Fuel Processing Technology 130 (2015) 214–223

Contents lists available at ScienceDirect

Fuel Processing Technology

j ourna l homepage: www.e lsev ie r .com/ locate / fuproc

Adsorptive desulfurization using different passivated carbonnanoparticles by PEG-200

Rahimeh Naviri Fallah a,⁎, Saeid Azizian b, Amarendra Dhar Dwivedi a, Mika Sillanpää a

a Laboratory of Green Chemistry, LUT Savo Sustainable Technologies, Lappeenranta University of Technology, Sammonkatu 12, FI-50130 Mikkeli, Finlandb Department of Physical Chemistry, Faculty of Chemistry, Bu-Ali Sina University, Hamedan 65167, Iran

⁎ Corresponding author. Tel.: +358 503481188.E-mail addresses: Minafallahh@yahoo.com, Rahimeh.N

http://dx.doi.org/10.1016/j.fuproc.2014.10.0180378-3820/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 4 September 2014Received in revised form 9 October 2014Accepted 14 October 2014Available online xxxx

Keywords:Carbon nanoparticlesDesulfurizationLiquid fuelAdsorption

Carbon nanoparticles dispersed in aqueous phase (CNPs) were synthesized using different poly(ethylene glycol)(PEG 200) assistant solution methods including microwave pyrolysis and alkali-assisted heating, ultrasonica-tion, and refluxing. The CNPs (b5 nm) were used for the selective removal of sulfur thiophenic compounds(benzothiophene (BT), dibenzothiophene (DBT) and dimethyldibenzothiophene (DMDBT)) from liquid modelfuels including aromatic compounds (naphthalene (NP) and 1-methylnaphthalene (MNP)). The presence ofPEG in the preparation procedure of CNP was crucial and adsorption capability was related to the passivationby PEG. The decomposition of PEG at high temperature introduced more acidic functional groups on the carbonframework and also promoted the formation of porous CNP nanostructures because of hydrophobic PEGaggregations. Adsorption experiments showed the order for CNPs with different method as: microwaveirradiation N heating N refluxing N ultrasonication. The highest efficiency of CNPs by microwave-assisted tech-nique for sulfur removal was examined due to the higher decomposition of PEG andmore porosity of CNP nano-structures which were produced at higher temperature. Moreover the CNPs were regenerable by a solvent-washing process. This study showed that passivated CNPs are very effective for rapid desulfurization of liquidfuels, operating at ambient conditions with the ease of separation and regeneration.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

The harmful effects of sulfur oxide (SOx) pollution on environ-ment have imposed stringent sulfur regulation on automotive fuelsglobally [1]. Thus, in recent years the reduction of the sulfur content infuel to reach a sulfur level of less than 10 ppmhas receivedmuch atten-tion in many countries. The demand for producing low-sulfur fuelsimposes significant challenges to current desulfurization methods andthe development of new technologies. To reach this low level, it is re-quired that the highly refractory S-containing compounds, such asbenzothiophenes and most refractory substituted dibenzothiophenes(e.g. 4,6 dimethyldibenzothiophene) are removed from fuels [2–4].Sulfur removal from various liquid hydrocarbon fuels has conductedby the catalytic hydrodesulfurization (HDS) process at 300–400 °C andhigh hydrogen pressure with high hydrogen consumption in refineries.Unfortunately, when this level of deep desulfurization is attempted, thecatalytic hydrodesulfurization process conditions become very severe,leading to increased process cost [5,6].

Various new approaches to deep desulfurization of liquid hydrocar-bon fuels have been reported in the literature [5,7–9]. Adsorption bynovel adsorbents due to their high capacity and selectivity has highly

aviri.Fallah@lut.fi (R.N. Fallah).

attracted researchers [10–14]. Plenty of research has been conductedto find adsorbentmaterials that are highly selective towards sulfur com-pounds, in the presence of coexisting aromatic hydrocarbons and ole-fins that account for a large excess in the fuel. Activated carbon [11,15,16], zeolite [13,17] and mesoporous materials [6] are the well-knownadsorbents. Nano-adsorbents such as nano γ-alumina [18,19] and load-ed nanoparticle adsorbents [20] have been recently reported for the re-moval of organosulfur compounds.

Carbon nanoparticles (CNPs), a novel class of recently discoverednanocarbons, consist of discrete, quasispherical nanoparticles withsizes below 10 nm [21]. CNPs have attracted considerable attention be-cause of their abundant photo-physical or chemical properties, low costand goodwater-solubility. During thepast fewyears,much progress hasbeen made in the synthesis, properties and applications of CNPs. CNPshave been synthesized by various methods such as the candle soot oxi-dation, carbohydrate carbonization, thermal decomposition of smallmolecules, pyrolysis of polymers, microwave-based assisted pyrolysisand electrochemical methods [22].

The strong and tunable fluorescence, capable surface functionalities,and the variety of simple, fast and cheap synthetic routes promotes thewide application of CNPs in catalysis, bio-imaging, optoelectronics,sensors, surface-enhanced Raman scattering and as replacements fortoxic metal-based quantum dots (QDs) and photoactive nanocatalysts[21,23]. Han et al. have successfully constructed different complex

Table 1Adsorbents synthesis: method preparation, optimized conditions and materials used.

