Direct separation of minor actinides from high level liquid waste by Me2-CA-BTP/SiO2-P adsorbent

Directly separating minor actinides (MA: Am, Cm, etc.) from high level liquid waste (HLLW) containing lanthanides and other fission products is of great significance for the whole nuclear fuel cycle, especially in the aspects of reducing long-term radioactivity and simplifying the post-processing separation process. Herein, a novel silica-based adsorbent Me2-CA-BTP/SiO2-P was prepared by impregnating Me2-CA-BTP (2,6-bis(5,6,7,8-tetrahydro-5,8,9,9-tetramethyl-5,8-methano-1,2,4-benzotriazin-3-yl)pyridine) into porous silica/polymer support particles (SiO2-P) under reduced pressure. It was found Me2-CA-BTP/SiO2-P exhibited good adsorption selectivity towards 241Am(III) over 152Eu(III) in a wide nitric acid range, acceptable adsorption kinetic, adequate stability against γ irradiation in 1 and 3 M HNO3 solutions, and successfully separated 241Am(III) from simulated 3 M HNO3 HLLW. In sum, considering the good overall performance of Me2-CA-BTP/SiO2-P adsorbent, it has great application potential for directly separating MA from HLLW, and is expected to establish an advanced simplified MA separation process, which is very meaningful for the development of nuclear energy.

As the development of nuclear energy, spent nuclear fuel in storage has amounted to around 266 000 t of heavy metal (HM) and is accumulating at a rate of around 7000 t HM/year, which has long-term potential radioactivity threats to the environment and must be managed safely and efficiently. Although the industrialized PUREX (Plutonium Uranium Recovery by EXtraction) process based on TBP can recover 99.5% of uranium and plutonium from spent nuclear fuel, the resulting high-level radioactive liquid waste (HLLW) containing most of the fission products (FP) and minor actinides (MA: Am, Cm, etc.) reserves most of the radioactivity, especially the α-emitters MA, which are the main contributors of radiotoxicity after three centuries storage 1 . Partitioning MA and the other long-lived FP from HLLW and then transmuting them into short-lived or stabile nuclides, which is the so-called Partition and Transmutation strategy (P&T), can reduce the time isolated from the environment from over 20,000 years to 300-500 years [2][3][4][5] . However, as the content of lanthanides (Ln), which accounts for about 1/3 of the FP, is much higher than that of MA in order of magnitude, and some Ln have large neutron absorption cross sections and are neutron poisons, Ln should be separated from MA before the transmutation step 5,6 . Considering MA and Ln dominantly exist as trivalent cations in solution with comparable radii and coordination numbers, mutual separation of MA(III) and Ln(III) is very difficult, which requires high selectivity for the separation materials [7][8][9][10][11] . Moreover, as the acid in HLLW produced by PUREX process was 2-5 M HNO 3 and HLLW is highly radioactive with α, β, γ radioactivity, it further requires that the materials used for MA separation shall have good acid and irradiation resistance stability.
In a word, more challenging processes for directly separating MA from HLLW are eagerly needed to be developed. Separating MA from HLLW by one step process has been proposed, such as 0.2 M CMPO + 1.4 M TBP in n-dodecane used in SETFICS process 5 , 0.015 M CyMe 4 -BTBP + TODGA or DMDOHEMA in TPH/Octanol used in the 1-cycle SANEX process 5,13 , TODGA + water-soluble SO 3 -Ph-BTBP in TALSPEAK process 31 . As the processes mentioned above involved more than one kind of extractant, it is inconvenient for the organic phase regeneration. Wei. etc. proposed an compact and effective process based on extraction chromatographic technology named MAREC process (Minor actinide extraction by chromatographic process) as shown in Fig. 1S 32 whose purpose is to directly separate MA(III) from PUREX raffinate HLLW containing fission products Ln and other FPs through single column packed with high-efficiency microporous adsorbent, e.g., BTPs/SiO 2 -P. If success, it will significantly simplify the MA separation process and is of great significance for the whole nuclear fuel cycle. Efforts to realize the single-column MAREC process from simulated HLLW have been made by using isoHex-BTP/SiO 2 -P by co-authors 33,34 , but isoHex-BTP/SiO 2 -P turned out to be easy to lose the adsorption ability when it accepted γ irradiation in 3 M HNO 3 solution in our previous research 35 , so more advanced materials are needed to be developed. Furthermore, MAREC process uses almost no organic solvent, avoiding the third phase formation and large amount of secondary organic waste accumulation.
In this work, Me 2 -CA-BTP/SiO 2 -P adsorbent was prepared by impregnating 5 g Me 2 -CA-BTP dissolved in CH 2 Cl 2 into 10 g stable microporous silica/polymer composite support (SiO 2 -P) particles under reduced pressure. The adsorbent properties of selectivity, kinetics, γ-irradiation stability was studied by batch adsorption experiments. Finally, the feasibility of using Me 2 -CA-BTP/SiO 2 -P adsorbent to directly separate MA(III) from 3 M HNO 3 simulated HLLW through single-column MAREC process was evaluated.

