Bioresource Technology 161 (2014) 310–319

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Hydrothermal conversion of urban food waste to chars for removal of textile dyes from contaminated waters Ganesh K. Parshetti, Shamik Chowdhury, Rajasekhar Balasubramanian ⇑ Department of Civil and Environmental Engineering, National University of Singapore, 1 Engineering Drive 2, Singapore 117576, Republic of Singapore

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

 HTC of food waste yields novel

hydrochar materials for dye adsorption.  Microstructure and chemical properties of hydrochars were evaluated.  Substantial increase in adsorption with increasing functional groups on hydrochar.  Adsorption mechanism was elucidated using kinetic and thermodynamic analysis.  An ANN-based model is also developed to predict adsorption capacity of hydrochar.

a r t i c l e

i n f o

Article history: Received 7 January 2014 Received in revised form 14 March 2014 Accepted 17 March 2014 Available online 26 March 2014 Keywords: Food waste Hydrochars Dyes Adsorption Artificial neural network

a b s t r a c t Hydrothermal carbonization of urban food waste was carried out to prepare hydrochars for removal of Acridine Orange and Rhodamine 6G dyes from contaminated water. The chemical composition and microstructure properties of the synthesized hydrochars were investigated in details. Batch adsorption experiments revealed that hydrochars with lower degree of carbonization were more efficient in adsorption of dyes. Operational parameters such as pH and temperature had a strong influence on the dye uptake process. The adsorption equilibrium data showed excellent fit to the Langmuir isotherm. The pseudo-second-order kinetic model provided a better correlation for the experimental kinetic data in comparison to the pseudo-first-order kinetic model. Thermodynamic investigations suggested that dye adsorption onto hydrochars was spontaneous and endothermic. The mechanism of dye removal appears to be associated with physisorption. An artificial neural network (ANN)-based modelling was further carried out to predict the dye adsorption capacity of the hydrochars. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction The adsorption characteristics of biochars (pyrogenic black carbon derived from incomplete combustion of biomass) have been investigated for the removal of a variety of contaminants, including heavy metals, nutrients, aromatic compounds, pesticides and

⇑ Corresponding author. Tel.: +65 65165135; fax: +65 67744202. E-mail address: [email protected] (R. Balasubramanian). http://dx.doi.org/10.1016/j.biortech.2014.03.087 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.

herbicides in wastewater (Cabrera et al., 2011; Chen and Chen, 2009; Inyang et al., 2012; Yao et al., 2011; Zheng et al., 2010). Hydrothermal carbonization (HTC) is gaining worldwide attention as a viable technology for conversion of waste biomass to carbonrich, energy-dense, value-added solid hydrochar materials (Sevilla et al., 2011). HTC is carried out at relatively low temperatures (180–350 °C) in the presence of water under autogenous pressures, and does not require the drying of traditionally wet feedstocks for waste conversion. Also, a large percentage of the carbon in the original feedstock actually remains integrated within the hydrochar,

G.K. Parshetti et al. / Bioresource Technology 161 (2014) 310–319

311

Nomenclature A Ca Ce Co Ct E Ea DG# DGO DH# DHO DHX h KC KF KL k kB ki k1 k2 m N n p

Arrhenius constant equilibrium dye concentration on the adsorbent (mg L1) equilibrium dye concentration in solution (mg L1) initial dye concentration (mg L1) dye concentration in solution at any time, t (mg L1) mean free energy of adsorption per mole of adsorbate (kJ mol1) activation energy (kJ mol1) free energy of activation (kJ mol1) Gibbs free energy change (kJ mol1) standard enthalpy of activation (kJ mol1) enthalpy of reaction (kJ mol1) isosteric heat of adsorption (kJ mol1) Planck’s constant (6.626  1034 J s) distribution coefficient for adsorption Freundlich isotherm constant (mg g1) (L g1)1/n Langmuir isotherm constant (L mg1) reaction rate constant Boltzmann constant (1.380  1023 J K1) intra-particle diffusion rate constant (min1) pseudo-first-order rate constant (min1) pseudo-second-order rate constant (g mg1 min1) mass of adsorbent (g) number of data points characteristic constant related to degree of favorability of adsorption number of parameters in the model equation

ultimately minimizing greenhouse gas emission (Parshetti et al., 2013). Hydrochars differ from biochars in that they are generally less aromatic, consisting of mostly alkyl moieties (Cao et al., 2009). Work conducted by Sun et al. (2011, 2012) showed that hydrothermally produced hydrochars have the potential to adsorb a greater number of polar and nonpolar organics in water more effectively than biochar, demonstrating the use of hydrochars as a useful adsorbent for decontamination of wastewaters. However, mostly agricultural wastes and by-products are employed as feedstock’s for hydrochar production. A wide range of natural and waste materials are currently available for the production of hydrochar-based adsorbents, among which food waste is of considerable interest. According to the Global Food Losses and Food Waste report published in 2011 by the Food and Agriculture Organization of the United Nations, about 1.3 billion tonnes, or one third of the total food produced in the world for human consumption are lost or wasted every year throughout the various stages of the food supply chain (production, processing, distribution, consumption, and disposal). According to the waste statistics from Singapore’s National Environment Agency (NEA), the annual generation of food waste in Singapore was 542,700 tonnes in 2006 and reached about 640,500 tonnes in 2010, which is around 10% of the total waste output in Singapore (Rajagopal et al., 2013). A recent report indicated that the amount of food waste generated in Singapore hit a record high of 703,200 tonnes in 2012 (NEA, 2012). Conventional disposal methods such as landfills and ocean dumping emit methane, which has 20–25 times stronger warming effect than CO2 on a molecular basis, thereby contributing significantly to climate change (Nahman et al., 2012). HTC as a hydrothermal conversion technique has the potential to overcome many of the challenges associated with the conventional disposal methods and/or biological treatment of discarded food. Carbonization offers the advantages of smaller footprint/land area required,

qe qe,exp qe,cal qe,pred qm qt R R2 DS# DSO s T t V Xi Xmax Xmin Xnorm

equilibrium dye uptake per g of adsorbent (mg g1) experimental equilibrium dye uptake per g of adsorbent (mg g1) calculated equilibrium dye uptake per g of adsorbent (mg g1) ANN predicted equilibrium dye uptake per g of adsorbent (mg g1) maximum adsorption capacity (mg g1) amount of dye adsorbed at any time, t (mg g1) universal gas constant (8.314 J mol1 K1) correlation coefficient standard entropy of activation (J mol1 K1) entropy of reaction (J mol1 K1) number of neurons per input value in a layer temperature (K) time (min) volume of test solution (mL) input or output variable X maximum value of variable X minimum value of variable X normalized value of variable X

