Biomass derived palygorskite–carbon nanocomposites: Synthesis, characterisation and affinity to dye compounds

Applied Clay Science 114 (2015) 617–626

Contents lists available at ScienceDirect

Applied Clay Science journal homepage: www.elsevier.com/locate/clay

Biomass derived palygorskite–carbon nanocomposites: Synthesis, characterisation and affinity to dye compounds Binoy Sarkar a,b,⁎, Erming Liu a,b, Stuart McClure a, Jayaraman Sundaramurthy c, Madapusi Srinivasan c, Ravi Naidu a,b,d,⁎⁎ a

Centre for Environmental Risk Assessment and Remediation, University of South Australia, Mawson Lakes, SA 5095, Australia CRC for Contamination Assessment and Remediation of the Environment, P.O. Box 486, Salisbury, SA 5106, Australia Department of Chemical and Biomolecular Engineering, Faculty of Engineering, National University of Singapore, Blk E5, Engineering Drive 4, Singapore 117576, Singapore d Faculty of Science and Information Technology, University of Newcastle, Callaghan, NSW 2308, Australia b c

a r t i c l e

i n f o

Article history: Received 3 May 2015 Received in revised form 17 June 2015 Accepted 1 July 2015 Available online 16 July 2015 Keywords: Nanocomposite Palygorskite Graphitised carbon Mesopores Adsorption Pore diffusion

a b s t r a c t Clay minerals can act as a uniform dispersion medium for nano-sized carbon particles. However, literature on the preparation and characteristics of palygorskite–carbon nanocomposites is scant. Using a hydrothermal carbonisation technique this study developed two nanocomposites on fibrous palygorskite from starch: the first without a post-synthesis treatment (Composite 1); and the second with an activation at 550 °C for 3 h (ramp at 10 °C min−1) under CO2 environment (200 mL min−1) (Composite 2). A uniform dispersion of nanoscale carbon spheres was formed on partially destroyed palygorskite structures. Composite 2, which indicated the formation of graphitised carbon nanoparticles, generated a 17-fold greater specific surface area than Composite 1 and also created micro- and mesopores in its structure. The nanocomposites, especially in Composite 1, contained organic surface functional groups (C\\H, C_C, C_O) and indicated variable affinity to cationic and anionic dye compounds. While Composite 2 adsorbed a larger amount of anionic orange II dye (23 mg g−1), Composite 1 adsorbed more cationic methylene blue (46.3 mg g−1). Isothermal and kinetic modelling of the adsorption data indicated that in addition to electrostatic attraction for methylene blue adsorption on both nanocomposites, a pore diffusion mechanism was involved and the boundary resistance was greater for orange II than methylene blue adsorption. Being a material developed from green biomass (starch) and an abundant natural resource (palygorskite), these nanocomposites have immense potential for application in environmental remediation including in situ immobilisation of contaminants in soil. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Clay minerals have wide potential applications for treating environmental contaminants. Due to their unique properties such as layered structures, large surface areas, high cation exchange capacity (CEC) and enormous environmental stability, clay minerals are a popular choice for cleaning diverse organic and inorganic pollutants in the environment (Churchman et al., 2006; Sarkar et al., 2012b). Worldwide availability and hence low cost also make these minerals appropriate for environmental applications. Both naturally occurring clay minerals and their modified products can be used for treating contaminated water, soils and air. Modification of clay minerals by physical, chemical

⁎ Correspondence to: B. Sarkar, Centre for Environmental Risk Assessment and Remediation, University of South Australia, Mawson Lakes, SA 5095, Australia. ⁎⁎ Correspondence to: R. Naidu, Faculty of Science and Information Technology, University of Newcastle, Callaghan, NSW 2308, Australia. E-mail addresses: [email protected] (B. Sarkar), [email protected] (R. Naidu).

http://dx.doi.org/10.1016/j.clay.2015.07.001 0169-1317/© 2015 Elsevier B.V. All rights reserved.

or biological processes improves their efficiency in removing contaminants many-fold (Churchman et al., 2006). Among the various clay modifications, anchoring of carbonaceous materials, which are abundant in oxygen-containing groups and exist on the surface of solid adsorbents such as clay minerals, has emerged as a new class of adsorbent for contaminant removal (Anadão et al., 2011, 2014; Chen et al., 2011; Kumar et al., 2011; Ai and Li, 2013). Clay minerals or zeolite can act as a uniform dispersion medium for nano-sized carbon particles and thus ideally help improve the composite's reactivity to target pollutant compounds (Katsuki et al., 2005; Anadão et al., 2011; Wu et al., 2014). Uniform dispersion of carbon nanoparticles onto a suitable support is required in order to apply the composite under in situ conditions. This can refer, for example, to the remediation of contaminated ground water because such a composite largely prevents: (a) nanoparticle agglomeration; and (b) unwanted displacement from the target site. Other high end applications of carbon–clay mineral materials include catalysis, biocatalyst support, adsorbents for chromatography and haemosorbents in medicine (Anadão et al., 2011).