Results
Adsorption selectivity. To study adsorption selectivity, the effects of initial HNO3 concentration (pH = 4-4 M) on Me 2 -CA-BTP/SiO 2 -P adsorption towards 241 Am(III) and 152 Eu(III) were evaluated. Figure 2 illustrates the distribution coefficients (K d ) of 241 Am(III) and 152 Eu(III) and the separation factors of 241 Am(III) over 152 Eu(III) (SF Am/Eu ) changing as a function of HNO 3 concentration. The distribution factors K d of 241 Am(III) and 152 Eu(III) increased as the HNO 3 concentration increased until up to 2 M HNO 3 solution which are explained by the adsorption of BTP towards 241 Am(III) needing the participation of NO 3 −28 . Then with further increase of the acid, the distribution factors K d of 241 Am(III) and 152 Eu(III) decreased. The decreasing distribution factors are explained by metal ions ( 241 Am(III) and 152 Eu(III)) and nitric acid competing adsorption for BTP. The whole adsorption mechanism may be as shown in Equation (1) and (2) similar to CA-BTP 28 . Three NO 3 − complexes are incorporated into the coordination sphere so as to maintain neutral due to metal iron's trivalent oxidation state Overall, Me 2 -CA-BTP/SiO 2 -P exhibited good adsorption selectivity towards 241 Am(III) with the uptake rate of 241 Am(III) over 96% and SF Am/Eu over 75 in a wide nitric acid range of 1-3 M, which was high enough for effective separation 241 Am(III) from 152 Eu(III). Furthermore, according to previous work 29, 36 , Me 2 -CA-BTP/SiO 2 -P showed almost no adsorption ability towards most typical FP except Pd(II) in 3 M HNO 3 . The results above indicates that Me 2 -CA-BTP/SiO 2 -P has the potential to directly separate MA from high nitric acid (3 M) HLLW. The adsorption ability and the applicable acidity range of Me 2 -CA-BTP/SiO 2 -P towards Am(III) has been improved and enlarged compared with CA-BTP which only exhibited good adsorption ability towards Am(III) in 0.5-1 M HNO 3 and CyMe 4 -BTBP whose adsorption ability towards Am(III) increased as HNO 3  Stability evaluation. Considering the radioactive nuclides in HLLW emit vast amounts of radiation, e.g., α, β, γ, the adsorbent irradiation stability is a serious concern in practical applications for HLLW reprocessing.

Directly separate MA(III) from HLLW by single-column MAREC process.
To examine the separation of MA(III) from fission products, a hot test using simulated 3 M HNO 3 HLLW containing typical FP (Sr, Y, Zr, Mo, Ru, Pd, La, Ce, Nd, Sm, Eu, Gd, Dy, 5 mM respectively) and 241 Am(III) (500 Bq/mL) was carried out at 35 °C using a glass column (5 mm in inner diameter and 500 mm in length) packed with 5 g Me 2 -CA-BTP/SiO 2 -P with the results shown in Fig. 5 and Table 1. As can be seen, Me 2 -CA-BTP/SiO 2 -P showed very poor or almost no adsorption towards lanthanides and most other typical fission products, such as Sr, Y, Zr, Mo, Ru. These elements flowed out with the feed solution in step B and the following 3 M and 0.1 M HNO 3 solution in step C and D. On the other hand, Pd, well known as a "soft" metal ion which has strong electron acceptance ability, was adsorbed onto Me 2 -CA-BTP/SiO 2 -P and could not be effectively eluted from the adsorbent by reducing the HNO 3 concentration. A complexing agent, thiourea, which has strong complexation affinity with Pd, was attempted to desorb the adsorbed Pd. As shown in Fig. 5, Pd can be effectively eluted off by 0.01 M HNO 3-0.1 M thiourea in step E. Meanwhile, 241 Am was strongly adsorbed by Me 2 -CA-BTP/SiO 2 -P adsorbent, then it was eluted off using 0.001 M HNO 3 -0.01 M DTPA as an eluent with 241 Am recovery yield of 95.87% and the other typical fission products less than 3% except Dy. But considering there is almost no Dy or other Ln heavier than Dy in HLLW, the effects caused by Dy are almost negligible. In a word, a successful separation between Am and the various typical fission products including lanthanides has been achieved, which is an important step toward a simplified direct process for separation MA from HLLW.  In a word, Me 2 -CA-BTP/SiO 2 -P has great application potential in the single-column MAREC process, and is expected to establish an advanced simplified MA(III) separation process.

