ARTICLE pubs.acs.org/Langmuir
Functionalization of Regenerated Cellulose Membrane via Surface Initiated Atom Transfer Radical Polymerization for Boron Removal from Aqueous Solution Yu-Ting Wei, Yu-Ming Zheng, and J. Paul Chen* Department of Civil and Environmental Engineering, National University of Singapore, 10 Kent Ridge Crescent, Singapore 119260
bS Supporting Information ABSTRACT: In this study, an adsorptive membrane was prepared for efficient boron removal. Poly(glycidyl methacrylate) was grafted on the surfaces of the regenerated cellulose (RC) membrane via surface-initiated atom transfer radical polymerization, and N-methylglucamine was used to further react with epoxide rings to introduce polyhydroxyl functional groups, which served as the major binding sites for boron. The pristine and modified membranes were characterized by X-ray photoelectron spectroscopy (XPS), attenuated total reflection-Fourier transform infrared spectroscopy (ATR-FTIR), dynamic water contact angle measurement, and scanning electron microscopy. It was shown that the designed functional groups were successfully grafted onto the RC membrane, and surface modification contributed to higher boron binding capability. The optimal pH range for boron adsorption was 48. Under a neutral pH condition, the maximum adsorption capacity of the modified membrane was determined to be 0.75 mmol/g, which was comparable with those of commercial resins. Studies of electrolyte influence indicated the formation of inner-sphere surface complexes on the membrane surface. The ATR-FTIR and XPS analyses showed that secondary alcohol and tertiary amine groups were mainly involved in boron adsorption, and tetrahedral boron complexes were found on the membrane surface.
1. INTRODUCTION Boron naturally occurs in the aqueous environment, especially in seawater. As an essential micronutrient, it plays a critical role for both plant growth and human health. However, if boron content is excessive in irrigation water, the plants will show certain toxicity symptoms, such as poor growth and even plant decay and expiration.1 For human beings, overexposure of boron may result in acute toxicity, including nausea, vomiting, diarrhea, kidney damage, and reproductive and nervous system diseases.1,2 Seawater has increasingly become an important water source. Reverse osmosis (RO) is the most used technology for seawater desalination, which is based on the chemical potentials of solutes. It can efficiently reject most of salts, such as sodium and chloride. However, RO is inefficient in boron removal, which is likely due to the lower concentration of borate ions in seawater.35 Hence, desalination would need a multiple-stage RO system or polymerassisted filtration.68 Such systems, however, require high energy for the seawater desalination process. In recent years, use of an adsorptive membrane has emerged as an alternative approach to remove contaminants from aqueous solution.9,10 The membrane carries specific functional groups on the surfaces, which can reject the contaminants through a series r 2011 American Chemical Society
of interactions, such as charged neutralization, ion exchange, or surface complexation. Therefore, even though the pore size of such a membrane is much larger than the size of the contaminants, it can still exhibit high retaining efficiencies. Besides, it is favorable to some extent for increasing the permeate flux and meanwhile greater reducing the energy consumption. Most research on boron uptake has demonstrated that the vicinal polyalcohol functional groups are the most efficient ligands for the complexation of boron in aqueous solutions.6,11 Hence, our aim was to prepare an adsorptive membrane that possessed such functionalities for efficient boron removal. Due to the presence of hydroxyl groups, regenerated cellulose (RC) membrane is considered to be a versatile platform for surface modification. Among the chemical modification techniques, the atom transfer radical polymerization (ATRP) developed by Wang and Matyjaszewski has been proven to be versatile for the polymerization of various vinyl monomers, such as acrylonitrile, acrylates, methacrylates, and styrene.12 It does not require the Received: January 13, 2011 Revised: March 28, 2011 Published: April 21, 2011 6018
dx.doi.org/10.1021/la200154y | Langmuir 2011, 27, 6018–6025
Langmuir
ARTICLE
Figure 1. Schematic illustration of the surface modification of RC membrane.
