Shell biomass material supported nano-zero valent iron to remove Pb2+ and Cd2+ in water

Nanoscale zero-valent iron (NZVI) has a high adsorption capacity for heavy metals, but easily forms aggregates. Herein, preprocessed undulating venus shell (UVS) is used as support material to prevent NZVI from reuniting. The SEM and TEM results show that UVS had a porous layered structure and NZVI particles were evenly distributed on the UVS surface. A large number of adsorption sites on the surface of UVS-NZVI are confirmed by IR and XRD. UVS-NZVI is used for adsorption of Pb2+ and Cd2+ at pH = 6.00 in aqueous solution, and the experimental adsorption capacities are 29.91 and 38.99 mg g−1 at optimal pH, respectively. Thermodynamic studies indicate that the adsorption of ions by UVS-NZVI is more in line with the Langmuir model when Pb2+ or Cd2+ existed alone. For the mixed solution of Pb2+ and Cd2+, only the adsorption of Pb2+ by UVS-NZVI conforms to the Langmuir model. In addition, the maximum adsorption capacities of UVS-NZVI for Pb2+ and Cd2+ are 93.01 and 46.07 mg g−1, respectively. Kinetic studies demonstrate that the determination coefficients (R2) of the pseudo first-order kinetic model for UVS-NZVI adsorption of Cd2+ and Pb2+ are higher than those of the pseudo second-order kinetic model and Elovich kinetic model. Highly efficient performance for metal removal makes UVS-NZVI show potential application to heavy metal ion adsorption.


Preparation of the undulating venus shell carrier
After washing with deionized water, the UVS were immersed in 0.01 mol l −1 HNO 3 solution for 1 h to remove the soluble impurities attached to the surfaces. Then, the shells were cleaned with deionized water and placed in a drying oven for 24 h. The washed and dried UVS were crushed and passed through a 140 mesh sieve, and the resulting powder was placed in a muffle furnace and calcined at 450°C for 60 min [1]. Finally, the optimized UVS carrier was obtained.

Preparation of the supported adsorbent (UVS-NZVI)
The supported adsorbent was prepared by the liquid-phase reduction process, and the synthesis process compared with NZVI-GAC of Li et al. has been simplified and improved [24]. The whole synthetic process was performed in the helium atmosphere. Firstly, the UVS (6.00 g) and FeCl 2 · 4H 2 O (5.49 g) were placed into a three-necked open flask, adding ethanol/water solution (200 ml, 30%) and stirring for 30 min. Subsequently, the NaBH 4 solution (100 ml, 1.00 mol l −1 ) was added dropwise into the mixture, constantly stirring for 30 min. Finally, the black solid (UVS-NZVI) was isolated by suction filtration and dried. The reduction reaction is as formula (2.1) [25] ð2:1Þ

Characterization of UVS-NZVI
The microstructures of the UVS and UVS-NZVI were imaged on S-4800 scanning electron microscopy (Hitachi, Japan) and H-9500 transmission electron microscopy (Hitachi, Japan). Fourier transform infrared (FTIR) spectra were obtained on a Tensor27 FTIR spectrometer (Bruker, Germany). X-ray diffraction (XRD) was recorded on a D/Max-2400 powder diffractometer (Rigaku, Japan), operating at 40 kV and 40 mA. The specific surface area and pore structure were measured on a specific surface area and porosity analyser (ASAP 2460 analyser, Micromeritics, USA). UVS-NZVI was degassed in vacuum and characterized by N 2 sorption at 77 K. The specific surface area was calculated by the Brunauer-Emmett-Teller (BET) method. Pore diameter were calculated by Barrett-Joyner-Halenda (BJH) method [26,27].

