The Application of Self-Assembled of Meso-2,3-Dimercaptosuccinic Acid-Polydopamine-Zinc Oxide for Trace Cadmium Analysis

Nanomaterials have been applied in the analysis of trace contaminants in aquatic samples; however, the adsorption affinity of heavy metal ions to metallic oxide nanoparticles is low. Meso-2,3-dimercaptosuccinic acid-polydopamine-zinc oxide (DMSA/PDA/ZnO) was synthesized through a facile self-assembly process for enhancing the adsorption capacity of ZnO to cadmium cation (Cd(II)). A method for the analysis of trace Cd(II) in water samples was developed while using dimercaptosuccinic acid/Polydopamine/zinc oxide (DMSA/PDA/ZnO) as an adsorbent and atomic absorption spectrometry (AAS) as a means of determination. The results demonstrate that the adsorption ratio of DMSA-PDA-ZnO to 1 mg/L Cd(II) is around 95% at pH 6.0–8.0 with a contact time of 31 h, and the enrichment rate can be more than 98% with the dosage of DMSA-PDA-ZnO exceeding 1.0 g/L. The adsorption isotherm of Cd(II) to DMSA/PDA/ZnO fits the Langmuir equation well, and the saturated adsorption of DMSA-PDA-ZnO to Cd(II) is 28.5 mg/g. The detection limit (3σ) of this method is 0.1 μg/L, with a relative standard deviation of 1%. Additionally, this method presents satisfactory performance with a recovery rate of 86−90% when applied to the detection of Cd(II) in environmental water samples.


Introduction
The measurement of aquatic chemical substrates is important for research on water quality assessment, human health, and water pollution control. Among the chemical substrates, cadmium (Cd) is a well-known toxic and persistent pollutant that can cause both short-term and long-term health effects [1,2]. Besides foods, drinking water is the major source for human exposure to Cd(II), and the guideline value for Cd in drinking water is 3 µg/L, as set by the World Health Organization (WHO) [3]. The concentrations of Cd in unpolluted water are low, and the reported concentrations of Cd were at (or below) the µg/L level [4,5]. High values of tens of µg/L were also reported in some water samples [6]. A low concentration of Cd in natural water and drinking water samples requires high-sensitivity analytical methods for the precise concentration measurement. Graphite furnace atomic absorption spectrometry (GFAAS) [7], flame furnace atomic absorption spectrometry (FAAS) [8], colorimetric determination with dithizone [9,10], inductively coupled argon plasma mass spectrometry

Characterization of DMSA/PDA/ZnO
The X-ray powder diffraction (XRD) spectra of ZnO, PDA/ZnO, and DMSA/PDA/ZnO were measured with a Dmax-rA powder diffractometer (Rigaku, Japan), with Cu Kα as a radiation source at a scanning rate of 2 • /min. Contents of elements (C, H, N, and S) in these materials were quantified by an elemental analyzer (Vario Cube; Elementar, Langenselbold, Germany). Scanning electron micrograph (SEM) was taken with a FEI-QUANTA 200 microscope, and the transmission electron microscopy (TEM) images were recorded on a JEOL JEM 2010HT microscope (Japan Electronics, Tokyo, Japan) at an accelerating voltage of 200 kV. The FT-IR spectra of sample pellets with KBr were recorded on a Nicolet 5700 Fourier transform infrared (FTIR) spectrometer (Thermo Electron Corporation, Waltham, America). The dynamic light scattering (DLS) volume distribution and surface zeta potential (ζ) of ZnO, PDA/ZnO, and DMSA/PDA/ZnO (0.1 g/L H 2 O suspension) was measured with a Zetasizer nano device (Malvern, UK).

Adsorption of Cd(II) on DMSA/PDA/ZnO
250 mL working solution containing 1−40 mg/L Cd(II) was added to a flask, followed by the addition of a certain amount of ZnO related adsorbent (0.1−1.0 g/L). Afterwards, the solution pH was adjusted (3.0-9.0) and the resulting solution was shaken at 25 • C. Solution samples of 10 mL were withdrawn at different time intervals, and then centrifuged at 7000 r/min. for 20 min. A certain amount of supernatant was transferred into a volumetric flask and 0.5 mL HCl was added, and the solution was diluted to 10 mL with ultrapure water.
The Cd(II) concentration was then determined by on a Hitachi Z-5000 Polarized Zeeman Atomic Absorption Spectrophotometer (Hitachi High-Technologies Corporation, Tokyo, Japan). The initial concentration and concentration at different adsorption times were expressed as c 0 and c t , respectively. c e and q e is the concentration of Cd(II) in the water phase and adsorbent after reaching adsorption equilibrium, respectively. c e and q e that were obtained from different Cd(II) initial concentrations (1−40 mg/L) were used to investigate the adsorption isotherm.