stringent experimental conditions that the anionic or cationic polymerization requires.1315 Much research has been conducted to functionalize the RC membranes surface via ATRP for different applications. For example, Singh and co-workers prepared a polymeric membrane adsorbent by growing surface-tethered and charged polymer nanolayers on the surfaces of RC membranes. The prepared membranes exhibited a high binding capability for lysozyme.16 A high-capacity weak-anion-exchange membrane was designed by Bhut and coauthors for the protein separation through surface-initiated ATRP. It was found that the protein static adsorption capacities increased with an increase in polymerization time and finally achieved a plateau value of approximately 66 mg/mL.17 Besides, a p-vinylbenzylsulfobetaine was grafted on RC membranes by Liu and co-workers using ATRP for blood compatibility improvement, and the designed membranes showed an improved resistance to nonspecific protein adsorption and platelet adhesion.18 However, to the best of our knowledge, there has been no research on the surface-initiated ATRP to functionalize RC membranes for boron removal application. In the present work, an adsorptive membrane was prepared by grafting poly(glydicyl methacrylate) (PGMA) on a RC membrane via the ATRP technique and further introducing polyhydroxyl groups through ring-opening reactions with N-methylglucamine (NMDG) on the membrane. In our early studies, the NMDG exhibited high affinity for boron and arsenic.19,20 The chemical functionality, wettability, and physical morphology of the pristine and modified membranes were characterized by attenuated total reflection-Fourier transform infrared spectroscopy (ATR-FTIR), X-ray photoelectron spectroscopy (XPS), water contact angle (WCA), and scanning electron microscopy (SEM), respectively. In addition, the adsorption performance as a function of pH, boron
initial concentration, and ionic strength was investigated. The adsorption mechanism was finally studied by ATR-FTIR and XPS. The objective of this study was to obtain a high-performance functional membrane and to better understand the underlying mechanism for the boron uptake.
2. MATERIALS AND METHODS 2.1. Materials. Regenerated cellulose membrane with a diameter of 47 mm and average pore size of 1.0 μm was purchased from Whatman Inc. Tetrahydrofuran (THF, HPLC grade) and methanol (HPLC grade) from TEDIA, 2-bromoisobutyryl bromide (2-BIB, >97%) from Fluka as well as triethylamine (TEA, >99%) from Sigma-Aldrich were used for the introduction of polymerization initiator. Glycidyl methacrylate (GMA, g97%) supplied by Fluka was used in the surface graft polymerization on RC membranes. Chemicals used for GMA polymerization included N,N-dimethylformamide (DMF, HPLC grade) from TEDIA, 2,20 -bipyridine (BPY, g99%) and copper(I) bromide (CuBr, 98%) from Sigma-Aldrich, and acetone (HPLC) grade from TEDIA. N-Methyl-D-glucamine (NMDG, g97%) from Fluka was used for further functionalization of grafted polymers. Boric acid (H3BO3, 99.6%) from Fisher Scientific was used to prepare the stock solutions of boron. Sodium perchlorate (NaClO4, 98%) used as electrolyte backgrounds was analytical grade purchased from Sigma-Aldrich. Nitric acid (HNO3, g65%) and sodium hydroxide (NaOH, g99%) from Merck were used for pH adjustment. 2.2. Membrane Surface Modification. The RC membrane was first soaked in methanol for 1015 min to remove glycerine and rinsed with ultrapure water to remove the residual methanol. The modification of the pretreated membrane mainly consisted of three steps (Figure 1): (1) surface initiation, (2) surface-initiated atom transfer radical polymerization, and (3) NMDG functionalization. 6019
dx.doi.org/10.1021/la200154y |Langmuir 2011, 27, 6018–6025
Langmuir 2.2.1. Surface Initiation. The objective of the surface initiation (the first step) was to introduce the initiator onto the RC membrane to obtain the initiated membrane (termed RC-Br). The RC membrane was immersed in anhydrous THF for 10 min. The membrane was removed from the THF and dried for 5 min before being placed in the solution for the initiation. The initiator 2-BIB was immobilized by the reaction with hydroxyl groups on the RC membrane to covalently immobilize bromo ester initiator groups to the membrane surface. A typical solution consisted of RC membranes, 2-BIB, TEA, and anhydrous THF, in which the molar ratio of 2-BIB to TEA was kept at 1:1. The mixture for the reaction was gently stirred at 0 °C for 2 h and then at the room temperature overnight. The initiator-functionalized membrane (RC-Br) was washed thoroughly with methanol and ultrapure water. 