Fabrication of Core-shell Type Alginate/CNT Composite Adsorbent Beads by Combined Method of Magnetic-eld/electrospray: Effect of CNT Orientation on Adsorption of Methylene Blue and Environmental Applications

In the present work, carbon nanotubes (CNT)/alginate composite adsorbent was prepared using combined method of magnetic �eld/electrospray. The effects of magnetic and electric �elds on CNT orientation in the polymeric matrix of adsorbent, and consequently on the methylene blue adsorption were investigated. The adsorbents characteristics and adsorption performance of the six types samples were obtained using eld-emission scanning electron microscopy (FE-SEM) and Ultraviolet-visible (UV-Vis) spectroscopy, respectively. The FE-SEM image observations indicated that under the setup’s operating condition, the magnetic �eld could partially orient the CNTs in the polymer matrix while the electrical �eld had a little effect. The obtained results clearly represented the positive effects of CNT orientation on enhancing the adsorption of methylene blue into the CNT/alginate composite adsorbents. The egg-box model in the CNT/alginate core-shell type beads could explain the adsorption performance of methylene blue solution onto the six types fabricated beads.


Introduction
Industrial development, rapid urbanization, and increasing agricultural products have promoted the need for treating unconventional water sources such as wastewaters.The wastewater treatment, in addition to restoring water to the consumption cycle, also, can prevent environmental pollution.The adsorption processes by particulate solid adsorbents show potential as one of the most powerful ways for the elimination of pollutants in wastewater due to subjecting large adsorption surface area [1][2][3][4].In the recent century, the development of the adsorption technology is owed to nanotechnology and fabrication of the new nanocomposites [5][6][7][8].
Lately, the challenge of using CNTs in nanocomposite materials has been shifted from production, puri cation, and surface modi cation methods to sorting, arrangement, and controlling the orientation of the CNTs in the different substrates [19][20][21][22][23].There are some chemical, electrical, mechanical, and magnetic methods for aligning the CNTs in a polymeric substrate [19,[24][25][26]].In the mechanical methods, the shear forces on CNT-reinforced nanocomposites would control the CNTs position and their arrangement.Despite simplicity, however, the alignment effectiveness of the mechanical methods is much lower than other techniques [26].In the chemical methods, the dense arrays of vertically aligned CNTs, so-called CNT forests, are used.The oriented CNT forest is stabilized using capillarity-induced wetting with unmodi ed complex thermosets or some epoxies [19,27].
The orientation of the CNTs under electrical eld has been reported according to the electrophoresis and dielectrophoresis phenomena.The CNTs, due to the electrical properties such as the quasi-onedimensional and symmetrical tubular structures, can be aligned in parallel structures using a high electrical eld.This method can orient the CNTs in low viscosity polymeric substrates, but, applying the high electric eld magnitude limits the possibility of its large-scale employment [9,[28][29][30][31][32].
Calculations and some experimental works found that the CNTs were paramagnetic in the direction of their long axes, and tend to align parallel to the ambient magnetic eld [28,33,34].The magnetic susceptibility of the CNTs is highly dependent on the magnetic eld direction and temperature.For a successful orientation, the CNTs must prevail on the thermal energy of Brownian motion and the rotation resistance in the viscous polymer solution [35,36].Although there are a number of investigations reporting the CNT orientation under the magnetic eld, nevertheless, many questions have arisen about the capability of this method in various materials and applications [30,33,[35][36][37].Therefore, more detailed investigations must be carried out regarding the CNTs' alignment in a polymer matrix under the magnetic eld and their applications such as in the adsorption of heavy metals or organic pollutants from wastewaters.It seems that in case of synthesizing aligned CNTs dispersed in a polymer matrix, this method might be feasible as an energy e cient and cheap method.
In this study, a novel setup of magnetic eld/electrospray was employed to orient the CNTs in the porous calcium alginate-CNTs (CA-CNTs) composite adsorbent beads.To the best knowledge of authors there is no study focusing on combining magnetic and electric elds with aims to align CNTs dispersed in the polymeric adsorbent beads, as well as applying them to eliminate aqueous organic dyes.To investigate the adsorption performance of the prepared absorbents, the sodium alginate-CNTs droplets were electrosprayed in calcium chloride gelling agent to obtain CA-CNTs core-shell beads, while the magnetic eld on sol ow was applied before electrospray.The composite adsorbents were then characterized by Field Emission Scanning Electron Microscope (FESEM), Fourier Transform Infrared spectroscopy (FTIR), and Ultraviolet and Visible spectroscopy (UV-vis).The adsorption studies was done to nd out the effects of magnetic and electrical elds on the CNT orientation and, subsequently, methylene blue (MB) removal.