618

B. Sarkar et al. / Applied Clay Science 114 (2015) 617–626

Clay–carbon composite material can be synthesised by pyrolysis of a carbon precursor hosted on the clay surface or intercalated in the clay layers. However, this process involves large energy consumption and extreme heat that can destroy clay minerals' crystalline structure. Also, increased awareness about sustainable and green nanotechnology has made the development of clay–carbon nanocomposite from renewable green precursors critical. The hydrothermal carbonisation (HTC) process has recently proved to be effective for clay–carbon nanocomposite synthesis using biomass carbon precursor (Hu et al., 2010; Chen et al., 2011; Li et al., 2012; Zhang et al., 2014). This is considered to be a greener approach because it consumes less energy (carbon deposits at low temperature) and uses renewable carbon precursors such as glucose, sucrose and cellulose. Since carbon nanoparticle deposition takes place at a comparatively lower temperature in the HTC process compared to carbon nanotube synthesis in the high-temperature chemical vapour deposition (CVD) technique, the former provides numerous oxygen-containing groups to the nanocomposite surfaces (Qian et al., 2006; Li et al., 2012, 2014). Consequently it would be a good adsorbent for water purification (Xu et al., 2008). The properties of clay–carbon nanocomposites would depend on both the clay mineral characteristics and carbon precursor type. Since palygorskite is a non-swellable type clay mineral, its carbon nanocomposite may impart greater permeability than a swellable montmorillonite composite for removing contaminants under flow conditions when used in a barrier (Sarkar et al., 2010a, 2012a). However, information on the preparation and characteristics of palygorskite–carbon nanocomposites is scarce in the literature (Chen et al., 2011; Wu et al., 2011; Liu et al., 2013). Therefore, this study aims to: firstly, develop nanocomposites from starch using fibrous palygorskite; and secondly, characterise the products and study their affinity to organic dye compounds (orange II and methylene blue) in aqueous solutions. Removing dye compounds from contaminated water is important because many of them are toxic, carcinogenic and can damage the environment and human health (Sarkar et al., 2011; Ai and Li, 2013). 2. Experimental section 2.1. Materials and reagents A palygorskite (Grade 050F) originally from Western Australia was supplied by Hudson Resources Limited and was used as received, without further purification. The CEC of this clay mineral was 17.0 cmol (p+) kg−1 and it had a specific surface area of 89 m2 g−1. Sodium exchanged palygorskite was prepared by dispersing 20 g clay in 1 L of 1 N NaCl (Sigma-Aldrich) solution. The mixture was agitated on a magnetic stirrer at room temperature for 48 h, following which the sediment was separated through centrifugation at 15,000 rpm for 10 min. The sediment was washed 4 times with deionised water, dried at 60 °C, ground into powder and stored in an air tight container. Soluble starch was used as the renewable carbon precursor and zinc chloride (ZnCl2) served as the catalyst for carbon deposition (both supplied by Merck Pte Ltd, Singapore). The dye compounds: orange II (4-(2-hydroxy-1naphthylazo) benzenesulfonic acid sodium, C16H11N2NaO4S, MW: 350.33), and methylene blue (3,7-bis(dimethylamino)phenazathionium chloride, C16H18ClN3S·3H2O, MW: 373.90), were supplied by SigmaAldrich, Australia.

to a Teflon lined hydrothermal bomb and sealed properly. It was kept in a pre-heated oven at 250 °C for 3 h. Following cooling, the material was heat activated under two conditions: (1) heated at 105 °C in an oven for 3 h in presence of air (Composite 1); and (2) heated at 550 °C for 3 h (ramp at 10 °C min− 1) under CO2 environment (200 mL min−1) (Composite 2). 2.3. Materials characterisation 2.3.1. Surface area and pore size distribution N2 adsorption–desorption experiments were performed at liquid nitrogen temperature (−196 °C) on a Micromeritics Gemini 2380 surface analyser. Samples were outgassed at 70 °C for 12 h under high vacuum. Specific surface area was determined with the Brunauer–Emmett–Teller (BET) method and microporous volume was evaluated using the t-plot method. Average pore diameter was estimated via the Barrett– Joyner–Halenda (BJH) method. 2.3.2. X-ray diffraction (XRD) Powdered nanocomposite samples were pressed in stainless steel sample holders for XRD analysis. XRD patterns of the nanocomposites were acquired using CuKα radiation (λ = 1.5418 Å) on a PANalytical, Empyrean X-ray diffractometer operating at 40 kV and 40 mA between 5 and 80° (2θ) at a step size of 0.016° (100 s scan step time) with a fixed 0.5° divergence slit and 1° anti-scatter slit. The basal spacing was calculated from the 2θ value using Bragg's equation, nλ = 2d sinθ. 2.3.3. Scanning electron microscopy (SEM) Nanocomposite samples were fixed on a double-sided tape, and coated with 20 nm of carbon or 5 nm platinum by evaporation using a Quorum QT150ES coating system. Samples were examined using a FEI Quanta 450 FEG Environmental Scanning Electron Microscope equipped with an EDAX Apollo EDX detector. Images were taken in high vacuum mode and with a 20 kV accelerating voltage using an Everhart–Thornley Detector (ETD) and a solid state Back Scattered Electron detector (BSED). EDAX spectra, X-ray maps and line-scans were acquired for selected areas on the samples. EDAX maps were acquired overnight; line scans were derived from the EDAX maps. EDAX point spectra were acquired for 100 s at a point. 2.3.4. Transmission electron microscopy (TEM) Nanocomposite samples were mounted on copper grid and observed under a JEOL JEM – 2100F Transmission Electron Microscope with a 200 kV accelerating voltage. 2.3.5. Particle size The palygorskite and nanocomposites were suspended in 0.1 M NaCl solution, agitated for 24 h on an end-over-end shaker and ultrasonicated for 30 min. Size of the uniformly dispersed particles was determined by using a Nicomp™ 380 DLS Particle Sizer.

2.2. Preparation

2.3.6. Fourier transformed infrared spectroscopy (FTIR) The nanocomposite samples (about 1% w/w) were ground and mixed homogenously with dehydrated KBr and pressed into discs for FTIR analysis. Infrared (IR) spectra were obtained using an Agilent Cary 600 Series FTIR Spectrometer. Spectra over the 4000–400 cm−1 range were obtained by the co-addition of 64 scans with a resolution of 4 cm−1 and a mirror velocity of 0.6329 cm s−1.

Carbon nanocomposites were synthesised using the HTC method (Chen et al., 2011). In a typical procedure, 15 g starch was dissolved in 200 mL suspension containing 3 g sodium–palygorskite in deionised water at 50 °C by stirring it magnetically for 3 h. A gel was formed and separated by filtration through Whatman No. 1 filter paper. Zinc chloride (zinc equivalent to 400% CEC of the palygorskite) was added to the gel and mixed homogeneously. The mixture was then transferred

2.3.7. Raman spectroscopy The nanocomposites were incubated statically on a silicon wafer. Raman spectra were collected by a WITec Confocal Raman Microscope (Alpha 300RS, Germany) equipped with a 532 nm laser diode (b60 mW). A CCD detector (cooled to approximately – 60 °C) was used to collect Stokes Raman signals under a 100 × objective (Nikon) at room temperature over the wavenumber range of 0–2500 cm−1

B. Sarkar et al. / Applied Clay Science 114 (2015) 617–626

with an integration time of 1 s for measurement of single spectra. The spectral resolution was 1 cm−1. The 520 cm−1 line of a silicon wafer was used for calibrating the Raman spectra.