Adsorbents Method preparation Optimized conditions Materials used

Time(min)

Power/temperature Glucose Distilled water(mL)

Sodium hydroxide PEG 200(mL)

Micr.GHP-9 min Microwave irradiation 9 600 W and 2450 Hz 1 g 5 – 15Micr.HP-9 min Microwave irradiation 9 600 W and 2450 Hz – 5 – 15Micr.HPN-5 min Microwave irradiation 5 600 W and 2450 Hz – 5 Basic solution 15Micr.HPN-9 min Microwave irradiation 9 600 W and 2450 HZ – 5 Basic solution 15HeatingGHPN-1.5 h Heating in water bath 90 90 °C 0.7 g 5 Basic solution 15RefluxingHPN-6 h Heating with refluxing 360 120 °C – 10 3 g 50UltrasonicGHPN-4 h Sonication in ultrasonic bath 240 150 W 10 mL of glucose

solution (1 M)– 10 ml of NaOH

solution (1 M)60

Fig. 1. UV/Visible absorption spectra of different CNPs.

215R.N. Fallah et al. / Fuel Processing Technology 130 (2015) 214–223

photocatalysts such as TiO2/CQDs, SiO2/CQDs, Fe2O3/CQDs, ZnO/CQDs,and Cu2O/CQDs, and applied these in organic contaminant photodeg-radation with visible or near infrared light excitation [23].

The application of CNPs as an adsorbent for fuel was first reported byour research group [24,25]. In our previous work we demonstrated theefficiency of CNPs for the adsorption of thiophenic sulfur compoundsfrom simulated model diesel fuel. Because of the immiscibility of CNPsaqueous solution in organic phase, it can be used as an adsorbent forremoval of sulfur compounds from liquid fuel. Due to the presence ofcarboxylic acid groups at the surface of CNPs, they have excellentwater solubility and suitability for subsequent functionalization withvarious organic, polymeric, inorganic, or biological species [21].

In our previous works, the synthesized CNPs in aqueous mediawere presented as a novel adsorbent for the removal of sulfur from liq-uid fuel and the importance of presence of PEG in preparation of carbonnanoparticles which have capability for desulfurizationwas not clear. Inthe present study, it is found that only CNPs that are produced in thepresence of PEG can remove the organosulfur compounds. To furtherclarify the dominant role of PEG in the production of nanoparti-cles with desulfurization ability and the adsorption mechanism of sul-fur compounds by CNPs, we synthesized CNPs by different, rapidand low cost methods. The applied methodologies were microwavepyrolysis [26], alkali assisted heating [22], alkali assisted ultrasonica-tion [27] and alkali assisted refluxing [28]. The synthesized CNPs were

Scheme 1. The preparation, passivation

then used for the adsorption of the thiophenic compounds benzothio-phene (BT), dibenzothiophene (DBT), 4,6-dimethyldibenzothiophene(DMDBT), naphthalene (NP) and 1-methylnaphthalene (MNP) from

and adsorption process by CNPs.

Fig. 2. FTIR spectrum of different CNPs.

Fig. 4. Size histogram of different CNPs.

216 R.N. Fallah et al. / Fuel Processing Technology 130 (2015) 214–223

model diesel fuel solutions. In addition, the following objectives of thiswork are (i) to investigate the potential of synthesized carbon nanopar-ticles by differentmethods as adsorbents for desulfurization (ii) to eval-uate the mechanism of adsorption of sulfur compounds by CNPs whichhave been produced in the presence of PEG.

2. Experimental

2.1. Preparation of the adsorbent

Different preparation methods with some modifications havebeen utilized for the production of the dispersed CNPs in aqueousphase. These include irradiation in a microwave oven in different con-ditions and compositions [26], heating in a water bath under alkaliconditions [22], sonication in an ultrasonic bath under alkali con-ditions [27], and alkali assisted refluxing [28] methods. The resultantsamples according to preparation method and conditions are referredto as micr.GHP-9 min, micr.HP-9 min, micr.HPN-5 min, micr.HPN-9 min,heatingGHPN-1.5 h, refluxingHPN-6 h and ultrasonicGHPN-4 h, where micro-wave irradiation (micr.), heating, refluxing and ultrasonication (ultra-sonic) indicate the preparation method and G, H, P and N denoteglucose, distilled water, PEG 200 and sodium hydroxide, respectively.In Table 1 and Scheme 1, different types of adsorbent with the

Fig. 3. TEM image of different CNPs, a) micr.GHP-9 min, b) micr.HP-9 min, c) micr.HPN-5 min

preparation method, preparation conditions and the applied reactantfor their generation are listed.

+D-Glucose and PEG-200 were supplied by Merck Co. Extra purereagents of BT, DBT, DMDBT (purity N 98%), NP (99%), MNP (99%), n-heptane (N99%), sodium hydroxide and ethylbenzene were purchasedfrom Sigma-Aldrich Co. The chemicals were used without furtherpurification.

