Chemical Engineering Journal 211–212 (2012) 224–232

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier .com/locate /ce j

Application of carbonized hemp fibers as a new solid-phase extraction sorbent foranalysis of pesticides in water samples

Marija Vukcevic a,⇑, Ana Kalijadis b, Marina Radisic a, Biljana Pejic a, Mirjana Kostic a, Zoran Lausevic b,Mila Lausevic a

a Faculty of Technology and Metallurgy, University of Belgrade, Karnegijeva 4, 11000 Belgrade, Serbiab Laboratory of Physics, Vinca Institute of Nuclear Sciences, University of Belgrade, Mike Petrovica Alasa 12-14, P.O. Box 522, 11001 Belgrade, Serbia

h i g h l i g h t s

" Production of cheep pesticides sorbents using the waste hemp fibers as row material." Production parameters affect the materials morphology and sorption properties." Carbonized and activated carbons from hemp fibers were successfully used for pesticide preconcentration." Efficiency of activated hemp fibers comparable with commercial cartridges.

a r t i c l e i n f o

Article history:Received 11 July 2012Received in revised form 10 September2012Accepted 12 September 2012Available online 29 September 2012

Keywords:PesticidesSorptionSolid-phase extractionShort hemp fibersCarbonizationSurface characteristics

1385-8947/$ - see front matter � 2012 Elsevier B.V. Ahttp://dx.doi.org/10.1016/j.cej.2012.09.059

⇑ Corresponding author. Tel.: +381 113303647; faxE-mail addresses: marijab@tmf.bg.ac.rs (M. Vuk

(A. Kalijadis), marinar@tmf.bg.ac.rs (M. Radisic), bilkostic@tmf.bg.ac.rs (M. Kostic), zoranl@vinca.rs (Z.(M. Lausevic).

a b s t r a c t

There is a growing interest in utilization of abundantly available materials, bio-mass or industrial byprod-ucts, as precursors for the preparation of carbon materials. Short hemp fibers, acquired as waste from tex-tile production, were used as low-cost precursor for production of carbon materials as a sorbent in thesolid-phase extraction, for pesticide analysis in water samples. Different carbon materials were preparedby carbonization of unmodified and chemically modified hemp fibers. Activation of carbonized materialswith potassium hydroxide improves sorption properties of carbonized hemp fibers by increasing the spe-cific surface area (up to 2192 m2/g) and the amount of surface oxygen groups. The following parametersthat may affect the solid-phase extraction procedure efficiency were optimized: different elution solventsand the pH value of pesticide solution. Extracts were analyzed by liquid chromatography–tandem massspectrometry technique. For this study pesticides belonging to the different chemical classes were cho-sen. Obtained results indicate that carbonized and activated hemp fibers could be successfully appliedas a solid-phase sorbent for the pesticide analysis in water samples.

� 2012 Elsevier B.V. All rights reserved.

1. Introduction

Activated carbons with high surface area and pore volumes canbe prepared from variety of carbonaceous materials such as coal,coconut shell, wood, agricultural or industrial wastes. In the recentyears, there is a growing interest in utilization of the low-cost andabundantly available waste materials as precursors for the prepa-ration of carbon materials [1]. The usage of the waste materialsrepresents a special way of recycling and producing useful prod-ucts. At the same time the cost of waste disposal are minimized.The possibility of using different type of biomass has already been

ll rights reserved.

: +381 113370387.cevic), anaudovicic@vinca.rsjanap@tmf.bg.ac.rs (B. Pejic),Lausevic), milal@tmf.bg.ac.rs

tested for production of the carbon materials [1–11]. Among otherbiomass types, Reed and Williams [12] have used hemp fibers forobtaining activated carbon. Hemp fibers are lignocellulosic materi-als, traditionally used for textile production. In our previous workwe have show that, due to their specific structure and presenceof the surface functional groups, hemp fibers have good sorptioncharacteristic, especially toward heavy metals [13,14]. For thatinvestigation we have used short hemp fibers that represent awaste in textile industry.

The possibility of producing carbon materials with high specificsurface areas, microporous structure, high adsorption capacity anddegree of surface reactivity brings the variety of application forthese materials. Different carbon materials have been widely usedas sorbents in the solid phase extraction (SPE) which is an efficientand economical sample preparation technique for preconcentra-tion of the target analyt. This method has been previously appliedto the determination of many pesticides in natural water and crops

M. Vukcevic et al. / Chemical Engineering Journal 211–212 (2012) 224–232 225

due to its substantial advantages such as providing higher concen-tration factors, decreasing sample preparation time, reducing costs,and requiring less solvent [15–17].

