RSC Advances

PAPER

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article OnlineView Journal | View Issue

Catalytic activati

aNanotechnology and Catalysis Research C

zaira.chowdhury@um.edu.my; zaira.chow

79676959; +60 1 02675621bKumoh National Institute of Technology (K

† These authors contributed equally to th

Cite this: RSC Adv., 2016, 6, 102680

Received 10th December 2015Accepted 5th July 2016

DOI: 10.1039/c5ra26189a

www.rsc.org/advances

102680 | RSC Adv., 2016, 6, 102680–1

on and application of micro-spherical carbon derived from hydrothermalcarbonization of lignocellulosic biomass: statisticalanalysis using Box–Behnken design

Z. Z. Chowdhury,*a S. B. Abd Hamid,†a Md. M. Rahmana and R. F. Rafique†b

In this study, activated carbon was produced by physico-chemical activation of hydrothermally carbonized

(HTC) material derived from the dried stem of Corchorus olitorius commonly known as jute (JS), using

potassium hydroxide (KOH) as an activation agent. The activation process was optimized using Box–

Behnken factorial design (BBD), with outcome of 29 different experiments under predefined conditions.

Four different parameters, namely activation temperature (x1), activation time (x2), ratio of char to KOH

(x3) and CO2 flow rate (x4), were optimized with respect to their influence on maximum removal

percentage for divalent cations of Cu(II) (Y1) and carbon yield (Y2). All the four process parameters had

strong positive effect on adsorption capacity up to a certain limit; beyond which it started to decline.

The specific surface area of the hydrochar (HTC) was enhanced substantially after the activation process.

Scanning electron microscopy (SEM) revealed that the morphology of the JS based hydrochar (JSC)

changed significantly after KOH impregnation and activation under the flow of CO2 gas. The Langmuir

maximum monolayer adsorption capacity for Cu(II) cations was 31.44 mg g�1. Equilibrium isotherm data

were well followed by Freundlich and Temkin models. Due to an increase in temperature, the Langmuir

maximum monolayer adsorption capacity, qm (mg g�1) and Freundlich constant, KF increased

successively, representing an endothermic nature of the adsorption.

1. Introduction

The imminent shortage of non-renewable fossil fuels hascommenced the search for waste lignocellulosic biomass toyield bio-carbon for wastewater treatment. As the most abun-dant component in nature, lignocellulosic biomass is regardedas an almost innite bio-resource which has been used forthousands of years to produce structured as well as surfacefunctionalized carbon materials for versatile applications.1,2

Thus extensive attention has been given for synthesizing bio-char and hydrochar from conventional pyrolysis and hydro-thermal carbonization (HTC) process.3–7 HTC is a comparativelymodern carbonization process where thermochemical conver-sion of biomass is performed at comparatively low temperatures(150–350 �C) under autogenously created pressures. The HTCprocess takes place through a series of hydrolysis, condensa-tion, decarboxylation, and dehydration reactions using deion-ized water as “green” solvent or acid/base or metallic ion as

enter (NANOCAT), Malaysia. E-mail: dr.

dhury76@gmail.com; Tel: +60 3

IT), Gumi, South Korea

is work.

02694

catalyst.8,9 Application of HTC process is very common owing toits simplicity in design, less expenses, and energy efficiencies.Additionally, it can also be considered as “green technique”when water is used as catalytic medium for carbonizationprocess.8 Throughout the hydrothermal carbonization, thelinkages between the carbohydrate rich fractions that is cellu-lose and hemicellulose with lignin are disrupted. Accordinglya complex cyclic-addition, aldol reactions as well as condensa-tion reactions take place to produce hydrochar with relativelyhigh carbon content. A number of lignin based oligomers withsome sugars are dissolved in liquid phase which later on can beused for extraction of some value added products like bio-plastics and biofuels. Since milder conditions are sufficient forinitial carbonization, the process is energy efficient. Further-more, it does not involve pre-drying of the raw biomass.10

Overall the process is exothermic which releases ample heat tosubsidize one-third of the energy required for completecarbonization process.10 The process is environmentally benignas it is spontaneous whereby the huge proportion of carbonpresent in original biomass is retained aer the carbonizationprocess.1 It does not produce toxic gases and tar like pyrolysiswhich needs additional separation technique.11 HTC char hasgood self-binding properties which is advantageous for pellet-isation.12 Finally, the economical feasibility of HTC process

This journal is © The Royal Society of Chemistry 2016

Table 1 Independent variables for activation of JSC with their codedand actual level

Variables Code Units

Coded and actualvariable levels

�1 0 +1

Temperature x1 �C 600 800 1000Activation time x2 Hour 1 2 3Ratio (hydrochar : KOH) x3 — 0.5 1.25 2CO2 gas ow rate x4 mg L�1 50 75 100

Paper RSC Advances

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article Online

compared to other carbonization process is that, it can beconducted without any gas ow.12

Previous literature stated that, hydrochar prepared frombiomass residues have less surface area with micro-poroustexture.1,13 The surface morphological features can be differentlike nano-spheres or nano-bers based on preparation condi-tion.13 It becomes hydrophilic due to presence of surface func-tional groups. Hydrochar contains polyfuranic structures wherethe furan rings are inter-connected by aliphatic chain withoxygen groups. However with increasing temperature, thearomaticity increases but the aliphatic and oxygen containinggroups decreases.14 Due to existence of oxygenated linkinggroups inside the nal carbon matrix, it contains a lot of reac-tive functional groups and the structure becomes more hydro-philic than the pyrolyzed one. Thus the usage of hydrochar asinitial precursors to produce activated carbon can offer anexpedient approach to yield porous carbonaceous materialswith desired surface chemistry to adsorb cations from water.Very few reports are available to observe the activation effect ofHTC materials with different activating agent and study thephysicochemical characteristics of the porous activated carbon.This leads to extensive research for developing promisingcarbonaceous materials for treating waste effluents before dis-charging.15 Presence of various metals such as manganese (Mn),mercury (Hg), lead (Pb), cadmium (Cd), arsenic (As), copper(Cu) etc. in waste water is potentially dangerous. Among theseheavy metals, copper is identied as one of the most toxicmetals due to its' bio-accumulating properties. Several indus-tries such as paper and pulp, fertilizer, wood preservatives,reneries, metal cleaning and plating bath etc. are generatingcopper containing waste water. Excessive consumption ofcopper than the necessary limit may cause severe mucosalirritation, renal and hepatic damage, capillary damage andgastrointestinal irritation. This might be the reason for prob-able necrotic changes in kidney and liver of the living organ-isms. According to World Health Organization (WHO),maximum acceptable limit for Cu(II) concentration in drinkingwater should be 1.5 mg L�1.16 Thus it is essential to developsuitable adsorbent materials to eliminate excess Cu(II) metalsfrom aqueous effluents before domestic supply.

The conventional method for preparing carbonaceousadsorbent materials by keeping one factor varying while theother variables involved at an unstipulated persistent level doesnot illustrate the collective inuence of all the variables. In thatcase, the optimization process of a multivariable system typi-cally outlines single-variable at a predened phase. This needsa large number of experiments to be performed and fails toillustrate the interaction effects between the factors. This makesthe overall process time consuming to dene the optimumlevels. Furthermore, it is not reliable enough for economicalprocess design. Different types of statistical design can providemathematical models, which can elucidate the interactioneffects among the variables.17 Among these methods, responsesurface methodology (RSM) based on Box Behnken Design(BBD) is one of the most appropriate methods employedrecently in various elds. RSM is a collection of mathematicaland statistical methods which can be employed to appraise the

This journal is © The Royal Society of Chemistry 2016

virtual importance of numerous affecting variables even in themanifestation of complicated interactions. Taking these intoaccount, we have explored the feasibility of using hydrochar(JSC) prepared from dried jute stalk powder (JS) to prepareactivated carbon (JSCAC) for elimination of Cu(II) cations fromaqueous phase. In BBD design, the experimental variables aretypically varied between two levels. This approach is consideredsystematic and describes nonlinear dependencies for a partic-ular variable and interactions between the variables.18 BBD isslightly more efficient than the central composite design (CCD)but much more efficient than the three-level full factorialdesigns.19 In this research, BBD design is applied to access theeffect of temperature, time, KOH ratio and CO2 gas ow rate foractivation of JSC. Adsorption experiments were conducted usingthe resultant carbon (JSCAC). Batch equilibrium data wereanalyzed using Langmuir, Freundlich and Temkin isothermmodel.