619

(mg L−1), V is the volume of liquid phase (mL), and M is the mass of nanocomposite (g). 2.5. Analysis of dyes

2.3.8. Surface charge Zeta potential values of a 0.05% suspension (w/v) of the palygorskite and nanocomposites in Milli-Q water were determined by a zeta potential analyser (Nicomp™ 380 ZLS, USA).

The initial and equilibrium dye concentrations were measured on an Agilent 8453 UV–Vis Spectrophotometer. Orange II and methylene blue were analysed against deionised water blank using a quartz cuvette at 486 nm and 665 nm wavelengths, respectively.

2.4. Adsorption of dyes 3. Results and discussion Adsorption of two organic dyes – methylene blue (cationic dye) and orange II (anionic dye) onto the nanocomposites – was evaluated in batch experiments with three replications. In a typical procedure for studying the adsorption isotherm, a portion of 0.05 g nanocomposite was put into a 50 mL polypropylene centrifuge tube to which 10 mL of dye solutions having concentrations up to 500 mg L−1 was added. The mixture was equilibrated on an end-over-end shaker for 24 h in the dark at room temperature. Preliminary experiments confirmed that adsorption equilibrium was achieved within 2 h of agitation. Following equilibration the suspension was centrifuged at 5000 rpm for 15 min (Multifuge 3S-R, Hevaeus). The clear supernatant was analysed for concentrations of respective dye compounds. Adsorption kinetics was studied by equilibrating 0.25 g nanocomposite with 50 mL of 50 mg L−1 respective dye solutions. The experiments were conducted in a 100 mL glass beaker in the dark at room temperature using a magnetic stirrer. A portion of 2 mL suspension was collected at specified time intervals (up to 2 h), filtered through a 0.45 μm nylon filter and analysed for respective dye concentrations. The amount of dyes removed from the aqueous solution was calculated using the following equation: qm ¼ V ðCi −Ce Þ=ðM x1000Þ

ð1Þ

where, qm is the amount of dye removed from the liquid phase (mg g−1), Ci is the initial liquid phase concentration of the dye (mg L−1), Ce is the equilibrium liquid phase concentration of the dye

3.1. Characterisation of nanocomposites 3.1.1. Surface area and pore size distribution Nitrogen sorption isotherm of pristine palygorskite (Fig. 1) showed representative type IV curves with H3 hysteresis loops, meaning the sample had typical slit-type mesopores generated by the interparticle porosity of plate or fibre-like morphology. A steep step occurred at approximately P/P0 = 0.80–0.98 owing to capillary condensation. The pore-size distribution curves calculated from this sample's adsorption branch confirmed a broad pore size distribution from 1 nm to above 100 nm (Fig. 1). Since palygorskite possesses a fibrous morphology, this broad pore size distribution could be attributed to the interparticle voids of the randomly stacked fibrous clay (Liu et al., 2012a). The BET surface area of pristine palygorskite and palygorskite heated at 550 °C were 89 and 98.2 m2 g−1, respectively. For Composite 1 (Fig. 1), the nitrogen uptake decreased over the whole pressure range. The hysteresis loop (close to H3 type) stayed at 0.8–1.0 of relative pressure and indicated a similar slope ratio to that of pristine palygorskite. The poresize distribution curve of Composite 1 showed the distribution from 10 nm to above 100 nm still remaining, which obviously originated from the randomly stacked clay fibres. The surface area of Composite 1 was 15 m2 g−1. On the other hand, the isotherm of Composite 2 (Fig. 1) took on a shape resembling a combination of Type I and IV adsorption. Type I isotherm is given by microporous solid and therefore the composite

Fig. 1. Nitrogen sorption isotherms and pore size distributions for (a) pristine palygorskite (b) Composite 1, (c) Composite 2, and palygorskite heated at 550 °C.

620

B. Sarkar et al. / Applied Clay Science 114 (2015) 617–626

14400 Na-Paly@550 deg C

Intensity (counts)

10000

6400 Composite 2

3600 Composite 1

1600 Na-Paly

400 10

20

30

40

50 2Theta (°)

60

70

80

Fig. 2. X-ray diffraction patterns of Na–palygorskite, Na–palygorskite heated at 550 °C and palygorskite–carbon nanocomposites.

material will contain both micro and mesopores. The pore-size distribution curve of this sample contained two parts: one ranged from 1 nm to 10 nm while the other varied from 10 nm to 100 nm size. As carbon is a typical micro-porous material, the first form could be attributed to the carbon material incorporated on the clay particles. This feature was absent in the palygorskite sample heated at 550 °C without carbon precursor (Fig. 1). Moreover, the pore size distribution of Composite 2 almost coincided with that for Composite 1, whereas that of heat treated (550 °C) palygorskite was similar to the pristine palygorskite. This indicates that the large pores would remain after calcination at high temperature. Compared to Composite 1, Composite 2 had a larger surface area (259 m2 g−1). The clogging of pores by carbon precursor (starch) might have caused the suppression of mesopores in the composites relative to those found in the raw palygorskite. However, mesopores were partially recovered in Composite 2 as a result of the heat treatment and structural changes in palygorskite as well as carbon nanoparticles. 3.1.2. XRD patterns The 110 plane of Na-palygorskite reflected at 2θ = 8.27° (Fig. 2). It corresponded to a d-value of 1.07 nm. As indicated by the reflection at 2θ = 12.38° (0.71 nm) (Fig. 2), the sample contained hydrated oxides of sodium and magnesium between the layers (Araújo Melo et al., 2002). Stronger reflections also appeared at 2θ = 20.92° (0.42 nm), 2θ = 19.96° (0.44 nm) and 2θ = 23.99° (0.37 nm) (Fig. 2), which represented the 040, 121 and 221 planes of palygorskite, respectively.