Calculation method for adsorption amount
The total concentrations of Cd 2+ and Pb 2+ were determined by a flame atomic absorption spectrophotometer (TAS-990, Purkinje General, Beijing). The adsorption amount of Cd 2+ and Pb 2+ in aqueous solution by the UVS-NZVI is calculated using the formula (2.2) where Q (mg g −1 ) was the amount of adsorption, C 0 (mg l −1 ) was the initial concentration of Cd 2+ and Pb 2+ before adsorption and C (mg l −1 ) was the concentration of Cd 2+ and Pb 2+ after adsorption. V (l) was the volume of the solution, while m (g) was the mass of the UVS-NZVI. Prepare lead standard series solutions with mass concentrations of 5, 10, 20, 30 and 40 mg l −1 with 0.15 mol l −1 dilute nitric acid solution. Measure the absorbance of the standard series of solutions by flame atomic absorption method; draw the standard curve of A Pb (absorbance)-C Pb (mass concentration of Pb 2+ ). Measurement conditions: gas flow rate 1500 ml min −1 , height 5.00 mm, position 0.50 mm, working lamp current 2.00 mA, spectral bandwidth 0.40 nm and negative high voltage 300 V.
The standard curve of Pb 2+ is shown in figure 1, the standard curve equation: y = 0.00906 + 0.02038x, R 2 = 0.99983. Within the range of Pb 2+ mass concentration of 5-40 mg l −1 , the linear relationship is good.
Prepare cadmium standard series solutions with mass concentrations of 0, 1, 2, 3, 4 and 5 mg l −1 with 0.15 mol l −1 dilute nitric acid solution. Measure the absorbance of the standard series of solutions by flame atomic absorption method; draw the standard curve of A Cd (absorbance)-C Cd (mass concentration royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 7: 201192 of Cd 2+ ). Measurement conditions: gas flow rate 1300 ml min −1 , height 5.00 mm, position 0.50 mm, working lamp current 2.00 mA, spectral bandwidth 0.40 nm and negative high voltage 300 V.
The standard curve of Cd 2+ is shown in figure 2, the standard curve equation: y = 0.01471 + 0.13171x, R 2 = 0.9968. Within the range of Cd 2+ mass concentration of 0-5 mg l −1 , the linear relationship is good.

Adsorbent dosage experiment
Six samples of 100 ml 20 mg l −1 Pb 2+ solutions were prepared with the pH 6.00, and the adsorbent dosage set to 0.01 g, 0.02 g, 0.03 g, 0.04 g, 0.05 g and 0.06 g, respectively. The mixture was shaken at 30°C for 2 h. After filtration, the filtrate was measured for absorbance, the average value was determined with three parallel measurements, and the adsorption amount was calculated using formula (2.2).
Cd 2+ solution was used for the above experiment, and other experimental conditions were the same.

PH gradient setting experiment
Six samples of 100 ml Pb 2+ solutions were prepared with a concentration of 20 mg l −1 . After adjusting the value of pH from 4.00 to 6.50 (at 30°C, the solubility product of Pb(OH) 2 is about 1 × 10 −17 ; according to the concentration of Pb 2+ in the solution, calculations show that when the pH of the solution is slightly greater than 7.00, Pb 2+ begins to precipitate), 0.04 g of UVS-NZVI was added into each solution. Then, the mixture was shaken at 30°C for 2 h. After filtration, the filtrate was measured for absorbance, the average value was determined with three parallel measurements, and the adsorption amount was calculated using formula (2.2). Nine samples of 100 ml Cd 2+ solutions were prepared with a concentration of 20 mg l −1 . The value of pH was adjusted from 4.00 to 8.00 and 0.04 g of UVS-NZVI was added into each solution. The subsequent processes were the same as Pb 2+ solutions.   royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 7: 201192 The pH of each group of solutions was adjusted with dilute solutions of NaOH and HNO 3 by pH meter (FE28-Standard, Mettler Toledo, Switzerland).

Coexistent ion experiment
Seven samples of 100 ml 20 mg l −1 Pb 2+ solutions and seven samples of 100 ml 20 mg l −1 Cd 2+ solutions were prepared with the pH 6.00, and the Cu 2+ concentration set to 0, 2, 4, 6, 8, 10 and 12 mg l −1 , respectively. Of UVS-NZVI, 0.04 g was added into each solution. Then, the mixture was shaken at 30°C for 2 h. After filtration, the filtrate was measured for absorbance, the average value was determined with three parallel measurements, and the adsorption amount was calculated using formula (2.2).
Change Cu 2+ to Ni 2+ in the above process, and other experimental conditions are the same.