2.2. Method

2.2.1. Adsorbent characterizationAll the sampleswere characterized by various techniques. The ultra-

violet–visible (UV–vis) absorption spectra of CNPs were obtained using

, d) micr.HPN-9 min, e) heatingGHPN-1.5 h, f) refluxingHPN-6 h and g) ultrasonicGHPN-4 h.

Fig. 5. Removal percentage of total sulfur contents of MDF1 as a function of time and at different runs for different CNPs, including micr.GHP-9 min, micr.HP-9 min, micr.HPN-5 min andmicr.HPN-9 min.

217R.N. Fallah et al. / Fuel Processing Technology 130 (2015) 214–223

a PerkinElmer Lambda 45 UV/Vis spectrophotometer. Fourier transforminfrared spectroscopy (FTIR) was carried out on a Bruker Vertex 70 FTIRspectrometer using the attenuated total reflectance method (ATR) at aresolution of 4 cm−1. High-resolution transmission electronmicroscopy(HRTEM) was performed using a double-aberration corrected JEOL2200FS (JEOL, Japan) microscope equipped with a field emission gun(FEG) operated at 200 kV. The size distribution was determined byMalvern Zetasizer, Nano Series.

2.2.2. Adsorption experimentsThe adsorption experiments were conducted in a stirred batch

system at ambient temperature. In order to evaluate the adsorptionfrom liquid phase, two model diesel fuels (MDF1 and MDF2) wereprepared. MDF1 contained equal concentrations of benzothiophene(BT), dibenzothiophene (DBT) and 4,6-dimethyldibenzothiophene(DMDBT) in n-heptane. In theMDF2, in addition to the three thiophenicsulfur compounds, two aromatic compounds naphthalene (NP) and 1-methylnaphthalene (MNP) were added with equal concentration. Theconcentration of each component was about 3.12 mmol/L (equal to100 mg/L of sulfur). Thus, the total sulfur concentration in the MDFswas 300 mg/L. In all adsorption experiments, 10 mL of the preparedMDF solution was added to each bottle containing 10 g of CNP aqueoussolution and the two phase system was magnetically stirred at roomtemperature for 40 min. At different interval times, the samples weretaken from the organic phase to measure the residual content of sulfurcompounds and aromatics in the MDF solution. The concentration ofBT, DBT and DMDBT as a total sulfur concentration was determinedusing UV–vis spectroscopy. The absorbance was measured at 232 nm.The residual concentration of thiophenes in heptane after adsorption

of BT, DBT and DMDBT in the presence of NP and MNP (from MDF2)were determined using a gas chromatograph (GC) (Agilent 6890 N),equipped with a flame ionization detector (FID). The compounds wereseparated by a HP-5 capillary column 30m in length, 0.25 mm in inter-nal diameter, andwith 0.25mmoffilm thickness. The oven temperaturewas initially set at 80 °C for 1 min and then ramped at 10 °C/min to200 °C and held at 200 °C for 1 min, followed by a ramp at20 °C/minfrom 200 to 260 °C, then ramped immediately at 5 °C/min to 300 °C.In these analyses, ethylbenzene was used as an internal standard forthe quantitative analysis of each compound.

3. Results and discussion

3.1. Preparation and characterization of adsorbents

As previously reported, the aqueous solution of CNPs which wasprepared by a microwave assisted method [26] was successfully usedfor the separation of sulfur compounds from a simulated liquid fuel.In this route, a clear solution which was prepared by dissolving glu-cose and PEG-200 in distilled water was applied to produce a darkbrown aqueous solution of CNPs by irradiating in a microwave oven.For more details see Table 1.

These prepared carbon nanoparticles showed excellent capacity,selectivity and rapid feasibility for decreasing sulfur content of liquidfuel due to their immiscibility in the organic phase [24,25]. CNPs canbe synthesized by various facile methods. Here, in addition to micro-wave pyrolysis approach, CNPs were prepared by heating, refluxingand ultrasonicating techniques which have been reported before andthen tested for separation of sulfur from fuel.

Fig. 6. Removal percentage of total sulfur content of MDF1 as a function of time and at different runs for different CNPs, including heatingGHPN-1.5 h, refluxingHPN-6 h and ultrasonicGHPN-2 h.

218 R.N. Fallah et al. / Fuel Processing Technology 130 (2015) 214–223

It was expected that all samples would show the same capability fordesulfurization as CNPs produced by the first method (micr.GHP-9 min).Interestingly, two samples i.e. heatingGHPN-1.5 h and ultrasonicGHPN-4 h,were not able to remove sulfur compounds from fuel andRefluxingHPN-6 h sample was the only sample that had sulfur removal

Table 2Comparison of obtained results for prepared CNPs.