Greek letters b coefficient related to the mean free energy of adsorption (mol2 J2) e Polanyi potential (J mol1) = RT ln (1 + 1/Ce) 2 v Chi-square test

efficient conversion of mixed wastes, and greater waste volume reduction. As HTC is a thermochemical technique, mixed wastes are not as significant an operational issue as in composting and anaerobic digestion. In addition, because of the moisture requirement, food waste is more suitable for conversion via HTC than other common dry thermal conversion techniques. The overarching objective of this study was to (i) prepare solid carbonaceous hydrochars by HTC of food waste at two different temperatures i.e. 250 and 350 °C, (ii) examine the microstructure and chemical composition of the produced hydrochars by ultimate analysis, scanning electron microscopy (SEM), Brunauer–Emmett– Teller (BET) surface area analysis, X-ray photoelectron spectroscopy (XPS) and inductively coupled plasma optical emission spectrometry (ICP-OES) and (iii) evaluate the potential practical application of the hydrochars as an adsorbent for removal of cationic dyes from contaminated waters. Acridine Orange (AO) and Rhodamine 6G (R6G) were selected as the model pollutants due to their extensive use in textile manufacturing, leather processing, printing inks and lithography, and their reported mutagenic effects in humans (Sadhasivam et al., 2007; Zubieta et al., 2012). The adsorption behavior was modelled using standard isotherm equations and the associated kinetics and thermodynamics were analyzed. Finally, the experimental adsorption data were used to predict the dye adsorption profile by applying an artificial neural network (ANN)-based model, because of its ability to capture the non-linear relationships existing between variables in complex systems. 2. Methods 2.1. Materials Food waste was collected from the National University of Singapore restaurants. Visual observation of the food waste indicated

312

G.K. Parshetti et al. / Bioresource Technology 161 (2014) 310–319

that the waste consisted of a variety of cooked food (e.g., chicken, seafood, French fries, vegetables, leftovers like rice and gravy), uncooked food (e.g., fruit peels, vegetable parts) and condiments (e.g., salad dressing, ketchup, cocktail sauce). The collected waste was weighed and immediately separated out for bones, eggshells, plastic utensils, etc. because of processing limitations. Following separation, the food waste was mixed and homogenized with a food grade blender and stored at 4 °C prior to HTC. All chemicals used in this study were of analytical reagent grade. Acridine Orange (AO) (C.I. 46005, CAS No. 10127-02-3, FW: 370.0, MF: C17H20ClN3 0.5ZnCl2) and Rhodamine 6G (R6G) (C.I. 45160, CAS No. 989-388, FW: 479.02, MF: C28H31N2O3Cl) were procured from Sigma–Aldrich Corporation and used as received. Their chemical structures are shown in Fig. S1 (see Supplementary information). Dye stock solutions (500 mg L1) were prepared by dissolving 0.5 g of AO or R6G in 1 L distilled water. Experimental dye solution of desired concentration was prepared by further dilution of the stock solution with suitable volume of distilled water. The initial solution pH was adjusted with 0.1 (M) HCl and 0.1 (M) NaOH solutions using a digital pH meter (PH 700, Eutech Instruments, Singapore) calibrated with standard buffer solutions. 2.2. Hydrothermal carbonization: apparatus and procedure HTC of food waste was carried out in a 500 mL Parr stirred pressure batch reactor (Model 4575, Germany; Heater power: 1000 W). The reactor was run with, 25% of solid (as-received food waste) waste at two different temperatures of 250 and 350 °C with a constant residence time of 20 min. The reactor was sealed and heated to the desired reaction temperature with the help of an electric furnace. After the desired residence time, the heater was turned off and the reactor was rapidly cooled to quench the reaction. Residence time is defined as the time the reactor is held and operated at the desired reaction temperature, excluding preheating and cooling time (see Supplementary information, Fig. S2). The reactor was then opened to collect the reaction products. The solid carbonaceous product (denoted as hydrochar) was washed with deionized water and dried in air. The char product was finally weighed independently and stored in a sealed container until analysis. The resulting hydrochar materials were denoted as FWH-250 and FWH-350, respectively, depending on the reaction temperature. It should be noted that the average energy demand of the hydrochar production process is 3 MJ (based on electricity consumption). However, following carbonization, a portion of water is removed from the material via gravity drainage and the wet recovered solids are air-dried, both of which require no energy, ultimately decreasing the net energy requirements. 2.3. Characterization of hydrochar The moisture content, volatile solid content and total solid content of the food waste was measured using a gravimetric technique. Ultimate analysis of FWH-250 and FWH-350 was carried out to determine the weight percent of different elements (C, H, N, O, and S) by using an elementary analyzer (vario MACRO cube, ELEMENTAR, Germany). The concentration of trace elements was determined using ICP-OES (Perkin Elmer, ELAN 6100, USA) analysis. The surface morphology of the hydrochar materials was determined by SEM (JEOL JSM-6700F Oxford Inca Energy 400). The BET method was utilized to calculate the specific surface area of the prepared hydrochar materials. Nitrogen adsorption–desorption isotherm data at 77 K was recorded for this purpose using a N2 adsorption analyzer (Quantachrome Instruments, NOVA 4200e). XPS analysis of the samples was carried out using a VG ESCA 220i-XL Imaging (Thermo VG Scientific Ltd., UK). For further details

on these characterization methods, please see Supplementary information. 2.4. Batch adsorption experiments Batch tests were conducted to evaluate the effect of pH (3.0–9.0, step size: 1) and temperature (293–313 K, step size: 10 K) on the adsorption rate of AO and R6G by hydrochar materials. In each experiment, 100 mL dye solution (50 mg L1) was taken in a 250 mL glass-stoppered Erlenmeyer flask. A weighed amount of adsorbent (0.05 g) was added to the solution. The flask was then agitated at a constant speed (150 rpm) and temperature in an incubator shaker (LM-575RD, Yihder Technology Co., Ltd., Taiwan). At the end of the adsorption period, the adsorbent was separated from the solution by filtration through Whatman 0.45 mm filter paper. The residual liquid-phase dye concentration was analyzed by using a UV/Vis spectrophotometer (U 2800, Hitachi, Japan) at maximum absorbance wavelength (kmax) of 490.5 nm for AO and 530.0 nm for R6G, respectively. Kinetics of adsorption was determined by analyzing the adsorptive uptake of dye from aqueous solution at different time intervals. Adsorption equilibrium studies were conducted at different temperatures (293, 303, and 313 K, respectively) by agitating 100 mL dye solution of different concentrations (10–100 mg L1) with 0.05 g of adsorbent for 2 h. For calculation of removal efficiency and adsorption capacity, please see Supplementary information. 2.5. Theory Details of equilibrium isotherms, kinetic models, and thermodynamic equations are provided in the Supplementary information. 3. Results and discussion 3.1. Hydrochar characterization 3.1.1. Chemical characteristics An ultimate analysis of the food waste used in this study and the hydrochars (FWH-250 and FWH-350) thus derived from it is presented in Table 1. The initial chemical composition of the food waste is provided because the diets from different regions around the world are different dramatically; so the applicable scope of the results from this research can be defined by the composition of food wastes. The prepared hydrochars contained C and O as their most important components. FWH-250 had a slightly lower C content, but a relatively higher O content in comparison to FWH-350. FWH-250 also had higher contents of H and N than those of FWH-350. Furthermore, O/C and H/C atomic ratios are very low for FWH-350 which may be due to exothermic oxidation of C and H to CO2, CO, and H2O at higher temperatures (Sevilla et al., 2011). These finding suggest that FWH-250 is less carbonized; therefore problem of organics likely to be released during the intended application of the hydrochar will be less (Becker et al., 2013). It is interesting to note that hydrochars with higher degrees of carbonization is highly desirable, particularly in applications such as fuel development. However, in this study hydrochars with lower degree of carbonization but increased functionality was important since the produced material was to be applied for adsorption studies. Inorganic composition of FWH-250 and FWH350, determined by ICP-OES studies, is also summarized in Table 1. Platinum (Pt) concentration was the highest in the samples, followed by calcium (Ca), sodium (Na), potassium (K) and iron (Fe). A detectable level of molybdenum (Mo) was also present, but only in FWH-250. In general, hydrochars with less inorganic contents are