Adsorption thermodynamic experiment
Seven samples of 100 ml Pb 2+ solutions and seven samples of 100 ml Cd 2+ solutions were prepared, and the concentrations were 5, 10, 20, 30, 50, 75 and 100 mg l −1 , respectively. Then 0.04 g of UVS-NZVI was added to each solution and the mixture was shaken at 30°C for 2 h. After filtration, the filtrate was measured for absorbance, and the average value was determined with three parallel measurements. The adsorption amount was calculated using formula (2.2). Also, seven kinds of different concentrations of mixed solutions of Pb 2+ and Cd 2+ were prepared and adjusted to pH of 6.00. The following steps are consistent with the above experiment. The Langmuir and Freundlich adsorption isotherm models were adopted to evaluate the isotherm experiment data.
Langmuir and Freundlich adsorption isotherm models were given as formulae (2.1) and (2.2) [28,29] q e ¼ q max bC e 1 þ bC e ð2:3Þ and where q e (mg g −1 ) is the equilibrium adsorption capacity, q max (mg g −1 ) is the maximum adsorption capacity, C e (mg g −1 ) is the equilibrium concentration, b is the constant, K f is the Freundlich constant, and n f is the concentration index. Five samples of 100 ml 20 mg l −1 Pb 2+ solutions and five samples of 100 ml 20 mg l −1 Cd 2+ solutions were prepared. Then 0.04 g of UVS-NZVI was added to each solution and the mixture was shaken at 20, 30, 40, 50 and 60°C for 2 h, respectively. After filtration, the filtrate was measured for absorbance, and the average value was determined with three parallel measurements. The adsorption amount was calculated using formula (2.2).
The Van't Hoff equation was used to fit the adsorption capacity at different temperatures to obtain the thermodynamic parameters ΔH and ΔS of Pb 2+ and Cd 2+ . Calculate ΔG at different temperatures by Gibbs equation. Van't Hoff and Gibbs equations were given as formula (2.5) and (2.6) [30,31] DG ¼ DH À TDS ð2:6Þ and where C s is the concentration of the solid surface at the adsorption equilibrium and C e is the concentration in the solution at the adsorption equilibrium.

Kinetic adsorption experiment
A 50 mg l −1 Pb 2+ solution and 50 mg l −1 Cd 2+ solution were prepared. Then 0.02 g of UVS-NZVI was added to 100 ml of Pb 2+ or Cd 2+ solution and the mixture was shaken at 30°C. Determining the concentration of Pb 2+ and Cd 2+ every 5 min within the adsorption time of 5-60 min, the adsorption amount was calculated using formula (2.2). The pseudo first-order, pseudo second-order and Elovich kinetic models were used to determine the rate of the adsorption process. Pseudo first-order and pseudo second-order kinetics models were given as formula (2.8) where q t (mg g −1 ) is the adsorption capacity at time t (min), and q e (mg g −1 ) is the equilibrium adsorption capacity. And k 1 (min −1 ) was the pseudo first-order kinetic rate constant, and k 2 (g min mg −1 ) was the pseudo second-order kinetic rate constant. Elovich kinetic model was given as formula (2.10) [34] where β E (g mg −1 ) is the desorption constant related to the activation energy of chemisorption and α E (mg (g min) −1 ) is the initial adsorption rate.

Recyclability of UVS-NZVI
The reusability of UVS-NZVI was tested by repeated Pb 2+ and Cd 2+ adsorption and desorption cycles for five consecutive cycles. First, 100 ml of 50 mg l −1 Pb 2+ or Cd 2+ solution was prepared. Then 0.02 g of adsorbent was added. The mixture was shaken at 30°C for 2 h. The desorption process was in 30 ml ultrapure water for 10 h. The filtrate during each cycle was measured for absorbance. The average value was determined with three parallel measurements, and the adsorption amount was calculated using formula (2.2).
The preparation and adsorption processes of UVS-NZVI are shown in figure 3.