In this work, short hemp fibers obtained as a waste from textileindustry was used for production of carbon material samples withdifferent surface characteristics. Surface characteristics of carbonmaterial depend on both carbon precursor nature and parametersof production [1]. For that reason, different carbon materials wereobtained by chemical modification of origin short hemp fiber fol-lowed by carbonization and activation. Carbonized and activatedhemp fibers obtained in this way were used as a sorbent in the so-lid-phase extraction for pesticide analysis in water samples. Thefollowing parameters that may affect the solid-phase extractionprocedure efficiency were optimized: different solvents used forpesticide elution from carbonized and activated hemp fibers sur-face and the pH value of the pesticide aqueous solution. Extracts,obtained after SPE procedure, were analyzed by liquid chromatog-raphy–tandem mass spectrometry technique. For this study pesti-cides belonging to the different chemical classes as triazine(atrazine, simazine, propazine), neonicotinoid (imidocloprid, ace-tamiprid, thiamethoxam), carbamate (carbofuran, methomyl),organophosphate (monocrotophos, dimethoate, malathion, ace-phate), hydroxyanilide (fenhexamid), diacylhydrazine (tebufenoz-ide) and phenylurea (linuron) were chosen. The ability of usingcarbonized and activated short hemp fibers as a sorbent in SPE pro-cedure was comparatively evaluated with two commercialcartridges.

2. Experimental

2.1. Material

Fibers used in this investigation as a starting material forcarbonization, were short hemp fibers obtained from ITES Odzaci,Serbia. Short hemp fibers were used as received. For easiermanipulation short hemp fibers were cut to the length of fewcentimeters. Chemical composition of used fibers is: 1.50% watersolubles, 0.69% fats and waxes, 1.39% pectins, 78.15% cellulose,6.06% lignin, 10.72% hemicelluloses.

2.2. Preparation of carbon materials samples

The amount of hemp fibers structural components, especiallylignin, hemicelluloses and cellulose, may affect surface characteris-tics of carbonized materials [18]. In order to obtain a row materialwith different characteristics, short hemp fibers were chemicallymodified as it is described in the literature [13]. The progressive re-moval of the hemicelluloses was brought by treating the fiberswith 17.5% NaOH solution, while the lignin was progressively re-moved by treating hemp fibers with 0.7% NaClO2. The samples ob-tained by chemical modification along with the original (asreceived) short hemp fiber were then carbonized at 1000 �C underconstant nitrogen flow (150 cm3/min), with the heating rate of5 �C/min. The isothermal time at maximum carbonization temper-ature was 30 min. After carbonization, five samples denoted Ch1,ChL5, ChL60, ChH5 and ChH45 (as it is shown in Fig. 1), were ob-tained. Further, carbonized unmodified short hemp fibers (Ch1)were chemically activated using KOH as an activating agent inthe 2-step process [5]. Carbonized fibers were mixed with KOHpallets in two different KOH:Ch1 weight ratio (1:1 and 2:1). Activa-tion process was carried out in an electrical furnace under the con-stant nitrogen flow (150 cm3/min) with the heating rate of 5 �C/min up to the 900 �C. Final temperature was maintained for30 min. The resulting products after activation were thoroughlywashed with tap water and finally distilled water to remove the

residual KOH until the pH value of the eluted water ranged from6 to 7. In this way, two activated carbon samples (denoted ACh1and ACh2) were obtained. The scheme of production and denote-ment of all samples are shown in Fig. 1.

2.3. Surface characteristics

2.3.1. Scanning electron microscopySurface structure and morphology were studied by scanning

electron microscopy (SEM JEOL JSM-6610LV).

2.3.2. Porous properties of carbonized and activated short hemp fibersAdsorption and desorption isotherms of N2 were measured on

carbonized and activated short hemp fibers at �196 �C, usingporosimeter Micromeritics ASAP 2020, Surface and Porosity Ana-lyzer (Micromeritics Instrument Corporation, US). The total porevolume (Vtotal), micropore volume (Vmicro) and mesopore includingexternal surface area (Smeso), were obtained from the adsorptiondata, using the manufacturer‘s software ASAP 2020 V3.05 H. Poresize distribution was estimated by applying BJH method [19] tothe desorption branch of isotherms and mesopore surface andmicropore volume were estimated using the high resolution as plotmethod [20]. The surface area corresponding to the micropores(Smicro) was obtained from the difference between SBET and Smeso.

2.3.3. Surface oxygen groupsTemperature-programmed desorption (TPD) in combination

with mass spectrometry was used to investigate the nature andthermal stability of carbonized and activated short hemp fiberssurface oxygen groups. The TPD profiles were obtained using a cus-tom-built set-up, consisting of a quartz tube placed inside an elec-trical furnace. Sample was outgassed in the quartz tube andsubjected to TPD at a constant heating rate of 10 �C/min to900 �C under high vacuum. The amounts of CO and CO2 releasedfrom the carbon sample (0.1 g) were monitored using an Extorr300 quadrupole mass spectrometer (Extorr Inc., USA).

2.4. Solid phase extraction procedure

2.4.1. Preparation of the SPE cartridgesThe 0.2 g of each carbon samples was packed into the empty

cartridge. The polypropylene upper and lower frits were placedat each end of the cartridge to hold the sorbent packing in place.Next, the outlet tip of cartridge was connected to a Visiprep™SPE Vacuum Manifold and the inlet end of it was connected to aPTFE suction tube whose other end was inserted into pesticidessample solution. The possibility of using carbonized and activatedshort hemp fibers as sorbents in SPE procedure was tested by usinga stock solution of fifteen pesticides mixture. Concentration of eachpesticide in solution was 1 ng/mL.