313

G.K. Parshetti et al. / Bioresource Technology 161 (2014) 310–319 Table 1 Organic and inorganic elemental analyses of food waste, FWH-250 and FWH-350.

⁄

Item

Food waste

FWH-250

FWH-350

Organic (%) C H N S O O/C H/C Solids yield (%) Carbon densification

35.2 ± 1.76 5.6 ± 0.28 1.5 ± 0.07 0.50 ± 0.02 57.2 ± 2.86 1.21 ± 0.06 1.90 ± 0.09 – –

52.23 ± 2.61 5.70 ± 0.28 3.38 ± 0.16 0.50 ± 0.02 38.1 9 ± 1.90 0.54 ± 0.02 1.30 ± 0.06 51 ± 2.55 1.10 ± 0.05

56.75 ± 2.83 4.31 ± 0.21 2.07 ± 0.10 0.50 ± 0.02 36.37 ± 1.81 0.48 ± 0.02 0.91 ± 0.04 36 ± 1.8 1.20 ± 0.06

Inorganic (lg g1) K Na Mg Cr Mn Fe Co Ni Cu Zn Al Pb Mo Ag Cd In Ga Ba Ca Li Ti Bi B

0.36 ± 0.012 0.40 ± 0.010 0.10 ± 0.002 ND ND 0.19 ± 0.004 ND ND ND 0.10 ± 0.003 0.10 ± 0.001 ND ND ND ND ND ND ND 0.52 ± 0.007 0.10 ± 0.001 ND ND 0.10 ± 0.002

0.20 ± 0.010 0.34 ± 0.009 0.10 ± 0.002 ND ND 0.17 ± 0.001 ND ND ND 0.10 ± 0.003 0.10 ± 0.002 ND ND ND ND ND ND ND 0.42 ± 0.007 0.10 ± 0.003 ND ND 0.10 ± 0.001

0.15 ± 0.011 0.42 ± 0.009 0.10 ± 0.001 ND ND 0.11 ± 0.002 ND ND ND 0.10 ± 0.002 0.10 ± 0.002 ND ND ND ND ND ND ND 0.21 ± 0.008 0.10 ± 0.003 ND ND 0.10 ± 0.002

ND, not detected.

preferred as adsorbents as they are not likely to alter the chemical composition of the wastewater being treated, significantly. The relatively low concentration of inorganic elements in the hydrochar samples derived from HTC of food waste suggests that they can be used as adsorbents for remediation of environmental waters.

stover. However, differences in the specific surface area and pore dimension of the hydrochar samples in the present study may be due to collapse of the pores in FWH-350 at higher temperature (Wang et al., 2011). The surface composition and element characterization of the hydrochars were further analyzed using XPS. As shown in Fig. S5a and b (see Supplementary information), the XPS survey scan spectra of the hydrochars exhibited the typical peaks of C1s, O1s and N1s at around 285, 533 and 397 eV, respectively. However, no characteristic peaks for inorganic elements were observed, which might be due to undetectable levels of these species in the samples at depths of up to a few nanometers, beyond which X-rays cannot penetrate. Fig. S6a and b (see Supplementary information) illustrate the high-energy resolution carbon C1s XPS spectra of FWH-250 and FWH-350. Deconvolution of the C1s peak for both the samples showed the presence of more than one peak, corresponding to different carbon-based functional groups. The main peak at 284.5–285.0 eV was likely contributed from a graphitic structure corresponding to CAC/CAH. The other peaks at around 286.4–286.9 and 288.6–289.1 eV can be presumably attributed to CAO and OAC@O, respectively. A shoulder appearing at 286.4– 286.9 eV may be due to CAOAC (ether) and CAOH (alcohol) functional groups. The peak at 288.6–289.1 eV can be ascribed to C@O (carbonyl) bonds of ketones and aldehydes or ACOOH (carboxylic acid) and ACOOR (ester) functional groups. The O1s spectra (see Supplementary information, Fig. S6c) of FWH-250 and FWH-350 both showed distinct peaks at 532.84 eV which suggest that the high binding energy states of O such as chemisorbed oxygen were present in the hydrochars (Kijima et al., 2011). In the case of N1s spectra (see Supplementary information, Fig. S6d), FWH-250 and FWH-350 showed peaks for typical NAH bond at 400.55 eV. The percent atomic composition of the hydrochar samples, as determined from these spectrums, is given in Table 2. FWH-350 had a greater proportion of C1s CAC/CAH in comparison to FWH-250. However, the fraction of CAO, OAC@O, OAC/O@C and NAH was remarkably high for FWH-250. It is thus apparent that FWH-250 had a more diverse distribution of surface functional groups compared to FWH-350. 3.2. Dye adsorption studies