2. Materials and methods2.1. Materials

2.1.1. Preparation of feedstock (JS). The primary biomassfeedstock used for these experiments was dried jute stalk (JS). Asreceived, the JS sample was vigorously washed with hot deion-ized water to remove dust and dried in an oven at 80 �C for 24hours. To promote more effective homogeneous mixing andstirring, the sample was crushed using a ball mill. Thepowdered sample was sieved to remove large particles, washed,dried and stored in a sealed bottle for additional processing.

2.1.2. Hydrochar (JSC) preparation. The waste lignocellu-losic residues of JS are not utilized properly; usually they aredecomposed or burnt aer extraction of the ber. Duringpreparation of hydrochar 7 g of JS powder wasmixed with 70mLof deionized water and stirred until it was homogeneouslymixed. Thus prepared sample was transferred into a Teon-lined autoclave (100 mL). The temperature was ramped to 220�C and carbonized for 4 h. The resulting char sample (JSC) waswashed several times with de-ionized water until a neutral pHwas obtained. The hydrochar (JSC) thus obtained was notinitially activated and stored in air-tight vessels for preliminarycharacterization studies.

2.1.3. Activation of hydrochar (JSC) to prepare activatedcarbon (JSCAC). JSC was further impregnated with predenedratio of KOH based on basic design matrix of BBD (Table 2).During impregnation, 100 mL deionized water was mixed with

RSC Adv., 2016, 6, 102680–102694 | 102681

RSC Advances Paper

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article Online

the char (JSC) and KOH pellets. The resultant mixture washeated at 90–100 �C for 4 h. The sample was kept overnight inan oven at 90 �C. The dried sample was further pyrolyzed toinitiate activation process to yield nal product with enlargedsurface area. The activation process was carried out in thepresence of CO2 gas ow at pre-dened conditions (Tables 1 and2). The activated carbon sample (JSCAC) thus obtained waswashed vigorously in order to remove unreacted KOH attachedwith the surface. The carbon samples got aer two stage ofphysicochemical activation were dried and stored for furthercharacterization and isotherm studies.

2.2. Methods

2.2.1. Isotherm and thermodynamic characterization forsingle solute system. Requisite amount of CuSO4$5H2O saltwere dissolved in 1000 mL of deionized water to prepare thestock solution of Cu(II) cations having concentration of 1000 mgL�1. The stock solution was further diluted to obtain test solu-tions having concentration of 50, 60, 70, 80, 90 and 100 mg L�1

for isotherm studies. For removal percentage calculation,highest solution concentration, 100 mg L�1 was agitated with0.2 g of JSCAC sample (Table 2). Adsorption experiment wascarried out at (30� 1) �C around 250 rpm. Prior to agitation, the

Table 2 Experimental responses for Cu(II) removal percentages (Y1) and

Standard order RunType ofpoint Reaction variables (actua

Order No. PointTemperature(x1) (�C)

Tim(h)

1 1 IBFact 600 12 2 IBFact 1000 13 10 IBFact 600 34 11 IBFact 1000 35 17 IBFact 800 26 3 IBFact 800 27 4 IBFact 800 28 5 IBFact 800 29 6 IBFact 600 210 7 IBFact 1000 211 8 IBFact 600 212 9 IBFact 1000 213 12 IBFact 800 114 13 IBFact 800 315 14 IBFact 800 116 15 IBFact 800 317 16 IBFact 600 218 17 IBFact 1000 219 18 IBFact 600 220 19 IBFact 1000 221 20 IBFact 800 122 21 IBFact 800 323 22 IBFact 800 124 23 IBFact 800 325 24 Center 800 226 25 Center 800 227 27 Center 800 228 28 Center 800 229 29 Center 800 2

102682 | RSC Adv., 2016, 6, 102680–102694

solution pH was adjusted to 5.5 for isotherm characterization ofthe equilibrium system. The amount of Cu(II) cations adsorbedonto the surface of JSCAC was calculated by the followingequation:20,21

qe ¼ ðC0 � CeÞVW

(1)

here, qe (mg g�1) denotes the quantity of cation loaded atequilibrium; C0 is the initial concentration of the targetedcation; Ce (mg L�1) is the residual equilibrium concentrationsof metal ions at liquid phase; V (L) represents the volume of thesingle solute solution; and W (g) is the mass of JSCAC used. Forcalculating removal percentages at different experimentalcondition based on Table 2, following equation was used:

% removal ¼ ðC0 � CeÞC0

� 100 (2)

2.2.2. Experimental design. Box Behnken Design (BBD)based on RSM method was used here to examine the impact ofdifferent effective factors on Cu(II) cations adsorption efficiency.Simultaneously it can determine the most appropriate permu-tation of variables ensuing in maximum removal efficiency withhighest process yield. The Box Behnken Design (BBD) with four

carbon yield (Y2) for JSCAC

l factors) Removal% Yield%

e (x2) Ratio(x3)

Flow rate (x4)(mg L�1) Y1 Y2

1.25 75 82.33 46.881.25 75 86.32 30.761.25 50 83.99 39.991.25 50 84.99 26.660.50 100 85.88 41.092 100 91.88 40.890.50 50 85.78 35.882 50 86.09 31.781.25 100 81.89 45.661.25 100 88.23 31.091.25 75 81.66 41.891.25 75 85.03 26.760.50 75 84.89 38.990.50 75 86.98 36.772 75 90.09 41.992 75 88.78 33.980.50 75 80.99 48.480.50 75 88.29 29.062 50 83.46 42.892 50 87.20 28.791.25 100 86.90 43.881.25 100 89.88 39.891.25 75 87.89 34.771.25 75 81.67 32.881.25 75 90.88 37.441.25 75 91.09 37.481.25 75 90.77 37.051.25 75 91.08 37.291.25 75 91.98 36.99

This journal is © The Royal Society of Chemistry 2016

Table 3 Statistical parameters for model accuracy test

Statistical parameters

Removal% Yield%

Y1 Y3

Standard deviation, SD% 1.04 1.10Correlation coefficient, R2 0.952 0.978Adjusted R2 0.904 0.965Mean 86.79 37.14Coefficient of variation, CV 1.19 2.96Adequate precision 14.19 33.17

Paper RSC Advances

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article Online

factors, involving ve replicates at the center point for assess-ment of residual errors was used to obtain the basic designmatrix (Table 2). It was used to develop a second-order poly-nomial model with linear, interaction and quadratic terms. Forthe activation of JSC, four important effective parameters suchas temperature (x1), time (x2), ratio (x3) and CO2 ow rate (x4)were considered as the independent variables, while theremoval efficiencies of Cu(II) (Y1) and yield (Y2) were theresponses (dependent variable). Overall the optimizationprocess contains three major steps of statistically designing andperforming the experiments, determining the coefficients ina mathematical model and analyzing the adequacy of the modelaccording to following eqn (3):27

Y ¼ f(x1, x2, x3, x4.xn) (3)

where, Y is the response of the system and xi is the variablesunder investigation for the process. Temperature was variedfrom 600 to 1000 �C, time from 1–3 h, char with KOH ratio 0.5 to2 and CO2 ow rate was changed from 50 to 100 mg L�1. Thelow, middle, and high levels of each variable were indicated as�1, 0, and +1, respectively. It is essential to nd a proper esti-mation for the precise functional relationship between theindependent variables and the selected responses for optimi-zation.21 The number of experiments (N) required for BBDdesign is expressed by:

N ¼ 2k(k � 1) + C0 (4)

where k is number of factors and C0 is the number of centralpoints. This method is less time consuming, when k is not toolarge.16 The order of experimental run was randomized and theresponses were used to develop empirical models that cancorrelate the responses of removal efficiencies with yield usinga second-degree polynomial equation as given by eqn (5):

Y ¼ b0 þXn

i¼1

b1x1 þXn

i¼1

b11xi2 þ

Xn

i¼1

Xn

j. 1

bijxixj (5)

where, Y is the predicted response, b0 the constant coefficient,b1 the linear coefficients, bij the interaction coefficients, bii thequadratic coefficients and xi, xj are the coded values of theadsorption.27 For four variables, the suggested number ofexperiments at the center point is ve and the total number ofexperiments (N) required are 29 (Table 2). Table 2 summarizesthe coded and actual levels of four independent variablestogether with the responses obtained for BBD experimentaldesign matrix.

