2.1 Materials

All the chemicals (including TU and nitrates of Pd(II), Ce(III), Sr(II), La(III), Dy(III), Eu(III), Sm(III), etc.) were of analytical grade. The extracting agent, isoBu-BTP/SiO₂-P, was prepared as described in the previous study ^[24]. 0.5g of isoBu-BTP was impregnated into 1.0g silica/polymer composite support (SiO₂-P). In other words, the content of the extracting agent was 33.3% of the total mass of the adsorbent. The SiO₂-P support with pore size of 0.6 μm, pore fraction of 0.69, and mean diameter of 50 μm, contains a macroreticular styrene–divinylbenzene copolymer (SDB) which was immobilized in porous silica (SiO₂) particles (Shanghai Haiyuan Chemical Technology Co., Ltd., China). Fig. 1(a) and Fig. 1(b) show the chemical structure of isoBu-BTP and SEM image of isoBu-BTP/SiO₂-P, respectively.

Fig. 1.Full-screen View

Chemical structure of isoBu-BTP(a) and SEM image of isoBu-BTP /SiO₂-P(b)

2.2 Batch adsorption and batch desorption experiments

The adsorption of isoBu-BTP/SiO₂-P adsorbent towards metal ions and the desorption of Pd(II) from loaded adsorbent were tested by batch experiments. In batch adsorption experiments, the solutions were prepared by dissolving metal nitrates into required concentration of nitric acid. For each batch experiment, 0.1 g adsorbent was combined with 5 mL aqueous solution in a glass vial with a screw cap. The mixture in the vial was shaken mechanically at 120 rpm at 298 K for a pre-determined time in a water bath. The aqueous phase was filtrated through a membrane filter with 0.45 μm pore size. The concentrations of metal ions in solution were determined by an inductively coupled plasma atomic emission spectrometry (ICP-AES, Shimadzu ICP-7510, Japan).

The adsorption amount per unit adsorbent, Q (mmol g^-1), is calculated by Eq. (1):

where C_o and C_e (mmol L^-1) are the concentration of metal ions in the aqueous phase before and after adsorption, respectively; V(L) and W(g) indicate the volume of the aqueous phase and the weight of the dry adsorbent, respectively.

In batch desorption experiments, the loaded adsorbent was obtained by filtrating the adsorbent after 24 h’s adsorption towards Pd(II). In consideration of the hydrolysis of Pd(II), the eluting solutions were made by dissolved TU in a low concentration of nitric acid. 0.1 g loaded adsorbent was contacted with 5 mL TU eluting solution and separated from the aqueous solution after 24 h. Then, the metal ion concentration was measured as described above.

The desorption performance was evaluated by using desorption efficiency, E (%), which was calculated by Eqs. (2) and (3).

Q_d is the desorption amount in mmol g^-1; C_d indicates desorbed Pd(II) concentration in eluting solution in mmol L^-1.

2.3 Stability evaluation

To investigate the stability of the isoBu-BTP/SiO2–P adsorbent against nitric acid, 0.1g adsorbent and 5 mL nitric acid solutions were mixed into a glass vial, which was shaken mechanically at 120 rpm at a designed contact time. After filtrated and dried, the stability of the adsorbent was evaluated by batch adsorption experiments.

2.4 Column experiment

A glass column with a dimension of 5 mm in inner-diameter and 500 mm in effective length packed with 4.3 g of isoBu-BTP/SiO₂-P adsorbent was used for the column experiment. The column experiment was performed at 298 K by circulating the thermostated water through the water jacket of the column. The flow rate was controlled to 0.1 mL/min using a NPG-50UL Model pressure gage (Nihon Seimitsu Kagaku Co., Ltd., Japan) and a 2GN15K Model pressure limiter (Oriental Motor Co., Ltd., Japan). Before the experiment, a 15 mL HNO₃ (3 mol L^-1) solution was passed through the column to reach pre-equilibrium state. Then, the feed solution and different kinds of eluents subsequently flowed through the column. Effluents were collected every 15 min by an EYELA DC-1500 Model auto-fractional collector (Tokyo Rikakikai Co., Ltd., Japan). The concentrations of metal ions in the initial solution (C₀) and effluents (C) were determined by ICP-AES.