Adsorbents Adsorbates % Removal S-capacity of the CNPs dispersedin aqueous solution (g S/L)

Micr.GHP-9 min BT 80 2.5DBT 83DMDBT 56

Micr.HP-9 min BT 81 2.5DBT 83DMDBT 56

Micr.HPN-5 min BT 79 2.5DBT 83DMDBT 56

Micr.HPN-9 min BT 78 2.5DBT 82DMDBT 54

HeatingGHPN-1.5 h BT 56 1.7DBT 57DMDBT 20

RefluxingHPN-6 h BT 40 1.2DBT 40DMDBT 10

UltrasonicGHPN-4 h BT 30 0.9DBT 29DMDBT 10

ability comparable to micr.GHP-9 min. All characterization methods in-cluding FTIR, UV/Vis and TEM image confirmed the generation of carbonnanoparticles by all routes. It can be suggested here that the only gener-ation and the presence of carbon nanoparticle in the aqueous phase isnot sufficient for the removing of sulfur compounds and the particularfunctionality of these materials or types of initial reactant is important.According to the technique that was presented by Bhunia [22],a heatingGHPN-1.5 h sample was synthesized by dissolving glucose inwater, the pH of the solution was made basic by the addition of NaOHsolution, and the resulting solution was heated in a water bath. The ob-tained CNPs which are directly soluble in water did achieve some sulfurremoval, but the removal percentagewas only 5%. This processwas test-edwithout using glucose but carbonnanoparticleswere not formed. It islikely related to lower reaction temperature which is not sufficient todecompose the PEG molecule to carbon.

This result was the same for the aqueous soluble CNPs preparedby Li and coworkers (ultrasonicGHPN-4 h) [27]. They synthesizedmonodispersed carbon nanoparticles directly from glucose solutionand NaOH solution by ultrasonic treatment. It should be mentionedhere that without using glucose, preparation of CNPs by this methodaswell as by heatingmethod is not possible. Finally, as it has beenmen-tioned, the CNPs which were synthesized by microwave irradiation(micr.GHPN-9 min) and refluxing (refluxingHPN-6 h) procedures could re-move sulfur from heptane. In later methodology, the fluorescent carbonnanoparticles were efficiently synthesized by refluxing PEG, distilledwater and sodiumhydroxide. Based on these observations, it is assumedthat the presence of PEG is essential for the preparation of efficient CNPsfor removing sulfur compounds from fuel.

Fig. 7. Variation of removal percentage of BT, DBT and DMDBT from heptane with a) mass of glucose, b) volume of distilled water and c) volume of PEG 200.

219R.N. Fallah et al. / Fuel Processing Technology 130 (2015) 214–223

Recently it has been reported that many organic molecules couldserve the purpose of surface passivation of CNPs [29]. Surface passiv-ation is the most promising technique for enhancing the photolu-minescence (PL) properties of CNPs. After passivation the CNPs exhibitminimum toxicity, so they are used as a bioimaging agent againsthuman red blood cells (RBCs) and the human liver carcinoma Hep G2cell line [30]. Polymers, such as diamine-terminated oligomeric poly-(ethylene glycol) (PEG1500N) and poly (propionylethyleneimine-co-ethyleneimine) (PPEI-EI), have been widely used as passivating agent[31,32]. Recently the nontoxic polymer PEG-200 has been usedwith dual purpose, both as the source of CNPs and afterward as a passiv-ating agent (self-passivated CNPs) [30,33]. Our results in separationmedia showed that only passivated CNPs are capable of the uptake ofsulfur fromorganic phase. In order to confirm this claim,we have slight-ly adapted the previous CNP production methods (heating andultrasonicating in a water bath) by adding PEG 200 to the preparationprocess. In the Scheme 1 schematic diagram for preparation, passivationand adsorption by CNPs is presented. All methodologies and details arelisted in Table 1. The synthesized dispersed CNPs in aqueous phase havebeen characterized by various techniques and the results are given inFigs. 1–4.

Fig. 1 presents the typical UV–Visible absorption spectra of the pre-pared CNPs. As shown, CNP samples prepared by different routes givea characteristic peak in the 225–300 nm regions. These peaks indicatedthe typical absorption of an aromatic pi system, which is similar to thatof polycyclic aromatic hydrocarbons [27].

Fourier transform infrared (FTIR) spectroscopywas utilized to deter-mine the functional groups of the prepared CNPs (Fig. 2). All the sam-ples displayed many common characteristics in their spectra. Thepeaks around 3400, 1700, and 1100 cm−1 respectively correspond tothe vibrations of C\OH, C_O, and C\O bonds, and the peaks at about3000 and 1500 cm−1 are due to the C\H stretch of polycyclic aromatichydrocarbons [28]. In addition a characteristic peak appeared around1400 cm−1 which clarified theO\Hbending [30]. FTIR analysis showedthat the surfaces of the CNPswere significantly coveredwith hydrophil-ic groups (\OH and\COOH)which are produced from the decomposi-tion of PEG at elevated temperature and are covalently bonded to thecarbon frameworks, improving the hydrophilicity and stability of thenanoparticles in aqueous systems [27]. The existence of these functionalgroups on the CNP surface can contribute in removing sulfur from liquidfuel via acid–base interactions [24,25].