2.4.2. Optimization of the SPE procedureThe optimization of the SPE procedure, as a sample preparation

method, is an important process to achieve the highest enrichmentefficiency and the best recovery. The following parameters thatmay affect the SPE procedure efficiency were optimized: differentelution solvents and the pH values of the pesticide aqueous solu-tion. For the optimization of elution solvent, four different elutionsolvents were used: methanol, dichloromethane, acetonitrile andmethanol–dichloromethane (1:1). For this experiment, 100 mL ofdeionized water (without pH adjustment) was spiked with theworking pesticides standard solution in order to achieve the con-centration of 1 ng/mL for each analyte in the solution. Followingthe standard procedure [15], the SPE cartridges were precondi-tioned with 5 mL of selected elution solvent followed by 5 mL ofdeionized water. Spiked water samples were loaded at the flow

Fig. 1. The scheme of carbonized and activated hemp fibers production.

226 M. Vukcevic et al. / Chemical Engineering Journal 211–212 (2012) 224–232

rate of 1 mL/min. The cartridges were then dried under vacuum for10 min and analytes were eluted with selected elution solvent upto extract volume of 15 mL. Extracts were evaporated to drynessand reconstituted with 1 mL of methanol. The final extracts werefiltered through 0.45 lm polyvinylidene difluoride (PVDF) filters,acquired from Roth (Karlsruhe, Germany), into the autosamplervials and analysed using HPLC–MS/MS method [21]. After theselection of the elution solvent, the effect of sample pH adjustmentprior to extraction was evaluated. The tested pH-values were: 4.5,6.0 and 8.0. For this experiment, deionized water (100 mL) wasspiked with working pesticides standard solution to achieve con-centration of 1 ng/mL for each analyte in the solution. After thatthe pH of pesticides solution was adjusted to the selected values.Extracts were obtained and then analyzed in the same way as forthe previous experiment. On the basis on recovery studies optimalelution solvent and the pH value was selected.

2.4.3. HPLC–MS/MS analysisSensitive analytical method based on SPE followed by HPLC–

MS2 was developed for simultaneous analysis of 15 agriculturalpesticides that belong to eight different chemical classes.

Surveyor HPLC system (Thermo Fisher Scientific, USA) was usedfor the separation of the analytes on the reverse-phase ZorbaxEclipse XDB-C18 column, 75 mm long, 4.6 mm i.d. and 3.5 lm par-ticle size (Agilent Technologies, USA). The mobile phase consistedof methanol (A), water (B) and acetic acid (C). Gradient changedas follows: 0 min 33% A, 66% B, 1% C; 7.5 min 58% A, 41.4% B,0.6% C; 25 min 100% A; 27.01 min 33% A, 66% B, 1% C. The flow rateof the mobile phase was 0.5 ml/min. An aliquot of 10 ll of theaqueous solution was injected into HPLC system. Quadrupole iontrap mass spectrometer, LCQ Advantage (Thermo Fisher Scientific,USA), was used for detection and quantification of pesticides. Theelectrospray ionization technique was used and all pesticides wereanalyzed in the positive ionization mode. Mass chromatogram ofthe pesticides is given in Fig. 2.

3. Results and discussion

3.1. Surface characteristics of carbonized and activated hemp fibers

3.1.1. The influence of preparation process parameters on themorphology of carbonized and activated short hemp fiber samples

The changes in chemical composition incurred as the result ofalkali and oxidative chemical treatment are described in our previ-

ous paper [13]. Through hemp fibers treatment with 17.5% NaOHhemicelluloses were progressively removed and their content de-creased for approximately 70%, while treatment with 0.7% NaClO2

progressively reduced lignin for about 50% in relation to unmodi-fied fibers. During both types of treatment removal of target com-pound, lignin or hemicelluloses, is proportional to the modificationtime. Removal of lignin or hemicelluloses from short hemp fiber bychemical modification changed both chemical and structural prop-erties of the hemp fibers. During the oxidation treatment lignin isselectively removed, resulting in more homogenous middle lamel-la due to the gradual elimination of micro-pores and the less rigidcell wall. Removal of lignin is accompanied by fibrillation and for-mation of new capillary spaces in inter-surfacial layer betweenpartially separated elementary fibers. In the other hand, hemicellu-loses are deposited in amorphous areas of fiber structure and occu-py spaces between the fibrils in primary and secondary walls.When the hemicelluloses were gradually removed, by alkalinetreatment, inter-fibrillar regions become less dense and rigid andthereby make the fibrils more capable of rearrangement [22–24].Both treatments lead to liberation of the elementary fibers, whichare more pronounced in the case of alkaline treatment.

After carbonization all samples retain fibrous structure of theprecursor fibers. As it can be seen from the SEM photographs ofcarbonized samples, samples ChL5 and ChL60 (Fig. 3b and c,respectively) have visible fibrils on the sample surface due to pre-viously removed lignin from middle lamella. Furthermore, it can benoted that the main feature of the ChH5 and ChH45 structure(Fig. 3d and e, respectively) is pronounced fibrillation which isespecially noticeable at the micrograph of sample ChH45(Fig. 3e). Compared to the morphology of non-activated sample,Ch1 (Fig. 3a), activated hemp fiber samples, ACh1 and ACh2(Fig. 3f and g) have pronounced longitudinal cracks along the fiberwhich is probably the consequence of activation.