2.5 Cyclic voltammetry studies

The cyclic voltammetry of the desorption solution (10 mL) was recorded at a platinum plate (Φ=2mm) working electrode at 298 K. The electrolyte was made by dissolving Pd(NO₃)₂•2H₂O into the appropriate concentration of nitric acid and TU in accordance with the desorption solution. A platinum plate (15 mm × 15 mm × 0.2 mm) was used as the counter electrode and a KCl-saturated Ag/AgCl electrode (0.197 V vs. SHE) acted as reference electrode. Before each electrochemical experiment, the working electrode surface was polished with 1.0, 0.3, and 0.05 micro gamma-alpha alumina powder in a sequence to a mirror finish and then washed in 3 mol L^-1 HNO₃, acetone, ethanol, and distilled water in an ultrasonic cleaner for 10 min, respectively. At the same time, the solutions were deoxygenated by ventilating pure argon for 10 min. All the electrochemical experiments were controlled with an electrochemical workstation CorrTestTM model CS1350 equipped by electrochemical testing and analysis tool Version 4.3 software.

2.6 Electrodeposition studies

Electrolysis experiments were conducted on a titanium plate (50mm×50mm×2mm) in desorption solution (100 mL). The electrolyte was made by dissolving Pd(NO₃)₂•2H₂O into 0.1 mol L^-1 nitric acid and the appropriate concentration of TU. A platinum plate (15 mm×15 mm×0.2 mm) and a KCl-saturated Ag/AgCl electrode (0.197 V vs. SHE) acted as the counter electrode and the reference electrode, respectively. After electrolysis, the electrolytes were examined by UV–VIS absorption spectrum and the concentrations of Pd(II) were determined by ICP-AES. The morphology and the composition of the electrodeposited palladium crystallizations was characterized by a low vacuum ultra-high resolution field emission scanning electron microscope (LV UHR FE-SEM, OVA NanoSEM 230) with a large area silicon drift detector energy-dispersive spectrometer (Large Area SDD EDS, AZtec X-Max 80).

3.1 Adsorption studies

Figure 2 shows the adsorption behavior of isoBu-BTP/SiO₂-P towards the typical FPs contained in HLLW in different initial HNO₃ concentrations. As can be seen, Pd(II) showed a strong adsorption at all the examined HNO₃ concentrations. However, the adsorbent showed no or weak adsorption towards Ce(III), Zr(IV), Mo(VI), Sr(II), Ru(III), Nd(III), Y(III) and La(III). Dy(III), Eu(III), Sm(III), and Gd(III) exhibited a little stronger adsorption onto the adsorbent in ≥0.5mol L^-1 HNO₃ solution. So, the high selectivity of isoBu-BTP/SiO₂-P towards Pd(II) is rather obvious. Pd(II) is a well-known soft metal ion, and the adsorption is considered to result from the complex formation with the soft ligand, isoBu-BTP ^[16].

Fig. 2.Full-screen View

(Color online) Effect of initial HNO₃ concentration on the adsorption capacity of some typical FPs in HLLW onto isoBu-BTP/SiO₂-P adsorbent (metal ions: 1 mmol L^-1, adsorption conditions: phase ratio: 0.1 g/5 mL, temperature: 298 K, contact time: 24 h).