Fig. 3 shows the transmission electron microscopy (TEM) images ofthe prepared CNPs. These TEM images illustrated that the particles arewell dispersed, spherical and extremely small in size. From the particlesize distribution, the average size of the obtained CNPs was determinedto be between 1 and 10 nm, as shown in Fig. 4.

3.2. Adsorption experiments

In order to compare the adsorption efficiency of different passivatedCNPs for the removal of thiophenic sulfur compounds from liquid fuel,the adsorption of BT, DBT and DMDBT from two different simulated

Fig. 8. Removal percentage of different components from MDF2 as a function of time by the used CNPs, including micr.GHP-9 min, micr.HP-9 min, micr.HPN-5 min and micr.HPN-9 min.

220 R.N. Fallah et al. / Fuel Processing Technology 130 (2015) 214–223

diesel fuels (MDF1 and MDF2) onto dispersed CNPs in aqueous phasewas carried out as a function of time. All the adsorption experimentswere conducted using a batch method.

Figs. 5 and 6 show the obtained result for the removal of total sulfur(as BT, DBT and DMDBT) from MDF1 by various CNPs in three conse-quence runs. In all cases the produced CNPs have sulfur uptake ability.These figures make it clear that the removal percentage is not thesame for all samples. The trend of the performance of different CNPsin desulfurization process was micr.GHP-9 min ≈ micr.HP-9 min ≈micr.HPN-5 min ≈ micr.HPN-9 min N heatingGHPN-1.5 h N refluxingHPN-6 h N

ultrasonic GHPN-2 h. Sulfur removal performance is strongly dependenton the preparation method and process conditions. These results alsohave been briefly clarified in Table 2. Apparently the presence of PEGin the process of CNP preparation enhances their efficiency for desulfur-ization. Almost the same amount of PEG was used in all the preparationprocedures. In the present work, PEG and distilled water were appliedwith a ratio of 3:1 (volume/volume) for the production of CNPs by allmethodologies. This ratio was determined by optimization experimentswhich are discussed further in detail below.

The amount of raw materials which is used for the production ofCNPs by themicrowave assistedmethod (micr.GHP-9 min) was optimizedaccording to their ability to remove sulfur compounds (BT, DBT andDMDBT) from heptane. The impact of the initial values of glucose, PEGand distilled water on the sulfur removal efficiency was also investigat-ed. In all these experiments, the microwave exposure time (9 min) andmicrowave power (600 W) were kept constant.

Fig. 7a illustrates the effect of the weight of glucose (1–4 g) onthe removal of three thiophenic sulfur compounds (BT, DBT andDMDBT) under constant conditions of PEG (10 mL) and of distilledwater (10 mL). It is concluded that increasing the weight of glucosefrom1 to 4 g does not impact on BT, DBT andDMDBT removal efficiency,

which remains almost constant. When the amount of glucose increases,not all of it changes into carbon nanoparticles and it remains in the op-erating environment. Thus the lowest amount of glucose is the opti-mum weight for preparing CNPs (micr.GHP-9 min). It should be notedhere that CNPs can also be produced using PEG without glucose. Ashas beenmentioned, in this case the PEG can act as both a carbon sourceand a passivating agent (micr.HP-9 min). It was also found that the addi-tion of NaOH reduces the time needed for the construction of CNPs(micr.HPN-9 min and micr.HPN-5 min). The variation in the amount of dis-tilledwater (5–20mL)was also evaluated (see Fig. 7b).With an increas-ingwater volume, the sulfur removal amount decreased for BT and DBTfrom 74% to 70% and for DMDBT from 44% to 40%, respectively. With anincreasing microwave oven temperature, a great amount of the watervaporized and the volume of final solution was almost half the volumeof the initial solution. When the water volume increased, evaporationdramatically increased,wasting a great amount of solution and reducingthe final amount of the prepared CNPs. Therefore theminimum volumeof distilled water (5 mL) was chosen as the optimal amount.

Fig. 7c shows the influence of the applied PEG on the removal rate ofthiophenic compounds in the preparation of carbon nanoparticles(8–20 mL). As this figure indicates, at 15 mL of PEG the removal per-centage of BT and DBT increased from 60% to 65% and for DMDBTfrom35% to 45%, respectively. Thereafter the removal amount remainedconstant. Therefore, to reduce the amount of chemical substances used,15 mL of PEG was selected as the optimum dose for the preparation ofCNPs. The obtained optimum ratio for PEG to H2O in the microwavemethod was 3:1, so this ratio was applied to all the preparationmethods. It is necessary to determine the optimum amount for eachprocedure separately. This work, however, aimed to investigate the ef-fect of the presence of PEGon the preparation of CNPs and their efficien-cy in the removal of sulfur compounds from the organic phase.

Fig. 9. Removal percentage of different components from MDF2 as a function of time by the used CNPs, including heatingGHPN-1.5 h, refluxingHPN-6 h and ultrasonicGHPN-2 h.

221R.N. Fallah et al. / Fuel Processing Technology 130 (2015) 214–223

Generally this portion for PEG to H2O was extended to the entiresystem.