2.2.2. Surface-Initiated Atom Transfer Radical Polymerization. The second step was the surface-initiated atom transfer radical polymerization, by which the PGMA was grafted onto the initiated membrane (RCBr) to obtain PGMA grafted membrane (RC-PGMA). The initiators GMA and BPY were added into a mixed solvent of DMF and ultrapure water. The mixture was deoxygenated with argon for 30 min; and CuBr was then introduced to the mixture. The flask was sealed and the ATRP reaction proceeded for a predetermined period of time. Subsequently, the RC membrane with surface-grafted poly(glycidyl methacrylate) (denoted as RC-PGMA) was removed from the solution and washed thoroughly with excess acetone and ultrapure water to remove various impurities. 2.2.3. NMDG Functionalization. The final step was the NMDG functionalization that led to its immobilization on the RC-PGMA to obtain the functionalized membrane (RC-MG). The RC-PGMA was treated by excess NMDG in DMF. After the reaction, the mixture was cooled, washed with ultrapure water, dried, and finally stored for the adsorption and membrane characterization studies. 2.3. Surface Analyses. ATR-FTIR, XPS, SEM, and WCA were used to characterize the surface properties of functionalized RC membrane. 2.3.1. Attenuated Total Reflection-Fourier Transform Infrared Spectroscopy. ATR-FTIR was used to provide information on the surface chemistry of unmodified, initiator-functionalized, GMA-grafted, and NMDG-modified membranes. It was also employed to elucidate the adsorption mechanism of NMDG-modified membranes. The instrument was IRPrestige-21 (Shimadzu) equipped with a single reflection ATR accessory with ZnSe crystal. The spectra were recorded within the range of 6004000 cm1 with a resolution of 4 cm1 over 32 scans. 2.3.2. X-ray Photoelectron Spectroscopy. Different types of membranes were analyzed using X-ray photoelectron spectroscopy (ULVACPHI Quantera SXM), with a monochromatic Al KR X-ray source. The applied voltage and total output power were 15 kV and 25 W, respectively. For wide scan spectra, an energy range from 0 to 1100 eV was used with a pass energy of 112 eV and step size of 0.1 eV. The high-resolution scans were conducted according to the peak being examined with a pass energy of 55 eV and step size of 0.05 eV. To compensate for the charging effects, all spectra were calibrated with graphitic carbon as the reference at a binding energy (BE) of 284.6 eV. The XPS results were collected in binding energy forms and fitted using a nonlinear least-squares curve fitting program (XPSPEAK41 software). 2.3.3. Scanning Electron Microscopy. The surface morphology of pristine and modified membranes was visualized by an SEM instrument (JEOL, JSM-5600 V). During the SEM analysis, the samples were initially mounted on the sample studs using double-sided adhesive tape and then coated with platinum for production of electrical conductivity during test. The SEM analysis enables the direct observation of the changes in the surface microstructures due to chemical modification. 2.3.4. Contact Angle. Water contact angles of the membranes were measured and calculated in dynamic mode at ambient temperature. The
ARTICLE
Figure 2. ATR-FTIR spectra of (a) RC, (b) RC-PGMA, and (c) RC-MG. automated contact angle goniometer (DSA20B, Kruss) was equipped with software-controlled dosing system and contact angle determination software. For each measurement of contact angle value, at least three points from different parts of the membrane surface were recorded and averaged. The curves of contact angle versus drop age were plotted to compare the relative hydrophilicity of the membranes at different reaction stage. 2.4. Boron Adsorption. In the pH effect study, boron solutions at various initial pH values were adjusted by nitric acid or sodium hydroxide. The RC-MG membrane was added to boron solution with a known concentration. All mixtures were shaken at room temperature for 7 d. The samples were taken at the end of the experiments to analyze boron concentration by an inductively coupled plasma emission spectrometer (ICP-OES; Perkin-Elmer Optima 7300). Both initial and equilibrium pH values were measured by an ORION 920Aplus pH meter. Adsorption isotherm experiments were conducted by adding a dosage of 0.5 g/L pristine and modified membranes in high-density polyethylene (HDPE) bottles with varying initial boron concentrations. The initial pH of the solution was adjusted to be 7. All bottles were shaken at room temperature for 7 d to reach equilibrium. At the end of the experiments, 20 mL samples were collected and filtered for the boron concentration analysis by the ICP-OES. The experiments to determine the effect of ionic strength on boron removal were performed by adding 20 mg of modified membranes into a vessel containing 40 mL of 0.47 mM boron solution. The pH of the solutions was adjusted every 4 h with dilute HCl or/and NaOH solution to designated values in the range of 410 during the adsorption process. Finally, the equilibrium pH was measured and the supernatant was filtered for boron concentration analysis.