3.1.2. Microstructure properties FE-SEM images of FWH-250 and FWH-350 showed the presence of interconnected spherical particles ranging from 700 nm to 2 lm (see Supplementary information, Fig. S3a and b). Food waste consisted primarily of organic matter of plant or animal origin containing water and carbohydrates (glucose, fructose, sucrose, cyclodextrins, celluose and starch). The HTC process is suitable to generate colloidal carbonaceous spheres from the carbohydrates (Hu et al., 2010). The dehydration and fragmentation of sugars give rise to different soluble products such as furfural-like compounds, organic acids, and aldehydes. which then undergo polymerization or condensation reactions, forming the final spherical carbonaceous material (Hu et al., 2010). In addition, N2 adsorption–desorption isotherms were also employed to investigate the specific surface area and the pore structures of the prepared hydrochar samples (see Supplementary information, Fig. S4). According to the IUPAC classification of adsorption isotherms, the hydrochars in this study showed the Type II isotherm behavior, which is characteristic of non-porous materials. The BET surface areas of FWH-250 and FWH-350 were found to be 6.07 and 1.01 m2 g1, respectively. The BJH pore diameters (6.29 and 1.25 nm for FWH-250 and FWH-350, respectively) and pore volumes (0.064 and 0.001 cm3 g1 for FWH-250 and FWH-350, respectively) confirmed that the hydrochars obtained from HTC of food waste are non-porous materials, which may explain their low surface area. A similar trend was previously reported by Fuertes et al. (2010) for hydrochar prepared from the HTC of corn

3.2.1. AO and R6G removal by FWH-250 and FWH-350 Preliminary adsorption tests were performed in order to ascertain the dye removal capability of the prepared hydrochar materials. It can be observed in Fig. 1a that both the hydrochars were

Table 2 Characterization of surface elemental composition of FWH-250 and FWH-350 as determined by XPS. Sample

Elements

Binding energy (eV)

Assignment

Composition (At%)

FWH-250

C, N, and O

285.0 286.4 288.6 400.6 532.8 285.0 286.4 288.6

C1s CAC/CAH C1s CAO C1s OAC@O N1s NAH O1s OAC/O@C C1s CAC/CAH C1s CAO C1s OAC@O

60.1 14.9 4.7 4.8 15.6 75.2 18.0 6.8

285.0 286.6 288.5 400.0 532.2 285.0 286.6 288.5

C1s CAC/CAH C1s CAO C1s OAC@O N1s NAC O1s OAC/O@C C1s CAC/CAH C1s CAO C1s OAC@O

78.2 5.3 3.2 3.9 9.5 90.2 6.1 3.7

C only

FWH-350

C, N, and O

C only

314

G.K. Parshetti et al. / Bioresource Technology 161 (2014) 310–319

effective adsorbents for removal of AO and R6G from the aqueous solutions. However, FWH-250 had a higher dye adsorption potential than that of FWH-350. This might be because of the larger specific surface area and pore dimensions of FWH-250 than FWH-350 as inferred from BET surface area and porosity measurements. Another reason for the high dye uptake by FWH-250 can be attributed to its greater distribution of functional groups, as demonstrated by XPS studies. Overall, it is interesting to note that hydrochars with lower degree of carbonization contain significant graphitic portions, has more diverse functional groups and a greater specific surface area, and is thus more efficient for adsorption of dyes. Henceforth, batch experiments were performed to evaluate the dye adsorption characteristics of FWH-250 only.

3.2.2. Effect of pH The solution pH is an important monitoring parameter governing an adsorption process. Therefore, in the current investigation, the effect of pH on the dye adsorption efficiency of FWH-250 was examined. The results of the pH study are shown in Fig. 1b. It can be seen that the dye removal rate by FWH-250 increased with an increase in pH of the dye solutions and reached a maximum at around pH 8.0 (99.01% for AO and 92.54% for R6G). Further increase in pH (>8.0), did not significantly change the dye removal capacity. It can be explained that FWH-250 retained its remarkable high dye adsorption efficiency at alkaline pH primarily through electrostatic interaction between the cationic dye molecules and negatively charged surface. However, decreased pH dramatically reduced the adsorption capacity, which can be due to the protonation of the dye in acidic medium and the presence of excess H+ ions that compete with the cationic dye molecules for adsorption sites (Lin et al., 2013).

(a)

3.2.4. Adsorption isotherms In the present study, the data obtained from the equilibrium adsorption experiments, at temperatures of 293, 303 and 313 K, were analyzed using the Langmuir, Freundlich and D–R isotherms. The model constants, determined using linear regression analysis, are presented in Table 3. Based on the error function values as shown in Table 3, it is evident that the Langmuir isotherm showed an excellent fit to the experimental equilibrium data than the other isotherm models (see Supplementary information, Fig. S7a and b). The adsorption of AO and R6G occurs at specific homogeneous sites on the

AO R6G

80 60 40 20 0

FWH-250

(b)

100 90

Dye Removal (%)

Dye Removal (%)

100

3.2.3. Effect of temperature Fig. 1c and d show the dye adsorption profile of FWH-250 over the temperature range 293–313 K. The dye uptake capacity increased with rising temperature, indicating that dye removal by adsorption onto FWH-250 was kinetically controlled by an endothermic process. An increase in temperature results in an increased mobility of the dye molecules and a decrease in the retarding forces acting on the molecules. This possibly enhances the adsorption capacity of the adsorbent at higher temperatures. In addition, both AO and R6G showed a fast adsorption rate during the first 30 min of dye-adsorbent contact, which may be explained by an increased availability in the number of dye binding sites on the adsorbent surface. At higher contact time, the rate of adsorption appeared to slow down, gradually leading to equilibrium. Such a behavior can be attributed to the decrease in the number of binding sites available for adsorption. The equilibrium time for the maximum dye uptake was 60 min beyond which the amount of dye adsorbed did not show any time-dependent change. The rapid uptake of pollutants and establishment of equilibrium in a short period signified the potential of FWH-250 for its application in the removal of textile dyes.

80 70 60

AO R6G

50 40 30

FWH-350

2

3

4

5

6

7

8

9

10

pH 120

120

(c)

100

80

qt (mg/g)

qt (mg/g)

80 60 293 K 303 K 313 K

40 20 0

(d)

100

60 293 K 303 K 313 K

40 20

0

20

40

60

80

t (min)

100

120

140

0

0

20

40

60

80

100

120

140

t (min)

Fig. 1. (a) Results of preliminary adsorption tests for removal of Acridine Orange and Rhodamine 6G by FWH-250 and FWH-350 (Co = 50 mg L1; m/V = 0.05 g/0.1 L; agitation speed = 150 rpm). (b) Effect of solution pH on the dye removal efficiency of FWH-250 (Co = 50 mg L1; m/V = 0.05 g/0.1 L; T = 303 K; agitation speed = 150 rpm). Effect of temperature on the adsorption rate of (c) Acridine Orange and (d) Rhodamine 6G on FWH-250 (Co = 50 mg L1; pH = 8.00; m/V = 0.05 g/0.1 L; agitation speed = 150 rpm).