Experimental
Materials Sodium-alginate (SA) was provided by Sigma-Aldrich Co. (USA).The multi-walled carbon nanotubes (CNT-95% purity) with external diameters between 20 to 45 nm and lengths between 20 to 30 µm were supplied by US Research Nanomaterials Inc. (USA).Figure 1 depicts a FE-SEM image of the CNTs.
Three types of CNTs containing hydroxyl (-OH), carboxyl (-COOH), and amine (-NH 2 ) functional groups were used.FTIR spectra of the CNT powders containing the mentioned functional groups are shown in Fig. 2. As reported in the literature, vibration of hydroxyl groups from intermolecular hydrogen-bonded OH:OH or unbound or free hydroxyl of phenol appears around 3200-3600 cm − 1 .Furthermore, the absorbance bands at 1640 − 1630 cm − 1 are related to the bending or deformation of the hydroxyl groups [38][39][40].In Fig. 2b, the absorbance peaks around 1700-1800 cm − 1 , and the peak at 1556 cm − 1 resulted from C = O and C-O stretching, respectively.These two absorbance bands and a broadband peak at 3400 cm − 1 are attributed to the -COOH groups on the external surface of CNTs [37,38,41].For CNT-NH 2 (Fig. 2c), the in-plane bending vibration of N-H appears at 1799 cm − 1 .The peaks at 1115 cm − 1 and 3408 cm − 1 were ascribed to the stretching vibration of C-N and the stretching vibration of N-H, respectively [21,41,42].
Methylene blue (MB), as a cationic dye, was selected as adsorbate and provided by Merck Co. (Germany)

Preparation of the SA-CNT suspension
To obtain a most stable SA-CNT suspension, 0.5 wt.% CNT (calculation according to the weight of the SA polymer), containing each of functional group, was initially dispersed in the doubled distilled water.After sonicating for 2 h (34 kHz, Elmasonic P 30 H, Germany), the samples were then centrifuged under 6000 rpm for 45 min (Hettich® EBA 20, Germany) to separate the possible aggregated CNTs.The supernatant liquid was sonicated again for 20 min to obtain partially stable CNT suspension, and 1.5 wt.% SA solution was then added to the latter suspension.The resultant colloidal solution was stirred for 5 h using a magnetic stirrer (Hei-Standard-Germany).At last the sedimentation rate test was conducted on the suspensions for a period time of 16 h to characterize the optimum stable sample.