After carbon deposition through hydrothermal treatment and heat activation, the 110 plane reflection partially disappeared in Composite 1 and completely disappeared in Composite 2. However, the reflections of 040, 121 and 221 planes of palygorskite remained unaltered, indicating that the palygorskite structure was preserved to some extent after carbon deposition in both composites. The hydrated sodium and magnesium oxides remained unaffected in Composite 1, but largely disappeared in Composite 2 because the latter was activated at higher temperature. The most intense reflection in the XRD pattern at 2θ = 26.69° (0.33 nm) (Fig. 2) confirmed the presence of quartz impurity in palygorskite (Önal and SarIkaya, 2009) and it remained unaffected by carbon deposition. Other impurities such as dolomite (2θ = 30.93°; 0.29 nm) and halite (2θ = 31.70°; 0.28 nm) disappeared in the composites due to heat treatment because these reflections were also absent in the heat treated sample (550 °C) (Fig. 2). The reflection due to halite (2θ = 31.76°) disappeared in both the composites, while dolomite peak was affected at high temperature activation only (Fig. 2). The XRD patterns indicated no obvious difference among the palygorskite and composites except elimination of 110 palygorskite plane, halite and dolomite peaks, demonstrating that the crystalline structure of palygorskite does not completely disappear in the nanocomposites and it plays the role of a template (Chen et al., 2011). No pattern of crystalline carbon (which may appear at 2θ = 43°) was clearly observed in the composites because: (a) they might be obscured by strong palygorskite peaks; and (b) the degree of crystallinity was poor

Fig. 3. SEM images of palygorskite–carbon nanocomposites; (a) Composite 1, and (b) Composite 2.

B. Sarkar et al. / Applied Clay Science 114 (2015) 617–626

Fig. 4. Frequency distribution of pore sizes derived from SEM images of palygorskite–carbon nanocomposites.

(Cui et al., 2006; Hu et al., 2008; Wu et al., 2011). However, spherical crystalline carbon particles were observed in the SEM and TEM images of the nanocomposites. 3.1.3. Morphological properties The SEM images of Composite 1 and Composite 2 showed uniform distribution of carbon particles on the palygorskite surfaces (Fig. 3). The frequency distribution of pore sizes (5–100 nm) was obtained by processing the SEM images with Fiji open source image processing program (Fig. 4). A pore size distribution having a lesser small-sized (b10 nm) and middle-sized (15–100 nm) pores was observed in Composite 1. Composite 2 on the other hand revealed a pore size distribution having a consistently higher frequency of small-sized (b 10 nm) and middle-sized (50–100 nm) pores relative to Composite 1. These results are in agreement with the pore size distribution results obtained in N2 sorption studies. Thus, SEM analysis further confirmed the formation of micro and mesopores in the composites. The TEM images further supplement the SEM results (Fig. 5). Similar to the crystallinity patterns obtained in the XRD study, SEM and TEM analyses further confirmed that Composite 1 retains the crystalline structure of palygorskite better than Composite 2. In dynamic light scattering (DLS) measurement, Composite 2 exhibited a narrower particle size distribution having average particle diameter around 170 nm, whereas Composite 1 exhibited a broader particle size distribution with greater average particle diameter (Table 1). 3.1.4. Elemental analysis and mapping The aim of hydrothermal process was to functionalise the raw palygorskite with carbon-containing moieties. An EDAX elemental analysis of the pristine palygorskite and its hydrothermally treated product

621

in the presence of starch clearly exhibited deposition of carbon in the clay mineral structure (Fig. 6). The initial small amount of carbon appearing in the palygorskite was contributed mainly by the presence of dolomite admixture as evidenced in the XRD study. Following subtracting the initial carbonate carbon content, the hydrothermal product appeared to contain 68.4 ± 6.7% carbon. Since the palygorskite was exchanged with sodium (Na) before the hydrothermal treatment, the product showed an abundance of Na instead of calcium (Ca) (Fig. 6). The appearance of platinum (Pt) in the hydrothermal product was due to coating of the sample with Pt prior to SEM-EDAX analysis (Fig. 6). The product also contained Zn in its structure (Fig. 6). Zn was contributed in the product as a result of ZnCl2 addition in the hydrothermal reaction mixture. The EDAX mapping of key elements in Composite 2 proved the presence of carbon (C), oxygen (O), aluminium (Al), silicon (Si) and zinc (Zn) (Fig. 7). The structural elements of palygorskite (Si and Al) were present, but their proportions were not maintained as should they normally be in a 2:1 type palygorskite structure because of the high temperature treatment during the composite's preparation. However, the line scan study revealed the key feature of the composite in which the carbonaceous particles were closely associated with the modified palygorskite structure. In the selected 5 μm length of the line scan map, the carbon particles were present with various degrees of oxygenation also indicating the formation of spherical particles of comparatively pure carbon in the composite. As discussed later in this paper, these carbon particles provided Raman spectra similar to a graphitised carbon structure (Fig. 9). The distribution of Zn in the composite was more uniform than other elements in the map area, thereby indicating its role of deciphering catalytic activity to palygorskite in the conversion of carbon substrate into spherical graphitised carbon particles.

3.1.5. Functional group analysis For Na-palygorskite, the IR bands appearing in the 1200–400 cm−1 region are due to the vibrations of the aluminosilicate framework (Fig. 8a) (Gionis et al., 2006). The bands at 1034 cm−1 and 536 cm−1 were attributed to the stretching of Si\\O\\Si bond in palygorskite tetrahedral sheet whereas bands at 470 cm− 1 were designated to the bending of O\\Si\\O bond (Fig. 8a) (Giustetto and Chiari, 2004; Gionis et al., 2006). The bands at 1648 and 3446 cm− 1 were attributed to OH-bending modes of the coordinated and zeolitic water associated with palygorskite (Fig. 8b), while the bands at 3620 and 3549 cm− 1 were attributed to the OH-stretching modes of the hydroxyl groups associated with Al and Si (Fig. 8c) (Gionis et al., 2006). The IR bands at 3697 cm− 1 were due to the vibration of OH-stretching modes in Mg-OH in the palygorskite structure (Frost et al., 1998). The IR spectra of the nanocomposites, which showed new characteristic bands for organic components, were very different from the initial Na–palygorskite (Fig. 8). Bands at the 536–470 cm−1 regions appeared in the composite

Fig. 5. TEM images of palygorskite–carbon nanocomposites; (a) Composite 1, and (b) Composite 2.