315

G.K. Parshetti et al. / Bioresource Technology 161 (2014) 310–319 Table 3 Isotherm constants for dye adsorption onto FWH-250 at different temperatures. Dye

T (K)

Langmuir qm (mg g

Freundlich

1

)

KL (L mg

1

)

R

2

2

v

KF (mg L

1

Dubinin–Radushkevich 1 1/n

) (L g

)

2

n

2

v

R

qm (mg g1)

b (mol2 J2)

E (kJ mol1)

R2

v2

8

AO

293 303 313

60.249 75.187 79.365

0.798 2.955 6.631

0.992 0.996 0.997

0.518 0.439 0.387

97.006 111.301 130.978

4.314 5.444 7.552

0.955 0.958 0.927

4.217 4.105 6.531

84.504 90.035 98.032

2.84  10 1.79  108 1.42  108

4.193 5.285 5.920

0.981 0.985 0.989

1.856 1.633 1.454

R6G

293 303 313

51.056 62.344 71.428

0.968 2.493 4.250

0.990 0.996 0.995

0.772 0.448 0.485

82.264 98.920 113.156

2.461 3.832 5.159

0.932 0.927 0.901

5.548 6.679 6.984

68.277 80.639 87.413

8.70  108 5.59  108 4.16  108

2.379 2.988 3.457

0.976 0.983 0.978

2.362 1.527 2.211

adsorbent surface and that once a dye molecule occupies a site no further adsorption can take place at that site, thereby forming a monolayer. The maximum adsorption capacity increased from 60.24 mg g1 at 293 K to 70.36 mg g1 at 313 K for AO while it increased from 51.05 mg g1at 293 K to 71.42 mg g1 at 313 K for R6G. Table 4 compares the maximum AO and R6G adsorption capacity of FWH-250 with some alternative adsorbents reported in the literature. Since the maximum amount of dye uptake per unit weight of adsorbent varies as a function of different process parameters (temperature, pH, contact time, etc.), the experimental conditions applied in those studies are also listed in Table 4. It is clearly evident that FWH-250 has a higher dye adsorption capacity than many of the other reported adsorbent materials. Therefore, the hydrothermal conversion of food waste to hydrochar is a thoughtful attempt to meet wastewater discharge and water reuse requirements. The Freundlich model constant n gives a measure of favorability of adsorption: favorable (1 < n < 10), irreversible (n = 1), unfavorable (n < 1). In the current investigation, the values of n suggest that the dye uptake by FWH-250 can be considered to be a favorable adsorption process. Additionally, an increase in the value of n with increasing temperature reflects stronger interaction between the adsorbate and the adsorbent at higher temperatures. Furthermore, the D–R isotherm constant, b, can be used to determine the mean free energy of adsorption per mole of the adsorbate (E, kJ mol1), which in turn gives an idea about the type of adsorption i.e. physisorption or chemisorption. E can be computed as follows:

1 E ¼ pffiffiffiffiffiffi 2b

ð1Þ

If the magnitude of E is between 8 and 16 kJ mol1, the adsorption process is supposed to proceed via chemisorption, while for values of E < 8 kJ mol1, the adsorption process is of physical nature. Based on this information, it can be said that the adsorption of AO and R6G onto FWH-250 is a simple physisorption process (Table 3). 3.2.5. Adsorption kinetics The experimental kinetic data of adsorption of dyes onto FWH250 were examined using the pseudo-first-order and pseudo-second-order equations. The model constants were calculated from the intercept and slope of the plots between log (qe–qt) versus t

and t/qt versus t for pseudo-first-order and pseudo-second-order, respectively. The parameters thus obtained along with the R2 and v2 values are listed in Table S1 (see Supplementary information). The exceedingly high R2 values and relatively low v2 values confirm that the dye adsorption data were well represented by pseudo-second-order kinetics over the entire adsorption period (see Supplementary information, Fig. S7c and d). Also, the theoretical qe,cal values showed a good agreement with the experimental qe,exp values for both AO and R6G, demonstrating the applicability of this model. The values of k2 increased with an increase in temperature (Table S1), suggesting that higher temperatures favor the dye uptake process by increasing the adsorption rate and capacity. 3.2.6. Activation parameters Ea for adsorption of AO and R6G by FWH-250 was computed from the slope of the Arrhenius plot (see Supplementary information, Fig. S8a) and is summarized in Table 5. The magnitude of Ea provides information about the type of adsorption mechanism. Low activation energies (5–40 kJ mol1) are characteristics of physical adsorption, while higher activation energies (40– 800 kJ mol1) indicate chemical adsorption (Jalil et al., 2012). In the present study, the obtained Ea values suggest that the uptake of AO and R6G by FWH-250 is a physical adsorption process. This finding is in agreement with the results from the D–R isotherm studies. Therefore, the affinity of FWH-250 for both dyes may be ascribed to van der Waals forces and electrostatic attractions between the surface functional groups and dye molecules. The Eyring equation (see Supplementary information, Eq. (S10)) was further applied to calculate the activation parameters (DH#, DS#, and DG#) of dye adsorption. DH# and DS# were estimated from the slope and intercept of the plot of ln (k2/T) versus 1/T (see Supplementary information, Fig. S8b). DG# at different temperatures was determined using Eq. (S11). The values of DH#, DS#, and DG# thus obtained are listed in Table 5. The large positive DG# values at all temperatures reflect the existence of an energy barrier for adsorption of AO and R6G. It further implies that an energy input is required for the reactant molecules to have enough kinetic energy in order to overcome this energy barrier and the adsorption process to take place (Mahmood et al., 2011). The positive values of DH# ascertain that considerable energy is required for adsorption of dyes by FWH-250. The negative DS# values

Table 4 Comparison of the maximum dye removal capacity of FWH-250 with other reported low-cost adsorbents. pH

AO

7.8–8.0 296 1.8–9.0 – Room temp. 1000–2000

Autohydrolyzed pine sawdust Magnetically modified fodder (Kluyveromyces fragilis) yeast cells Esterified soybean hull FWH-250

Temp. (K)

Conc. range (mg L1) Contact time (h) Adsorption capacity (mg g1) References

Dye Adsorbent

6.0 8.0

– 313

50–500 10–100

R6G Trichoderma harzianum mycelial waste 8.0 Almond shells 8.0 Activated carbon 7.0 FWH-250 8.0

303 – 333 313

10–50 100–1000 5–45 10–100

168 3

18.78 62.22

7 2

238.1 79.36

2 2 48 2

3.40 32.6 44.7 71.42

Sidiras et al. (2011) Safarik et al. (2007) Gong et al. (2008) This study Sadhasivam et al. (2007) Senturk et al. (2010) Annadurai et al. (2001) This study

316

G.K. Parshetti et al. / Bioresource Technology 161 (2014) 310–319

Table 5 Activation parameters, thermodynamic properties and isosteric heats for dye adsorption onto FWH-250. Activation parameters Dye

T (K)

DG# (kJ mol1)

DH# (kJ mol1)

DS# (J mol1 K1)

Ea (kJ mol1)