Magnetic eld/electro-spray combined method
Figure 3 shows the picture and a schematic drawing from the lab scale setup of magnetic-eld/electrospray used in this work.The setup includes two hydraulic syringes, a microinfusion pump (WZ-60C2, China), and a calcium chloride solution collector container.Brie y, the SA-CNT solution ow was conducted to a nozzle spinneret (18G) using the microinfusion pump at a ow rate of 50 mL/h.In the tests, a 0-25 kV DC power supply (LD Didactic GmbH, Germany) was used, where its positive charge was connected to the solution's spinneret tip and the collector container was earthed.The distance between spinneret tip and the collector top surface was adjusted to 7.5 cm, and the DC voltage was keep constant at 18 kV to obtain 2.4 V/cm electric eld intensity.The SA-CNT was sprayed into the calcium chloride gelling solution and mixed for 10 min to encapsulate the SA-CNT by a solid layer of the CA-CNT.These sample beads were labeled as CA-CNT ES .
To fabricate the gel type calcium alginate beads by the magnetic-mechanical spray (CA-CNT M/MS ) or magnetic-electrospray (CA-CNT M/ES ) methods, two magnetic plates of Neodymium (N42 grade with 20mm*40mm*50mm dimensions) were used.In this regard, before the spraying process, the SA-CNT solution was subjected to the uniform magnetic eld of strength 318 mT for 2 h.
The beads labeled as CA-CNT MS , were obtained by dripping the SA-CNT solutions under no magnetic and electric elds into gelling agent of calcium chloride.These beads were prepared to compare the effect of mechanical spraying (MS) with the solely electrospray (ES) or combination of magnetic eld (M) and electrospray methods on their adsorption performance.
All the formed CA beads were thoroughly washed with doubled distilled water to make them chloride free.
Finally, CA beads were stored under water for further use.
To study the effect of the CNT presence in the adsorbent bead on the methylene blue (MB) adsorption, two control samples of the CA MS and the CA ES were also fabricated without CNT by the mechanical and the electrospray methods, respectively.Some beads were sieved from gelling solution and were frozen in liquid nitrogen for the freeze-drying and then microscopic analysis.

Characterization methods
The surface functional groups of CNTs were determined by a PerkinElmer FTIR spectrophotometer between 400 and 4000 cm − 1 wavenumbers.The dispersion of CNTs in each bead was evaluated using an optical microscope (OLYMPUS) and a eld-emission scanning electron microscopy (FESEM, MIRA3, TESCAN).To prepare the FE-SEM samples and obtain images from the surface or cross-section of the beads, the adsorbents were initially freeze-dried (Dorsatech, Iran) and some of them were then broken.To characterize the adsorption e ciency of MB onto SA-CNT composite, the concentration of MB was characterized in the bulk of solution by the T80/T80 + UV-Vis spectrophotometer (PG Instruments Ltd.).
In the adsorption experiments, the MB solution (50 mg/lit) was prepared and then added to a gel type adsorbent beads at a 3:1 volume ratio.The contact times were set as 0, 5, 30, 60, 120, 240, 480, 960 min and the adsorption of the MB was reported.
One-way analysis of variance (ANOVA) with a Tukey Post hoc test was used for statistical analysis of the MB adsorption onto the fabricated absorbents.For p-value < 0.05 the differences were considered signi cant.

Results And Discussion
Dispersion and distribution of the CNTs in the polymer solution Figure 4 shows the results of the sedimentation rate test for the prepared SA-CNT samples.Comparison of the images clearly show that for the suspension containing CNT-NH 2 , the stability conditions are disturbed after 4 h, and the CNTs are precipitated as large clusters.This phenomenon occurred for the CNT-OH and CNT-COOH samples after 16 h and 8 h, respectively.These observations imply that the CNT-COOH and CNT-OH samples present relatively more stability than the CNT-NH 2 sample in aqueous suspension.The similar results have been reported in other works, but in different settling rate for CNTs functionalized with carboxyl and hydroxyl groups [40,43,44].
Figure 5 shows the CA-CNT adsorbents produced by the stable and unstable suspension of CNT.As can be seen in the gure, the dark spots (white arrows) indicate the presence of CNT clusters demonstrating the poor dispersion of the CNT in the adsorbent matrix.
The size of beads produced by electrical and mechanical spraying Figure 6 depicts the beads' size distributions of the CA MS and CA ES adsorbents, which are produced in similar conditions.As shown, the bead diameter spans for the samples produced by electrical and mechanical spraying are 180-580 µm and 1900-2140 µm, respectively.This indicates that the size mode of beads produced by electrospray is smaller as compared to those beads produced by mechanical spray.These results are in agreement with other works representing the effect of electric eld on the size reduction of sprayed droplets [45][46][47][48].