3. RESULTS AND DISCUSSION 3.1. Surface Characterization. Surface properties of original membranes and modified membranes were characterized by ATR-FTIR and XPS to elucidate the synthetic reactions. 3.1.1. ATR-FTIR Analysis. Figure 2 presents typical ATR-FTIR spectra for unmodified RC membranes, RC-PGMA, and RC-MG. For the pristine RC membrane (spectrum a), the broad peak located at 3361 cm1 is due to the OH stretching vibration.21 The spectrum also clearly shows the major peaks of the saccharine structure, which include asymmetric stretching of COC at 6020
dx.doi.org/10.1021/la200154y |Langmuir 2011, 27, 6018–6025
Langmuir
ARTICLE
Table 1. Elemental Surface Composition of Pristine and Modified RC Membranes Determined from XPS element (atom %) sample
C
O
RC
55.4
44.6
RC-Br
57.8
41.6
RC-PGMA RC-MG
70.2 62.9
29.8 33.2
N
Br
0.6 3.9
1159 cm1 as well as skeletal vibration involving the CO stretching at 1062 and 1024 cm1.18,22 The band at 896 cm1 characterizes the amorphous regions of RC membranes.23,24 Due to the low content of initiator in comparison to the substrates (RC membrane), the characteristic peaks for RC-Br are not visible (data not shown). Following the polymerization, a new peak appears at 1726 cm1 in the RC-PGMA (spectrum b), which is attributed to the ester CdO stretching vibrations in COO.25,26 The vibration at 1452 cm1 corresponds to the CH2 scissoring band of PGMA.26 Three obvious characteristic bands of epoxide groups at 906, 848, and 759 cm1 are also found in the spectrum.27,28 Meanwhile, the intensity of hydroxyl characteristic peak at 33003400 cm1 for RC membranes greatly decreases, which can be explained as follows. The hydroxyl groups are involved in the polymerization. On the other hand, the surface of membrane is covered by a layer of PGMA after the polymerization. Both can result in the decrease in intensity of OH. All the above-mentioned changes of spectra support the successful grafting of PGMA from the surface of RC membranes. Comparison of spectra b and c shows that three typical characteristic peaks for epoxide groups disappear after further functionalization with the NMDG, implying the successful conversation of the epoxide groups. In addition, new peaks appear at 1058 and 3377 cm1, which can be assigned to the amine CN stretching vibrations and OH stretching vibrations, respectively.21 The results provide clear evidence of the successful ring-opening reactions to introduce polyhydroxyl and tertiary amine functional groups. 3.1.2. XPS Analysis. XPS analysis was used to analyze the surface composition variations of the pristine and modified RC membrane surfaces. Table 1 summarizes the elemental compositions determined by the XPS study on the different surfaces. As shown, after surface-initiated reaction, the content of carbon (C 1s) is slightly increased, while the content of oxygen (O 1s) is slightly decreased. Simultaneously, a small amount of bromine (Br 3d, 0.6%) appears with a binding energy of around 70 eV (Supporting Information, Figure 1b).29 Figure 1c,d of the Supporting Information provides the narrowly scanned C 1s diagrams; the spectrum of RC and RC-Br can be deconvoluted into three individual component peaks, which come from different groups and overlap each other. The peaks of RC at 284.6, 286.2, and 287.6 eV are mainly attributed to CC/CH, CO, and CCdO species, while peaks of RC-Br at 284.6, 286.2, 287.6, and 288.6 eV can be assigned to CC/CH, CO, CCdO, and OCdO groups.22 It is clear that a new peak at 288.6 eV appears in the C 1s of RC-Br. All these changes indicate that the initiator is successfully anchored on the membrane surface. After the ATRP reaction, the quantity of carbon increases, whereas that of oxygen greatly decreases. As shown in Table 1, the C/O atomic ratio is approximately 7:3 (70.2%: 29.8%),
Figure 3. XPS spectra of RC-PGMA (a and c) and RC-MG (b and d).
Figure 4. Water contact angle versus drop age for pristine and modified RC membranes.
which agrees well with the corresponding theoretical ratio in GMA. The C/O ratio further suggests that the RC membrane is covered by the grafted PGMA layer. With the further NMDG functionalization, the content of oxygen remarkably increases to 33.2% and that of nitrogen on the surface of RC-MG increases to 3.9%. However, the content of nitrogen on the RC-PGMA is zero. Figure 3b,d demonstrates the wide scan and C 1s narrow scan of the RC-MG. In comparison with the spectra of RC-PGMA, a new signal of N 1s appears, which originates from amine groups in the NMDG. Besides, a new peak component at 285.6 eV, attributed to CN species, appears in the narrow scan of C 1s in RC-MG. The results from XPS analyses are thus consistent with the ATR-FTIR results and confirm the successful grafting of polyhydroxyl and tertiary amine groups onto RC membranes. 6021
dx.doi.org/10.1021/la200154y |Langmuir 2011, 27, 6018–6025
Langmuir
ARTICLE
Figure 5. SEM images of RC (a and b) and RC-MG (c and d) membranes: (a) top view 2500, (b) cross section view 2500, (c) top view 2500, and (d) cross section view 2500.