2.2.3. Characterization. JS, JSC and JSCAC were character-ized by several analytical techniques to identify the surface prop-erties of the prepared samples. The surfacemorphological changesof nal products were characterized using a eld emission scan-ning electron microscope (FE-SEM) (Zeiss SUPRA 35VP; Germany).The BET surface area, micro pore, meso pore volume and pore sizedistribution of the samples were analyzed using TriStar II BETSurface Area Analyzer. Prior to nitrogen gas adsorption, thesamples were outgassed under vacuum at 300 �C for 4 h to removemoisture content. Surface area and pore diameter was calculated

This journal is © The Royal Society of Chemistry 2016

by Brunauer–Emmett–Teller (BET) method, whereas microporevolume was obtained by the t-plot method. The crystalline struc-ture of the samples was analyzed using X-ray diffraction (XRD,Bruker AXSD8 Avance, Germany) at 40 kV and 40 mA using Cu-Karadiation sources. Thermogravimetric analysis coupled witha differential thermal analyzer (DTA) (Mettler Toledo Star SW901,Japan) was carried out to determine the thermal stability of thesamples under a 5 mLmin�1 nitrogen ow. In the TGA analysis, 7mg of each sample was heated under a N2 ow at 1000 �C witha heating rate of 5 �C min�1. Ultimate analysis (ElementalAnalyzer-PerkinElmer, Series II 2400) was carried out to calculatethe percentage of carbon, hydrogen, and nitrogen in JS, JSC andthe nal carbon, JSCAC.

3. Results and discussion3.1. Statistical analysis for model validation

The level of process variables with the design matrix withoutcome of 29 numbers of experiments and the outputresponses are shown in Tables 1 and 2 respectively. Five repli-cate runs were performed at the center point for the estimationpure error.

Three tests are necessary to estimate the accuracy of theprepared model for statistical analysis. In this context, threedifferent tests explicitly sequential model sum of squares, lackof t tests and model summary statistics were observed in thisresearch. Cubic model was not suggested for removalpercentage (Y1) and yield (Y2) due to insufficient points forcalculating the coefficients for this kind of the model. Based onsequential model sum of the squares and model summarystatistics, quadratic model showed the best t with the experi-mental data having the lowest standard deviation, the highestcorrelation coefficient R2 and adjusted R2 values for Cu(II)removal percentages (Y1). 2FI model containing constant termwith positive sign, variables (x1, x2, x3 and x4) and interactionterms (x1x2, x2x3, x3x4 and x4x1) with co-efficient was proposedby the soware for yield (Y2) calculation. The model articulatedby eqn (6), represents the removal percentage of Cu(II) cationsby prepared carbon JSCAC (Y1) in terms of their coded values.The nal yield of the carbon sample (Y2) is expressed by eqn (7).Both the response equations were derived as a function oftemperature (x1), retention time (x2), hydrochar with KOH ratio(x3) and the rate CO2 gas ow during the activation process (x4).

RSC Adv., 2016, 6, 102680–102694 | 102683

Fig. 1 Studentized residuals versus actual values for experimental run (a) removal percentages, Y1 of Cu(II) cations (b) yield percentages (Y2) ofJSCAC.

RSC Advances Paper

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article Online

% Cu(II) removal ¼ +92.16 + 2.15x1 � 0.18x2 + 1.22x3 � 1.38x4� 0.75x1x2 � 0.89x1x3 � 0.74x1x4� 0.85x2x3 � 2.30x2x4 � 1.42x3x4� 4.66x1

2 � 2.12x22 � 1.42x3

2 � 2.36x42

(6)

% yield ¼ +37.19 � 7.76x1 � 2.26x2 � 0.59x3 � 3.48x4 + 0.70x1x2+ 0.62x1x3 + 0.67x1x4 � 1.45x2x4 � 0.97x3x4 (7)

Table 4 Analysis of variance (ANOVA) and lack of fit test for removal% o

Source Sum of squares Degree of freedom

Model 298.91 14x1 55.21 1x2 0.38 1x3 17.98 1x4 22.80 1x1

2 140.93 1x2

2 29.15 1x3

2 13.08 1x4

2 36.17 1x1x2 2.24 1x1x3 3.17 1x1x4 2.21 1x2x3 2.89 1x2x4 21.16 1x3x4 8.09 1Residuals 15.05 14Lack of t 14.14 10Pure error 0.91 4

102684 | RSC Adv., 2016, 6, 102680–102694

The coefficients containing independent variables oftemperature (x1), time (x2), ratio of KOH (x3) and CO2 ow rate(x4) denote the inuence of that specic factor on the responsesof removal percentage of Cu(II) (Y1) and yield (Y2). Multiplicationvalues between the two factors and others with second orderterms exhibit the interaction and quadratic effect, respec-tively.24,25 A positive sign in front of the constant terms indicatesa synergistic effect, whereas a negative sign indicates anantagonistic effect.26 The high values of regression coefficient,

f Cu(II) cations onto JSCAC (Y1)

Mean square F value Prob > F Comments

21.35 19.86 <0.0001 Signicant55.21 51.36 <0.00010.38 0.35 0.5626

17.98 16.73 0.001122.80 21.21 0.0004

140.93 131.10 <0.000129.15 27.12 0.001013.08 12.17 0.036036.17 33.64 0.00102.24 2.08 0.17133.17 2.95 0.10812.21 2.05 0.17402.89 2.69 0.1234

21.16 19.68 0.00068.09 7.53 0.01581.081.41 6.19 0.0470 Signicant0.23

This journal is © The Royal Society of Chemistry 2016

Table 5 Analysis of variance (ANOVA) and lack of fit test for yield% (Y2)

Source Sum of squares Degree of freedom Mean square F value Prob > F Comments

Model 951.40 10 95.19 78.63 <0.0001 Signicantx1 721.99 1 721.99 596.67 <0.0001x2 61.20 1 61.20 50.58 <0.0001x3 4.24 1 4.24 3.50 0.0777x4 145.39 1 145.39 120.16 <0.0001x1x2 1.95 1 1.95 1.16 0.2209x1x3 1.56 1 1.56 1.29 0.2707x1x4 1.78 1 1.78 1.47 0.2406x2x3 8.38 1 8.38 6.93 0.0169x2x4 1.10 1 1.10 0.91 0.3525x3x4 3.80 1 3.80 3.14 0.0932Residuals 21.78 18 1.21Lack of t 21.57 14 1.54 29.30 0.0025 SignicantPure error 0.21 4 0.053

Paper RSC Advances

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article Online

R2 for all the two responses (Table 3) indicated that the pre-dicted results for model tting were reasonably similar with theexperimental observations. The R2 value for eqn (6) and (7) are0.952 and 0.978 which is nearer to unity reecting better suit-ability of the developed models. The values obtained for co-efficient of variation (CV) and standard deviations were smallwhich shows reproducibility of the developed models. Thevalues of adequate precision were calculated based on signal tonoise ratio.24,25 For effective model simulation, adequate preci-sion should be larger than 4. The adequate precision observedfor percentage Cu(II) removal and yield were 14.19 and 33.77.Thus it can be concluded that the experimental data obtainedhere is statistically signicant to navigate the design.26

Fig. 1a and (b) illustrate the studentized residuals versus theactual values obtained for each experimental run for removalpercentage (Y1) and yield (Y2) respectively. The data obtainedaer residual analysis are arbitrarily distributed about zero.This demonstrates that the variance of the experimental inter-pretations is almost constant for 29 experimental run. All of thedata points observed are found to fall within the range of +3 to�3. This conrms that response transformation is not neededfor the experimental design applied here.27

Fig. 2 3-D response surface contour plots of the combined effects of (a)Cu(II) cations onto JSCAC (Y1) when the other variables are at center po

This journal is © The Royal Society of Chemistry 2016

Tables 4 and 5 illustrate the result of ANOVA analysis forremoval percentage (Y1) and yield (Y2) respectively, where the F-test revealed that these regressions are statistically signicant at95% condence level.