Protocol for Preconcentration of Cd(II) with of DMSA/PDA/ZnO
A 100 mL water sample was added to a flask containing 1.0 g/L DMSA/PDA/ZnO, and then the solution was shaken at 25 • C for 3 h. After that, the dispersion was filtrated while using a 0.45 µm Appl. Sci. 2020, 10, 2462 4 of 12 membrane. The membrane was washed and diluted to 10 mL with 5% HCl. FAAS directly determined the concentration of Cd(II) in the solution mentioned.

Data Analysis
The adsorption ratio (R, %) was calculated by Equation (1), The adsorption isotherm was fitted by Equations (2) and (3), where q e and q max (mg/g) are the equilibrium and maximum adsorption of Cd(II) on DMSA/PDA/ZnO, and K L (L/mg) is the Langmuir adsorption constant.
K F and 1/n represent the Freundlich adsorption constant and the unitless linearity parameter, respectively.
Limit of detection (LOD) was the calculated concentration to a signal of three times standard deviation of blank sample.

Data Analysis
The adsorption ratio (R, %) was calculated by Equation (1), The adsorption isotherm was fitted by Equations (2) and (3), where qe and qmax (mg/g) are the equilibrium and maximum adsorption of Cd(II) on DMSA/PDA/ZnO, and KL (L/mg) is the Langmuir adsorption constant.
KF and 1/n represent the Freundlich adsorption constant and the unitless linearity parameter, respectively.
Limit of detection (LOD) was the calculated concentration to a signal of three times standard deviation of blank sample. Figure 1 presents the XRD patterns of ZnO, DPOA/ZnO, and DMSA/DPOA/ZnO for 2θ ranging from 10 to 80°. All the three ZnO relevant materials exhibited the same XRD pattern, indicating the structure did not change during the coating process. Peaks that were observed at 31.7, 34.4, 36.2, 47.5, 56.5, 62.8, 66.3, 67.9, and 69.0° were indexed to hexagonal wurtzite (ZnO; JCPDS Data Card no. 36-1451). By using the Williamson-Hall method [28], it was found that the obtained average crystallite size of ZnO was 23, 23, and 25 nm for bare ZnO, PDA/ZnO, and DMSA/PDA/ZnO, respectively. The corresponding strains are 0.88%, 0.88%, and 1.01%.   The ZnO that was used in this study has a relative wide size distribution range (16−36 nm) with the majority at 22−26 nm ( Figure S1, Supplementary Materials). The size of the ZnO core was not changed after modification by PDA and DMSA. The size values from TEM analysis are close to the diameter that was estimated from XRD analysis. However, these values were smaller than their corresponding hydrodynamic diameter that was obtained from DLS analysis ( Figure S2), suggesting both barely and modified ZnO tend to form agglomerates. When compared with the TEM images of ZnO (Figure 3a,b), it could be clearly observed that the outer surface of PDA/ZnO was capped by an amorphous coating, which was assigned as the attached PDA layer (Figure 3c,d). The PDA layer was approximately 12-15 nm in thickness. It also demonstrated that the "carbohydrate" layer became thicker after further coating with DMSA (Figure 3e,f). The ZnO that was used in this study has a relative wide size distribution range (16−36 nm) with the majority at 22−26 nm ( Figure S1, Supplementary Materials). The size of the ZnO core was not changed after modification by PDA and DMSA. The size values from TEM analysis are close to the diameter that was estimated from XRD analysis. However, these values were smaller than their corresponding hydrodynamic diameter that was obtained from DLS analysis ( Figure S2), suggesting both barely and modified ZnO tend to form agglomerates. When compared with the TEM images of ZnO (Figure 3a,b), it could be clearly observed that the outer surface of PDA/ZnO was capped by an amorphous coating, which was assigned as the attached PDA layer (Figure 3c,d). The PDA layer was approximately 12-15 nm in thickness. It also demonstrated that the "carbohydrate" layer became thicker after further coating with DMSA (Figure 3e,f).