3.1.3. Water Contact Angle Analysis. The relative hydrophilicity or hydrophobicity for four surfaces was investigated with a dynamic mode. Figure 4 gives the curves of water contact angle as a function of drop age. As shown, due to the presence of massive hydrophilic functional groups in pristine RC membrane, its initial water contact angle is very low and quickly decreases to about 4° within 1.3 s. After the anchoring of initiator, the initial water contact angle of RC-Br sharply increases to 130°, and the decrease of contact angle with drop age is very slow. These changes imply a transformation from hydrophilic to hydrophobic of membrane surfaces, which is due to the immobilization of hydrophobic initiator. However, after the ATRP reaction, it can be found that the initial contact angle of these grafted surfaces (RC-PGMA) becomes smaller and finally declines to nearly 30°. With further ring-opening reaction, compared to RC-Br, the attenuation of contact angle with drop age for RC-MG is much faster, which decrease to only 9° after 5 s. Such a tendency suggests that despite the high hydrophobicity of the surface initiators, the containing of hydrophilic hydroxyl groups still brings hydrophilicity to the RC-MG membrane. 3.1.4. Surface Morphology of Membranes. Figure 5 and Supporting Information Figure 2 show the SEM images of the top and cross-sectional views for pristine and modified RC membranes at different magnification. Comparison of the top views in Figure 5a and Supporting Information Figure 2a with Figure 5c and Supporting Information Figure 2c indicates different surface morphologies between the virgin and grafted RC substrates. A relatively dense layer forms on the RC membrane surface, and little porous structure is observed, which is possibly due to the fact that the grafted polymer fills the pore spaces of RC membranes.30 Besides, the cross-sectional views in Figure 2b,d of the Supporting Information show that the thickness of membranes increases after chemical
Figure 6. Effect of pH on the boron removal by RC-MG (experimental conditions: [B]o = 0.46 mM, m = 0.5 g/L, T = 293 K, contact time = 7 d).
modification. The polymer is grafted on the membrane outer surface and the pore surfaces within the bulk of the membranes. Due to the polymerization, the pore size of the RC-MG membranes would become slightly smaller than that of the RC membrane (1.0 μm), which is demonstrated in Figure 2ad of the Supporting Information. The internal pores of the membrane seem not to decrease significantly after the modification, as shown in Figure 5b,d. The mass transfer resistance of boron is not anticipated to increase dramatically. Hence, an equilibrium time such as a few days would be sufficient to achieve complete adsorption, which is comparable to the adsorbent developed by a similar methodology.20 An extended 6022
dx.doi.org/10.1021/la200154y |Langmuir 2011, 27, 6018–6025
Langmuir
Figure 7. Adsorption isotherms of boron removal by pristine and modified RC membranes (experimental conditions: pH = 7, m = 0.5 g/L, T = 293 K, contact time = 7 d).
adsorption time of 7 days in this study was used in order to ensure complete uptake of the boron from the aqueous solution. The variation in surface structures would affect the flux of membranes. However, due to its nature of controllability in degree of polymerization, the ATRP technique can be used to effectively control the pore sizes and thickness of the prepared membranes in order to achieve the desired performance. 3.2. Adsorption Studies. Batch adsorption experiments were conducted to evaluate the adsorption performance of the functionalized membrane for boron removal under different conditions. 3.2.1. Effect of pH. Figure 6 shows the effect of pH on the boron adsorption. The modified membrane efficiently removes the boron from the aqueous solution over a wide range of pH. The adsorption capacity of approximately 0.5 mmol/g can be achieved at an initial pH 48. However, the boron uptake starts to decrease from pH 9 and drops over 50% when the pH increases to 11. It has been reported that, at a total concentration of less than 25 mmol/L, the major species are only in the forms of H3BO3 and B(OH)4, and the polyanionic species do not exist.31 The hypothesized reactions are given in eqs 14, where R*OH represents the functional group on the modified membrane RC-MG.
