization and applications, J. Organomet. Chem. 692 (2007) 2861–2886. [41] N. El-Ashgar, M. Abdel-Latif, Synthesis and applications of a new polysiloxaneimmobilized macrocyclic ligand system, Anal. Lett. 41 (2008) 3074–3087. [42] G.G. Talanova, K.B. Yatsimirskii, O.V. Kravchenko, Peculiarities of K2PdCl4 and K2PtCl4 complexation with polymer-supported dibenzo-18-crown-6, Ind. Eng. Chem. Res. 39 (2000) 3611–3615. [43] C. Fontàs, E. Anticó, F. Vocanson, R. Lamartine, P. Seta, Efficient thiacalix[4]arenes for the extraction and separation of Au(III), Pd(II) and Pt(IV) metal ions from acidic media incorporated in membranes and solid phases, Sep. Purif. Technol. 54 (2007) 322–328. [44] H.J. Schneider, A.K. Yatsimirsky, Selectivity in supramolecular host-guest complexes, Chem. Soc. Rev. 37 (2008) 263–277. [45] M.J. Deetz, M. Shang, B.D. Smith, A macrobicyclic receptor with versatile recognition properties: simultaneous binding of an ion-pair and selective complexation of dimethylsulfoxide, J. Am. Chem. Soc. 122 (2000) 6201–6207. [46] E.P. Horwitz, M.L. Dietz, S. Rhoads, C. Felinto, N.H. Gale, J. Houghton, A leadselective extraction chromatographic resin and its application to the isolation of lead from geological samples, Anal. Chim. Acta 292 (1994) 263–273.

46

F. Bai et al. / Separation and Purification Technology 106 (2013) 38–46

[47] F. Bai, G. Ye, G. Chen, J. Wei, J. Wang, J. Chen, New macrocyclic ligand incorporated organosilicas: co-condensation synthesis, characterization and separation of strontium in simulated high level liquid waste, React. Funct. Polym. 73 (2012) 228–236. [48] E.P. Horwitz, M.L. Dietz, D.E. Fisher, Separation and preconcentration of strontium from biological, environmental, and nuclear waste samples by

extraction chromatography using a crown ether, Anal. Chem. 63 (1991) 522– 525. [49] S. Lagergren, About the Theroy of So-Called Adsorption of Soluble Substances, vol. 24, Kungliga Svenska Vetenskapsakademiens. Handlingar, 1898. [50] Y.S. Ho, G. Mckay, A comparison of chemisorption kinetic models applied to pollutant removal on various sorbents, Process. Saf. Environ. 76 (1998) 332–340.