Structural Characterization of ZnO, DPOA/ZnO, and DMSA/DPOA/ZnO
The results of the TEM and SEM suggested that PDA and DMSA were both attached to the ZnO surface and that assembly behavior resulted in morphological changes of ZnO through the formation of the "carbohydrate" layer (Vide Infra). The results of the TEM and SEM suggested that PDA and DMSA were both attached to the ZnO surface and that assembly behavior resulted in morphological changes of ZnO through the formation of the "carbohydrate" layer (Vide Infra).

FTIR Spectra of ZnO and DMSA/PDA/ZnO
The sharp peak around 434 cm −1 were assigned as the Zn-O stretching bond, as shown in Figure  4 (Black line) [29]. Two narrow peaks around 1580 cm −1 and 1380 cm −1 in the FTIR spectrum of bare ZnO were attributed to deformation vibrations of the adsorbed water and the in-plane angular deformation of the surface OH groups, respectively. However, these two peaks might also from residues of the substrates used for ZnO synthesis (Section 2.2). New peaks were found in the FTIR spectrum of DMSA/PDA/ZnO, which suggested the existence of other compounds other than ZnO. Peaks around 1604 cm −1 , 1272 cm −1 , and 1039 cm −1 were from the vibration of attached PDA [30,31]. No peaks belonging to DMSA were obtained in FTIR spectrum of DMSA/PDA/ZnO, which might be due to its low amount (Vide Infra).

FTIR Spectra of ZnO and DMSA/PDA/ZnO
The sharp peak around 434 cm −1 were assigned as the Zn-O stretching bond, as shown in Figure 4 (Black line) [29]. Two narrow peaks around 1580 cm −1 and 1380 cm −1 in the FTIR spectrum of bare ZnO were attributed to deformation vibrations of the adsorbed water and the in-plane angular deformation of the surface OH groups, respectively. However, these two peaks might also from residues of the substrates used for ZnO synthesis (Section 2.2). New peaks were found in the FTIR spectrum of DMSA/PDA/ZnO, which suggested the existence of other compounds other than ZnO. Peaks around 1604 cm −1 , 1272 cm −1 , and 1039 cm −1 were from the vibration of attached PDA [30,31]. No peaks belonging to DMSA were obtained in FTIR spectrum of DMSA/PDA/ZnO, which might be due to its low amount (Vide Infra).

Quantification of Attached DMSA and PDA in DMSA/PDA/ZnO
There is little C, H, and S in bare ZnO, which might be the residual of reactants after solvents wash, as listed in Table 1 (see Section 2.2). The contents of C, N, and H are much higher in PDA/ZnO relative to ZnO, indicating the coating of an organic layer. The molar ratio of C:N:H is 8.1:1:10.6, which is close to the molar ratio in dopamine (8:1:11). Therefore, the attached layer was assigned as polydopamine in PDA/ZnO, and the mass content of PDA is 10.8% in PDA/ZnO. After further