Characterization of adsorbents
The porous lamellar structure of UVS is shown in figure 4a,b. More channels inside the UVS might be opened due to the acidic conditions in the preparation process, which increased the pore volume and    royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 7: 201192 internal specific surface area of UVS [35]. The SEM images of NZVI and UVS-NZVI are presented in figure 5. From figure 5a,b the NZVI particles can be seen clearly and had an average diameter of 60 nm. However, these particles are aggregated due to lack of support material [36]. According to figure 5c,d, NZVI particles are uniformly dispersed on the UVS surface without obvious accumulation, and the size of particles is relatively small, indicating that NZVI loaded onto the UVS successfully [37]. The structural information provided by TEM is in agreement with SEM analysis. From figure 6a,b, massive NVZI particles are aggregated, while in figure 6c,d, flocculation (NZVI or derivatives of Fe 0 ) is distributed around the UVS structure evenly [38].
The XRD patterns of UVS and UVS-NZVI are shown in figure 7. For UVS (a), the characteristic peaks are attributed to aragonite CaCO 3 and calcite CaCO 3 , and most of the characteristic peaks are calcite CaCO 3 [39]. For UVS-NZVI (b), except for the 2θ = 57.48°attributed to calcite CaCO 3 characteristic peaks, the rest are characteristic peaks of Fe and its compounds. As shown, the strong peaks at 44.81°a nd 50.93°are present on UVS-NZVI composite, corresponding to Fe 0 [40]. In addition, some minor peaks of Fe compounds, for example, Fe 2 O 3 (55.23°) and Fe(OH) 3 (64.01°), are also observed, which might result from oxidation and corrosion during the synthesis and drying process [41,42]. Figure 8 exhibits FTIR of the UVS and UVS-NZVI. For UVS (a), the 710.58, 860.18 cm −1 , etc. are the characteristic peaks of CO 2À 3 , indicating that the main component of UVS is CaCO 3 [43]. The peak at 860.18 cm −1 is blue shifted by 5.68 cm −1 compared with the peak at 854.50 cm −1 of aragonite CaCO 3 , which shows that UVS also contains a small amount of calcite CaCO 3 , and the remaining peaks are essentially identical with those of aragonite CaCO 3 . Based on the FTIR spectrum of UVS, there are no other functional groups in addition to CO 2À 3 [39,44]. Therefore, calcium carbonate did not decompose at 450°C in this experiment.
After UVS was loaded with NZVI (b), various peaks of iron compounds are all obvious. The peak at 3353.91 cm −1 can be attributed to O-H association with Fe or its oxyhydroxide [45]. The peaks at 1276.71 and 1344.56 cm −1 can be attributed to bands associated with Fe 3 O 4 , Fe 2 O 3 and FeOOH formation surrounding Fe 0 , as well as ethanol added in synthesizing the adsorbents [46]. In addition, 1130.26, 946.31 and 823.57 cm −1 are the characteristic peaks of −OH in Fe(OH) 3 , and 618.56 cm −1 is the characteristic peak of Fe 2 O 3 [47,48]. The FTIR spectrum indicates the slight oxidation of Fe 0 , and the abundant oxygen functional groups could act as available adsorption sites, which could provide free pairs of electrons to interact with the empty orbital of metal ions (Pb 2+ and Cd 2+ ). This characterization result is basically consistent with the XRD, indicating the existence of zero-valent iron on the surface of UVS. royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 7: 201192 Nitrogen sorption isotherm and pore size distribution are shown in figures 9 and 10. The specific surface area and pore diameter of UVS-NZVI calculated by BET method and BJH method are 83.77 m 2 g −1 and 3.54 nm, respectively, which indicates that the porous structure of shell material makes nano-zero-valent iron have a larger contact area with the solution, and the material may have higher adsorption performance. The effects of different adsorbent dosage on the adsorption of Pb 2+ and Cd 2+ by UVS-NZVI are shown in figure 11. For Pb 2+ and Cd 2+ , when the amount of adsorbent is less than 0.04 g, the adsorption capacity  royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 7: 201192 remains almost unchanged. When the amount of adsorbent is greater than 0.04 g, the adsorption capacity drops significantly. This may be due to the ion concentration being too small, the added adsorbent has not reached its optimal adsorption effect. Therefore, in order to achieve the best removal effect and the best utilization of the adsorbent, the amount of adsorbent in subsequent experiments will be 0.04 g.   In strong acidic conditions, Fe 0 was corroded to some extent, which made UVS-NZVI poor adsorption effect on metal ions. As a result, within 4.00 < pH < 5.50, the equilibrium adsorption capacity was less than 20 mg g −1 , whether it is Pb 2+ or Cd 2+ , and the pH increased significantly after adsorption. Within pH ≥ 5.50, adsorption performance improved obviously.