3.1.2. Porous properties and surface oxygen groupsPorous properties of all carbonized and activated hemp fibers,

are listed in Table 1. Specific surface area values, calculated byBET equation (SBET), for all samples lie within 388.6 – 2192 m2/g.All non-activated samples, except sample ChH45, have lower spe-cific surface area than sample Ch1. It seems that chemical modifi-cation of the origin short hemp fiber prior to carbonization causedthe decrease in specific surface area of resulted carbonized sam-ples, which is accompanied with decrease in both micropurousand mesoporous surface. Nevertheless, from the values of the

Fig. 2. Mass chromatogram of pesticides solution.

M. Vukcevic et al. / Chemical Engineering Journal 211–212 (2012) 224–232 227

Smicro/SBET and Vmicro/Vtotal ratios and from the average pore size(rp), it can be seen that chemical modification prior to the carbon-ization slightly increases the microporosity and decreases averagepore size of the carbonized samples. As it was observed in the lit-erature [18], the amount of lignin, hemicelluloses and cellulose in

the carbon precursor affects the specific surface area of carbonizedmaterials. Lignin has been found to be effective in creating pores,as evident from the work by Kennedy et al. [25]. Furthermore,the BET surface area was found to be highest for the carbon mate-rials obtained from carbon precursors with highest lignin content

Fig. 3. SEM photographs of all tested samples: (a) Ch1, (b) ChL5, (c) ChL60, (d) ChH5, (e) ChH45, (f) ACh1 and (g) ACh2.

228 M. Vukcevic et al. / Chemical Engineering Journal 211–212 (2012) 224–232

[12]. In view of that, as it is shown in Table 1, short hemp fibermodified by removing the lignin, after carbonization gave samplesChL5 and ChL60 with lower specific surface area.

For the samples ChH5 and ChH45 specific surface area increasedwith the increasing amount of hemicelluloses removed (Table 1).

During the alkali treatment of origin short hemp fibers, the crystalstructure of cellulose, named as cellulose I (Cell I), is transformedinto cellulose II (Cell II) [26,27]. The polymorphic transformationof Cell I to Cell II depends on alkali concentration and the time oftreatment [28]. In the chemical treatment used for obtaining the

Table 1Porous properties and amounts of CO and CO2 evolving surface oxygen groups of carbonized and activated short hemp fibers samples.

Sample SBET

(m2/g)Smeso

(m2/g)Smicro

(m2/g)Smicro/SBET

Vtotal

(cm3/g)Vmicro

(cm3/g)Vmicro/Vtotal

rp

(nm)CO evolving groups(mmol/g)

CO2 evolving groups(mmol/g)

CO + CO2

(mmol/g)

Ch1 518.5 132.2 386.3 0.75 0.291 0.180 0.62 2.24 1.718 2.192 3.910ChL5 428.6 93.8 334.8 0.78 0.208 0.156 0.75 1.94 2.641 0.812 3.453ChL60 388.6 86.1 302.5 0.78 0.194 0.140 0.72 1.99 4.364 1.851 6.215ChH5 425.9 88.9 337.0 0.79 0.207 0.157 0.76 1.94 3.513 1.119 4.632ChH45 573.5 137.6 435.9 0.76 0.290 0.203 0.70 2.02 2.054 0.613 2.667ACh1 673.0 47.0 626.0 0.93 0.403 0.327 0.81 1.74 4.044 3.282 7.326ACh2 2192 81.0 2111 0.96 1.203 1.059 0.88 1.78 5.598 5.660 11.259

Fig. 4. TPD spectra of carbonized short hemp fibers samples: (a) CO and (b) CO2

desorption profile.

M. Vukcevic et al. / Chemical Engineering Journal 211–212 (2012) 224–232 229

carbon precursor for sample ChH45 the concentration of NaOH andthe time of treatment was high enough to provide appropriate con-ditions for this polymorphic transformation. This transformation ofcellulose I into more reactive cellulose II is probably the reason forhigh specific surface area of sample ChH45. Also, Khezami et al.[29] have suggested that microporosity of carbonized material isdue to cellulose and not due to lignin and hemicelluloses.

However, taking in consideration all mentioned above and com-plexity of the structure and composition of hemp fibers, a simplerelation between the lignin, hemicelluloses and cellulose contentsversus the specific surface area could not be demonstrated.

Porous properties given in Table 1, show that activation withKOH at 900 �C provides samples with very high specific surfacearea. Also, the formation of micropores and decrease of averagepore size is even more pronounced for the samples obtained byKOH activation. Higher amount of hydroxide used for obtainingsample ACh2 significantly increases specific surface area of thissample compared to ACh1. Stronger activation i.e. increased ratioof KOH, open up the porous structure and increases SBET [30]. Theseresults are in agreement with those in the literature [31]. Also, spe-cific surface area of sample ACh2 is much higher compared to thevalues of SBET reported in the literature and obtained for activatedcarbons derived from hemp fibers [12].