AO

293 303 313

78.942 80.734 82.520

26.614

178.615

29.131

R6G

293 303 313

68.685 70.315 71.946

20.903

163.079

23.420

Thermodynamic parameters Dye

T (K)

DGO (kJ mol1)

DHO (kJ mol1)

DSO (J mol1 K1)

AO

293 303 313

22.621 25.538 27.144

43.422

226.289

R6G

293 303 313

16.226 18.924 20.017

39.057

189.633

Isosteric heat of adsorption Dye

qe (mg g1)

DHX (kJ mol1)

R2

AO

5 10 15 20 25

92.182 93.136 94.296 95.677 97.371

0.989 0.991 0.985 0.983 0.988

R6G

5 10 15 20 25

70.430 71.832 73.551 75.712 78.515

0.990 0.988 0.989 0.987 0.989

indicate that the adsorption goes through the formation of an activated complex, suggesting that uptake of AO and R6G by FWH-250 is an associated mechanism. 3.2.7. Thermodynamic parameters The Gibb’s free energy change (DGO) for the adsorption of AO and R6G by FWH-250, calculated using Eq. (S13) (see Supplementary information), is listed in Table 5. DHO and DSO values were determined from the slope and intercept of the plot of DGO versus T and are also presented in Table 5. Negative values of DGO indicate that adsorption of dyes onto FWH-250 is spontaneous and feasible at all the studied temperatures. The increase in the absolute values of DGO with rising temperature suggests that adsorption of both dyes onto FWH-250 is more favorable at higher temperatures. The positive values of DHO further confirm that the adsorption reaction of AO and R6G is endothermic. Furthermore, the positive DSO values suggest increased randomness at the solid/solution interface with some structural changes in the adsorbate and the adsorbent and an affinity of FWH-250 towards AO and R6G. The positive DSO also corresponds to an increase in the degree of freedom at the solid–liquid interface. 3.2.8. Isosteric heat of adsorption In the present study, the isosteric heats of adsorption of AO and R6G were calculated at constant surface coverage (qe = 5, 10, 15, 20, and 25 mg g1) using the Clausius–Clapeyron equation (Eq. S15). For this purpose, the equilibrium liquid-phase dye concentration (Ce) at a constant amount of dye adsorbed was determined from the Langmuir adsorption isotherm parameters at different temperatures. DHX was then calculated from the slope of the plot of ln Ce versus 1/T for different amounts of AO and R6G onto FWH-250 (see Supplementary information, Fig. S9a and b). The R2 values of the isosteres and the corresponding DHX values are

shown in Table 5. Generally, DHX for the absolute physical adsorption is less than 80 kJ mol1 and for chemical adsorption it ranges between 80 and 400 kJ mol1 (Chowdhury et al., 2011). Therefore, in this study, the adsorption of R6G should be considered to be physical adsorption, while that of AO should be regarded as chemical adsorption but dominated by physical adsorption, since the DHX values were not significantly greater than 80 kJ mol1. In addition, DHX increased steadily with an increase in qe indicating that FWH-250 had heterogeneous surfaces. The dependence of DHX on surface coverage can be due to adsorbate–adsorbate interaction followed by adsorbate–adsorbent interaction. At lower qe values, adsorbate–adsorbate interactions are strong resulting in low heats of adsorption. With the increase in qe, adsorbate–adsorbate interaction decreases due to an increase in the number of available binding sites. Adsorption occurs on the most active sites resulting in high heats of sorption. The variation of DHX with surface loading can also be attributed to the possibility of having lateral interactions between the adsorbed dye molecules. 3.2.9. Regeneration and reuse Regeneration of the used adsorbent is of crucial importance as it helps to determine the reusability of an adsorbent which in turn contributes to evaluating the effectiveness and economic feasibility of the adsorption process. Desorption experiments were, therefore, performed by agitating (150 rpm) the dye-loaded adsorbent with 50 mL of ethanol at 303 K for 1 h. It was observed that for both dyes, the percentage desorption was almost 98%. This implies that the adsorption of AO and R6G onto FWH-250 is a reversible process and of physical nature. In order to examine the reusability of the adsorbent, adsorption/desorption cycles were carried out five times. The results indicate that FWH-250 can be used repeatedly without much loss in in the total adsorption capacity of the two dyes (Fig. 2).

317

G.K. Parshetti et al. / Bioresource Technology 161 (2014) 310–319

3.2.10. Treatment of simulated textile wastewater In an effort to examine the practical utility of FWH-250 in the treatment of real wastewater containing AO and R6G, further experiments were conducted to remove these dyes from simulated textile wastewaters prepared based on the composition of real textile wastewaters reported by Sahinkaya (2013) (see Supplementary information, Table S2). As much as 90.47% of AO and 84.19% of R6G could be removed from the simulated effluent by the adsorbent. Compared to the results obtained from controlled laboratory experiments using pure aqueous dye solutions, a decrease in adsorption capacity was observed, which could be due to the presence of other contaminants that compete with the dye molecules for a limited number of binding sites available on the surface of the adsorbent. However, the results obtained suggest that FWH250 can still be used as an effective adsorbent for removal of basic dyes from real wastewaters.

Adsorption Capacity (mg/g)

3.2.11. Dye removal mechanism Three important mechanisms, namely van der Waals forces, electrostatic interactions, and hydrogen bonding, have been proposed to explain the adsorption behavior of AO and R6G dyes onto FWH-250. van der Waals attraction forces are normally predominant for gas or vapor adsorption on the surface of solid substances, which may also be significant for adsorption of AO and R6G by FWH-250, as suggested by the calculated Ea values. However, considering only van der Waals forces in the process of dye adsorption may not be applicable as electrostatic interactions and hydrogen bonding can also be involved between the dye and the hydrochar. Molecular electrostatic interactions can occur in an adsorption process depending on the surface charge of the adsorbent and the degree of ionization and speciation of the adsorbate, which in turn is governed by the solution pH. In the present study, the observed pH-dependent adsorption of AO and R6G on FWH-250 suggests that electrostatic attraction forces exist at higher pH values resulting in high dye uptake capacity. Compared to van der Waals forces, the attachment of dye molecules to the surface functional moieties via electrostatic attraction is also much stronger and significant. Hydrogen bonding can also play an important role in adsorption when organic molecules or carbon-based adsorbents have certain functional groups such as ACOOH, AOH, and ANH2. XPS spectral analysis showed the existence of CAO, OAC@O, ACOOH and NAH groups in FWH-250. These groups can act as hydrogen-bonding donors and form hydrogen bonds with AO and R6G. Nevertheless, hydrogen bonding may not contribute significantly to the high