When the pH is below 8, the neutral forms of boron are primarily present in solution (pKa = 9.2 at 25 °C).32 Boric acid can be adsorbed onto the RC-MG through the complexation with the surface functional groups (OH), and H2O and/or Hþ may simultaneously release according to eqs 13.33,34 At higher pH values (e.g., pH >9), borate anions become the predominant forms. Direct competition between the hydroxyl ions and borate ions (eq 4) results in less boron removal. This observation is consistent with the finding reported by Wang and co-workers.35
ARTICLE
It can be found that pH changes during adsorption are not very obvious. Under an acidic condition, the equilibrium pH increases slightly higher, which is possibly due to the protonation of tertiary amine and hydroxyl groups on the surface of RC-MG. The optimal pH within a wide range plays an important role for actual application. It means that no pH adjustment is necessary, as the majority of water supplies usually have a range of 6.58.5. In addition, the used RC-MG membrane may be regenerated at higher pH, such as pH above 11 according to Figure 6. 3.2.2. Adsorption Isotherm. Figure 7 illustrates the adsorption isotherms of pristine and modified membranes. As shown, the RC-MG exhibits a much higher uptake of boron, while the adsorption capacity of RC nearly approaches zero. In other words, the surface modification definitely increases the boron removal capability. Direct graphic maximal adsorption capacity (corresponding to the isotherm plateau) is determined to be 0.75 mmol/g for the RC-MG, which is comparable with those of commercial resins, such as Amberlite IRA 743.20 Both Langmuir and Freundlich isotherms were used to analyze the experimental data of RC-MG. It was found that the Freundlich isotherm given below was more suitable for describing the adsorption behavior of RC-MG qeq ¼ Kf Ceq 1=n
ð5Þ
where Kf is Freundlich constant and 1/n is the heterogeneity coefficient. The Freundlich isotherm is used to describe an adsorption where the surface is heterogeneous with adsorption sites that have different energies. The better fitting by the Freundlich isotherm clearly demonstrates the existence of heterogeneous adsorption in the uptake of boron by the RC-MG. The result further indicates a multilayer adsorption in the uptake. 3.2.3. Effect of Ionic Strength. In the experiment of effect of ionic strength on boron adsorption, the ionic strength was varied for 2 orders of magnitude, from 1 to 100 mM NaClO4. As shown in Figure 3 of the Supporting Information, the adsorption peak is found at pH 6 in the presence of 1 mM NaClO4; an increase in ionic strength leads to a shift in the position of pH edge toward the alkaline region (pH 7 to 8). In addition, at low pH, the boron adsorption slightly decreases with increasing ionic strength, while the effect of ionic strength is reversed in the high pH regions. At low pH, the surface is usually positively charged; an increase in ionic strength causes a decrease in the potential in the plane of adsorption and thus less boron removal. At high pH, an increase in ionic strength can provide more cations (e.g., Naþ) near the negatively charged surface, which causes an increase in the potential in the plane of adsorption.36 It is favorable for the uptake of borate anions. A similar phenomenon was observed on the adsorption of boron on the oxides and clays.36 The observation of the influence of ionic strength on the adsorption may provide some insight to distinguish between inner-sphere and outer-sphere adsorption. Generally, the adsorption by the formation of outer-sphere complexes is weakly bonded and highly dependent upon the ionic strength, since the background electrolyte anions such as ClO4 may also form outer-sphere complexes via charged interaction and directly compete with adsorbed ions for surface binding sites. Conversely, if the uptake is due to the formation of inner sphere complexes, the adsorption shows little sensitivity to ionic strength or responds to higher ionic strength with greater adsorption.37 As a consequence, the results given in Figure 3 of 6023
dx.doi.org/10.1021/la200154y |Langmuir 2011, 27, 6018–6025
Langmuir the Supporting Information imply that the boron may form inner-sphere complexes at the water/solid interface. 3.3. Mechanism Study. Figure 4 of the Supporting Information shows the ATR-FTIR spectra of the RC-MG before and after the boron adsorption. As shown in curve a (for the pristine RC membrane), the peak at 892 cm1 is attributed to the CO stretching vibration which originates from secondary alcohols (CHOH).21 After the boron adsorption (curve b), the most obvious finding is that the peak at 892 cm1 greatly diminishes, and meanwhile, a new peak appears at 958 cm1, which is assigned to the tetrahedral B asymmetric stretching band.34,38 It is therefore confirmed that the boron is successfully adsorbed onto the membrane surface by the formation of tetrahedral complexes. After the boron adsorption, the OH band shifts from 3379 to 3396 cm1, and the CN stretching vibration in tertiary amine and CO stretching vibration in polyhydroxyl groups shift from 1058 to 1070 cm1, respectively. The minor shifts in the peaks indicate that the amine and hydroxyl groups are possibly involved in the boron adsorption. The widescan spectrum of boron-loaded membranes in Figure 5 of the Supporting Information clearly shows a peak around 190 eV. This further confirms the successful attachment of boron species onto the modified membrane. Boron is normally removed by the RO technology that requires a higher energy consumption due to the working pressure of 5070 bar. As the concentrations of sodium and chloride are much higher than the concentration of boron, most of both sodium and chloride can be rejected by the RO membrane and the removal of boron is negligible. The average pore size of the RC membrane prior to the modification is 1.0 μm. The SEM results given in Figure 2ad of the Supporting Information show that the pore size of the modified membrane becomes slightly smaller. The RC membrane serves as a “carrier” of functional groups, which can greatly adsorb the boron. A pressure of around 1 bar is needed in the filtration. The modified membrane would outperform the conventional RO membrane due to a great energy saving in the operations.