F-Test values representing the deviation in data about themean were determined to check the competence of the models.The model F-value for removal percentage of Cu(II) cations andyield were 19.86 and 78.63 respectively which inferred that thesemodels were signicant. However, the values of Prob > F valuesfor all these two responses were less than 0.05, which speciedthat the model terms considered here for response analysiswere signicant.23,24 For the response of removal percentage, Y1,the linear factors of temperature (x1), ratio (x3), ow rate (x4) aswell as their quadratic terms (x1

2, x32, x4

2) were signicant. Theinteraction terms of time with ow rate (x2x4) and ratio withow rate (x3x4) were also the signicantmodel terms. Comparedto temperature (x1) and ow rate (x4), other linear terms of time(x2) and ratio (x3) had moderate inuence on removalpercentage (Y1). The interaction inuence of time and CO2 owrate (x2x4) was more prominent (owing to the highest value of F)than the other interaction factors related to the percentageremoval, Y1.

temperature (x1) and time (x2) (b) time (x2) and ratio (x3) on % removal ofints.

RSC Adv., 2016, 6, 102680–102694 | 102685

Fig. 3 3-D response surface contour plots of the combined effects of (a) temperature (x1) and ratio (x3) (b) temperature (x1) and CO2 flow rate (x4)on % removal of Cu(II) cations onto JSCAC (Y1) when the other variables are at center points.

RSC Advances Paper

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article Online

For yield percentages (Y2), activation temperature (x1), acti-vation time (x2), and CO2 ow rate (x4) were signicant to theresponse. However, the interaction inuence of x2x3 and x3x4had greatest impact on yield percentages. This phenomenoncan be observed from Table 2 (run 9 and 17). Experimental run 9and 17 were using same temperature of 600 �C with residencetime of 2 hours. In run 17, comparatively lesser ratio (0.50) withow rate (75 mg L�1) were used than run 9 (ratio 1.25, ow rate100). Interaction effect of ratio and ow rate (x3x4) was prom-inent to give lesser yield of 45.66% in run 9 compared to run 17(48.89%). Like removal percentages (Y1), temperature (x1) hadthe highest impact on yield (Y2) calculation.

3.2. Inuence of functional variables on Cu(II) uptake

The physicochemical characteristics of an ideal activatedcarbon should ensure high removal efficiencies. In that case,the carbon sample (JSCAC) should have a large BET surface area

Fig. 4 3-D response surface contour plots of the combined effects of (aJSCAC (Y2) when the other variables are at center points.

102686 | RSC Adv., 2016, 6, 102680–102694

with appropriate pore size distribution which can facilitate forthe liquid phase adsorption process. The inuence of the fourvariables on the removal percentage of Cu(II) cations wereinvestigated by RSM using three-dimensional response surfacewith two dimensional contour plots. From the response surfaceanalysis, it was observed that all the four variables had cumu-lative effect in elevating the removal efficiencies (Y1) up toa certain level.

Fig. 2(a) shows the interactions between temperature (x1)and time (x2) on removal efficiencies (Y1) when the other vari-ables of ratio (x3) and CO2 ow rate (x4) were xed at centerpoint (x3 ¼ 1.25, x4 ¼ 75 mg L�1). Fig. 3b illustrates the effect oftime (x2) and ratio (x3) onto removal percentages (Y1) whentemperature (x1) and CO2 ow rate (x4) were xed at center point(x1 ¼ 800 �C, x4 ¼ 75 mg L�1). The values on the x and Y-axis arethe real values.

Fig. 2(a) and (b) show that higher temperature and longerretention time enhances the removal efficiencies up to a certain

) temperature (x1) and time (x2) (b) time (x2) and ratio (x3) on % yield of

This journal is © The Royal Society of Chemistry 2016

Fig. 5 3-D response surface contour plots of the combined effects of (a) temperature (x1) and ratio (x3) (b) temperature (x1) and CO2 flow rate (x4)on % yield of JSCAC (Y2) when the other variables are at center points.

Paper RSC Advances

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article Online

level giving concave pattern of the RSM plots. The highest valueof removal percentage recorded was about 91.98% (run 6:temperature (x1) ¼ 800 �C, time (x2) ¼ 2 h ratio (x3) ¼ 1.25 andow rate (x4) ¼ 75 mg L�1). This is because, during the activa-tion process, the proliferation in temperature and residencetime would release more volatile component as a result ofintensied dehydration and elimination reactions. Longerretention time would facilitate more contact between charmatrix with KOH and CO2 gas to enhance the surface area.26,28,29

Fig. 3a shows the interactions between temperature (x1) andratio (x3) on removal efficiencies (Y1) when the other variables oftime (x2) and CO2 ow rate (x4) were xed at center point (x2 ¼ 2h, x4 ¼ 75 mg L�1). Fig. 3b illustrates the effect of temperature(x1) and CO2 ow rate (x4) onto removal percentages (Y1) whentime (x2) and ratio (x3) were xed at center point (x2 ¼ 2 h, x3 ¼1.25). From the curvature of these plots, it reveals that thetemperature (x1) and ratio (x3) of impregnating base is propor-tional with the removal percentages. It appears that, up toa xed limit; the increase of temperature and ratio of activatingbase in presence of CO2 gas ow will enhance the reaction ratebetween KOH and hydrochar resulting more porous texturewhich is desirable for adsorption. Thus when ration (x3) wasincreased with temperature up to a certain extent removalpercentage (Y1) of Cu(II) cations was increased. Aer certainlimit, enhancing temperature (x1) and ratio (x3) was unfav-ourable for adsorption process. This was expected as too hightemperature or KOH concentration might destroy some surface

Table 6 Numerical optimization for process design

Temperature Time Ratio Gas ow Percentage remo

(x1) (x2) (x3) (x4) (Y1)

(�C) (h) — mg L�1 Predicted E

752.93 1.86 2.00 51.23 90.55 9

This journal is © The Royal Society of Chemistry 2016

functional groups. It might destroy the porous texture ofprepared carbon by thermal annealing which will abolish thewalls of pores reducing the pore volume.30,31

Previous literature showed that, the activation time hadmoderate to minimum effect on enhancing the surface area andpore structure of the activated carbon prepared from cassavapeel whereas the pore distribution in terms of micro and meso-pore percentages were strongly inuenced by KOH impregna-tion ratio and temperature.28,29 Too short activation time andtoo lower temperature might not be sufficient to enhance thesurface area up to a greater extent.28,29 Lower temperature willallow the volatile matters to reside inside the pores impedingthe pore development. At very high temperature and longerresidence time with high ratio of KOH, the carbon skeletonmight burn out to produce more ash which also can contributein lower removal efficiency. Too long contact time for activationis not suggested as it might increase the liquid yield, lessencarbon yield with successive decrease in removal efficiency. Inaddition to that, longer contact time also evidently meansgreater expenses as the high activation temperature needs beretained for a longer period of time.32 This trend was alsoobserved in previous literature during preparing activatedcarbon for removal of Zn(II) cations using oil palm fronds.33