As it has been mentioned, the surface of CNPs is full of hydrophilicgroups such as hydroxyl and carboxylic groups, which may be due tothe decomposition of PEG molecules with an increasing temperature.The introduction of such functional groups is a contributory factor inthe adsorption of sulfur compounds by CNPs via acid–base interaction,because of the basic properties of thiophenic compounds. The higherobserved total sulfur removal efficiency trend of micr.GHP-9 min,micr.HP-9 min, micr.HPN-5 min and micr.HPN-9 min samples might be due tothe higher temperature in the microwave oven and higher amount ofPEG decomposition in these samples. It should be noted here thatchanging the reaction conditions in this method (omitting the glucoseand using NaOH) does not affect their efficiency. It was also observedthat the ultrasonic GHPN-2 h sample prepared by sonication at room tem-perature showed the lowest removal efficiency.

To further investigate the role of PEG and themechanism of adsorp-tion by CNPs, the adsorption of BT, DBT, DMDBT, NP and MNP fromMDF2 on different CNP sampleswas analyzed. The results are illustratedin Figs. 8 and 9. These adsorption experiments were done in threeconsecutive runs of all CNP samples without regeneration. The curvespresented in these figures belong to the first run in each case. Allsamples showed similar trends for each component. The observedtrend is DBT ≈ BT N NP N MNP≈ DMDBT. This trend reveals not onlythat acid–base interaction between acidic site on CNPs and slightlybasic sulfur compounds contributed to the adsorption process, butalso that other factors such as polarity and molecular size of adsorbatesaffect the process of sulfur compound removal from liquid fuel via

dispersive interaction. In spite of the higher basicity of DMDBT, itsadsorbed amount is the lowest. It seems that in addition to acid–baseinteraction, other factors such as dispersive interaction in small porescan govern the amount of sulfur removed by CNPs. The order of criticaldiameter of the molecules or the area of the molecular plane increasesin the order of NP b BT bMNP b DBT b DMDBT [7]. Thus the adsorptionmechanism can be proposed by considering the adsorbate size andthe interaction between adspecies and CNPs. The presence of acidicfunctional groups on CNPs, resulting from both the decomposition ofPEG and probably glucose molecules and the basic characteristic ofthiophenic compounds, confirm the strong electrostatic interaction be-tween sulfur molecules and CNP samples. The existence of small poreson carbon samples can also be explained by the fact that only passivatedCNPs can remove sulfur.

Recently, Jia et al. published a simple PEG assistant solution syntheticmethod for the preparation of porous ZnO micro/nanostructures [34].They demonstrated that the presence of hydrophobic groups of PEG re-sulted in the existence of microheterogeneities and PEG aggregations inthe mixed solution, which act as a soft template and play a key role inthe formation of the porous structure of ZnO. They also stated that thehigh reaction temperature resulted in a decreased hydrogen bondingbetween the PEG and water molecules, increased hydrophobic interac-tions between the PEG chains and the increased pore size of the finalZnO products. Based on their results, it is assumed that such a phenom-enon might also occur in our systems.

The adsorption experiments clarified that only CNPswhich had beenprepared in the presence of PEGwere able to remove sulfur compounds.These results suggested that porous CNP nanostructureswere formed in

Fig. 10. Removal percentage of different compounds from MDF2 at different runs for regenerated CNP samples (micr.GHP-9 min and micr.HP-9 min).

222 R.N. Fallah et al. / Fuel Processing Technology 130 (2015) 214–223

a mixture of distilled water and PEG. It is important to mention that allthese conclusions are based on adsorption data and not on the micro-scopic images. As Jia et al. have already observed, hydrophobic PEG ag-gregations play a key role in pore formation. These aggregations arerelated to the hydrogen bonding strength between PEG and waterwhich is heavily temperature dependent: at higher temperatures, thehydrogen bonding between PEG and water decreases and more poresare formed. These pores are most likely to be small in size, because al-though DMDBT has more basicity, it cannot be adsorbed effectively,which is probably due to its larger size and possession of two methylgroups, which give it more steric hindrance. The preferential adsorptionof thiophenic compounds can therefore be explained by their size andthe interaction between adsorbate and adsorbent. It should be notedhere that the mechanism of different methods for the preparation ofCNPs and their porosity needs to be studied in future work.

3.3. Adsorbent regeneration

For the evaluation of CNP regeneration fresh heptane as a sol-vent was used to regenerate the used CNPs. The micr.GHP-9 min andmicr.HP-9 min samples were tested for regeneration after saturationwith MDF2model fuel. The first adsorption experiments were conduct-ed for three runs, and then the used carbon samples and 10 mL ofheptane was shaken for 40 min at 50 °C. This procedure was repeatedthree times with fresh heptane. After regeneration the adsorption ex-periments were repeated for three further runs and the obtained resultsare presented in Fig. 10. The sulfur removal efficiency of both samplesdecreased by about 20% after regeneration. This could be explained bythe fact that the adsorbed sulfur compounds on CNPs can easily be re-moved using a fresh solvent at room temperature.