Adsorption
In the adsorption of Pb 2+ by UVS-NZVI, compared with Fe 2+ /Fe 0 , Pb 2+ /Pb 0 has the higher electrode potential. Therefore, Pb 2+ can be reduced to Pb 0 by Fe 0 and the Pb 0 was continually adsorbed onto the iron oxide layer. Within pH ≥ 5.50, the chemical equation of the reaction is as formula (3.1) [6,49] Pb 2þ þ Fe 0 þ 2OH À ! Pb 0 þ FeðOHÞ 2 : ð3:1Þ According to the chemical equation, increasing OH − concentration is conducive to the reaction, and the adsorption process reduces the pH of the solution. As a result, for Pb 2+ (figure 12a), the equilibrium adsorption amount increased with increasing pH in the pH range of 5.5-6.5. The adsorption effect of UVS-NZVI for Pb 2+ was optimal at pH = 6.00, and the pH of each solution decreased after adsorption.
For Cd 2+ , due to the ion exchange between Cd 2+ and H + on the surface of UVS-NZVI, pH of the solution decreased [50]. In addition, Cd 2+ could coprecipitate with Fe 2+ to form metal hydroxides over the external surfaces of adsorbents. Within pH ≥ 5.50, the chemical equation of the reaction is as formula (3.2) [51,52] nCd 2þ þ ð1 À nÞFe 2þ þ 2OH À ! Cd n Fe ð1ÀnÞ ðOHÞ 2 : ð3:2Þ  royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 7: 201192 Therefore, increasing pH facilitated the adsorption of Cd 2+ by UVS-NZVI (figure 12b). The equilibrium adsorption amount first raised and then stayed the same approximately in the pH range of 5.50-8.00. The adsorption effect of UVS-NZVI for Cd 2+ was optimal in pH ≥ 6.00.
Due to the acidic environment, the Fe 0 reacted with H + , which limited its adsorption effectiveness. As a result, the optimal pH for the adsorption by UVS-NZVI is selected as pH = 6.00 in subsequent experiments.

Influence of coexisting ion on the adsorption of Pb 2+ and Cd 2+ ions
The effects of different concentrations of Cu 2+ and Ni 2+ on the adsorption of Pb 2+ and Cd 2+ by UVS-NZVI are shown in figures 13 and 14.
According to figure 13, the presence of Cu 2+ has a great influence on the adsorption of Pb 2+ and Cd 2+ . Since the oxidability of Cu 2+ is similar to that of Pb 2+ and slightly greater than that, as the concentration of Cu 2+ increases, the adsorption performance of UVS-NZVI for Pb 2+ decreases to a certain extent. The oxidability of Cu 2+ is much greater than that of Cd 2+ , so as the concentration of Cu 2+ increases, the adsorption performance of UVS-NZVI on Pb 2+ decreases significantly, and when the concentration of Cu 2+ reaches 12 mg l −1 , the adsorption capacity of Cd 2+ is only 1.93 mg g −1 , which shows that Cd 2+ is hardly adsorbed.
The effects of Ni 2+ on the adsorption of Pb 2+ and Cd 2+ by UVS-NZVI are shown in figure 14. Since the oxidability of Pb 2+ is much greater than that of Ni 2+ , with the increase of Ni 2+ concentration, the adsorption capacity of Cd 2+ hardly changes, and there is almost no adsorption competition between the two ions. The oxidability of Ni 2+ is close to that of Cd 2+ . As the concentration of Ni 2+ increases, the adsorption capacity of Cd 2+ decreases to a certain extent, but UVS-NZVI has still a significant adsorption capacity for Cd 2+ .