Progressive increase in temperature (x1) and ow rate (x4)facilitate the volatilization process of the low molecular weightorganic substances which consequently results in highersurface area of the prepared carbon (JSCAC).34–36 Increment in

val Percentage yield

(Y2)

xperimental Error Predicted Experimental Error

1.75 1.30 43.02 42.87 0.348

RSC Adv., 2016, 6, 102680–102694 | 102687

Table 7 Isotherm types based on separation factor RL

Value of RL Magnitude Types of isotherm

RL > 1 Greater than one UnfavorableRL ¼ 1 Equal to one Linear0 < RL < 1 Between zero to one FavorableRL ¼ 0 Zero Irreversible

RSC Advances Paper

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article Online

temperature enhances the activation process which is endo-thermic and leads to the removal of disorganized carbon.37 Thiswill expose more aromatic sheets of carbon to react with CO2

gas.36 It was observed that, an increase in ow rate (x4) untilmoderate temperature (x1) enhances the removal efficiencies.But, at a point where the ow rate (x4) or the retentiontemperature (x1) was too high, the removal efficienciesdecreased. This is due to the extreme burn-off of carbon. Thus itcan be concluded that increasing temperature (x1) and ow rate(x4) beyond the optimum limit will destroy the pore structureresulting lower BET surface area with less removal efficiencies(Y1). However, at very low CO2 ow rate and reaction time, theremoval percentage was lower. This happens due to insufficientcontact time between KOH impregnated hydrochar (JSC) andthe CO2 gas which adversely affects the quality of the preparedcarbon (JSCAC).38

3.3. Inuence of functioning variables on yield percentages

Fig. 4a reveals the cumulative inuence of two independentfactors temperature (x1) and time (x2) on percentage yield ofcarbon (Y2) where KOH ratio (x3) and CO2 gas ow (x4) were xedat center point. Fig. 4b showed the combined impact of time (x2)and KOH ratio (x3) over yield percentage (Y2), where thetemperature (x1) and ow rate (x4) were kept at center point.

In this research, all the four parameters investigated hadsynergistic inuence on percentage yield (Y2). It was found thatwhen the activation temperature (x1), retention time (x2), ratio(x3) and ow rate of CO2 (x4) were increased, the yieldpercentage decreased signicantly. Temperature (x1) hadhighest impact on yield percentage (Y2) which was earlier shownby the highest F value of 543.08, as shown by Table 5, whereasactivation ratio (x3) was less noteworthy compared to activationtemperature (x1), time (x2) and ow rate (x4). The lowest yieldwas obtained when temperature was at 1000 �C for 3 h usingKOH ratio 1.25 under the ow rate of 75 mg L�1 (run 22) asshown by Table 2. The enhancement of temperature, contacttime, ratio and ow rate will intensify the diffusion of basic

Table 8 Model constants for Langmuir, Freundlich and Temkin equatio

CationTemp.(�C)

Langmuir isotherm

qmax

(mg g�1)KL

(L mg�1) R2 RL

Cu(II) 30 31.44 0.0037 0.98 0.7250 31.74 0.0024 0.88 0.8170 31.84 0.0014 0.93 0.88

102688 | RSC Adv., 2016, 6, 102680–102694

medium inside the hydrochar matrix leading to conversion ofcarbon with ash residues. This would further initiate thedisintegration of the crystalline region of hydrochar and formcarbon with amorphous texture which is desirable for adsor-bent preparation. This observation is consistent with theprevious work carried out for activation of cassava peel,28 whereactivation temperature was found to play a vigorous role on thecarbon yield, whereas activation time showed least impact onyield.

However, using higher reaction temperature (x1) with higherratio (x3) would result in lower yield percentages (Y2) (Fig. 5a).The increase in temperature and ow rate together wouldgradually enhance more dehydration and elimination reactionsto release additional volatile organic fractions from hydrocharmatrix as gas and liquid products rather than solid activatedcarbon (Fig. 5b).

3.4. Optimizing operating conditions

For successful process design in case of adsorption process,relatively high removal efficiencies are expected for targetedpollutant under optimum condition. As observed from the basicdesign matrix that, when the values of process variables wereincreasing, percentage removal was increasing up to a certainlimit whereas yield was decreasing and vice versa. The numer-ical optimization menu was selected using Design Expert so-ware version 8.0.6 (STAT-EASE Inc., Minneapolis, LTS). In orderto optimize the adsorption process, the targeted goal was set asmaximum values for removal percentage and yield, while thevalues of the temperature, time, ratio and CO2 ow rate were setin the ranges being studied as depicted earlier by previousliterature.2 The error was evaluated and provided by Table 6.

3.5. Application of JSCAC for liquid phase adsorption:equilibrium isotherm studies

Isotherm parameters were determined using linear modelequations, i.e. Langmuir, Freundlich, and Temkin for temper-ature 30, 50 and 70 �C. Maximum monolayer sorption quantity,qm (mg g�1) and Langmuir constant KL (L mg�1) demonstratingthe binding energy for adsorption uptake were determined forCu(II) adsorption onto JSCAC using Langmuir model eqn (8):39

qe ¼ KLCe

1þ aLCe

(8)

The nonlinear form of eqn (8) was linearized to give eqn (9):

ns at different temperatures

Freundlich isotherm Temkin isotherm

KF (mg g�1)(L mg�1)1/n 1/n R2 B

KT

(L mg�1) R2

8.56 0.45 0.98 7.51 2.18 0.9810.76 0.42 0.92 5.65 6.42 0.9612.46 0.51 0.97 5.70 8.39 0.99

This journal is © The Royal Society of Chemistry 2016

Fig. 6 X-ray diffractograms of jute stick powder (JS), hydrochar (JSC)and hydrochar activated carbon (JSCAC).

Fig. 7 SEM images of (a) jute stick powder (JS), (b) hydrochar (JSC) (c)

Paper RSC Advances

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article Online

Ce

qe¼ 1

qmaxKL

þ 1

qmax

Ce (9)

The dimensionless factor, RL, of the Langmuir model can becalculated using eqn (10):

RL ¼ 1

1þ KLC0

(10)

In this present research, RL values are determined for thehighest initial concentration (100 mg L�1). RL was determinedto get specic idea about the types of isotherm, as illustrated byTable 7.40

Freundlich isotherm provides insight about the surfaceheterogeneity. It shows the multilayer sorption characteris-tics of Cu(II) cations onto JSCAC. The empirical equationshows that the reactive sites over the carbon sample arescattered exponentially with the heat of the adsorptionmethod.41 The nonlinear form of Freundlich equation isgiven by:

qe ¼ KfCe1/n (11)

The above equation was linearized to determine theparameters and is represented by eqn (12):

ln qe ¼ ln Kf þ 1

nln Ce (12)

here, Kf (mg g�1) is the affinity factor of the cations towards thecarbon sample (JSCAC) and 1/n represents the intensity of theadsorption.

Equilibrium data was further analyzed by the Temkinisotherm and are expressed by:42

qe ¼ RT

bln KTCe (13)

This journal is © The Royal Society of Chemistry 2016

Eqn (13) can be linearized as:

qe ¼ RT

bln KT þ RT

bln Ce (14)

here, RT/b ¼ B (J mol�1) is Temkin constant; which is relatedwith the heat of the sorption process; whereas KT (L g�1) revealsthe binding constant at equilibrium contact time; R (8.314 Jmol�1 K�1) is the universal gas constant; and T (K) is theabsolute solution temperature.42

Freundlich and Temkin model showed higher R2 valuesreecting better linearity for Cu(II) cations at temperature 30, 50and 70 �C. The process was favorable as the magnitude forLangmuir separation factor, RL, and the Freundlich exponent, 1/n obtained here was below one. The Langmuir maximummonolayer capacity, qm was increasing with temperature (Table8). This observation is consistent with successive increase ofFreundlich affinity factor, KF with the increase of temperature.This reected that higher temperature had ensured moreremoval percentage of Cu(II) cations from waste water. Overall itshowed the endothermic nature of sorption.16 This phenom-enon was reported earlier for Mn(II) cations sorption onto rawand acid hydrolyzed corncob biomass and mangosteene fruitpeel based activated carbon.22,42

3.6. physicochemical characterization

3.6.1. X-ray diffraction (XRD) analysis. The X-ray diffractionpatterns of the raw jute stick powder (JS), hydrochar (JSC) andactivated carbon (JSCAC) samples are illustrated by Fig. 6. Twonarrow, sharp peaks at 2q values of around 16 and 22� for theraw biomass (JS) were observed due to presence of crystallinecellulose in the sample.43 For hydrochar (JSC), the intensities ofthese two peaks were decreased slightly, demonstrating partialcellulose degradation.44 The XRD pattern of JSC reected thathydrothermal carbonization could not disrupt the crystallinetexture of cellulose completely. Similar phenomenon wasobserved for sawdust, wheat straw and corn stalk basedhydrochar and KOH treated hydrochar by previous

hydrochar activated carbon (JSCAC).