622

B. Sarkar et al. / Applied Clay Science 114 (2015) 617–626

Table 1 Cumulative particle sizes of palygorskite and palygorskite–carbon nanocomposites obtained from number-weighted Gaussian distribution analysis. Distribution parameters

Palygorskite

Composite 1

Composite 2

25% of distribution 50% of distribution 75% of distribution 90% of distribution 99% of distribution 80% of distribution Mean diameter Variance Chi-squared

b589.3 nm b885.0 nm b1346.2 nm b1972.2 nm b3820.4 nm b1495.0 nm 1055.0 nm 0.406 59.550

b464.9 nm b667.0 nm b958.5 nm b1328.8 nm b2332.6 nm b1048.7 nm 769.6 nm 0.291 14.148

b121.4 nm b166.4 nm b241.5 nm b346.0 nm b659.0 nm b266.2 nm 169.9 nm 0.417 38.488

materials due to the presence of silica and Si\\O groups (Fig. 8a). This happened because of the heat treatment during material preparation and gradual loss of palygorskite's crystallinity. For Composite 2, the band was broader because palygorskite lost its structure more than Composite 1 and produced fine silica. Bands in the 1650–1350 cm−1 region clearly showed the elimination of zeolitic water in the nanocomposites due to the heat treatment (Fig. 8b). In Composite 2, which was prepared at higher temperature, dehydroxylation of Si and Al occurred which caused the disappearance of bands in the 3700–3600 cm−1 region (Fig. 8c). Partial dehydroxylation of Si and Al took place in Composite 1 too. Organic components attached to the palygorskite were traced to the nanocomposites in the 2950–2850 cm−1 region (Fig. 8c). These bands are attributable to the vibrations of the C\\H groups of the saturated alkyl hydrocarbons (Ai and Li, 2013; Wu et al., 2014). The new bands at 1700 and 1400 cm−1 corresponded to carbonyl group vibrations (Demir-Cakan et al., 2009; Chen et al., 2011; Ai and Li, 2013). The broad band at 1610 cm−1 could be assigned to C_C stretching vibrations (Demir-Cakan et al., 2009; Ai and Li, 2013; Wu et al., 2014). Therefore, the FTIR results confirmed that the nano-scale particles (as indicated by the morphological and elemental studies) which were deposited on the palygorskite surface contained organic molecules. The palygorskite–carbon nanocomposites were also characterised by studying their Raman spectra. As shown in Fig. 9, the broad bands in Raman spectra in both composites exhibited carbon particles of smaller crystalline size (Jawhari et al., 1995). Graphite-like carbon spectra (Jawhari et al., 1995) were obtained in Composite 2 which was heated at a higher temperature in a CO2 environment.

3.2. Adsorption of dyes The affinity of methylene blue and orange II dye compounds to the palygorskite–carbon nanocomposites was determined by studying the adsorption isotherms and kinetics. 3.2.1. Adsorption isotherm The adsorption data was best fitted to the Langmuir isotherm model which is expressed as follows (Eq. (2)): qm ¼ ðXm KL Ce Þ=ð1 þ KL Ce Þ

ð2Þ

where, qm is the amount of adsorbate adsorbed (mg g− 1), Ce is the adsorbate concentration in the solution phase (mg L− 1), KL is the Langmuir constant, and Xm is an estimate of the maximum adsorption capacity (mg g−1). Methylene blue had greater affinity to Composite 1 whereas orange II had greater affinity to Composite 2 (Fig. 10). This is attributed to the charge behaviour of the dye compounds as well as the nanocomposite surfaces. While methylene blue is a cationic dye, orange II is an anionic dye. The zeta potential values of Composite 1 and Composite 2 were −23.0 and −11.7 mV, respectively, against −18.9 mV of the unmodified palygorskite. Naturally, Composite 1 having a greater negative charge on its surface, exhibited a greater affinity to cationic methylene blue dye and vice-versa (Sarkar et al., 2012a). However, as indicated by the kinetic modelling results, the adsorption of dye compounds was governed not only by the charge behaviour of adsorbent surfaces, but also by the pore size distribution of the nanocomposites, especially for adsorption on Composite 2. The adsorption data fitted considerably well to the Langmuir isothermal model with R2 values (at 95% confidence level) greater than 0.99 for both compounds adsorbing on Composite 1 and methylene blue adsorbing on Composite 2 (Table 2). Orange II adsorbing on Composite 2 showed a slightly smaller R2 value (0.88). Fitting the data to the Langmuir model made it possible to calculate the maximum adsorption capacity values (Xm) (Table 2). As expected, Composite 1 provided a greater Xm value for methylene blue (46.3 mg g−1) and Composite 2 gave a greater value for orange II (23 mg g− 1). The Langmuir model can explain a saturated monolayer adsorption of dyes on the nanocomposite surfaces without any transmigration of the adsorbate molecules (Sarkar et al., 2010b). However, as indicated by the kinetics results, to

Fig. 6. EDAX spectra of palygorskite before and after hydrothermal carbonisation treatment.

B. Sarkar et al. / Applied Clay Science 114 (2015) 617–626

Fig. 7. EDAX line profile map of Composite 2.

some extent intra-particle diffusion of dye compounds occurred in Composite 2. 3.2.2. Adsorption kinetics Composite 1 removed methylene blue from the solution almost instantly; it showed an extremely fast kinetic reaction (Fig. 10c). As apparent from the curve slopes in Fig. 10d, Composite 2 exhibited slightly slower adsorption of orange II than Composite 1, despite the former resulting in a greater total dye adsorption at equilibrium. The adsorption kinetic data were fitted to intra-particle diffusion (parabolic diffusion) and pseudo-second order models as given by the following Eqs. (3) and (4), respectively: qt ¼ Kip t0:5 þ C

ð3Þ

t=qt ¼ 1=K2 qe 2 þ t=qe

ð4Þ

where, qt is the amount of dye adsorbed (mg g−1) at various times t, C is the intercept, Kip is the intra-particle diffusion rate constant (mg g−1 min−1/2), qe is the maximum adsorption capacity (mg g−1)