Adsorption thermodynamic analysis
The parameters and fitted plots of Langmuir and Freundlich adsorption isotherm models are listed in table 1 and figure 15a,b. The maximum adsorption capacity of UVS-NZVI for Pb 2+ in two different situations was 93.01 and 86.99 mg g −1 (figure 15a), respectively. In the solution containing only Pb 2+ , based on the higher determination coefficient (R 2 ), the Pb 2+ adsorption by UVS-NZVI was more in line with the Langmuir model, which indicates that the adsorption sites on the adsorbent are homogeneously distributed with a monolayer coverage of adsorption products [45,53,54]. The adsorption capacity of UVS-NZVI to Pb 2+ was hardly affected by the coexisting Cd 2+ in mixed solution and the adsorption of UVS-NZVI to Pb 2+ was still more consistent with Langmuir model compared to Freundlich. In addition, the Freundlich constant n f for single and mixed adsorption experiment were all between 0.1 and 1, indicating that the adsorption process of Pb 2+ was favourable [55,56].
The maximum adsorption capacity of UVS-NZVI for Cd 2+ was 46.07 mg g −1 in the solution containing only Cd 2+ . Because the correlation coefficient of Langmuir model (0.9797) fitting is greater than that of Freundlich model (0.9344), the Cd 2+ adsorption by UVS-NZVI was more in line with the Langmuir model. Nevertheless, the presence of Pb 2+ observably decreased the Langmuir adsorption capacity of Cd 2+ to 3.31 mg g −1 (figure 15b), which indicates the competition for active sites between them. The compound Pb-hydroxide on the surface of UVS-NZVI inhibited the formation of Cd-Fehydroxide in mixed solution and the surface cation exchange sites were completely occupied by Pbhydroxide precipitation and Pb [57][58][59]. Hence, there was almost no Cd 2+ adsorbed on the UVS-NZVI composite. In summary, Pb 2+ is far more competitive than Cd 2+ .
In separate adsorption experiment for Pb 2+ or Cd 2+ , the correlation coefficients of Langmuir fitting are 0.9911 and 0.9797, respectively, which are higher than the correlation coefficients of Freundlich fitting (0.9234 and 0.9344). Since the important assumption of the Langmuir model is monolayer adsorption, and typical chemical adsorption is also monolayer adsorption, it can be inferred that the adsorption of Pb 2+ or Cd 2+ by UVS-NZVI may be chemical adsorption.
Since the activated shell (UVS) has a porous structure and a larger specific surface area than other biomass materials, the adsorption capacity of UVS-NZVI has certain advantages. The maximum   Van't Hoff's fitting results and parameters for temperature changes are shown in figure 16 and table 3. According to the calculation results of thermodynamic parameters, the ΔH of UVS-NZVI for Pb 2+ and Cd 2+ adsorption are 29.33 and 53.89 kJ mol −1 , respectively, which means that the adsorption processes of UVS-NZVI for Pb 2+ and Cd 2+ are endothermic. According to the study of Ma et al. [64], the enthalpy change range of physical adsorption is between 2.10 and 20.90 kJ mol −1 , and the enthalpy change range of chemical adsorption is between 20.90 and 418.40 kJ mol −1 . Therefore, the  The calculation results of the above thermodynamic parameters fully prove that the adsorption process of Pb 2+ and Cd 2+ by UVS-NZVI is a spontaneous chemical adsorption process.