Quantification of Attached DMSA and PDA in DMSA/PDA/ZnO
There is little C, H, and S in bare ZnO, which might be the residual of reactants after solvents wash, as listed in Table 1 (see Section 2.2). The contents of C, N, and H are much higher in PDA/ZnO relative to ZnO, indicating the coating of an organic layer. The molar ratio of C:N:H is 8.1:1:10.6, which is close to the molar ratio in dopamine (8:1:11). Therefore, the attached layer was assigned as polydopamine in PDA/ZnO, and the mass content of PDA is 10.8% in PDA/ZnO. After further modification by DMSA, the content of sulphur (S) increased to 0.834% in DMSA/PDA/ZnO, which indicated the presence of DMSA. According to the ratio of molecular weight of S to DMSA, the attached DMSA is approximately 2.4% in DMSA/PDA/ZnO.   The adsorption of Cd(II) to DMSA/PDA/ZnO could be better fitted by the Langmuir isotherm model relative to Freundlich isotherm model, as listed in Table 2. The Langmuir isotherm is an empirical model that assumes monolayer adsorption, namely, that each adsorbate exhibits constant enthalpies and sorption activation energy [32]. The adsorption of Cd(II) on DMSA/PDA/ZnO followed the Langmuir adsorption isotherm model, as shown in Figure 6. The adsorption capacity  The adsorption of Cd(II) to DMSA/PDA/ZnO could be better fitted by the Langmuir isotherm model relative to Freundlich isotherm model, as listed in Table 2. The Langmuir isotherm is an empirical model that assumes monolayer adsorption, namely, that each adsorbate exhibits constant enthalpies and sorption activation energy [32]. The adsorption of Cd(II) on DMSA/PDA/ZnO followed the Langmuir adsorption isotherm model, as shown in Figure 6. The adsorption capacity (q max ) and adsorption constant (K L ) were 28.5 mg/g and 0.14 L/mg, respectively. When comparing q max reported in this study with the reported values for other adsorbents (4-137 mg/g) [22], q max reported in this study is in the medium rank. However, other adsorbents, like meso silica (MCM-48, SBA-15 etc.), could not be directly dissolved in acid solution.
order and pseudo-second-order kinetic model. The adsorption of Cd(II) to DMSA/PDA/ZnO could be better fitted by the Langmuir isotherm model relative to Freundlich isotherm model, as listed in Table 2. The Langmuir isotherm is an empirical model that assumes monolayer adsorption, namely, that each adsorbate exhibits constant enthalpies and sorption activation energy [32]. The adsorption of Cd(II) on DMSA/PDA/ZnO followed the Langmuir adsorption isotherm model, as shown in Figure 6. The adsorption capacity (qmax) and adsorption constant (KL) were 28.5 mg/g and 0.14 L/mg, respectively. When comparing qmax reported in this study with the reported values for other adsorbents (4-137 mg/g) [22], qmax reported in this study is in the medium rank. However, other adsorbents, like meso silica (MCM-48, SBA-15 etc.), could not be directly dissolved in acid solution.

Effects of Solution pH and Adsorbent Dosage on Adsorption of Cd on DMSA/PDA/ZnO
Aquatic pH could change the distribution of adsorbate and the surface charge of adsorbent ( Figure  S2), which could subsequently influence the adsorption efficiency. The surface of DMSA/PDA/ZnO was negatively charged after modification by DMSA and PDA ( Figure S3). The results are in line with the report that PDA is a negative polyelectrolyte [33]. The negative surface of DMSA/PDA/ZnO favored the binding of positive charged Cd(II) other than metal-ligand chelation. As shown in Figure 7a, aquatic pH had a significant influence on the adsorption of Cd(II) on DMSA/PDA/ZnO. The adsorption ratio increased from 70% to 98% for a pH increase from 3.0 to 7.0 and decreased to 80% at pH 9.0. The decrease in the adsorption ratio at pH 9.0 might also be due to the formation of CdOH + ion.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 9 of 13 Aquatic pH could change the distribution of adsorbate and the surface charge of adsorbent ( Figure S2), which could subsequently influence the adsorption efficiency. The surface of DMSA/PDA/ZnO was negatively charged after modification by DMSA and PDA ( Figure S3). The results are in line with the report that PDA is a negative polyelectrolyte [33]. The negative surface of DMSA/PDA/ZnO favored the binding of positive charged Cd(II) other than metal-ligand chelation. As shown in Figure 7a, aquatic pH had a significant influence on the adsorption of Cd(II) on DMSA/PDA/ZnO. The adsorption ratio increased from 70% to 98% for a pH increase from 3.0 to 7.0 and decreased to 80% at pH 9.0. The decrease in the adsorption ratio at pH 9.0 might also be due to the formation of CdOH + ion.    Figure 7b shows the effects of DMSA/PDA/ZnO dosage on the adsorption of Cd. The adsorption ratio of Cd increased from 86% to 94% with the increase of loaded DMSA/PDA/ZnO from 0.1 to 0.2 g/L. The adsorption ratio further increased to 98% with a further increase in DMSA/PDA/ZnO up to 1.0 g/L. The increase in adsorption rate with the increase of loaded DMSA/PDA/ZnO was likely to have been caused by the additional chelating sites for Cd(II). Therefore, a dosage of 0.5 g/L was adopted for Cd(II) preconcentration due to the accepted adsorption ratio and the relatively low consumption of acid for dissolution.
3.7. Performance of DMSA/PDA/ZnO in the Analysis of Ultra-Trace Cd(II)