4. Conclusion

The adsorption of three thiophenic sulfur compounds from modelfuel by various CNPs in aqueous solution was investigated. The resultspresented in this paper show that the presence of PEG 200 is necessaryfor thepreparation of CNP sampleswhich are efficient for the removal ofsulfur compounds. The methodology and CNP preparation process con-ditions selected were highly effective for desulfurization. Among thetested CNPs, the samples prepared by microwave pyrolysis exhibitedthe highest thiophenic sulfur removal from fuels due to introductionof more acidic functionality and porosity. The sulfur adsorption efficien-cy of the CNPs strongly depended on the adsorbent preparationmethodand critical dimension of the sulfur containing compounds. The present-ed data indicate that CNPs can be considered as promising candidatesfor desulfurization processes, mainly because of their low cost and rela-tive ease of preparation. Furthermore, CNPs can also be regeneratedusing a solvent-washing process.

References

[1] A. Nanoti, S. Dasgupta, A.N. Goswami, B.R. Nautiyal, T.V. Rao, B. Sain, Y.K. Sharma,S.M. Nanoti, M.O. Garg, P. Gupta, Mesoporous silica as selective sorbents for removalof sulfones from oxidized diesel fuel, Microporous Mesoporous Mater. 124 (2009)94–99.

[2] L. Wang, Y. Chen, L. Du, S. Li, H. Cai, W. Liu, Nickel-heteropolyacids supported on sil-ica gel for ultra-deep desulfurization assisted by Ultrasound and Ultraviolet, Fuel105 (2013) 353–357.

[3] F. Shi, M. Hammoud, L.T. Thompson, Selective adsorption of dibenzothiopheneby functionalized metal organic framework sorbents, Appl. Catal. B 103 (2011)261–265.

[4] M.A. Guzmán, R. Huirache-Acuña, C.V. Loricera, J.R. Hernández, J.N. Díaz de León, J.A.de los Reyes, B. Pawelec, Removal of refractory S-containing compounds from liquidfuels over P-loaded NiMoW/SBA-16 sulfide catalysts, Fuel 103 (2013) 321–333.

[5] C. Song, X. Ma, New design approaches to ultra-clean diesel fuels by deep desulfur-ization and deep dearomatization, Appl. Catal. B Environ. 41 (2003) 207–238.

[6] F. Subhan, B.S. Liu, Y. Zhang, X.G. Li, High desulfurization characteristic of lanthanumloaded mesoporous MCM-41 sorbents for diesel fuel, Fuel Process. Technol. 97(2012) 71–78.

[7] A. Zhou, X. Ma, C. Song, Effects of oxidative modification of carbon surface on the ad-sorption of sulfur compounds in diesel fuel, Appl. Catal. B Environ. 87 (2009) 190–199.

[8] A. Sengupta, P.D. Kamble, J. Kumar Basu, S. Sengupta, Kinetic study and optimizationof oxidative desulfurization of benzothiophene using mesoporous titanium silicate-1 catalyst, Ind. Eng. Chem. Res. 51 (2012) 147–157.

[9] S. Matsuzawa, J. Tanaka, S. Sato, T. Ibusuki, Photocatalytic oxidation ofdibenzothiophenes in acetonitrile using TiO2: effect of hydrogen peroxide and ultra-sound irradiation, J. Photochem. Photobiol. A Chem. 149 (2002) 183–189.

[10] T. Song, Z. Zhang, J. Chen, Z. Ring, H. Yang, Y. Zheng, Effect of aromatics on deephydrodesulfurization of dibenzothiophene and 4, 6-dimethyldibenzothiopheneover NiMo/Al2O3 catalyst, Energy Fuels 20 (2006) 2344–2349.

[11] M. Seredych, T.J. Bandosz, Selective adsorption of dibenzothiophenes on activatedcarbons with Ag, Co, and Ni species deposited on their surfaces, Energy Fuels 23(2009) 3737–3744.

[12] N. Abedin Khan, S. Hwa Jhung, Adsorptive removal of benzothiophene using porouscopper-benzenetricarboxylate loaded with phosphotungstic acid, Fuel Process.Technol. 100 (2012) 49–54.

[13] M. Montazerolghaem, A. Rahimi, F. Seyedeyn-Azad, Equilibrium and kinetic model-ing of adsorptive sulfur removal from gasoline by synthesized Ce–Y zeolite, Appl.Surf. Sci. 257 (2010) 603–609.

[14] J. Xiao, C. Song, X. Ma, Z. Li, Effects of aromatics, diesel additives, nitrogen com-pounds, andmoisture on adsorptive desulfurization of diesel fuel over activated car-bon, Ind. Eng. Chem. Res. 51 (2012) 3436–3443.

[15] J.-x. Guo, J. Liang, Y.-H. Chu, M.-C. Sun, H.-Q. Ying, J.-J. Li, Desulfurization activity ofnickel supported on acid-treated activated carbons, Appl. Catal. A Gen. 421–422(2012) 142–147.

[16] R. Naviri Fallah, S. Azizian, Removal of thiophenic compounds from liquid fuel bydifferent modified activated carbon cloths, Fuel Process. Technol. 93 (2012) 45–52.