100

AO

adsorption capacity of FWH-250. This is mainly because the number of functional groups on FWH-250 is too limited to make them effective hydrogen-bonding donors. Moreover, the functional groups of FWH-250 can also form hydrogen bonds with water molecules. Such hydrogen-bonding interactions are much stronger than those between the dye molecules and functional groups on the surface of FWH-250 as predicted by Efremenko and Sheintuch (2006). In general, the dye adsorption mechanism is complex which might also involve other factors such as dispersive force, induction force and hydrophobic interactions. 3.2.12. ANN modelling In recent years, ANN has emerged as a very powerful tool to forecast the behavior of a given system, to design new large-scale processes and to analyze existing processes in many areas of science and engineering (Celekli and Geyik, 2011; Celekli et al., 2012). In environmental engineering, ANN has been successfully used to develop water treatment models since the prediction of the output water quality from any water treatment plant is very difficult because of the continuous change in the input water quality (Nayak et al., 2006). The main advantage of using the ANN technique is that it does not require the understanding of the complex nature of the process under consideration to be explicitly described in mathematical form (Nayak et al., 2006). In this work, a three layer ANN model, as illustrated in Fig. 3 was developed for predicting the dye adsorption behavior of FWH-250 under different experimental conditions. Operational parameters viz., pH (3.0–9.0), temperature (293–313 K), and reaction time (0–120 min) were used as inputs to the model. The equilibrium dye adsorption capacity was selected as the desired output from the network. A hyperbolic tangent–sigmoid transfer function (tansig) with back-propagation algorithm at hidden layer and a linear transfer function (purelin) at output layer were applied. The Levenberg–Marquardt back-propagation algorithm was used for network training. The complete experimental data were divided into three sets— training (50% of data), validation (25% of data) and test set (25% of data). The split of the data into training, validation and test subsets was carried out to estimate the performance of the neural network for prediction of ‘‘unseen’’ data that were not used for training. In this way, the generalization capability of the ANN model was assessed. Further, to avoid the scaling effect of the parameter values, all of the data (Xi) were converted to normalized values (Xnorm) in the 0.1–0.9 range as follows:

 X norm ¼ 0:8

X i  X min X max  X min

 þ 0:1

R6G

80 60 40 20 0

1

2

3

4

5

Cycle Number Fig. 2. Adsorption capacity of FWH-250 towards Acridine Orange and Rhodamine 6G during repeated adsorption/desorption cycles. (Co = 50 mg L1; pH = 8.00; m/ V = 0.05 g/0.1 L; T = 303 K; agitation speed = 150 rpm).

Fig. 3. Topology of the ANN architecture.

ð2Þ

318

G.K. Parshetti et al. / Bioresource Technology 161 (2014) 310–319

1.0

and R6G from contaminated water. Optimization of the carbonization degree is crucial for the proficiency as adsorbent. Thermodynamic investigations suggest that the adsorption is spontaneous, endothermic and physical in nature. The developed ANN-based model can be used in designing an automated dye wastewater treatment plant. Overall, the present findings are of great importance for the environmental application of hydrochars derived from food waste for removal of dyes from large volumes of aqueous solutions.

(a)

MSE

0.8 0.6 0.4 0.2 0.0

Acknowledgements 0

2

4

6

8 10 12 14 16 18 20 22 24 26

The authors are thankful to the National Environment Agency (ETRP Grant) and National University of Singapore for the financial support extended to this study. The authors are also thankful to A⁄STAR’s Institute of Materials Research and Engineering (IMRE), Singapore for technical assistance.

Number of neurons in hidden layer

100

100

(b)

(c)

80 Appendix A. Supplementary data qe, pred (mg/g)

qe, pred (mg/g)

80

60

60

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2014.03. 087.

40

40

References

20 20

R2=0.9815

40

60

80

100

qe, exp (mg/g)

20

R2=0.9838

20

40

60

80

100

qe, exp(mg/g)

Fig. 4. Effect of the number of neurons in the hidden layer on network performance (a) and comparison of the experimental qe values with those predicted by the ANN model for adsorption of (b) Acridine Orange and (c) Rhodamine 6G by FWH-250.

All ANN calculations were carried out using Neural Network Toolbox of MATLAB Version 7.9 (R2009b). To determine the optimum number of hidden nodes, a series of topologies was used, in which the number of nodes was varied from 1 to 25. Each topology was repeated three times to avoid random correlation due to random initialization of the weights. The mean square error value (MSE) was used as the error function to measure the performance of the network according to the following equation:

MSE ¼

N 1X ðq  qe;exp Þ2 N i¼1 e;pred

ð3Þ

Fig. 4a presents the relationship between network error and the number of neurons in the hidden layer. MSE was found to be minimum just about 19 neurons. Therefore, the number of neurons in the hidden layer was selected as 19. Fig. 4b and c shows a comparison of the experimental and predicted qe values obtained by using the designed neural network model. The plots in these figures have high R2 values of 0.981 and 0.983 for AO and R6G, respectively, suggesting the reliability of the developed ANN model. Furthermore, the MSE for AO was calculated to be 0.029 and for R6G it was about 0.031. This further confirms that the neural network model reproduces the experimental equilibrium dye uptake capacity of FWH-250, within the experimental ranges adopted for model fitting. 4. Conclusion This study demonstrates that hydrochars prepared by HTC of food waste at 250 °C is an efficient adsorbent for removal of AO