Adsorption kinetics analysis
The effect of adsorption time for the removal of Pb 2+ and Cd 2+ ions by the UVS-NZVI was studied with ion concentration = 50 mg l −1 , and the results are shown in figure 17a,b. It was observed that the adsorption amount of metal ions was increased with an increase of the time in the range of 5-60 min whether it was Pb 2+ or Cd 2+ , and the adsorption amount was increased quickly in the first 30 min, then increased gradually. As a result, Pb 2+ reached adsorption equilibrium on the UVS-NZVI at 50 min and Cd 2+ at 45 min. The adsorption of ions increased rapidly at the initial stage due to the accessibility of huge amount of binding sites on the surface of the UVS-NZVI. Thus the percentage removal of ions significantly increased at lesser adsorption time. At higher adsorption time, the rate of adsorption increased slowly until equilibrium due to the available binding sites being gradually decreased; for that reason, it took a long time for the adsorption process to reach equilibrium [18].
The adsorption mechanism of the Pb 2+ and Cd 2+ ions on UVS-NZVI was analysed by using three adsorption kinetic models, namely pseudo first-order kinetic model (the reaction rate is linearly related to the concentration of a reactant, and this model is based on the fact that rate-determining step is a physical process), pseudo second-order kinetic model (the reaction rate is linearly related to  the concentration of two reactants, and this model is based on the fact that rate-determining step is a chemical reaction) and Elovich kinetic model (the Elovich model is suitable for processes with irregular data or with large activation energy) [65]. The fitting results are shown in figure 17a,b, and the kinetic parameters for the adsorption of Cd 2+ and Pb 2+ are listed in table 4. The fitted adsorption capacity (q e ) was compared with the equilibrium adsorption amount (q e,exp ) in the experiment. It can be observed from table 4 that the q e of pseudo first-order kinetic for the adsorption of Cd 2+ and Pb 2+ by UVS-NZVI was much closer to the q e,exp compared with other kinetic models. In addition, the determination coefficient of the pseudo first-order kinetic was higher than that of pseudo secondorder and Elovich kinetic model for both Cd 2+ and Pb 2+ , which indicates that the physical adsorption is the rate-determining step [66]. It may be possible that the huge specific surface area of UVS material increases the contact area significantly between NZVI and the solution compared with other carrier materials and this advantage plays a decisive role in the adsorption process, which agrees with Huang et al. [67]. Figure 18 shows the adsorption amount variation trend of multiple adsorption of Pb 2+ or Cd 2+ by UVS-NZVI in aqueous solution. After five recycles, the adsorption of Pb 2+ decreased from the initial 41.26 to 14.34 mg g −1 and the Cd 2+ from 23.09 to 9.02 mg g −1 . The reason was that the NZVI particles were exhausted by chemical reduction and the absorbed Pb 2+ and Cd 2+ ions were not completely resolved [42]. In adsorption of the fourth recycle, the UVS-NZVI adsorption amount of Pb 2+ was 20.39 mg g −1 and the Cd 2+ adsorption amount was 11.90 mg g −1 , which indicates that UVS-NZVI can still achieve about 50% effect of the initial adsorption after four recycles.

Conclusion
In summary, UVS material-supported NZVI was successfully synthesized. NZVI particles can be evenly distributed on the porous layered structured UVS surface, which can effectively prevent NZVI from reuniting. A large number of adsorption sites can be derived from a large amount of Fe 0 , a small amount of Fe 2 O 3 , Fe(OH) 3 and FeOOH on the surface of UVS-NZVI. The adsorption amounts of Pb 2+ and Cd 2+ both increased with the initial pH and then remained unchanged, and the adsorption effect of the two ions reached the peak at pH = 6.00 and the experimental maximum adsorption capacities of Pb 2+ and Cd 2+ were 29.91 and 38.99 mg g −1 , respectively. The presence of Cu 2+ has a significant effect on the adsorption of Pb 2+ and Cd 2+ . Ni 2+ has a great influence on the adsorption of Cd 2+ , but there is almost no adsorption competition with Pb 2+ . Thermodynamic studies showed that the adsorption of Pb 2+ and Cd 2+ by UVS-NZVI is a chemical adsorption. Kinetic studies demonstrated that the physical adsorption was the rate-determining step. Further, the recyclability experiment suggested that UVS-NZVI had a good recycling effect in treating Pb 2+ and Cd 2+ in wastewater. Overall, this study not only provides an effective application approach for shellfish biomass resources, but also promotes the wider application of NZVI adsorbent, which makes UVS-NZVI have certain application potential in the heavy metal ions adsorption field. royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 7: 201192 Data accessibility. The data supporting the findings of this study are available at the Dryad Digital Repository: https://doi. org/10.5061/dryad.gqnk98sjz [68].
Authors' contributions. Z.W., X.W., S.L., Y.W., Z.T. and Q.D. undertook the research. Z.W. and X.W. participated in every part of the experiment and wrote the manuscript, they should be considered co-first authors.
Competing interests. We declare we have no competing interests. Funding. This research was supported by the Natural Science Foundation of China (grant no. 21767008) and the Natural Science Foundation of Hainan Province (grant no. 219000).