3.2 Desorption studies

The Pd(II) desorption efficiency from loaded isoBu-BTP/SiO₂-P by TU is showed in Fig. 3. The effective elution of Pd(II) by TU from loaded BTP/SiO₂-P was achieved and is attributed to the fact that Pd(II) can form stable complexes with TU because of the soft atoms N and S contained in TU ^[22]. The desorption efficiency increased greatly by increasing the concentration of TU within 0.2 mol L^-1 and the desorption efficiency reached the maximum gradually. When the concentration of TU increased to 0.1 mol L^-1, 95% of palladium was desorbed. From Fig. 3, it can also be seen that the nitric acid concentration at 0.01 and 0.1 mol L^-1 in eluting solution have little effect on the desorption efficiency. As the weak adsorption performance of the other FPs presented in Fig. 2 were under 0.01 mol L^-1 HNO₃, 0.1 mol L^-1 HNO₃ was chosen as the acid concentration of the TU eluent solution in the latter column experiment to avoid the possible elution of the other FPs by low acid concentration.

Fig. 3.Full-screen View

(Color online) Effect of the concentration of TU on Pd(II) desorption from the loaded isoBu-BTP/SiO₂-P adsorbent (adsorption conditions: phase ratio: 0.1 g/5 mL, temperature: 298 K, time: 24 h, [Pd]: 10 mmol L^-1, 3 mol L^-1 HNO₃; desorption conditions: phase ratio: 0.1 g/5 mL, temperature: 298 K, time: 24 h).

3.3 Stability evaluation

Due to the strong acidic environment of HLLW, the stability of the adsorbent was investigated by a set of experiments performed in 3 mol L^-1 HNO₃. Fig. 4 shows the effect of contact time with nitric acid solution on the adsorption of Pd(II) onto isoBu-BTP/SiO₂-P. As can be seen, a small decrease in the adsorption amount of Pd(II) occurred after the contact with 3 mol L^-1 HNO₃, and the adsorption showed almost no change as the contact time increased. The results indicate the adsorbent was affected by high concentrations of nitric acid and very limited quantities of extractant would be leaked into aqueous solution from the adsorbent. In consideration of the adsorption performance in long-time contact with 3 mol L^-1 HNO₃ solution, isoBu-BTP/SiO₂-P could be used a few more times in the column chromatographic process.

Fig. 4.Full-screen View

Effect of contact time with nitric acid solution on the adsorption of Pd(II) onto isoBu-BTP/SiO₂-P. (adsorption conditions: adsorbent: phase ratio: 0.1 g/5 mL, [Pd]: 15 mmol L^-1, 3 mol L^-1 HNO₃, temperature: 298 K, contact time: 24 h, shaking speed: 120 rpm).

3.4 Column experiment

Based on the batch experimental results, the separation performance of Pd(II) from a 3 mol L^-1 HNO₃ solution containing the typical FPs was conducted using the isoBu-BTP/SiO₂-P adsorbent packed column. The concentration of metal ions in the feed solution was 1 mmol L^-1. After the feed solution passed through the column, the given volumes of 3 mol L^-1 HNO₃, 0.1 mol L^-1 HNO₃, and 0.1 mol L^-1 HNO₃-0.1 mol L^-1 TU were subsequently flowed through the adsorption column. The partitioning results are showed in Fig. 5.

Fig. 5.Full-screen View

(Color online) Results of Pd(II) separation from a simulated HLLW by an isoBu-BTP/SiO₂-P adsorbent packed column (A: dead volumn, B: 5 mL feed solution, C:30 mL 3 mol L^-1 HNO₃, D: 40 mL 0.1 mol L^-1 HNO₃, E: 30 mL 0.1 mol L^-1 HNO₃-0.1 mol L^-1 TU).

As can be seen, La(III), Rh(III), Ru(III), and Ce(III) showed no or very weak adsorption and flowed through the column with the feed and washing solution. The other FPs, including Pd(II), were adsorbed strongly by the isoBu-BTP/SiO₂-P adsorbent. These FPs, except Pd(II), were eluted by 0.1 mol L^-1 HNO₃. Pd(II) was effectively eluted by using a TU solution. The elution band of Pd(II) is narrow and sharp, reflecting a fast elution dynamic. So, a successful separation of palladium from a nitric acid medium was achieved with a recovery ratio of more than 70%. These results indicate the feasibility to separate palladium from HLLW by extraction chromatography using BTP/SiO₂-P adsorbent. To recover metallic palladium, further electrochemistry studies were conducted towards palladium desorption solution, and the results were discussed below.