FE-SEM examinations
As mentioned before, the freeze-drying route was utilized to evaluate the CNTs' distribution in the matrix of each adsorbent.Figure 7 illustrates the bead images before and after the freeze-drying.Besides, the SEM micrographs from the outer surface of the CA-CNT MS and CA-CNT ES beads are also shown in this gure.As can be seen from SEM images, the electrospray and freeze-drying of CA-CNT ES has led to more shrinkage with wrinkled surface, and smaller size of the beads as compared to CA-CNT MS .
Figure 8 shows the SEM images from the outer surfaces of the CA-CNT MS , CA-CNT ES , CA-CNT M/MS and CA-CNT M/ES freeze-dried beads.The preliminary observation of the surfaces morphology indicates that the surface grooves are directional especially for CA-CNT ES (Figs. 8c and 8d) as compared to CA-CNT MS (Fig. 8a).Furthermore, it seems that the traces of CNTs are seen on the surface of CA-CNT MS sample in which the nano-tubes are most probable to be uniformly distributed in the alginate structure (Fig. 8a).For others there is no sign of CNTs on the surface, and therefore, those might be embedded more inside the absorbent beads.In general, it is obvious from the surface morphology of beads that the magnetic and electric elds have affected the structural characteristics of the produced beads.
To investigate the effect of the electrical and the magnetic elds on the arrangement and the orientation of the CNTs in the polymeric matrix, SEM images were taken from the fractured surfaces of the CA-CNT ES , and CA-CNT M/ES samples.As shown in Figs.9a and 9b it can see that the CNTs have been distributed well in the polymer matrix.
For the CA-CNT ES sample, it looks that the CNTs have no speci ed orientation, and therefore, the electric eld has not been able to orientate the CNTs (Fig. 9a).This would be attributed to insu cient force applied on CNTs for their alignment within the viscous calcium alginate suspension as well as a very short retention time of suspension ow under the electrospray eld.For the CA-CNT M/ES sample, however, the CNTs have been partially oriented demonstrating the ability of the magnetic eld to orient the CNTs (Fig. 9b) under the speci c processing conditions.The similar result for the magnetic eld was also reported in the work of Shi et al. [49].