TPD provides quantitative information on the total number ofsurface oxygen groups. Surface oxygen complexes on carbon mate-rials decompose upon heating by releasing CO and CO2. TPD peaksof CO and CO2 at different temperatures are related to the bondstrength of the specific oxygen groups. Thus, the position of thepeak maximum at a defined temperature corresponds to a specificoxygen complex at the surface. For example, CO2 is released bydecomposition of carboxylic groups at 373–673 K [32–34] or lac-tone groups at 463–923 K [32,35]. Both CO and CO2 peaks originatefrom the decomposition of carboxylic anhydrides in the tempera-ture range of 623–900 K [32,33]. Phenols, ethers, carbonyls andquinones give rise to CO at 973–1253 K [34,36]. The quantities ofCO and CO2 released during the TPD experiments correspond tothe total amount of surface oxygen groups.

TPD profiles of CO and CO2 evolution for all tested samples areshown in Fig. 4. The TPD spectra of CO2 desorption profiles of alltested samples show an intensive peak at relatively high tempera-ture (from 890 K to 1073 K). For the sample ACh2 this peak is mostintensive and shifted to higher temperatures. It can be noted thatactivation process as well as the increased amount of activatingagent increase the intensity of the peak and shifts it to higher tem-perature, which suggests the stabilization of surface oxygengroups. For all tested samples, CO desorption profiles have a max-imum at the temperature which coincides with the maximum inCO2 desorption profile indicating the existence of anhydridegroups, which decompose upon heating by releasing both CO andCO2. Also, TPD peaks for CO evolution in this temperature rangecan also be attributed to phenols, ethers, carbonyls and quinones[37]. Less intensive peaks observed at lower temperature may bedue to thermal decomposition of carbonyl groups in a-substitutedketones and aldehydes [38].

The amounts of CO and CO2 released from the surface of carbon-ized and activated fiber samples were obtained by integration ofcorresponding TPD curves and presented in Table 1. It can beobserved that modification of the original short hemp fibers priorto carbonization affects the amount of oxygen groups at the surfaceof carbonized samples. For all samples modified prior to carboniza-tion amount of CO evolving groups increase while amount of CO2

evolving groups decrease compared to sample Ch1. It is interestingthat in the group of non-activated samples, sample ChL60 has thehighest amount of surface oxygen groups and sample ChH45 thelowest, which are totally opposite to the values of their specificsurface area. Also, it has been noticed that the extension of theoxidation treatment time leads to the increased amount of thesurface oxygen groups, while the extension of alkali treatmenttime leads to the reduced amounts of the functional groups. Inthe case of activated samples, as it was already mentioned,

Fig. 5. Recoveries of selected pesticides obtained by using different elutionsolvents.

Fig. 6. Recoveries of selected pesticides at different pH-values of pesticide aqueoussolution.

230 M. Vukcevic et al. / Chemical Engineering Journal 211–212 (2012) 224–232

increased amount of the activating agent increase the amount ofboth CO and CO2 evolving groups.

3.2. Carbonized and activated hemp fibers as a sorbent in solid phaseextraction

Based on previous characterization, i.e. good surface character-istics, sample ACh2 was used as a SPE sorbent for the optimizationof SPE procedure. Cartridges prepared with ACh2 filling andin combination with four different elution solvents were testedin the term of pesticides recovery. It is considered that SPE proce-dure has good efficiency when obtained recoveries fell within70–120% with relative standard deviation (RSD) 6 20% [15,17].The obtained pesticides recoveries are presented in Fig. 5.

As it can be seen from the Fig. 5, the most acceptable recoverieswere obtained by using mixture of dichloromethane–methanol(1:1). For all pesticides, except for acephate, imidocloprid, linuronand fenhexamid the recoveries were in the range 73.3–106.7%.According to these results, the mixture of dichloromethane-methanol was chosen as the optimal SPE elution solvent for simul-taneous determination of pesticides.

The next step in optimization of SPE procedure was the selec-tion of optimal pH values of pesticides water samples. For this pur-pose pH of the pesticides aqueous solution was adjusted to 4.5, 6

Table 2Recoveries of selected pesticides obtained using different carbonized hemp fibers as SPE c

Pesticide Recovery (%) (RSD, %)

Ch1 ChL5 ChL60 ChH5

Acephate 23.0 (14) 55.4 (1) 56.3 (7) 50.1 (14)Methomyl 24.6 (28) 38.9 (16) 69.6 (0) 19.2 (14)Thiamethoxam 62.9 (0) 93.1 (8) 90.7 (13) 89.5 (5)Monocrotophos 14.3 (11) 57.2 (17) 73.9 (0) 37.8 (20)Imidocloprid 64.2 (7) 88.8 (19) 95.2 (7) 95.2 (7)Acetamiprid 61.9 (14) 105.8 (12) 87.6 (15) 89.4 (0)Dimethoate 7.9 (15) 33.1 (11) 37.3 (2) 26.6 (0)Simazine 29.1 (2) 44.8 (4) 61.0 (3) 56.1 (20)Carbofuran 7.1 (2) 24.5 (14) 30.9 (5) 27.4 (7)Atrazine 20.9 (4) 40.5 (0) 43.5 (11) 42.7 (20)Propazine 22.1 (5) 41.5 (13) 37.2 (9) 38.1 (17)Linuron 14.3 (12) 7.1 (0) 55.2 (0) 64.2 (0)Malathion 2.8 (2) 12.2 (15) 24.0 (8) 9.2 (6)Tebufenozide 61.4 (10) 95.3 (1) 100.7 (15) 91.7 (12)Fenhexamid 40.4 (13) 71.1 (0) 84.3 (0) 62.5 (3)