4. CONCLUSIONS A regenerated cellulose membrane has been successfully modified for efficient boron removal from aqueous solution by a three-step surface modification method involving surfaceinitiated ATRP technology. The ATR-FTIR and XPS analyses confirmed the anchoring of ATRP initiator and grafting of PGMA polymers onto the RC membranes, as well as further functionalization by NMDG. Dynamic contact angle measurements demonstrated the hydrophilicity of the prepared membranes. SEM observations illustrated that surface modification slightly decreased the porous structure and increased the thickness of membranes. The modified membrane showed much higher binding capacity for the boron than the pristine RC membrane does and is comparable with commercially available resins. The optimal initial pH for boron uptake is in a wide range of 48, and the maximum adsorption capacity is around 0.75 mmol/g at a neutral pH. The result from the study of the effect of ionic strength demonstrated that the boron may form innersphere surface complexes at the water/solid interface. The ATRFTIR study revealed that polyhydroxyl and tertiary amine groups were mainly responsible for the boron removal; tetrahedral boron complex was present on the modified membrane surface.
ARTICLE
The XPS study further verified the successful attachment of boron. Therefore, the modified membrane would have a great potential as a novel separation material for boron removal from aqueous solutions.
’ ASSOCIATED CONTENT
bS
Supporting Information. XPS spectra of RC, RC-Br, and boron-loaded RC-MG membranes; SEM images of RC and RCMG membranes with magnification of 500; and ATR-FTIR spectra of RC-MG membranes before and after boron adsorption. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected];
[email protected].
’ ACKNOWLEDGMENT The authors would like to express their appreciation to Agency for Science, Technology and Research of Singapore (Grant No. 0 921 010 059, R-288-000-066-305) for the financial support of this study. ’ REFERENCES (1) Parks, J. L.; Edwards, M. Crit. Rev. Environ. Sci. Technol. 2005, 35 (2), 81–114. (2) Xu, Y.; Jiang, J. Q. Ind. Eng. Chem. Res. 2008, 47 (1), 16–24. (3) Prats, D.; Chillon-Arias, M. F.; Rodriguez-Pastor, M. Desalination 2000, 128 (3), 269–273. (4) Nadav, N. Desalination 1999, 124 (13), 131–135. (5) Glueckstern, P.; Priel, M. Desalination 2003, 156 (13), 219–228. (6) Geffen, N.; Semiat, R.; Eisen, M. S.; Balazs, Y.; Katz, I.; Dosoretz, C. G. J. Membr. Sci. 2006, 286 (12), 45–51. (7) Parschova, H.; Mistova, E.; Matejka, Z.; Jelinek, L.; Kabay, N.; Kauppinen, P. React. Funct. Polym. 2007, 67 (12), 1622–1627. (8) Kabay, N.; Sarp, S.; Yuksel, M.; Kitis, M.; Koseoglu, H.; Arar, O.; Bryjak, M.; Semiat, R. Desalination 2008, 223 (13), 49–56. (9) Zheng, Y. M.; Zou, S. W.; Nanayakkara, K. G. N.; Matsuura, T.; Chen, J. P. J. Membr. Sci. 2011, 374 (12), 1–11. (10) Han, W.; Bai, R. B. J. Appl. Polym. Sci. 2009, 115 (4), 1913– 1921. (11) Bicak, N.; Bulutcu, N.; Senkal, B. F.; Gazi, M. React. Funct. Polym. 