3.5 Cyclic voltammetry studies

Pd(II), nitrate ion, and TU tend to form complexes in an aqueous solution which would lead to a complex performance in the electrochemistry behavior of Pd(II). In order to understand the electrochemistry behavior of Pd(II), complexes formed in palladium desorption solution (10 mmol L^-1 Pd(II), 0.1 mol L^-1 TU, and 0.1 mol L^-1 HNO₃) and their potentials are discussed first.

At 0.1 mol L^-1 HNO₃, Pd (II) and a nitrate ion can mainly form one kind of complex ^[12] while Pd(II) and TU can form four kinds of complexes. Their formations in accordance with their stability constants are as follows ^[25]:

where β, β₁, β₂, β₃, β₄ are cumulative complex stability constants.

As the cumulative complex stability constants of Pd(NO₃)⁺ is far less than Pd(TU)_n²⁺, the complex of Pd(NO₃)⁺ formed in the solution was so little that it can be negligible. In consideration of the total concentration of Pd²⁺ (10 mmol L^-1) and TU (100 mmol L^-1), the identical mass balance can be described using the following equations:

After calculation, the results were (mmol L^-1): [Pd²⁺]=3.867×10^-15 mmol L^-1, [TU]=60 mmol L^-1, [PdTU²⁺]=5.83×10^-12 mmol L^-1, [Pd(TU)₂²⁺]=1.75×10^-8 mmol L^-1, [Pd(TU)₃²⁺]=4.19×10^-5 mmol L^-1, and [Pd(TU)₄²⁺] $\approx$ 10 mmol L^-1.

The reduction process of palladium (II) can be described using Eqs. (11)-(13):

The derivation of the value of reversible potential can be described as Eq. (14) ^[26]:

where E₂⁰ (V) is the reversible standard potential corresponding to each form of palladium complexes, E₃⁰ =0.915 V vs. SHE; R, T, and F are the gas constant (8.314 J mol^-1 K^-1), temperature (K), and Faraday constant (96485 C mol^-1), respectively.

The reversible potential for each species (E) in the solution can be obtained by using the Nernst equation and considering the real concentration of each complex. The calculation results are given in Table1. From the table, it is clear the dominant species in the desorption solution is Pd(TU)₄²⁺, and its reduction happens at about -0.019V.

Then, the cyclic voltammogram of palladium desorption solution was recorded at a platinum plate electrode (Φ=2mm) at a scan rate of 10 mV/s at 298 K, as shown in Fig. 6A. Three reduction peaks (a, b, and c) occurred continuously with small potential differences, which make it confused to figure out the Pd(II) reduction. To figure out the reaction that happened in the solution, the CVs of 10 mmol L^-1 Pd(II) in 0.1 mol L^-1 nitric acid (Fig. 6B) and 0.1 mol L^-1 TU in 0.1 mol L^-1 nitric acid (Fig. 6C) were obtained at the same scan rate and temperature.

Fig. 6.Full-screen View

Cyclic voltammograms of (A) 10 mmol L-1 Pd(II) in 0.1 mol L-1 nitric acid containing 0.1 mol L^-1 TU, (B) 10 mmol L^-1 Pd(II) in 0.1 mol L^-1 nitric acid, and (C) 0.1 mol L^-1 TU in 0.1 mol L^-1 nitric acid.