623

of dye at equilibrium, and K2 is the rate constant of pseudo-second order adsorption (g mg−1 min−1). In general, the adsorption data fitted very well to the pseudo-second order kinetic model with R2 values closer to 1 (Table 3). A good fit of the adsorption data to the pseudo-second order model indicated that both physical and chemical mechanisms controlled the adsorption process over the entire range of dye concentrations (Ho and McKay, 1998a; Özcan et al., 2005; Sarkar et al., 2010b). A remarkably greater K2 value of Composite 1 for methylene blue adsorption (1.01 g mg− 1 min−1) confirmed its instantaneous affinity to the dye compound. Composite 1 showed a greater K2 value (0.035 g mg−1 min−1) than Composite 2 for orange II adsorption as well. The higher qe values of Composite 2 combined with smaller respective K2 values (Table 3) were intriguing in that it could not be confirmed in the pseudo-second order kinetic model whether the adsorption process was controlled by diffusion. Consequently, the data were further analysed using the intra-particle diffusion model (Ho and McKay, 1998a). Interestingly, Composite 2, which produced a greater surface area than Composite 1 and showed the presence of mesopores, adsorbed both the dye compounds by maintaining the intra-particle diffusion equation. The kinetic adsorption data fitted to a plot of the square root of time (t1/2) vs the uptake (qt) (plot not shown) with R2 values 0.99 and 0.94 for methylene blue and orange II, respectively (Table 3), which confirmed that intra-particle diffusion was involved in the adsorption process of Composite 2 (Al-Asheh et al., 2003; Malik, 2003; Özcan and Özcan, 2004). However, the plots did not pass through the origin and they evolved some intercept (C) values (Table 3), which indicated that the intra-particle diffusion was not the only rate-controlling step. In fact other processes will also control the rate of adsorption, and all of which may operate simultaneously (Ho and McKay, 1998b; Özcan and Özcan, 2004; Özcan et al., 2005). The C value is larger for a greater boundary layer effect (Özcan and Özcan, 2004). The mobility of dye compounds into the particles can be increased by reducing the boundary layer resistance, for example through an increase in contact time and/or adsorbate concentration (Özcan and Özcan, 2004). Additionally, the boundary layer diffusion is also governed by adsorbent parameters such as external surface area, particle size, shape, density and porosity (Özacar and Şengil, 2003; Özcan and Özcan, 2004). In the current study, Composite 2 having micro and mesopores together with a smaller average particle size (Table 1) and a larger surface area contributed intra-particle diffusion resistance during the adsorption of dye compounds (Ho and McKay, 1998b; Kannan and Sundaram, 2001; Özacar and Şengil, 2003; Özcan and Özcan, 2004). The boundary resistance (C value) was greater for orange II adsorption than methylene blue adsorption (Table 3). Consequently, the intra-particle diffusion coefficient (Kip) was greater during methylene adsorption than orange II adsorption on Composite 2 (Table 3). The dye adsorption capacity of the synthesised palygorskite– carbon composites was superior or comparable to many low cost adsorbents such as biochar (Liu et al., 2012b; Cheng et al., 2013), activated carbon (Karagöz et al., 2008), coir pith carbon (Kavitha and Namasivayam, 2007), surfactant modified montmorillonite (Shin, 2008), modified zeolite (Alver and Metin, 2012), rice biomass (Rehman et al., 2012), and bottom ash and de-oiled soya (Gupta et al., 2006). 4. Conclusions Spherical carbon nanoparticles were synthesised on fibrous palygorskite clay mineral from starch using hydrothermal carbonisation technique. The nanocomposite, when activated with heat at 550 °C for 3 h (ramp at 10 °C min−1) under CO2 environment (200 mL min−1), provided graphite-like carbon nano-spheres which were uniformly hosted on the partially lost palygorskite structure. The conversion of starch into spherical graphitised particles was catalysed by Zn uniformly dispersed on the palygorskite surface. The heat-CO2 treatment resulted

624

B. Sarkar et al. / Applied Clay Science 114 (2015) 617–626

Fig. 8. FTIR spectra of Na–palygorskite and palygorskite–carbon nanocomposites in the region of (a) 1200–400 cm−1, (b) 2400–1200 cm−1, and (c) 4000–2400 cm−1 wave numbers.

in the creation of mesopores in the nanocomposite and provided a 17fold greater specific surface area compared to the nanocomposite prepared without such treatment. The nanocomposites contained oxygen-containing surface functional groups and showed variable affinity to cationic and anionic dye compounds. While the heat-CO2-treated nanocomposite adsorbed a larger amount of anionic orange II dye (23 mg g−1), its counterpart without such treatment adsorbed a higher

quantity of cationic methylene blue (46.3 mg g−1). In addition to electrostatic attraction for methylene blue adsorption on both the nanocomposites, pore diffusion mechanism was also involved and the boundary resistance was greater for orange II than methylene blue adsorption. Being a material developed from green biomass (starch) and an abundant natural resource (palygorskite), these nanocomposites have immense potential for application in environmental remediation,

Fig. 9. Raman spectra of the palygorskite–carbon nanocomposites.

B. Sarkar et al. / Applied Clay Science 114 (2015) 617–626

625

Fig. 10. Adsorption of dyes on palygorskite–carbon nanocomposites; (a) isotherm of methylene blue adsorption, (b) isotherm of orange II adsorption, (c) kinetics of methylene blue adsorption, and (d) kinetics of orange II adsorption; bars represent standard error at 95% confidence level, n = 3.

Table 2 Langmuir parameters for adsorption of dyes on palygorskite–carbon nanocomposites. Adsorbents

Methylene blue −1

Composite 1 Composite 2 a

Orange II −1

Xm (mg g

)

KL (L g

46.3 ± 3.19 17.8 ± 1.84

a

0.1942 ± 0.008 0.4086 ± 0.028

R

Xm (mg g−1)

KL (L g−1)

R2

0.9924 0.9947

8.5 ± 1.19 23.0 ± 2.27

0.0251 ± 0.004 0.0962 ± 0.009

0.9936 0.8807

2

)

Standard error at 95% confidence level, n = 3.