Annadurai, G., Juang, R.-S., Lee, D.-J., 2001. Adsorption of Rhodamine 6G from aqueous solutions on activated carbon. J. Environ. Sci. Health A 36, 715–725. Becker, R., Dorgerloh, U., Helmis, M., Mumme, J., Diakité, M., Nehls, I., 2013. Hydrothermally carbonized plant materials: patterns of volatile organic compounds detected by gas chromatography. Bioresour. Technol. 130, 621– 628. Cao, X., Ma, L., Gao, B., Harris, W., 2009. Dairy-manure derived biochar effectively sorbs lead and atrazine. Environ. Sci. Technol. 43, 3285–3291. Celekli, A., Birecikligil, S.S., Geyik, F., Bozkurt, H., 2012. Prediction of removal efficiency of Lanaset Red G on walnut husk using artificial neural network model. Bioresour. Technol. 103, 64–70. Celekli, A., Geyik, F., 2011. Artificial neural network (ANN) approach for modelling of removal of Lanaset Red G on Chara contraria. Bioresour. Technol. 102, 5634– 5638. Chen, B., Chen, Z., 2009. Sorption of naphthalene and 1-naphthol by biochars of orange peels with different pyrolytic temperatures. Chemosphere 76, 127–133. Cabrera, A., Cox, L., Spokas, K.A., Celis, R., Hermosín, M.C., Cornejo, J., Koskinen, W.C., 2011. Comparative sorption and leaching study of the herbicides fluometuron and 4-chloro-2-methylphenoxyacetic acid (MCPA) in a soil amended with biochars and other sorbents. J. Agric. Food Chem. 59, 12550–12560. Chowdhury, S., Mishra, R., Saha, P., Kushwaha, P., 2011. Adsorption thermodynamics, kinetics and isosteric heat of adsorption of malachite green onto chemically modified rice husk. Desalination 265, 159–168. Efremenko, I., Sheintuch, M., 2006. Predicting solute adsorption on activated carbon: phenol. Langmuir 22, 3614–3621. Fuertes, A.B., Arbestain, M.C., Sevilla, M., Macia-Agullo, J.A., Fiol, S., Lopez, R., Smernik, R.J., Aitkenhead, W.P., Arce, F., Macias, F., 2010. Chemical and structural properties of carbonaceous products obtained by pyrolysis and hydrothermal carbonisation of corn stover. Aust. J. Soil Res. 48, 618–626. Gong, R., Sun, J., Zhang, D., Zhong, K., Zhu, G., 2008. Kinetics and thermodynamics of basic dye sorption on phosphoric acid esterifying soybean hull with solid phase preparation technique. Bioresour. Technol. 99, 4510–4514. Hu, B., Wang, K., Wu, L., Yu, S.-H., Antonietti, M., Titirici, M.-M., 2010. Engineering carbon materials from the hydrothermal carbonization process of biomass. Adv. Mater. 22, 1–16. Inyang, M., Gao, B., Yao, Y., Xue, Y., Zimmerman, A., Pullammanappallil, P., Cao, X., 2012. Removal of heavy metals from aqueous solution by biochars derived from anaerobically digested biomass. Bioresour. Technol. 110, 50–56. Jalil, A.A., Triwahyono, S., Yaakob, M.R., Azmi, Z.Z.A., Sapawe, N., Kamarudin, N.H.N., Setiabudi, H.D., Jaafar, N.F., Sidik, S.M., Adam, S.H., Hameed, B.H., 2012. Utilization of bivalve shell-treated Zea mays L. (maize) husk leaf as a low-cost biosorbent for enhanced adsorption of malachite green. Bioresour. Technol. 120, 218–224. Kijima, M., Hirukawa, T., Hanawa, F., Hata, T., 2011. Thermal conversion of alkaline lignin and its structured derivatives to porous carbonized materials. Bioresour. Technol. 102, 6279–6285. Lin, L., Zhai, S.-R., Xiao, Z.-Y., Song, Y., An, Q.-D., Song, X.-W., 2013. Dye adsorption of mesoporous activated carbons produced from NaOH-pretreated rice husks. Bioresour. Technol. 136, 437–443. Mahmood, T., Saddique, M.T., Naeem, A., Mustafa, S., Zeb, N., Shah, K.H., Waseem, M., 2011. Kinetics and thermodynamic study of Cd(II), Co(II) and Zn(II) adsorption from aqueous solution by NiO. Chem. Eng. J. 171, 935–940.

G.K. Parshetti et al. / Bioresource Technology 161 (2014) 310–319 Nahman, A., de Lange, W., Oelofse, S., Godfrey, L., 2012. The costs of household food waste in South Africa. Waste Manage. (Oxford) 32, 2147–2153. Nayak, P.C., Rao, Y.R., Sudheer, K.P., 2006. Groundwater level forecasting in a shallow aquifer using artificial neural network approach. Water Resour. Manag. 20, 77–90. NEA Waste Statistics and Overall Recycling, 2012. . Parshetti, G.K., Hoekman, S.K., Balasubramanian, R., 2013. Chemical, structural and combustion characteristics of carbonaceous products obtained by hydrothermal carbonization of palm empty fruit bunches. Bioresour. Technol. 135, 683–689. Rajagopal, R., Lim, J.W., Mao, Y., Chen, C.L., Wang, J.Y., 2013. Anaerobic co-digestion of source segregated brown water (feces-without-urine) and food waste: for Singapore context. Sci. Total Environ. 443, 877–886. Sadhasivam, S., Savitha, S., Swaminathan, K., 2007. J. Environ. Manage. 85, 155–161. Safarik, I., Rego, L.F.T., Borovska, M., Mosiniewicz-Szablewska, E., Weyda, F., Safarikova, M., 2007. New magnetically responsive yeast-based biosorbent for the efficient removal of water-soluble dyes. Enzyme Microb. Technol. 40, 1551– 1556. Sahinkaya, S., 2013. COD and color removal from synthetic textile wastewater by ultrasound assisted electro-Fenton oxidation process. J. Ind. Eng. Chem. 19, 601–605. Senturk, H.B., Ozdes, D., Duran, C., 2010. Biosorption of Rhodamine 6G from aqueous solutions onto almond shell (Prunus dulcis) as a low cost biosorbent. Desalination 252, 81–87.

319

Sevilla, M., Macia-Agullo, J.A., Fuertes, A.B., 2011. Hydrothermal carbonization of biomass as a route for the sequestration of CO2: chemical and structural properties of the carbonized products. Biomass Bioenerg. 35, 3152–3159. Sidiras, D., Batzias, F., Schroeder, E., Ranjan, R., Tsapatsis, M., 2011. Dye adsorption on autohydrolyzed pine sawdust in batch and fixed-bed systems. Chem. Eng. J. 171, 883–896. Sun, K., Ro, K., Guo, M., Novak, J., Mashayekhi, H., Xing, B., 2011. Sorption of bisphenol-A, 17a-ethinyl estradiol and phenanthrene on thermally and hydrothermally produced biochars. Bioresour. Technol. 102, 5757–5763. Sun, K., Gao, B., Ro, K.S., Novak, J.M., Wang, Z.Y., Herbert, S., Xing, B.S., 2012. Assessment of herbicide sorption by biochars and organic matter associated with soil and sediment. Environ. Pollut. 163, 167–173. Wang, L., Guo, Y., Zou, B., Rong, C., Ma, X., Qu, Y., Li, Y., Wang, Z., 2011. High surface area porous carbons prepared from hydrochars by phosphorioc acid activation. Bioresour. Technol. 102, 1947–1950. Yao, Y., Gao, B., Inyang, M., Zimmerman, A.R., Cao, X., Pullammanappallil, P., Yang, L., 2011. Biochar derived from anaerobically digested sugar beet tailings: characterization and phosphate removal potential. Bioresour. Technol. 102, 6273–6278. Zheng, W., Guo, M., Chow, T., Bennett, D.N., Rajagopalan, N., 2010. Sorption properties of greenwaste biochar for two triazine pesticides. J. Hazard. Mater. 181, 121–126. Zubieta, C.E., Messina, P.V., Schulz, P.C., 2012. Photocatalytic degradation of acridine dyes using anatase and rutile TiO2. J. Environ. Manage. 101, 1–6.