Without the participation of TU, it is clear in Fig. 6B that the onset of Pd(II) reduction occured at a potential of 0.4 V and the peak appeared at 0.27 V. At a more negative potential, the reduction peak at -0.28 V and -0.4 V can be described as hydrogen adsorption on metallic palladium and hydrogen evolution reaction, respectively ^[14]. The oxidation peak occurred at -0.18 V which suggests the desorption of hydrogen from palladium. Compared with Fig. 6B, the hydrogen adsorption and desorption disappeared in the addition of TU, as presented in Fig. 6A. It is obvious that the reduction potential of Pd(II) shifts much towards negative, while a small negative shift of hydrogen evolution (Fig. 6A(a)) is also observed.

TU would react on the electrode in a nitric acid solution, as shown in Fig. 6C. A wide electrochemical window is exhibited in Fig. 6C due to the adsorption of TU on the surface of the platinum electrode in the first scan cycle. When it comes to a positive potential, the adsorbed TU was oxidized and the electrode surface was exposed to the solution. Then the Faradic current of TU series reduction occurred. As can be seen from Fig. 6C, the reduction potential of TU is very close to the calculated results of the Pd(II) reduction potential in Table 1. Therefore, TU would compete with Pd(II). In fact, TU has a very complex electrochemistry behavior in nitric acid solution. Eq. (15–17) give the main reactions related to the redox of TU in nitric acid ^[27]. Compared with Fig. 6C, similar results were observed from Fig. 6A that TU would be adsorbed on the electrode and hinder the occurrence of the Faradic current of palladium and TU series reduction in the first scan cycle. Then, several current peaks were observed.

After the above comparisons, the hydrogen evolution (Fig. 6A(a)), Pd(II) reduction(Fig. 6A(b)), and the redox of TU (Fig. 6A(c) and Fig. 6A(d)) are clear in Fig. 6A. The weak reduction wave at -0.4 V suggests the deposition of Pd(II) and the deposition is typically an irreversible process. TU leads to a negative shift of Pd(II) reduction potential and can be comparative to the reduction of hydrogen and Pd(II). These results indicates that it is possible to recover palladium in metallic form by direct electrolysis from the desorption solution, and the results are discussed below. However, further research is still needed to reduce the negative impact brought by TU.

3.6 Electrodeposition studies of Pd

The working voltage was selected based on the cyclic voltammogram (Fig. 6). All the electrowinning studies were conducted on a titanium plate (50 mm×50 mm×2 mm).

The UV–Visible absorption spectrum of 0.1 mol L-1 TU and effect of electrolysis time on the UV-Visible absorption of the electrolyte at -0.6V were examined, and the results are shown in Fig. 7. There are two obvious absorption peaks located at 234 and 197 nm attributed to the characteristic absorption of TU. Compared with the UV–Visible absorption spectra of TU, the only absorption reduction of electrolytes performed at 234 nm indicates that Pd(II) forms complexes with the sulfur atom and the nitrogen atom has no interaction with Pd(II). It is clear that the complexes had a strong absorbance in the vicinity of 290 nm. With the increasing of electrolysis time, the UV–Visible absorption of TU and palladium complexes both decrease, which indicates the deposition of palladium from palladium complexes accompanied by the consumption of TU on the electrode. The consumption of TU and palladium complexes was slow in the first hour and then became fast. After 6 h, the absorption of palladium complexes nearly disappeared, which indicates that most Pd(II) had been deposited.

Fig. 7.Full-screen View

(Color online) Effect of electrolysis time on the UV–VIS absorption spectrum of the electrolyte (initial solution: 10 mmol L^-1 Pd(II), 0.1 mol L^-1 HNO_3, and 0.1 mol L^-1 TU), T=298 K.