Table 3 Kinetic parameters for adsorption of dyes on palygorskite–carbon nanocomposites. Dyes

Methylene blue Orange II a

Adsorbents

Composite 1 Composite 2 Composite 1 Composite 2

Intra-particle diffusion

Pseudo-second order 2

C (mg g−1)

Kip (g mg−1 min−1/2)

R

qe (mg g−1)

K2 (g mg−1 min−1)

R2

49.9 ± 4.41a 17.9 ± 2.29 16.5 ± 2.21 24.3 ± 1.98

−0.008 ± 0.001 2.773 ± 0.975 0.530 ± 0.073 1.839 ± 1.112

0.8688 0.9936 0.4520 0.9391

49.8 ± 3.19 51.3 ± 5.27 21.2 ± 3.33 44.6 ± 1.49

1.01 ± 0.081 0.002 ± 0.001 0.035 ± 0.007 0.004 ± 0.001

1 0.9983 0.9997 0.9983

Standard error at 95% confidence level, n = 3.

and in particular the mesoporous material can be used for in situ immobilisation of persistent organic pollutants in soil (Duan et al., 2014). Acknowledgement This research was funded by CRC CARE through a Department of Defence (Australian Government) grant. BS is grateful to the University of South Australia for awarding him a New Appointee Grant. The authors also acknowledge Dr Cheng Fang for his assistance in collecting the Raman spectra of the materials. References Ai, L., Li, L., 2013. Efficient removal of organic dyes from aqueous solution with ecofriendly biomass-derived carbon@montmorillonite nanocomposites by one-step hydrothermal process. Chem. Eng. J. 223, 688–695. Al-Asheh, S., Banat, F., Abu-Aitah, L., 2003. Adsorption of phenol using different types of activated bentonites. Sep. Purif. Technol. 33 (1), 1–10.

Alver, E., Metin, A., 2012. Anionic dye removal from aqueous solutions using modified zeolite: adsorption kinetics and isotherm studies. Chem. Eng. J. 200–202, 59–67. Anadão, P., et al., 2011. Montmorillonite/carbon nanocomposites prepared from sucrose for catalytic applications. Appl. Clay Sci. 53 (2), 288–296. Anadão, P., Pajolli, I.L.R., Hildebrando, E.A., Wiebeck, H., 2014. Preparation and characterization of carbon/montmorillonite composites and nanocomposites from waste bleaching sodium montmorillonite clay. Adv. Powder Technol. 25 (3), 926–932. Araújo Melo, D.M., Ruiz, J.A.C., Melo, M.A.F., Sobrinho, E.V., Martinelli, A.E., 2002. Preparation and characterization of lanthanum palygorskite clays as acid catalysts. J. Alloys Compd. 344 (1–2), 352–355. Chen, L.-F., Liang, H.-W., Lu, Y., Cui, C.-H., Yu, S.-H., 2011. Synthesis of an attapulgite clay@ carbon nanocomposite adsorbent by a hydrothermal carbonization process and their application in the removal of toxic metal ions from water. Langmuir 27 (14), 8998–9004. Cheng, G., et al., 2013. Adsorption of methylene blue by residue biochar from copyrolysis of dewatered sewage sludge and pine sawdust. Desalin. Water Treat. 51 (37–39), 7081–7087. Churchman, G.J., Gates, W.P., Theng, B.K.G., Yuan, G., 2006. Chapter 11.1 Clays and Clay Minerals for Pollution Control. In: Bergaya, F., Theng, B.K.G., Lagaly, G. (Eds.), Developments in Clay Science. Elsevier, pp. 625–675. Cui, X., Antonietti, M., Yu, S.-H., 2006. Structural effects of iron oxide nanoparticles and iron ions on the hydrothermal carbonization of starch and rice carbohydrates. Small 2 (6), 756–759.

626

B. Sarkar et al. / Applied Clay Science 114 (2015) 617–626

Demir-Cakan, R., Baccile, N., Antonietti, M., Titirici, M.-M., 2009. Carboxylate-rich carbonaceous materials via one-step hydrothermal carbonization of glucose in the presence of acrylic acid. Chem. Mater. 21 (3), 484–490. Duan, L., Palanisamia, T., Liu, Y., Dong, Z., Mallavarapu, M., Kuchel, T., Semple, K.T., Naidu, R., 2014. Effects of ageing and soil properties on the oral bioavailability of benzo[a]pyrene using a swine model. Environ. Int. 70, 192–202. Frost, R.L., Cash, G.A., Kloprogge, J.T., 1998. ‘Rocky Mountain leather’, sepiolite and attapulgite—an infrared emission spectroscopic study. Vib. Spectrosc. 16 (2), 173–184. Gionis, V., Kacandes, G.H., Kastritis, I.D., Chryssikos, G.D., 2006. On the structure of palygorskite by mid- and near-infrared spectroscopy. Am. Mineral. 91 (7), 1125–1133. Giustetto, R., Chiari, G., 2004. Crystal structure refinement of palygorskite from neutron powder diffraction. Eur. J. Mineral. 16 (3), 521–532. Gupta, V.K., Mittal, A., Gajbe, V., Mittal, J., 2006. Removal and recovery of the hazardous azo dye acid Orange 7 through adsorption over waste materials: bottom ash and de-oiled soya. Ind. Eng. Chem. Res. 45 (4), 1446–1453. Ho, Y.S., McKay, G., 1998a. A comparison of chemisorption kinetic models applied to pollutant removal on various sorbents. Process Saf. Environ. Prot. 76 (4), 332–340. Ho, Y.S., McKay, G., 1998b. Kinetic models for the sorption of dye from aqueous solution by wood. Process Saf. Environ. Prot. 76 (2), 183–191. Hu, B., Yu, S.-H., Wang, K., Liu, L., Xu, X.-W., 2008. Functional carbonaceous materials from hydrothermal carbonization of biomass: an effective chemical process. Dalton Trans. (40), 5414–5423. Hu, B., et al., 2010. Engineering carbon materials from the hydrothermal carbonization process of biomass. Adv. Mater. 22 (7), 813–828. Jawhari, T., Roid, A., Casado, J., 1995. Raman spectroscopic characterization of some commercially available carbon black materials. Carbon 33 (11), 1561–1565. Kannan, N., Sundaram, M.M., 2001. Kinetics and mechanism of removal of methylene blue by adsorption on various carbons—a comparative study. Dyes Pigments 51 (1), 25–40. Karagöz, S., Tay, T., Ucar, S., Erdem, M., 2008. Activated carbons from waste biomass by sulfuric acid activation and their use on methylene blue adsorption. Bioresour. Technol. 99 (14), 6214–6222. Katsuki, H., Furuta, S., Watari, T., Komarneni, S., 2005. ZSM-5 zeolite/porous carbon composite: conventional- and microwave-hydrothermal synthesis from carbonized rice husk. Microporous Mesoporous Mater. 86 (1–3), 145–151. Kavitha, D., Namasivayam, C., 2007. Experimental and kinetic studies on methylene blue adsorption by coir pith carbon. Bioresour. Technol. 98 (1), 14–21. Kumar, A.S.K., Kalidhasan, S., Rajesh, V., Rajesh, N., 2011. Application of cellulose-clay composite biosorbent toward the effective adsorption and removal of chromium from industrial wastewater. Ind. Eng. Chem. Res. 51 (1), 58–69. Li, T., Shen, J., Li, N., Ye, M., 2012. Facile and novel hydrothermal preparation of functionalised carbon microspheres from glucose by using graphene sheets as a substrate. Mater. Lett. 89, 202–205. Li, T., Shen, J., Huang, S., Li, N., Ye, M., 2014. Hydrothermal carbonization synthesis of a novel montmorillonite supported carbon nanosphere adsorbent for removal of Cr (VI) from waste water. Appl. Clay Sci. 93–94, 48–55. Liu, E., Locke, A.J., Martens, W.N., Frost, R.L., Yang, X., 2012a. Fabrication of macromesoporous zirconia–alumina materials with a one-dimensional hierarchical structure. Cryst. Growth Des. 12 (3), 1402–1410.

