Shashi Prabha
Dubey
ab,
Amarendra Dhar
Dwivedi
b,
Mika
Sillanpää
b,
Young-Nam
Kwon
*a and
Changha
Lee
*a
aSchool of Urban and Environmental Engineering, Ulsan National Institute of Science and Technology (UNIST), 100 Banyeon-ri, Eonyang-eup, Ulju-gun, Ulsan 698-805, Republic of Korea. E-mail: kwonyn@unist.ac.kr; clee@unist.ac.kr; Tel: +82-52-217-2810 Tel: +82-52-217-2812
bLaboratory of Green Chemistry, LUT Chemtech, Lappeenranta University of Technology, Sammonkatu 12, 50130 Mikkeli, Finland
First published on 8th September 2014
Carbon spheres (CSs) have become a recent focus of attention in environmental remediation techniques. In this study, imine-functionalized CSs were synthesized from plant extract (peCSs) for the first time and effectively used in U(VI) removal from contaminated water. Plant extracts of Sorbaria sorbifolia were utilized for the synthesis of peCSs via a single-step hydrothermal carbonization, and the physico-chemical properties of the synthesized peCSs were characterized by spectroscopic analysis. The peCSs showed high nitrogen content (∼7.49%) due to the presence of naturally occurring cyanogenic glycosides and mesoporosity (8.31 nm). The plant extract concentration played an important role in determining the size of the peCSs, which ranged from 0.5 to 3.0 μm. The adsorption capacity (Qm) of peCSs for U(VI) (Qm ≈ 113 mg g−1) was higher than that of the glucose-derived CSs (Qm ≈ 57 mg g−1) and commercial powdered activated carbon (Qm ≈ 44 mg g−1). A plausible mechanism for the higher adsorption efficacy of peCSs was proposed via sorbate–sorbent interactions. The ionic strength (0.01 M to 1 M NaCl) showed the weakest effect on the U(VI) adsorption. The multiple adsorption–desorption cycling test revealed that the efficacy of peCSs does not significantly decrease after repetitive use.
Uranium is a radioactive heavy metal that can cause cancer in humans subjected to long-term exposure via drinking water. Naturally occurring uranium is usually found at low levels in aquifers. However, the presence of U(VI) at high concentrations is sometimes reported in ground water and surface water as a result of uranium mining and milling activities.12 Considering the irreparable damage caused by U(VI) in drinking water, the guidelines of the World Health Organization suggest 30 μg L−1 as the maximum allowable uranium concentration in drinking water.13 Meanwhile, the recovery of uranium from wastewater and seawater is another important issue for the nuclear fuel industry and related areas. From an economic and technical perspective, adsorptive removal processes are an efficient and viable option. Carbon materials such as activated carbon, carbon fibres, carbon nanotubes, graphene oxide and mesoporous carbon, along with minerals and magnetic metal oxides, have been explored for the removal and the recovery of U(VI).14–20 More recently, Zhang et al. explored the application of CSs in the selective adsorption of U(VI).21,22
The literature showed that hydrothermal carbonization generally utilizes 50–100% of the contents of carbohydrate solutions for the synthesis of CSs.1,23–26 Considering carbohydrate utilization in CS synthesis, our desire was to find a novel and simple method for the synthesis of CSs using natural plant extract, which can also help to reduce the increasing chemical burden of synthesis routes. For this purpose, the Sorbaria sorbifolia (Rosaceae family) plant was chosen due to the presence of bioactive cyanogenic glycosides in the aerial part of the plant, which is a natural nitrogen-containing sugar source.27,28 This plant is a deciduous perennial shrub and commonly grows in temperate areas of Asia, including Korea, Japan and northern China.
The purpose of the present study is to synthesize plant-extract-derived CSs (peCSs) using a green synthetic approach. To the best of our knowledge, there is no report available on the synthesis of CSs from plant extract via a hydrothermal route. Moreover, compared to conventional hydrothermal carbonization, our method for peCS fabrication has the following advantages: (i) the peCSs are fabricated by a completely green synthetic approach in the absence of toxic chemicals, which increases its biocompatibility and potential for medical applications; (ii) the peCSs are fabricated by a simple single-step synthesis using aqueous plant extract as the carbon source; and (iii) the synthesized peCSs exhibit excellent U(VI) adsorption, exceeding that of glucose-derived CSs (gCSs) and commercial powdered activated carbon (PAC).