and 8. As it is shown in Fig. 6, the best recoveries was obtainedat the pH-value of 6, therefore this pH-value was selected as theoptimal for extraction of selected pesticides from water samples.

The finally optimized analytical procedure for SPE of pesticideswas as follows: SPE cartridge (200 mg of carbonized and activatedhemp fibers as a sorbent) is preconditioned with 5 mL of metha-nol–dichloromethane mixture (1:1) followed by 10 mL of deionizedwater; 100 mL of the pesticide aqueous solution, with the pH-valueadjusted to 6, is applied to preconditioned cartridge at the flow rateof 1 mL/min; the cartridge is dried under vacuum for 10 min; thecartridge is eluted with methanol–dichloromethane mixture (1:1)to obtain 15 mL of eluted extract; extract is evaporated to drynessand reconstituted with 1 mL of methanol; the final extract is fil-tered through 0.45 lm PVDF filter into the autosampler vial andanalysed. Concentration of pesticides was determined by HPLC–MS/MS method. From obtained MS2 spectra of pesticides, mostabundant fragment ions were selected. The selected reaction mon-itoring (SRM) mode was used for quantification of all pesticides.

Following this SPE procedure, the possibility of using differentsamples of carbonized and activated short hemp fiber as SPEsorbents was tested. Recoveries obtained in this way were compara-tively evaluated with two commercial cartridges: SupelcleanENVI-18 (C18, 500 mg/6 mL; from Supelco, Bellefonte, PA, USA) andHyperSep�Hypercarb� (porous graphitic carbon, 500 mg/6 mL fromThermo Fisher Scientific, USA). Obtained results are given in Table 2.

artridges and two commercial ones.

ChH45 ACh1 ACh2 C18 Hypercarb�

43.3 (16) 75.4 (3) 22.4 (11) 2.3 (16) 52.3 (12)29.5 (7) 59.4 (15) 59.8 (15) 78.3 (11) 31.2 (14)91.9 (2) 86.7 (4) 73.1 (7) 23.5 (15) 87.3 (12)70.7 (10) 96.4 (10) 71.1 (5) 66.5 (7) 77.3 (12)85.3 (7) 37.2 (11) 71.3 (3) 78.3 (11) 84.6 (8)96.0 (18) 98.0 (16) 88.1 (3) 104.1 (9) 70.4 (7)43.2 (6) 95.4 (11) 76.2 (6) 69.3 (18) 84.0 (13)63.8 (1) 78.6 (17) 82.7 (8) 56.1 (7) 40.0 (2)43.0 (9) 83.0 (6) 101.7 (5) 76.4 (4) 71.7 (12)62.6 (4) 100.5 (8) 76.3 (1) 88.7 (9) 71.5 (10)66.7 (1) 87.7 (2) 73.4 (3) 86.4 (6) 57.5 (10)73.3 (0) 6.9 (16) 13.9 (2) 52.6 (3) 55.8 (17)34.8 (8) 83.5 (15) 80.1 (17) 65.5 (4) 55.0 (17)86.2 (1) 97.4 (6) 71.5 (7) 63.6 (3) 87.0 (17)81.8 (13) 35.4 (5) 37.8 (6) 86.7 (15) 25.1 (12)

Fig. 7. Recoveries of selected pesticides obtained using activated hemp fibers as SPE cartridges and two commercial ones.

M. Vukcevic et al. / Chemical Engineering Journal 211–212 (2012) 224–232 231

Sample Ch1 could not be used as sorbent for SPE cartridges dueto low recoveries. Carbonized hemp fibers modified prior to thecarbonization can be used for preconcentration of few pesticides:thiamethoxam, imidocloprid, acetamiprid, tebufenozide and fen-hexamid. Additionally, samples ChL60 and ChH45 can be used forpreconcentration of monocrotophos. As it is shown in Table 2. forextraction of linuron only ChH45 can be used. Activated hemp fi-bers sample ACh1 can be used for preconcentration of all examinedpesticides except for methomyl, imidocloprid, linuron and fen-hexamide. Activated sample ACh2, can be used for all examinedpesticides except for acephate, methomyl, linuron and fenhexa-mide. Recoveries obtained for activated samples are comparablewith those obtained for commercial cartridges (Fig. 7). In the caseof acephate, dimethoate, simazine, carbofuran, propazine, malationand tebufenozide recoveries obtained by activated hemp fiberswas even better than those obtained with commercial cartridges.