The effect of electrodeposition time on the recovery ratio for the electrodeposition of Pd(II) (10 mmol L^-1) in 0.1 mol L^-1 HNO₃ and 0.1 mol L^-1 TU at three different potentials (-0.5 V, -0.6 V, -0.7 V) is described in Fig. 8. As can be seen, the recovery ratio increased with increasing the electrodeposition time at different potentials. After 6 h’s electrolysis, both recovery ratio of palladium at -0.6 V and -0.7 V could reach 92%, higher than 85% at -0.5V. During the electrolysis, a small deposition rate of Pd(II) can be observed in the first hour due to the removal of TU from the electrode surface, which was consistent with UV–Visible absorption results. A higher deposition rate was accompanied by a higher working potential. However, the mean current efficiency for the electrodeposition of Pd at -0.7 V, -0.6 V, and -0.5 V were 5.0%, 7.6%, and 11.0%, respectively. The interpretation for the low efficiency could be that the reduction of H⁺, Pd(II), and TU happened at the same time due to their close reduction potential, which was consistent with the CV results.

Fig. 8.Full-screen View

(Color online) Effect of electrodeposition time on recovery ratio for the electrodeposition of Pd(II) from 0.1 mol L-1 HNO3 and 0.1 mol L-1 TU at different working potentials at 298 K.

Figure 9 shows the influence of TU concentration on the electrodeposition of Pd(II) (10mmol L^-1) at a constant applied potential of -0.6V. It is clear that the curves obtained from three TU concentrations (0.05, 0.1, 0.2 M) are close to each other. All the recovery ratios that could reach about 90% after 6 h’s electrolysis accompanied by and the mean current efficiency are 8%, 7.6%, and 6.5%, respectively. Lower current efficiency could be due to the competition of hydrogen evolution and the reduction of TU. The results indicate that TU makes a small difference to the recovery ratio in the range of this study concentration. However, TU greatly reduces the current efficiency due to the redox of TU series on electrodes. So the current efficiency of electrodeposition and rate of recovery of palladium can be improved and optimized by reducing or removing TU contained in the solution.

Fig. 9.Full-screen View

(Color online) Recovery ratio for the electrodeposition of Pd(II) from different TU concentrations in 0.1 mol L^-1 HNO₃ at -0.6V, T=298 K.

Figure 10 and Fig. 11 show the SEM and EDS characterization of palladium electrodeposition on the platinum electrode with E=-0.6V for 30min. From the SEM images (Fig. 10), a thick and relatively homogeneous and uniform distribution of palladium particles was observed. The clear dendrite morphology indicates a dendrite growth. In Fig. 11, the EDS pattern indicates that palladium was deposited successfully on the surface of the electrode. A small amount of sulphur would be deposited along with palladium during electrolysis, which would help palladium form a dense deposit ^[28]. The existence of sulphur could be removed by some technology, which still needs a further study.

Fig. 10.Full-screen View

The SEM images of the palladium deposit obtained from 10 mmol L-1 Pd(II) in 0.1 mol L^-1 nitric acid and 0.1 mol L^-1 TU.

Fig. 11.Full-screen View

(Color online) The EDS pattern of the palladium deposit obtained from 10 mmol L^-1 Pd(II) in 0.1 mol L^-1 nitric acid and 0.1 mol L^-1 TU.

3.7 Proposal of EC-ED process

The extraction chromatography method using the BTP/SiO₂-P adsorbent has shown a good performance in the separation of Pd(II) from nitric acid medium containing the typical FPs. Electrochemistry studies also verified the feasibility to recover palladium in metal form from desorption solution. According to these results, the conceptual flowchart of the EC-ED process for the recovery of palladium from HLLW was proposed, as shown in Fig. 12. HLLW flowed through a column packed with BTP/SiO₂-P adsorbent. Pd(II) and MA could be effectively separated by choosing appropriate eluents. Then, the separated Pd(II) could be recovered by the electrodeposition method after some pretreatment. The eluent for the separation of palladium, such as TU, could be the link between the extraction chromatography method and electrochemical approach. Due to the complexity of HLLW, some host tests and optimum conditions still need to be studied.

Fig. 12.Full-screen View

(Color online) The conceptual flowchart of the EC-ED process for the recovery of palladium from HLLW.