Methods
Reagents. FP element nitrates (FP: Sr(II), Zr(IV) and trivalent rare earths) and (NH 4 ) 6 Mo 7 O 24 •4H 2 O were commercial reagents of analytical grade. Pd(NO 3 ) 2 •2H 2 O was chemical pure with Pd(II) ≥ 39.5 wt% and chloridate ≤ 0.04 wt%. Ru(III) nitrosyl nitrate solution was in diluted nitric acid containing 1.5 wt% of Ru(III) with a density of 1.07 g·mL −1 . 241 Am(III) and 152 Eu(III) were from laboratory stock solution. Me 2 -CA-BTP was synthesized at laboratory with the purity of 97% according to the HPLC-MS test. All solutions were prepared with deionized water at 18 MΩ·cm resistance (DI water). Other agents such as nitric acid, dichloromethane, etc. were of analytical grade and used without further treatment.
Preparation of Me 2 -CA-BTP/SiO 2 -P adsorbent. The support SiO 2 -P with pore size of 0.6 μm, pore fraction of 0.69 and mean diameter of 50 μm was developed in the previous work 39 . P refers to macroreticular styrene-divinylbenzene copolymer (SDB) and is immobilized in porous silica (SiO 2 ) particles with the content of 17-18 wt% in SiO 2 -P. The synthesis procedure of Me 2 -CA-BTP/SiO 2 -P adsorbent was the same as ref. 40 . The synthesized Me 2 -CA-BTP/SiO 2 -P adsorbent overcame the limitation of low solubility of BTP in traditional diluents as the content of Me 2 -CA-BTP was as high as 32.0 wt% in the adsorbent 29 . The synthetic Me 2 -CA-BTP/SiO 2 -P adsorbent was characterized by high resolution field emission scanning electron microscope (SEM, Sirion 200, FEI COMPANY) and the SEM image is shown in Fig. 2S.

Batch adsorption experiments.
For batch adsorption experiments towards 241 Am(III) and 152 Eu(III), 0.1 g adsorbent was combined with 5 mL aqueous solution in a 12 mL glass vial with screw teflon cap. The mixture in the vial was shaken mechanically at 300 rpm at 25 °C for a certain time and the solid-liquid separation was realized by centrifugation. The radioactivity of 241 Am(III) and 152 Eu(III) in solution before and after adsorption were determined by high-purity germanium multichannel gamma spectrometer (CANBERRA) at 59.5 and 121.78 keV respectively, while the concentrations of non-radioactive FP were determined by inductively coupled plasma-optical emission spectrometer atomic emission spectroscopy (ICP-AES: Shimadzu ICPS-7510).
The distribution coefficient K d (mL·g −1 ), separation factor SF A/B and uptake rate E (%) which are key factors in solid-liquid adsorption are calculated by Equation (3), (4) and (5), respectively.  where A o , A e denote the radioactivity of metal ions in the aqueous phase before and after adsorption, respectively, Bq·mL −1 . V (mL) indicates the volume of aqueous phase and W (g) is the mass of dry Me 2 -CA-BTP/SiO 2 -P.
Column separation experiments. Column separation experiments for simulated HLLW solution were carried out using a glass column with 5 mm in diameter and 500 mm in length. 5 g Me 2 -CA-BTP/SiO 2 -P adsorbent was transferred to the column in the slurry state under atmosphere. The column volume (CV) of the adsorbent bed was 9.8 cm 3 . The column was kept at a constant temperature (35 °C) with water jacket. Prior to the separation experiment, the adsorbent in column was pre-equilibrium by passing 50 mL of 3 M HNO 3 . The simulated HLLW contained 5 m mol/L for each FP (Sr, Y, Zr, Mo, Ru, Pd, La, Ce, Pr, Nd, Sm, Eu, Gd, Dy) and 500 Bq/ mL 241 Am(III) in 3 M HNO 3 solution. All the mobile phase was pumped at 0.1 mL/min and the effluents were collected by a fractional collector. The metal concentrations in the effluents without radioactivity were determined by ICP-AES while 241 Am(III) was determined by high-purity germanium multichannel gamma spectrometer.
Statistical analysis. Each batch adsorption experiment was conducted in double parallel, and batch adsorption values used are means ± SD.
Data availability. The authors declare that all data supporting the findings of this study are available within the article and its Supplementary Information files.