Comparing the results obtained in the SPE experiments with thesurface properties and morphology of the carbonized and activatedsamples, it can be noted that the best SPE efficiency was achievedwith activated hemp fibers samples with the highest specific sur-face area and the amount of surface oxygen groups.

4. Conclusions

Chemical modification of hemp fibers, prior to carbonization, af-fects the specific surface area, amount of surface oxygen groupsand morphology of carbonized hemp fibers. Furthermore, activa-tion of carbonized materials with potassium hydroxide improvessorption properties of carbonized hemp fibers by increasing thespecific surface area (up to 2192 m2/g) and amount of surfaceoxygen groups. It was noted that the adsorption characteristicsare primary influenced by porous properties and the amount ofsurface oxygen groups, while morphology of examined samplesdoes not show direct influence on the SPE efficiency.

Results obtained from SPE experiments indicate that the acti-vated hemp fibers could be successfully applied as a solid-phasesorbent for the pesticides analysis in water samples. For same pes-ticides, recoveries obtained by these cartridges were even higherthan recoveries obtained by commercial cartridges.

Acknowledgments

The authors wish to thank the Ministry of Education andScience of the Republic of Serbia for financial support throughthe projects of Basic Research, Nos. 172007 and 172029 andPhysics and Chemistry with Ion Beams (III) No. 45006.

References

[1] E. Yagmur, M. Ozmak, Z. Aktas, A novel method for production of activatedcarbon from waste tea by chemical activation with microwave energy, Fuel 87(2008) 3278–3285.

[2] R. Xu, S. Xiao, J. Yuan, A. Zhao, Adsorption of methyl violet from aqueoussolutions by the biochars derived from crop residues, Bioresour. Technol. 102(2011) 10293–10298.

[3] S. Aber, A. Khataee, M. Sheydaei, Optimization of activated carbon fiberpreparation from Kenaf using K2HPO4 as chemical activator for adsorption ofphenolic compounds, Bioresour. Technol. 100 (2009) 6586–6591.

[4] X. Chen, G. Chen, L. Chen, Y. Chen, J. Lehmann, M.B. McBride, G. Anthony, A.G.Hay, Adsorption of copper and zinc by biochars produced from pyrolysis ofhardwood and corn straw in aqueous solution, Bioresour. Technol. 102 (2011)8877–8884.

[5] A.H. Basta, V. Fierro, H. El-Saied, A. Celzard, 2-Steps KOH activation of ricestraw: An efficient method for preparing high-performance activated carbons,Bioresour. Technol. 100 (2009) 3941–3947.

[6] G.H. Oh, C.R. Park, Preparation and characterisation of rice-Strow-based porouscarbons with high adsorption capacity, Fuel 81 (2002) 327–336.

[7] T. Yang, A.C. Lua, Characteristics of activated carbons prepared from pistachio-nut shells by potassium hydroxide activation, Micropor. Mesopor. Mater. 63(2003) 113–124.

[8] A.P. Carvalho, B. Cardoso, J. Pires, M.B. de Carvalho, Preparation of activatedcarbons from cork waste by chemical activation with KOH, Carbon 41 (2003)2873–2884.

[9] P.J.M. Carrott, M.M.L. Ribeiro Carrott, P.A.M. Mourao, Pore size control inactivated carbons obtained by pyrolysis under different conditions ofchemically impregnated cork, J. Anal. Appl. Pyrolysis 75 (2006) 120–127.

[10] M. Olivares-Marin, C. Fernandez-Gonzalez, A. Macias-Garcia, V. Gomez-Serrano, Preparation of activated carbons from cherry stones by activationwith potassium hydroxide, Appl. Surf. Sci. 252 (2006) 5980–5983.

[11] P. Nowicki, R. Pietrzak, H. Wachowska, Sorption properties of active carbonsobtained from walnut shells by chemical and physical activation, Catal. Today150 (2010) 107–114.

[12] A.R. Reed, P.T. Williams, Thermal processing of biomass natural fibre wastes bypyrolysis, Int. J. Energy Res. 28 (2004) 131–145.

[13] B. Pejic, M. Vukcevic, M. Kostic, P. Skundric, Biosorption of heavy metal ionsfrom aqueous solutions by short hemp fibers: effect of chemical composition,J. Hazard. Mater. 164 (2009) 146–153.

232 M. Vukcevic et al. / Chemical Engineering Journal 211–212 (2012) 224–232

[14] B. Pejic, M. Vukcevic, I. Pajic-Lijakovic, M. Lausevic, M. Kostic, Mathematicalmodeling of heavy metal ions (Cd2+, Zn2+and Pb2+) biosorption by chemicallymodified short hemp fibers, Chem. Eng. J. 172 (2011) 354–360.

[15] N. Dujakovic, S. Grujic, M. Radišic, T. Vasiljevic, M. Lauševic, Determination ofpesticides in surface and ground waters by liquid chromatography–electrospray–tandem mass spectrometry, Anal. Chim. Acta 678 (2010) 63–72.

[16] A.M. Rodrigues, V. Ferreira, V.V. Cardoso, E. Ferreira, M.J. Benoliel,Determination of several pesticides in water by solid-phase extraction,liquid chromatography and electrospray tandem mass spectrometry, J.Chromatogr. A 1150 (2007) 267–278.