Liu, Y., Zhao, X., Li, J., Ma, D., Han, R., 2012b. Characterization of bio-char from pyrolysis of wheat straw and its evaluation on methylene blue adsorption. Desalin. Water Treat. 46 (1–3), 115–123. Liu, W., et al., 2013. Kinetics and thermodynamics characteristics of cationic yellow X-GL adsorption on attapulgite/rice hull-based activated carbon nanocomposites. Environ. Prog. Sustainable Energy 32 (3), 655–662. Malik, P.K., 2003. Use of activated carbons prepared from sawdust and rice-husk for adsorption of acid dyes: a case study of Acid Yellow 36. Dyes Pigments 56 (3), 239–249. Önal, M., SarIkaya, Y., 2009. Some physicochemical properties of a clay containing smectite and palygorskite. Appl. Clay Sci. 44 (1–2), 161–165. Özacar, M., Şengil, İ.A., 2003. Adsorption of reactive dyes on calcined alunite from aqueous solutions. J. Hazard. Mater. 98 (1–3), 211–224. Özcan, A.S., Özcan, A., 2004. Adsorption of acid dyes from aqueous solutions onto acidactivated bentonite. J. Colloid Interface Sci. 276 (1), 39–46. Özcan, A.S., Erdem, B., Özcan, A., 2005. Adsorption of Acid Blue 193 from aqueous solutions onto BTMA-bentonite. Colloids Surf. A Physicochem. Eng. Asp. 266 (1–3), 73–81. Qian, H.-S., et al., 2006. Synthesis of uniform Te@carbon-rich composite nanocables with photoluminescence properties and carbonaceous nanofibers by the hydrothermal carbonization of glucose. Chem. Mater. 18 (8), 2102–2108. Rehman, M.S., Kim, I., Han, J., 2012. Adsorption of methylene blue dye from aqueous solution by sugar extracted spent rice biomass. Carbohydr. Polym. 90 (3), 1314–1322. Sarkar, B., Xi, Y., Megharaj, M., Krishnamurti, G.S.R., Naidu, R., 2010a. Synthesis and characterisation of novel organopalygorskites for removal of p-nitrophenol from aqueous solution: isothermal studies. J. Colloid Interface Sci. 350 (1), 295–304. Sarkar, B., et al., 2010b. Remediation of hexavalent chromium through adsorption by bentonite based Arquad® 2HT-75 organoclays. J. Hazard. Mater. 183 (1–3), 87–97. Sarkar, B., Xi, Y., Megharaj, M., Naidu, R., 2011. Orange II adsorption on palygorskites modified with alkyl trimethylammonium and dialkyl dimethylammonium bromide — an isothermal and kinetic study. Appl. Clay Sci. 51 (3), 370–374. Sarkar, B., Megharaj, M., Xi, Y., Naidu, R., 2012a. Surface charge characteristics of organopalygorskites and adsorption of p-nitrophenol in flow-through reactor system. Chem. Eng. J. 185–186, 35–43. Sarkar, B., et al., 2012b. Bio-reactive organoclay: a new technology for environmental remediation. Crit. Rev. Environ. Sci. Technol. 42, 435–488. Shin, W.S., 2008. Competitive sorption of anionic and cationic dyes onto cetylpyridiniummodified montmorillonite. J. Environ. Sci. Health A 43 (12), 1459–1470. Wu, X., Zhu, W., Zhang, X., Chen, T., Frost, R.L., 2011. Catalytic deposition of nanocarbon onto palygorskite and its adsorption of phenol. Appl. Clay Sci. 52 (4), 400–406. Wu, X., et al., 2014. Synthesis of clay/carbon adsorbent through hydrothermal carbonization of cellulose on palygorskite. Appl. Clay Sci. 95, 60–66. Xu, Y.-J., et al., 2008. Nanoarchitecturing of activated carbon: facile strategy for chemical functionalization of the surface of activated carbon. Adv. Funct. Mater. 18 (22), 3613–3619. Zhang, W., Mu, B., Wang, A., Shao, S., 2014. Attapulgite oriented carbon/polyaniline hybrid nanocomposites for electrochemical energy storage. Synth. Met. 192, 87–92.