2001, 47 (3), 175–184. (12) Yu, W. H.; Kang, E. T.; Neoh, K. G.; Zhu, S. P. J. Phys. Chem. B 2003, 107 (37), 10198–10205. (13) Yu, W. H.; Kang, E. T.; Neoh, K. G. Langmuir 2004, 20 (19), 8294–8300. (14) Jin, Y. Z.; Gao, C.; Kroto, H. W.; Maekawa, T. Macromol. Rapid Commun. 2005, 26 (14), 1133–1139. (15) Zheng, G. D.; Stover, H. D. H. Macromolecules 2003, 36 (6), 1808–1814. (16) Singh, N.; Wang, J.; Ulbricht, M.; Wickramasinghe, S. R.; Husson, S. M. J. Membr. Sci. 2008, 309 (12), 64–72. (17) Bhut, B. V.; Wickramasinghe, S. R.; Husson, S. M. J. Membr. Sci. 2008, 325 (1), 176–183. (18) Liu, P. S.; Chen, Q.; Liu, X.; Yuan, B.; Wu, S. S.; Shen, J.; Lin, S. C. Biomacromolecules 2009, 10 (10), 2809–2816. (19) Wei, Y. T.; Zheng, Y. M.; Chen, J. P. Water Res. 2011, 45, 2290– 2296. (20) Wei, Y. T.; Zheng, Y. M.; Chen, J. P. Water Res. 2011, 45, 2297– 2305. 6024
dx.doi.org/10.1021/la200154y |Langmuir 2011, 27, 6018–6025
Langmuir
ARTICLE
(21) Socrates, G. Infrared and Raman Characteristic Group Frequencies: Tables and Charts, 3rd ed.; Wiley: New York, 2001. (22) Liu, P. S.; Chen, Q.; Wu, S. S.; Shen, J.; Lin, S. C. J. Membr. Sci. 2010, 350 (12), 387–394. (23) Liu, S. L.; Zeng, J. A.; Tao, D. D.; Zhang, L. N. Cellulose 2010, 17 (6), 1159–1169. (24) Cao, Y.; Tan, H. M. J. Appl. Polym. Sci. 2006, 102 (1), 920–926. (25) Barringer, J. E.; Messman, J. M.; Banaszek, A. L.; Meyer, H. M.; Kilbey, S. M. Langmuir 2009, 25 (1), 262–268. (26) Arica, M. Y.; Akyol, A. B.; Bayramoglu, G. J. Appl. Polym. Sci. 2008, 107 (5), 2810–2819. (27) Chan, K.; Gleason, K. K. Langmuir 2005, 21 (25), 11773– 11779. (28) Ma, Z. Y.; Guan, Y. P.; Liu, H. Z. J. Polym. Sci., Polym. Chem. 2005, 43 (15), 3433–3439. (29) Pan, K.; Zhang, X. W.; Ren, R. M.; Cao, B. J. Membr. Sci. 2010, 356 (12), 133–137. (30) Pan, K.; Zhang, X. W.; Cao, B. Polym. Int. 2010, 59 (6), 733– 737. (31) Choi, W. W.; Chen, K. Y. Environ. Sci. Technol. 1979, 13 (2), 189–196. (32) Adams, R. M. Boron, Metallo-Boron Compounds, and Boranes; Wiley-Interscience: New York, 1964. (33) Simonnot, M. O.; Castel, C.; Nicolai, M.; Rosin, C.; Sardin, M.; Jauffret, H. Water Res. 2000, 34 (1), 109–116. (34) Su, C. M.; Suarez, D. L. Environ. Sci. Technol. 1995, 29 (2), 302–311. (35) Wang, L.; Qi, T.; Gao, Z.; Zhang, Y.; Chu, J. React. Funct. Polym. 2007, 67 (3), 202–209. (36) Goldberg, S.; Forster, H. S.; Heick, E. L. Soil Sci. Soc. Am. J. 1993, 57 (3), 704–708. (37) McBride, M. B. Clay. Clay Miner. 1997, 45 (4), 598–608. (38) Peak, D.; Luther, G. W.; Sparks, D. L. Geochim. Cosmochim. Acta 2003, 67 (14), 2551–2560.
6025
dx.doi.org/10.1021/la200154y |Langmuir 2011, 27, 6018–6025