[17] X. Yang, H. Zhang, Y. Liu, J. Wang, Y.C. Zhang, A.J. Dong, H.T. Zhao, C.H. Sun, J.Cui, Multiresidue method for determination of 88 pesticides in berry fruitsusing solid-phase extraction and gas chromatography–mass spectrometry:determination of 88 pesticides in berries using SPE and GC–MS, Food Chem.127 (2011) 855–865.

[18] P.J.M. Suhas, M.M.L. Carrott, Ribeiro Carrott, Lignin – from natural adsorbent toactivated carbon: a review, Bioresour. Technol. 98 (2007) 2301–2312.

[19] E.P. Barret, L.G. Joyner, P.P. Halenda, The determination of pore volume andarea distributions in porous substances. I. Computations from nitrogenisotherms, J. Am. Chem. Soc. 73 (1951) 373–380.

[20] K. Kaneko, C. Ishii, M. Ruike, H. Kuwabara, Origin of superhigh surface area andmicrocriystalline graphitic structures of activated carbons, Carbon 30 (1992)1075–1088.

[21] M. Radišic, S. Grujic, T. Vasiljevic, M. Lauševic, Determination of selectedpesticides in fruit juices by matrix solid-phase dispersion and liquidchromatography–tandem mass spectrometry, Food Chem. 113 (2009) 712–719.

[22] H. Wang, R. Postle, R. Kessler, W. Kessler, Removing pectin and lignin duringchemical processing of hemp for textile applications, Text. Res. J. 73 (2003)664–669.

[23] G. Buschle Diller, C. Fanter, F. Loth, Structural changes in hemp fibers as aresult of enzymatic hydrolysis with mixed enzyme systems, Text. Res. J. 69(1999) 244–251.

[24] M. Kostic, B. Pejic, P. Skundric, Quality of chemically modified hemp fibers,Bioresour. Technol. 99 (2008) 94–99.

[25] L.J. Kennedy, J.J. Vijaya, G. Sekaran, Effect of two-stage process on thepreparation and characterisation of porous carbon composite from rice huskby phosphoric acid activation, Ind. Eng. Chem. Res. 43 (2004) 1832–1838.

[26] A. Ishikawa, T. Okano, J. Sugiyama, Fine structure and tensile properties oframie fibres in the crystalline form of cellulose I, II, Illl and IVI, Polymer 38(1997) 463–468.

[27] S. Borysiak, J. Garbarczik, Applying the WAXS method to estimate thesupermolecular structure of cellulose fibers after mercerization, Fibres Text.East. Eur. 11 (2003) 104–106.

[28] J. Gassan, A.K. Bledzki, Alkali treatment of jute fibers: relationship betweenstructure and mechanical properties, J. Appl. Polym. Sci. 71 (1999) 623–629.

[29] L. Khezami, A. Chetouani, B. Taouk, R. Capart, Production and characterisationof activated carbon from wood components in powder: cellulose, lignin, xylan,Powder Technol. 157 (2005) 48–56.

[30] M.J.B. Evans, E. Halliop, J.A.F. MacDonald, The production of chemically-activated carbon, Carbon 37 (1999) 269–274.

[31] F.C. Wu, R.L. Tseng, C.C. Hu, Comparisons of pore properties and adsorptionperformance of KOH-activated and steam-activated carbons, Micropor.Mesopor. Mater. 80 (2005) 95–106.

[32] A. Kalijadis, M. Vukcevic, Z. Jovanovic, Z. Lauševic, M. Lauševic,Characterisation of surface oxygen groups on different carbon materials bythe Boehm method and temperature-programmed desorption, J. Serb. Chem.Soc. 76 (2011) 757–768.

[33] Y. Otake, R.G. Jenkins, Characterization of oxygen-containing surfacecomplexes created on a microporous carbon by air and nitric acid treatment,Carbon 31 (1993) 109–121.

[34] U. Zielke, K.J. Hüttinger, W.P. Hoffman, Surface-oxidized carbon fibers: I.Surface structure and chemistry, Carbon 34 (1996) 983–998.

[35] B. Marchon, J. Carrazza, H. Heinemann, G.A. Somorjai, TPD and XPS studies ofO2, CO2, and H2O adsorption on clean polycrystalline graphite, Carbon 26(1988) 507–514.

[36] D.M. Nevskaia, A. Santianes, V. Muñoz, A. Guerrero-Ruíz, Interaction ofaqueous solutions of phenol with commercial activated carbons: anadsorption and kinetic study, Carbon 37 (1999) 1065–1074.

[37] M. Domingo-Garcia, F.J. Lopez Garzon, M.J. Perez-Mendoza, On thecharacterization of chemical surface groups of carbon materials, J. ColloidInterf. Sci. 248 (2002) 116–122.

[38] G.S. Szymanski, Z. Karpinski, S. Biniak, A. Swiatkowski, The effect of the gradualthermal decomposition of surface oxygen species on the chemical andcatalytic properties of oxidized activated carbon, Carbon 40 (2002) 2627–2639.