The cyanogenic glycosides could be converted to imino glycosides during hydrothermal reaction by simple hydrolysis. Furthermore, polymerization and carbonization of the aqueous extract during hydrothermal treatment resulted in the formation of peCSs. A possible mechanism for the conversion of the cyanogenic group to an imine group has been proposed and is presented in Scheme 1. It is believed that three steps (I to III) are involved in the production of the imine-derived functionality: (i) activation of nitrile by a weak nucleophile (H2O) and protonation, increasing its electrophilicity, (ii) the attack by the nucleophilic centre of water (O) on the electrophilic C centre in CN and (iii) deprotonation of the oxygen from H2O, forming the imine-functionalized group, which led to the formation of imine-functionalized CSs after polymerization and carbonization.
The effect of plant extract concentration on the size of peCSs formed after hydrothermal treatment was investigated (Fig. 2). Keeping other experimental conditions and the solution volume constant, the plant extract concentration was varied from 100% to 50% to 25%. The particle size of the peCSs decreased with decreasing concentration of plant extract concentration. At 100% plant extract concentration, 1.5 to 3.0 μm peCSs were observed. However, at 50% and 25% plant extract concentrations, 1.0 to 2.0 μm and 0.5 to 1.0 μm peCSs were observed, respectively. Similar observations have been made in glucose-based CS synthesis.25 The rest of the study was performed using the peCSs obtained from 50% plant extract.
Fig. 2 Microscopic images of peCS samples at different plant extract concentrations: (a) 100%, (b) 50%, (c) 25%. |
Raman spectra recorded to analyse the nature of the peCS carbon structure showed two prominent peaks (ESI Fig. S1b†). The disorder (D) and graphitic (G) bands, representing sp3- and sp2-bonded C, respectively, were centred at 1360 and 1590 cm−1, respectively. It could be seen that the G peak was more intense than the D peak. The intensity ratio of the D and G band (ID/IG), used to estimate the order and disorder of the carbon network, was lower in peCSs (0.901), indicating the presence of fewer defects and a more extensive sp2-bonded carbon structural network.
The structural compositions of peCSs were studied by XPS. Survey spectra and detailed analyses of each element are correspondingly presented in Fig. 3a and b. A single, clear peak at 284.38 eV could be assigned to a sp2 hybrid C 1s core structure, indicating the presence of graphitic carbon, as already confirmed by the Raman spectrum. The presence of O 1s species was confirmed by the peak at 532.48 eV. XPS analysis was valuable for further confirming the nature of the N 1s component in peCSs. We observed a N 1s peak at 399.78 eV, which is very close to the expected value for imine N (399.3 eV).37 The synthetic procedure and IR analyses mentioned above also indicated the formation of imine N in the peCS samples.
Fig. 3 (a) XPS survey spectra of native peCSs and peCSs after U(VI) adsorption. (b) C 1s, O 1s, N 1s and U 4f regions of (i) native peCSs and (ii) peCSs after U(VI) adsorption. |
The XRD pattern exhibited two broad peaks corresponding to the (002) and (100) phases, confirming the presence of the amorphous carbon phase in peCS samples (ESI Fig. S2a†).38
N2 adsorption–desorption was utilized to measure the surface area and pore size of the samples. The observed isotherm curve is presented in ESI Fig. S2b.† The specific surface area measured from BET equilibrium is 16.77 m2 g−1. The average mesopore size calculated from the BJH method showed the presence of mesopores (8.31 nm) on the surface of peCSs.
Additionally, the thermal stability of peCSs was compared in air and nitrogen atmosphere by SDT analysis. Fig. 4a and b show the obtained TGA and DSC curves, respectively. Heating in the presence of air resulted in the decomposition/combustion of peCSs in temperature range of 250 to 600 °C, and the rapid weight loss of the material accelerated after 250 °C due to fast oxidation with further increasing temperature. Above 600 °C, the weight stabilized at 15% of the original, revealing the completion of the decomposition/combustion process of peCSs. In the DSC curve, two major exothermic peaks were observed at 316 and 409 °C, which may be due to the decomposition and combustion, respectively, of organic and amorphous C. In contrast, when the sample was heated under N2 atmosphere, no exothermic peak was observed, and the weight loss gradually slowed above 350 °C. This finding suggests that peCSs show higher thermal resistance in N2 than air atmosphere, having stabilized at 60% of the original weight above 800 °C.
To further differentiate the physico-chemical properties of peCSs and gCSs, FTIR, Raman, XRD, N2-BET isotherm, XPS and SDT analyses were performed. The obtained results are discussed in ESI Section A1 and Fig. S3–S4.† The major differences were due to the presence of N-functionality in the peCS samples, which is clearly absent in the gCS samples. The other characterizations results for the gCSs were similar to those for the peCSs.