RSC Adv., 2016, 6, 102680–102694 | 102689

Fig. 8 Nitrogen adsorption isotherms of (a) hydrochar (JSC) (b) hydrochar activated carbon (JSCAC).

RSC Advances Paper

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article Online

researchers.45 However, the crystalline cellulose was almostdestroyed aer activation at higher temperature. Higher pyrol-ysis temperatures yielded a broader peak at 2q values of around22 to 24� in JSCAC sample. This indicated the development ofatomic order in the activated sample at higher temperatures.44

The absence of sharp peaks demonstrated the amorphoustexture of the carbon sample.46 This was expected as amorphousphase materials show better performance as adsorbent due totheir high surface area and more active sites over the surface.47

Previous ndings also showed presence of amorphous carbonin coconut coir based activated carbon.48

3.6.2. SEM images analysis. The hydrochar obtained aercarbonization was brown in color which represents partiallycarbonized product. SEM images of raw JS powder, JSC andJSCAC are shown by following gures of Fig. 7a–c respectively.The images revealed interesting changes aer hydrothermalcarbonization (Fig. 7b). SEM image of raw biomass (JS) showedcomparatively rough, nonporous texture with some cracks andcrevices (Fig. 7a). In contrast to that, SEM images obtained aercarbonization (Fig. 7b) reveal substantial changes in surfacemorphology and texture. Lot of sphere like micro particles wasdeposited over the larger particles. This indicates that celluloseand hemicellulose were somewhat degraded during hydro-thermal carbonization. Thus somemicrospheres of carbon werefound to form.49 However, non-saccharide component ofbiomass i.e. lignin is chemically stable than the carbohydratefraction (cellulose and hemicellulose). It was partially degradedand the original skeleton of the particles were preserved.9,14

Aer activation, some conchoidal cavities with irregular shapedpores were observed over the surface of JSCAC (Fig. 7c). The

Table 9 Pore characteristics of hydrochar (JSC) and hydrochar activate

Sample

Specic surface area (m2 g�1) P

SBET SMicro SMeso Sext V

JSC 3.789 2.709 1.08 2.345 0JSCAC 856.89 698.76 158.13 559.98 0

102690 | RSC Adv., 2016, 6, 102680–102694

results demonstrated that a drastic morphological change hadtaken place due to KOH activation process in presence of CO2

gas. Similar observation was reported for activation of euca-lyptus wood based hydrochar.45 The pore structure was clearlyvisible aer activation which means that KOH impregnationand CO2 gas ow had sweep away the unwanted impuritiesblocking the pores (Fig. 7c).45 Consequently pore volume withsurface area was extensively increased aer the activationprocess. This was further supported by BET analysis of JSC andJSCAC samples.

3.6.3. BET surface area analysis. Fig. 8a and b display theN2 adsorption isotherms for hydrochar (JSC) and activatedhydrochar using KOH (JSCAC) while the textural parameters asdetermined from these isotherms are listed in Table 9.

According to IUPAC classication, the isotherm obtained forhydrochar (JSC) can be classied as intermediate between type Iand type II isotherm. This indicates a combination of micro-porous–mesoporous structure with less surface area.10 Aeractivation in presence of KOH and CO2 gas, the pore volumeincreased drastically with enhanced surface area 856.89 m2 g�1.The knee of the isotherm obtained for JSCAC appeared moreround at low pressure which is usually associated with widemicroporous texture.49 Around relative pressure of about 0.9,the curve showed upward trend. The amount of N2 gasadsorption was increased with increasing pressure. The shapeof the curve indicated a type II isotherm (Fig. 8b). This indicatedthe presence of micropores with certain volume of mesopores inJSCAC.25 The values displayed by Table 9 showed increase ofporosity and specic surface area of activated sample (JSCAC)compared to the hydrochar sample (JSC). Due to incomplete

d carbon (JSCAC)

ore volume (cm3 g�1)Pore diameter(nm) VMicro/VTotal%Total VMicro VMeso

.256 0.208 0.048 1.304 0.813

.889 0.689 0.200 5.896 0.775

This journal is © The Royal Society of Chemistry 2016

Fig. 9 (a) TGA and (b) DTG profile of (a) jute stalk powder (JS) (b) hydrochar (JSC) (c) hydrochar activated carbon (JSCAC).

Table 10 Ultimate analysis of JS, JSC and JSCAC

Ultimate analysis JS JSC JSCAC

Percentage carbon 39.81 51.55 87.04Percentagehydrogen

6.99 4.56 1.51

Percentage nitrogen 1.61 1.57 0.80Percentage oxygen 49.00 39.56 8.91Others 2.59 2.76 1.74H/C 0.175 0.088 0.017O/C 1.23 0.767 0.102

Paper RSC Advances

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article Online

carbonization during HTC process, the pores were partiallyblocked by decomposition products. Activation brings selectiveremoval of those components. In presence of CO2 gas ow, KOHmight react with these structural components over the surfaceof JSC and induced further depolymerization to increase thepore volume.

The surface area (SBET) of the hydrochar sample enhancedsignicantly aer activation. Due to activation in presence ofhigh temperature, KOH took part in enlargement the pore sizeand the pore diameter became 5.896 nm for the activatedsample (JSCAC). Synergistic effect of temperature, time KOHratio with CO2 gas ow had removed the unsaturated carbonatoms and gasied the micro-pore walls to produce more meso-pores.50

3.6.4. Thermogravimetric (TGA/DTG) analysis. Thethermal stability of (JS), hydrochar (JSC) and activated hydro-char (JSCAC) was investigated using the Thermogravimetricmethod. The thermal degradation curve for all the samplesshows several stages. The rst degradation step at approxi-mately 70 to 130 �C corresponds to the evaporation of adsorbedwater. The second degradation step for JS proceeds in thetemperature range of 200 to 300 �C for hemicellulose and 300 to400 �C for cellulose degradation respectively.

This journal is © The Royal Society of Chemistry 2016

It was reported earlier that, the degradation of hemi-celluloses partially overlaps with cellulose degradation in caseof biomass and lignin degradation proceeds within thetemperature range of 200 to 800 �C.51 Compared to JS, JSC andJSCAC showed higher thermal stability. As illustrated by Fig. 9,DTGmax of JS, JSC and JSCAC were at 346 �C, 356 �C and 360 �C,respectively. Among the three major constituents in lignocel-lulosic residues, the thermal stability of hemicelluloses was thelowest.52 Hydrothermal carbonization (HTC) removed hemi-celluloses and the amorphous region of the cellulose from JS,leading to an increase of thermal stability in JSC sample. Aeractivation, the sample (JSCAC) contains larger proportion ofstable form of carbon particles which is more heat resistant.This phenomenon is evident from elemental analysis alsowhere the percentage carbon in JS was initially 39.81% but aersuccessive treatment of HTC process and activation, it became87.04%.