[17] S. Nuntang, P. Prasassarakich, C. Ngamcharussrivichai, Comparative study on ad-sorptive removal of thiophenic sulfurs over Y and USY zeolites, Ind. Eng. Chem.Res. 47 (2008) 7405–7413.

[18] X. Yang, C. Cao, K.J. Klabunde, K.L. Hohn, L.E. Erickson, Adsorptive desulfurizationwith xerogel-derived zinc-based nanocrystalline aluminum oxide, Ind. Eng. Chem.Res. 46 (2007) 4819–4823.

[19] T. Zaki, N.H. Mohamed, M.I. Nessim, H. Abd El Salam, Characterization and applica-tion of nano-alumina sorbents for desulfurization and dearomatization of Suezcrude petrolatum, Fuel Process. Technol. 106 (2013) 625–630.

[20] A. Samadi-Maybodi, M. Teymouri, A. Vahid, A. Miranbeigi, In situ incorporation ofnickel nanoparticles into themesopores of MCM-41 bymanipulation of solvent–sol-ute interaction and its activity toward adsorptive desulfurization of gas oil, J. Hazard.Mater. 192 (2011) 1667–1674.

[21] D.R. Baker, P.V. Kamat, Photosensitization of TiO2 nanostructures with CdS quantumdots: particulate versus tubular support architectures, Adv. Funct. Mater. 19 (2009)805–811.

223R.N. Fallah et al. / Fuel Processing Technology 130 (2015) 214–223

[22] S. Kumar Bhunia, A. Saha, A. Ranjan Maity, S.C. Ray, N.R. Jana, Carbon nanoparticle-based fluorescent bioimaging probes, Sci. Rep. 3 (2013) 1473–1479.

[23] X. Han, Y. Han, H. Huang, H. Zhang, X. Zhang, R. Liu, Y. Liu, Z. Kang, Synthesis ofcarbon quantum dots/SiO2 porous nanocomposites and their catalytic abilityfor photoenhanced hydrocarbon selective oxidation, Dalton Trans. 42 (2013)10380–10383.

[24] R. Naviri Fallah, S. Azizian, Rapid and facile desulphurization of liquid fuel by carbonnanoparticles dispersed in aqueous phase, Fuel 95 (2012) 93–96.

[25] R. Naviri Fallah, S. Azizian, G. Reggers, S. Schreurs, R. Carleer, J. Yperman, Selectivedesulfurization of model diesel fuel by carbon nanoparticles as adsorbent, Ind.Eng. Chem. Res. 51 (2012) 14419–14427.

[26] H. Zhu, X. Wang, Y. Li, Z. Wang, F. Yang, X. Yang, Microwave synthesis of fluorescentcarbon nanoparticles with electrochemiluminescence properties, Chem. Commun.5118 (2009) 5118–5120.

[27] P. Bouras, E. Stathatos, P. Lianos, Pure versusmetal-ion-doped nanocrystallinetitaniafor photocatalysis, Appl. Catal. B Environ. 73 (2006) 51–59.

[28] R. Liu, H. Li, W. Kong, J. Liu, Y. Liu, C. Tong, X. Zhang, Z. Kang, Ultra-sensitive and se-lective Hg2+ detection based on fluorescent carbon dots, Mater. Res. Bull. 48 (2013)2529–2534.

[29] Y.P. Sun, B. Zhou, Y. Lin, W. Wang, K.A. Shiral Fernando, P. Pathak, M.J. Meziani, B.A.Harruff, X. Wang, H. Wang, P.G. Luo, H. Yang, M.E. Kose, B. Chen, L.M. Veca, S.Y. Xie,Quantum-sized carbon dots for bright and colorful photoluminescence, J. Am. Chem.Soc. 128 (2006) 7756–7757.

[30] S. Mitra, S. Chandra, S.H. Pathan, N. Sikdar, P. Pramanik, A. Goswami, Room temper-ature and solvothermal green synthesis of self passivated carbon quantum dots, RSCAdv. 3 (2013) 3189–3193.

[31] R. Liu, D. Wu, S. Liu, K. Koynov, W. Knoll, Q. Li, An aqueous route to multicolorphotoluminescent carbon dots using silica spheres as carriers, Angew. Chem. Int.Ed. 48 (2009) 4598–4601.

[32] L. Cao, X. Wang, M.J. Meziani, F. Lu, H. Wang, P.G. Luo, Y. Lin, B.A. Harruff, L.M. Veca,D. Murray, S.Y. Xie, Y.P. Sun, Carbon dots for multiphoton bioimaging, J. Am. Chem.Soc. 129 (2007) 11318–11319.

[33] A. Jaiswal, S.S. Ghosh, A. Chattopadhyay, One step synthesis of C-dots by microwavemediated caramelization of poly(ethylene glycol), Chem. Commun. 48 (2012) 407.

[34] Y. Ji, X.Y. Yu, T. Luo, M.Y. Zhang, J.H. Liu, X.J. Huang, PEG aggregation templated po-rous ZnO nanostructure: room temperature solution synthesis, pore formationmechanism, and their photoluminescence properties, Cryst. Eng. Commun. 15(2013) 3647–3653.