MB adsorption onto the beads
As shown in Fig. 10, MB adsorption onto the six fabricated adsorbent beads has been evaluated at different time steps.As seen, the adsorption of the MB is generally increased with the CA-CNTs beads, as compared to the bare CA sample.It is interesting to note that for the samples free from CNTs (i.e.CA MS and CA ES ), the MB adsorption is almost same at the early stages, while after about 60 min, it becomes higher for the CA MS sample as compared to CA ES one.In the presence of CNTs the adsorption trends get closer between CA-CNT MS and CA-CNT ES .However, as the effects of magnetic and electric elds, the MB elimination rate trend for CA-CNT M/ES beads is much better, and it achieves the highest adsorption e ciency among ve other samples.
SA droplets can acquire the shell and core gel type characteristics when carboxylate groups of the alginate are cross-linked by divalent cations such a calcium.This called the "egg-box" model for the gelation mechanism.The name "egg-box" is used because one divalent cation Ca + 2 interacts with four -COOH groups [50,51].The latter phenomenon, so-called external gelation, may lead to formation of mechanically stable gel network of CA shell around the SA solution core [50][51][52][53][54]. Figure 11 illustrates schematically the egg-box model and the cross sectional views of a produced SA-CA core-shell beads.
The MB species in the bulk of aqueous solution, surrounding the beads, are initially adsorbed onto the CA shell, followed by their molecular diffusions through the porous shell and then being transferred into SA core solution [55].The same mechanism would also prevail for the CA-CNTs beads.This mechanism has been well detailed in the drug release from calcium alginate beads [54,56,57].
As depicted in Fig. 10, the adsorption of MB for all embedded CNT samples initially exhibited a fast uptake.Wang et al, reported that the uptake process of MB for MWCNT is very fast and then showed a steady trend after the equilibration point [58].It can be concluded that at the early stages of the adsorption the physical interaction of the beads' surface with MB prevails, and then the migration process would proceed through shell intward the core [55].
To compare more precisely the combined effect of mechanical, magnetically, and electrical spraying, the adsorption trend of the MB species as a function of contact time for the rst hour are shown in Figs. 12 and 13, respectively.As seen in Fig. 12, the presence of CNTs and their orientation by magnetic eld (M) enhances the adsorption rate in the absence of electric eld and under mechanical spray (MS), where the beads are larger.The adsorption e ciency of these three samples have been calculated in 60 min, so that the CA-CNT M/MS sample adsorbs about 30% of the MB.However, this value is about 23% for CA-CNT MS and 19% for CA MS .It is noted that absorbance values of the samples are signi cant too (p-value < 0.05).
For the samples that electrospray route is employed (Fig. 13), there is a signi cant difference between the absorption rates of the CA-CNT ES and CA-CNT M/ES samples, as compared to the neat ones (p-value < 0.05).In fact, as an effect of electric eld in producing smaller beads, the adsorption mechanism were mainly controlled by the diffusive transport of the MB compounds within the pore network of the nanocomposites and the CNT's sites [16].Calculating the adsorption e ciency of the MB onto the CA-CN ES and CA-CNT M/ES beads indicates that in the rst hour, about 65% of the MB is adsorbed by the CA-CNT M/ES bead, and it is approximately 22% and 12% for the CA-CNT ES and neat beads, respectively.It must be noted that in 60 min, the absorbance values of the CA-CNT ES and CA-CNT M samples are not signi cant (p-value < 0.05).
After 1 hour, the slopes of the adsorption are decreased in all samples.The CA-CNT M/ES bead could remove about 88% of the MB concentration after 16h, and this value is about 67% for the same sample fabricated by the mechanical spraying.In this study, the equilibrium time for the CA-CNTs absorbent is longer than has been reported in other works [58][59][60].This could be due to differences in pore tortuosity of the nanocomposites.These observations con rm that the carbon nanotubes increase the e ciency of the MB adsorption, and their orientation in a polymeric structure can enhance the capability of the adsorption.The similar results have been reported in other experimental works [61,62].

Conclusions
In the present work, fabrication of CA-CNT composite adsorbents was investigated with an emphasis on the orientation of CNTs in a polymer matrix.In this regard, two mechanical and electrical spraying methods were used.To orient the carbon nanotubes, before the spraying process, the mixture containing the CNT was subjected to the magnetic eld.To achieve the high stability and uniformity of the CNT dispersion in the polymer matrix, the CNT with three functional groups (carboxyl, hydroxyl, and amine) was used.The results showed that CNT-COOH was more stable in sodium alginate and uniformly dispersed throughout the matrix when forming adsorbent beads.The morphological studies of the adsorbents showed that although the electric eld reduced the size of the adsorbent beads, compared to the adsorbents made by the mechanical spraying, it had no signi cant effect on the orientation of CNTs.
However, the FE-SEM results indicated that the magnetic eld could partially align the nanotubes in the polymer matrix.Comparing the adsorption of the MB by the fabricated adsorbents demonstrated that the presence of the CNTs in the adsorbent bead improved the adsorption e ciency.Furthermore, the oriented CNTs increased the MB adsorption, so that the MB adsorption e ciency for the samples made by the electrospray/magnetic eld was more than 85%.However, the highest adsorption for the same sample containing random CNT was about 67%.The results revealed that the orientation of the CNTs without increasing their concentration could improve the effectiveness of these nanoparticles, and led to fabricating the cheaper products with higher performance.

Figure 11 The
Figure 11