3.6.5. Ultimate analysis. Table 10 lists the ultimate analysisof biomass (JS), hydrochar (JSC) and hydrochar activated carbon(JSCAC). Aer hydrothermal carbonization, carbon contentinside the char matrix was increased, whereas hydrogen andoxygen content was decreased.52 It can be seen that both H/Cand O/C ratio decreased aer carbonization and activation.52,53

The H/C ratio decreases due to aromatization process duringhydrothermal carbonization. O/C ratio decreases for decarbox-ylation reaction during the process whereas both H/C and O/Cratios decrease due to dehydration reaction.54

4. Conclusions

This study explored the potential of utilizing dried jute stalk (JS)based hydrochar (JSC) to prepare activated carbon. Theprepared carbon sample (JSCAC) was used to eliminate Cu(II)cations from waste water. Activation conditions for the

RSC Adv., 2016, 6, 102680–102694 | 102691

RSC Advances Paper

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article Online

hydrochar were optimized. Preparation factors were optimizedusing response surface methodology (RSM) based on BoxBehnken Design (BBD). Activation temperature (x1), time (x2),ratio (x3) and CO2 ow rate (x4) had substantial inuence oneach response of removal efficiency as well as production yield.Model validation was done under optimum conditions. It wasperceived that the proposed models for Cu(II) removal (Y1) andyield percentages (Y2) were signicant. The deviation errorsobtained between predicted and experimental results for boththe models were around 1.30% and 0.348%, respectively. Theactivated carbon was amorphous with meso-porous texture.Aer activation, the carbon had surface area of 856.89 m2 g�1.The xed carbon content in the resultant carbon increasedsubstantially up to 87.04%. The equilibrium data were followedby linear regression analysis at different temperature usingLangmuir, Freundlich and Temkin isotherm showingmaximum monolayer sorption capacity about 31.44 mg g�1 forCu(II) cations. Better tting of Freundlich model showedheterogeneous surface with multilayer formation characteris-tics of Cu(II) cations onto JSCAC. The increasing trend ofLangmuir monolayer adsorption capacity with Freundlichaffinity factor with temperature clearly revealed that theadsorption process was endothermic in nature. The statisticaldesign methodology used for this research provides a feasibleand competent approach for Cu(II) cations removal whileminimizing the number of experiment.

Conflicts of interest

The authors declare no conict of interest.

Acknowledgements

The authors would like to thank BKP (BK054-2015) and HIRgrant (H-21001-F0032) of University Malaya, Malaysia and KoreaResearch Foundation, Korea (Research project number: 2013-218-017) for their cordial support in completing this work.

References

1 B. Hu, K. Wang, L. H. Wu, S. H. Yu, M. Antonietti andM. M. Titirici, Engineering carbon materials from thehydrothermal carbonization process of biomass, Adv.Mater., 2010, 22, 813–828.

2 Z. Z. Chowdhury, N. M. Julkapli, A. M. Ali,R. A. M. A. H. Abdul and A. Y. Wageeh, Microwave AssistedSynthesis, Characterization and Application of GraphiticBio-carbon using Two Level Factorial Designs, BioResources,2016, 11(2), 3637–3659.

3 B. Hu, S. H. Yu, K. Wang, L. Liu and X. W. Xu, Functionalcarbonaceous materials from hydrothermal carbonizationof biomass: an effective chemical process, Dalton Trans.,2008, 40, 5414–5423.

4 Z. Z. Chowdhury, M. Z. Karim and M. A. Ashraf, Inuence ofCarbonization Temperature on Physicochemical Propertiesof Biochar derived from Slow Pyrolysis of Durian wood

102692 | RSC Adv., 2016, 6, 102680–102694

sawdust (Durio zibethinus), BioResources, 2016, 11(2),3356–3372.

5 M. Sevilla and A. B. Fuertes, Chemical and structuralproperties of carbonaceous products obtained byhydrothermal carbonization of saccharides, Chem.–Eur. J.,2009a, 15, 4195–4203.

6 M. Sevilla and A. B. Fuertes, The production of carbonmaterials by hydrothermal carbonization of cellulose,Carbon, 2009, 47, 2281–2289.

7 M. M. Titirici, A. Thomas and M. Antonietti, Back in theblack: hydrothermal carbonization of plant material as anefficient chemical process to treat the CO2 problem?, NewJ. Chem., 2007a, 31, 787–789.

8 M. M. Titirici, A. Thomas, S. H. Yu, J. O. Muller andM. Antonietti, A direct synthesis of mesoporous carbonswith bicontinuous pore morphology from crude plantmaterial by hydrothermal carbonization, Chem. Mater.,2007b, 19, 4205–4212.

9 M. Sevilla and A. B. Fuertes, The production of carbonmaterials by hydrothermal carbonization of cellulose,Carbon, 2009b, 47, 2281–2289.

10 S. Roman, J. M. Valente Nabais, B. Ledesma, J. F. Gonzalez,C. Laginhas and M. M. Titirici, Production of low-costadsorbents with tunable surface chemistry by conjunctionof hydrothermal carbonization and activation processes,Microporous Mesoporous Mater., 2013, 165, 127–133.

11 S. Xiu, A. Shabazi, V. Shirley and D. Cheng, Hydrothermalpyrolysis of swine manure to bio-oil: effects of operatingparameters on products yield and characterization of bio-oil, J. Anal. Appl. Pyrolysis, 2010, 88, 73–79.

12 A. Amaya, N. Medero, N. Tancredi, H. Silva and C. Deiana,Activated carbon briquettes from biomass materials,Bioresour. Technol., 2007, 98, 1635–1641.

13 M. M. Titirici, M. Antonietti and N. Baccile, Hydrothermalcarbon from biomass: a comparison of the local structurefrom poly- to monosaccharides and pentoses/hexoses,Green Chem., 2008, 10, 1204–1212.

14 C. Falco, J. M. Sieben, N. Brun, M. Sevilla, T. Van derMauelen, E. Morallon, D. Cazorla-Amoros andM. M. Titirici, Hydrothermal carbons from hemicellulose-derived aqueous hydrolysis products as electrode materialsfor supercapacitors, ChemSusChem, 2013, 6, 374–382.

15 Z. Z. Chowdhury, S. B. Abd Hamid, R. Das, M. R. Hasan,S. M. Zain, K. Khalid and M. N. Uddin, Preparation ofcarbonaceous adsorbents from lignocellulosic biomass andtheir use in removal of contaminants from aqueoussolution, BioResources, 2013, 8(4), 6523–6555.

16 Z. Z. Chowdhury, S. M. Zain, A. K. Rashid and M. S. Islam,Preparation and characterizations of brous activatedcarbon from Kenaf for equilibrium adsorption studies ofcopper from waste water, Korean J. Chem. Eng., 2012, 29(9),1187–1195.

17 H. S. B. Abd, Z. Z. Chowdhury and M. Z. Karim, CatalyticExtraction of Microcrystalline Cellulose (MCC) from Elaeisguineensis using Central Composite Design (CCD),BioResources, 2014, 9(4), 7403–7426.

This journal is © The Royal Society of Chemistry 2016

Paper RSC Advances

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article Online

18 M. B. Baskan and A. Pala, A statistical experiment designapproach for arsenic removal by coagulation process usingaluminum sulfate, Desalination, 2010, 254, 42–48.

19 M. E. Fernandez, B. Ledesma, S. Roman, P. R. Bonelli andA. L. Cukierman, Development and characterization ofactivated hydrochars from orange peels as potentialadsorbents for emerging organic contaminants, Bioresour.Technol., 2015, 183, 221–228.

20 Z. Z. Chowdhury, M. Hasan, H. S. B. Abd, E. M. Shamsudin,S. M. Zain and K. Khalid, Catalytic Pretreatment of Biocharresidues derived from Lignocellulosic Feedstock forEquilibrium Studies of Manganese, Mn(II) cations fromaqueous solution, RSC Adv., 2015, 5(9), 6345–6356.

21 Z. Z. Chowdhury, S. M. Zain, A. K. Rashid and K. Khalid,Batch and Fixed bed Adsorption Studies of Lead(II) cationsfrom aqueous solutions onto granular activated carbonderived from Mangostana garcinia shell, BioResources, 2012,7(3), 2895–2915.

22 V. Gunaraj and N. Murugan, Application of response surfacemethodologies for predicting Weld base quality insubmerged arc welding of pipes, J. Mater. Process. Technol.,1999, 88, 266–275.