Linear Region
The linear region for the calibration curves was decided to reveal the dissolved DMSA/PDA/ZnO for the analysis of Cd(II) by FAAS. The dissolved matrix of DMSA, PDA, and ZnO did not influence the absorbance of Cd solution, as shown in Figure 8. However, the response of absorbance bent towards the X-axis when the concentration of Cd was higher than 2.0 mg/L. Therefore, the calibration curve for Cd was from 0−2.0 mg/L (insert Figure 8).

Accuracy of the Method
An environmental reference material (GBS07-1185-2000) with a Cd(II) concentration of 0.298 ± 0.01 mg/L was analyzed. The environmental reference material was diluted by a factor of 20. The reference material solution after dilution was determined by the experimental procedure that is shown in Section 2.5. The obtained Cd(II) concentration was 0.289 ± 0.004 mg/L, which was in good agreement with the certified value. The T-test results indicated that the analyzed value was not statistically significantly different from the certified value.

Limit of Detection
A 0.010 mg/L Cd(II) solution was enriched 20 times, according to Section 2.5, and then subjected to FAAS. The limit of detection (LOD) and the relative standard deviation (RSD, n = 6) of Cd were 0.1 g/L and 1.0% with a preconcentration factor of 10, respectively. The LOD that was obtained in this study was the same as that reported by using resin for preconcentration [14], both are one magnitude smaller than the value obtained by using amino-functionalized mesoporous sorbent for Cd(II) preconcentration [22].

Analysis of Cd(II) in Natural Water Samples
The above-mentioned method was used for the analysis of Cd(II) in lake and river water

Accuracy of the Method
An environmental reference material (GBS07-1185-2000) with a Cd(II) concentration of 0.298 ± 0.01 mg/L was analyzed. The environmental reference material was diluted by a factor of 20. The reference material solution after dilution was determined by the experimental procedure that is shown in Section 2.5. The obtained Cd(II) concentration was 0.289 ± 0.004 mg/L, which was in good agreement with the certified value. The T-test results indicated that the analyzed value was not statistically significantly different from the certified value.

Limit of Detection
A 0.010 mg/L Cd(II) solution was enriched 20 times, according to Section 2.5, and then subjected to FAAS. The limit of detection (LOD) and the relative standard deviation (RSD, n = 6) of Cd were 0.1 µg/L and 1.0% with a preconcentration factor of 10, respectively. The LOD that was obtained in this study was the same as that reported by using resin for preconcentration [14], both are one magnitude smaller than the value obtained by using amino-functionalized mesoporous sorbent for Cd(II) preconcentration [22].

Analysis of Cd(II) in Natural Water Samples
The above-mentioned method was used for the analysis of Cd(II) in lake and river water samples. Table 3 lists the results. The concentration of Cd in lake water was 0.6 µg/L and the Cd(II) concentration in the Yangtze River was below the LOD. The recoveries were 86% and 90%, respectively. Table 3. Determination of Cd(II) in water samples (sample volume: 100 mL, final volume: 10 mL, n = 3).

Sample
Added (

Conclusions
The innovation of this study was the synthesis of DMPA/PDA/ZnO and its application in the removal and preconcentration of Cd(II). DMPA/PDA/ZnO has high adsorption capacity (28.5 mg/g) to Cd(II), because the DMPA layer was capable of chelating Cd(II), and the adsorption of Cd(II) to DMPA/PDA/ZnO followed the pseudo-second-order kinetic model. The pre-concentrated Cd(II) on DMPA/PDA/ZnO can be dissolved in 5% HCl, which omitted the desorption process. The dissolved solution can be directly used for FAAS analysis. The results suggested that this method has many advantages, such as a simple preconcentration process (without desorption), low detection limit, and good precision. The detection limit was low (0.1 µg/L), and the proposed method can be applied to determine ultra-trace Cd(II) in environmental water samples with highly satisfactory results.

Conflicts of Interest:
The authors declare no conflict of interest.