The effect of pH on the behaviour of U(VI) adsorption is explained by the U(VI) speciation at varying pH and the surface characteristics of peCSs. Potentiometric titration was used to analyse the surface charge of peCSs at varying pH (2 to 10), and the results are presented in ESI Fig. S5.† The surfaces of the peCSs remained positive until pH 4.59, becoming negative with further increases in pH.
According to the pH-dependent speciation of U(VI) (ESI Fig. S6,† calculated using the constants reported in the literature39), uncomplexed uranyl cation (UO22+) is predominant at pH values lower than 4. At pH 4, uranyl hydroxo complexes ((UO2)(OH)+, (UO2)2(OH)22+ and (UO2)3(OH)5+) are observed for the first time. In the pH range of 7−8, the uncharged dihydroxo complex ((UO2)(OH)2) becomes the major species, and at pH above 8, negative species such as (UO2)3(OH)82− and UO2(OH)3− are dominant. Considering the electrostatic interactions between the U(VI) species and the peCS surface, the attraction forces should be maximized at approximately pH 5.5–6, which is slightly higher than the observed optimum pH for the U(VI) absorption (Fig. 5b). A possible explanation is that the nucleophilic reactivity of imine nitrogen on the peCS surface may favour the interaction with positive species of U(VI), lowering the optimum pH. The interaction between imine and U(VI) can also be dependent on pH, possibly affecting the pH-dependency of U(VI) absorption.
The dependency of U(VI) adsorption on ionic strength has been studied by different researchers. Moreover, a decrease with increasing ionic strength was observed in earlier reported works,40,41 but in present study, the U(VI) adsorption by peCSs showed a remarkable resistance against the effects of high ionic strength. Hence, peCSs can be successfully utilized for U(VI) recovery from seawater or wastewater sources containing ionic background substances.
Experiments were performed to study the effect of contact time on U(VI) removal by peCSs, gCSs and PAC at different time intervals varying from 0.25 to 18 h. However, the adsorption equilibrium attained at 6 h and longer contact times did not show significant changes in adsorption efficacy. The kinetic curves (qtversus t) for peCSs, gCSs and PAC are presented in Fig. 6a.
Comparison of the experimental adsorption capacities of peCSs, gCSs and PAC with the calculated adsorption capacities revealed that the pseudo-second-order model fits the experimental data well. Furthermore, the high correlation coefficient (close to unity) and F-value (larger number) for the pseudo-second-order over the pseudo-first-order model suggests that the pseudo-second-order model better fits the U(VI) adsorption on peCSs, gCSs and PAC surfaces (Table 1). The F-value is calculated from the ANOVA table.42 Microcal Origin 8.0 was used to calculate the F-value and the correlation coefficient.
Adsorbent | q e,exp (mg g−1) | Pseudo first order | Pseudo second order | ||||
---|---|---|---|---|---|---|---|
q e,cal (mg g−1) | R 2 | F-value | q e,cal (mg g−1) | R 2 | F-value | ||
peCSs | 18.32 | 17.46 | 0.845 | 375 | 19.40 | 0.926 | 797 |
gCSs | 15.30 | 13.67 | 0.822 | 318 | 15.35 | 0.922 | 731 |
PAC | 6.31 | 5.56 | 0.697 | 621 | 5.92 | 0.889 | 1706 |
Furthermore, transfer of adsorbate moieties on the adsorbent surface can be controlled through external mass transfer in non-porous surfaces or intra-particle diffusion to a porous surface material by adsorption on the interior sites of the adsorbent,43 which can be proved by the linearity of a plot of qtvs. t0.5. Intra-particle diffusion is the sole rate-limiting step if the curve passes through the origin.44 Because peCSs have large pores, which may facilitate the diffusion of adsorbate inside the pore, it appears that adsorption followed the intra-particle diffusion mechanism. The intra-particle diffusion rate constant (kipd) and correlation coefficient of peCSs and gCSs were calculated from the slope of the second linear section (Table 2 and the ESI Fig. S8†). Considering the multilinearity of these plots for U(VI) adsorption on peCSs, gCSs and PAC, three phases may be involved in the adsorption processes. In peCS, gCS and PAC samples, the first linear section revealed external surface diffusion, the second linear section represented intra-particle diffusion, and final linear section was an equilibrium adsorption stage. Because the curves did not pass through the origin, intra-particle diffusion could not be established as the only rate-determining step in the adsorption process on different adsorbent materials. Unuabonah et al. has also suggested that two or more steps govern the adsorption process when qtvs. t0.5 is multilinear.45 However, it was definitely one of the rate-determining steps in a specific time range, and this model was noticeably relevant for peCS-based U(VI) adsorption, which showed a higher correlation coefficient (0.973) than that for gCSs (0.933) and PAC (0.926).