23 J. C. Marcos, L. P. Fonseca, M. T. Ramalho andJ. M. S. Cabral, Application of surface response analysis tothe optimization of penicilin acylase purication inaqueous two-phase system, Enzyme Microb. Technol., 2002,31, 1006–1014.

24 F. Karacan, U. Ozden and S. Karacan, Optimization ofmanufacturing conditions for activated carbon fromTurkish by chemical activation using response surfacemethodology, Appl. Therm. Eng., 2007, 27, 1212–1218.

25 M. Z. Karim, Z. Z. Chowdhury, S. B. Abd Hamid andM. E. Ali,Optimizing the pretreatment condition for a-cellulose usinglewis acid catalyst, Sci. Adv. Mater., 2016, 8(3), 534–544.

26 Z. Z. Chowdhury, S. M. Zain, A. K. Rashid and K. Khalid,Process Variables Optimization for Preparation andCharacterization of Novel Adsorbent from LignocellulosicWaste, BioResources, 2012, 7(3), 3732–3754.

27 D. C. Montgomery, Design and Analysis of Experiments, JohnWiley and Sons Inc, New York, 2001, 5th edn, pp. 427–500.

28 Y. Sudaryanto, S. B. Hartono, W. Irawaty, H. Hindarso andS. Ismadji, High surface area activated carbon preparedfrom cassava peel by chemical activation, Bioresour.Technol., 2006, 97, 734–739.

29 Q. Cao, K. C. Xie, Y. K. Lv and W. R. Bao, Process effects onactivated carbon with large specic surface area from corncob, Bioresour. Technol., 2006, 97, 110–115.

30 E. Arenas and F. Chejne, The effect of the activating agentand temperature on the porosity development of physicallyactivated coal chars, Carbon, 2004, 42, 2451–2455.

31 A. Zolin, A. Jensen, K. Dam-Johansen and L. S. Jensen,Inuence of experimental protocol on activation energy inchar gasication: the effect of thermal annealing, Fuel,2001, 80, 1029–1032.

32 M. K. B. Gratuito, T. Panyathanmaporn, R.-A. Chumnanklang,N. Sirinuntawittaya and A. Dutta, Production of activatedcarbon from coconut shell: optimization using response

This journal is © The Royal Society of Chemistry 2016

surface methodology, Bioresour. Technol., 2008, 99, 4887–4895.

33 M. A. Zahangir, A. M. Suleyman and K. Noraini, Production ofactivated carbon from oil palm empty fruit bunches for removalof zinc, Twelh International Water Technology Conference,IWTC12, Alexandria, Egypt, 2008, pp. 373–383.

34 P. J. M. Suhas Carrott and M. M. L. Ribeiro Carrott, Lignin –

from natural adsorbent to activated carbon: a review,Bioresour. Technol., 2007, 98, 2301–2312.

35 R. C. Bansal, J.-B. Donnet and F. Stoeckli, Active Carbon,Marcel Dekker, Inc, New York, 1998, p. 482.

36 A. C. Lua, F. Y. Lau and J. Guo, Inuence of pyrolysisconditions on pore development of oil-palm-shell activatedcarbons, J. Anal. Appl. Pyrolysis, 2006, 76, 96–102.

37 H. Muhlen and K. H. Van Heek, in Porosity in Carbons:Characterization and Applications, ed. W. Patrick Jr,Halstead Press, London, 1995, p. 230.

38 A. C. Lua and J. Guo, Activated carbon prepared from oilpalm stone by one-step CO2 activation for gaseouspollutant removal, Carbon, 2000, 38, 1089–1097.

39 I. Langmuir, The adsorption of gases on plane surfaces ofglass, mica and platinum, J. Am. Chem. Soc., 1918, 40,1361–1403.

40 Z. Z. Chowdhury, S. B. Abd Hamid and S. M. Zain, BaseCatalytic Approach: A Promising Technique for Activationof Bio Char for Equilibrium Sorption Studies of Copper,Cu(II) ions in Single Solute System, Materials, 2014, 7,2815–2832.

41 H. M. F. Freundlich, Over the adsorption in solution, J. Phys.Chem., 1906, 57, 385–471; M. J. Temkin and V. Pyzhev,Recent modication to Langmuir isotherms, ActaPhysicochim. URSS, 1940, 12, 217–225.

42 I. A. Abideen, A. E. Ofudje, A. I. Mopelola andO. K. Sarafadeen, Equilibrium kinetics andthermodynamics studies of the biosorption of Mn(II) ionsfrom aqueous solution by raw and acid treated corn cobbiomass, Am. J. Appl. Sci., 2011, 6, 302–309.

43 T. Yang and A. C. Lua, Textural and chemical properties ofzinc chloride activated carbons prepared from pistachio-nut shells, Mater. Chem. Phys., 2006, 100, 438–444.

44 M. Keiluweit, P. S. Nico, M. G. Johnson and M. Kleber,Dynamic molecular structure of plant biomass-derivedblack carbon (biochar), Environ. Sci. Technol., 2010, 44(4),1247–1253.

45 K. Sun, J. Tang, G. Yanyan and H. Zhang, Characterization ofpotassium hydroxide (KOH) modied hydrochars fromdifferent feedstocks for enhanced removal of heavy metalsfrom water, Environ. Sci. Pollut. Res., 2015, DOI: 10.1007/s11356-015-4849-0.

46 W. Tongpoothorn, M. Sriuttha, P. Homchan, S. Chanthaiand C. Ruangviriyachai, Preparation of activated carbonderived from Jatropha curcas fruit shell by simple thermo-chemical activation and characterization of their physico-chemical properties, Chem. Eng. Res. Des., 2011, 8(9), 335–340, DOI: 10.1016/j.cherd.2010.06.012.

47 E. Tchomgui-Kamga, V. Alonzo, C. P. Nanseu-Njiki,N. Audebrand, E. Ngameni and A. Darchen, Preparation

RSC Adv., 2016, 6, 102680–102694 | 102693

RSC Advances Paper

Publ

ishe

d on

26

Aug

ust 2

016.

Dow

nloa

ded

by U

nive

rsity

of

Mal

aya

on 2

2/12

/201

6 04

:37:

12.

View Article Online

and characterization of charcoals that contain dispersedaluminum oxide as adsorbents for removal of uoridefrom drinking water, Carbon, 2010, 48, 333–343.

48 Y. Sharma and F. Gode, Engineering data for optimization ofpreparation of activated carbon from an economically viablematerial, J. Chem. Eng. Data, 2010, 55, 3991–3994.

49 W. Liu and G. Zhao, Effect of temperature and time onmicrostructure and surface functional groups of activatedcarbon bers prepared from liqueed wood, BioResources,2012, 7(4), 5552–5567.

50 C. Ishii, T. Suzuki, N. Shino and K. Kaneko, Structuralcharacterization of heat treated activated carbon bers, J.Porous Mater., 1997, 4(3), 181–186.

102694 | RSC Adv., 2016, 6, 102680–102694

51 M. G. Gronli, G. Varhegyi and C. Diblasi, Thermogravimetricanalysis and devolatilization kinetics of wood, Ind. Eng.Chem. Res., 2002, 41(17), 4201–4208.

52 Y. Zhang, S. Yang, W. Jian-Quan, Y. Tong-Qi and R. Sun,Preparation and Characterization of Lignocellulosic OilSorbent by Hydrothermal Treatment of Populus Fiber,Materials, 2014, 7, 6733–6747.

53 N. U. Saqib, M. Oh, W. Jo, S. Park and J. Lee, Conversion ofdry leaves into hydrochar through hydrothermalcarbonization (HTC), J. Mater. Cycles Waste Manage., 2015,DOI: 10.1007/s10163-015-0371-1.

54 F. Liu and M. Guo, Comparison of the characteristics ofhydrothermal carbons derived from holocellulose andcrude biomass, J. Mater. Sci., 2015, 50, 1624–1631.

This journal is © The Royal Society of Chemistry 2016