Adsorbent | k ipd (mg g−1 h−1/2) | C (mg g−1) | R 2 | F-value |
---|---|---|---|---|
peCSs | 5.44 | 3.48 | 0.973 | 214 |
gCSs | 3.85 | 3.17 | 0.933 | 84 |
PAC | 0.79 | 3.68 | 0.927 | 76 |
The effect of initial U(VI) ion concentration on adsorption by peCSs, CS and PAC was analysed. Isotherm curves were plotted between adsorption capacity (qe) and equilibrium concentration (Ce), as in Fig. 6b. Two parameter models (Langmuir and Freundlich) were investigated to fit the experimental data. The Langmuir isotherm yielded higher correlation coefficients and F-values than the Freundlich model for all three adsorbents over the entire U(VI) concentration range from 1 to 100 mg L−1. Hence, the maximum adsorption capacities (Qm) of the adsorbents were evaluated from the Langmuir isotherm model. Higher Qm was observed for peCSs (∼113 mg g−1) than gCSs (∼57 mg g−1) or PAC (∼44 mg g−1) under the optimized conditions (Table 3).
Adsorbent | Langmuir isotherm | Freundlich isotherm | ||||||
---|---|---|---|---|---|---|---|---|
Q m (mg g−1) | k L | R 2 | F-value | k f | n | R 2 | F-value | |
peCSs | 113.16 | 0.149 | 0.964 | 338 | 6599 | 3.25 | 0.935 | 186 |
gCSs | 57.16 | 0.066 | 0.909 | 153 | 552 | 2.73 | 0.885 | 120 |
PAC | 44.13 | 0.041 | 0.903 | 148 | 56 | 2.37 | 0.931 | 232 |
Higher U(VI) adsorption capacity of peCSs could possibly be explained by pore size (mesopore range) and chelating groups on the peCSs surfaces. Synthesized peCSs exhibited lower surface area (16.77 m2 g−1) than the gCSs (29.56 m2 g−1) sample but the average pore size of peCSs (8.31 nm) was larger than that of gCSs (6.78 nm), which increased corresponding intraparticle diffusion rate (5.44 mg g−1 h−1/2) compared to 3.85 mg g−1 h−1/2 (gCSs) and 0.79 mg g−1 h−1/2 (PAC). These results suggest that the high U(VI) adsorption capacity of peCSs is achieved by the facilitated diffusion of adsorbates into the material pores. Similar observations have been reported in literatures.46,47 In addition, the surface functional groups of peCSs could have intensive electrostatic and chemical interactions with U(VI) ions. The presence of N-containing groups (imine in this study) on the peCS surface form complexes with uranyl species.40 It is speculated that two imine-nitrogens with lone pair electrons coordinate one U(VI) as bidentate ligands. Accordingly, imine group-containing peCSs displayed higher U(VI) adsorption capacity than gCSs and PAC. The reduction in the N 1s peak in the U(VI)-adsorbed peCS sample in Fig. 3a suggested the involvement of an interaction between the N 1s species and U(VI), which facilitated the higher adsorption of U(VI) on the peCS surface relative to the gCS or PAC samples.
Moreover, the dimensionless separation factor (Sf) is generally used to express the characteristics of the Langmuir isotherm. It is reliable when the Sf value is greater than zero and less than one.48 The Sf values of peCSs, gCSs and PAC were calculated and presented in Table 4, showing that the Langmuir adsorption model was favoured for all three adsorbents (0 < Sf < 1).
Initial concentrations | Dimensionless separation factor | ||
---|---|---|---|
C 0 (mg L−1) | peCSs | gCSs | PAC |
1 | 0.870 | 0.938 | 0.961 |
2.5 | 0.728 | 0.858 | 0.907 |
5 | 0.573 | 0.752 | 0.829 |
10 | 0.402 | 0.602 | 0.709 |
20 | 0.251 | 0.431 | 0.549 |
30 | 0.183 | 0.336 | 0.448 |
40 | 0.144 | 0.275 | 0.379 |
50 | 0.118 | 0.233 | 0.328 |
100 | 0.063 | 0.132 | 0.196 |
Footnote |
† Electronic supplementary information (ESI) available: Details of sections A1, A2 and Fig. S1–S8. See DOI: 10.1039/c4ra06890d |
This journal is © The Royal Society of Chemistry 2014 |