Bioresource Technology 183 (2015) 93–100

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /bior tech

Bio-aviation fuel production from hydroprocessing castor oil promotedby the nickel-based bifunctional catalysts

http://dx.doi.org/10.1016/j.biortech.2015.02.0560960-8524/� 2015 Elsevier Ltd. All rights reserved.

⇑ Corresponding authors.E-mail address: weili@nankai.edu.cn (W. Li).

Siyang Liu, Qingqing Zhu, Qingxin Guan, Liangnian He, Wei Li ⇑Key Laboratory of Advanced Energy Materials Chemistry (Ministry of Education), Collaborative Innovation Center of Chemical Science and Engineering (Tianjin), College ofChemistry, Nankai University, Tianjin 300071, China

h i g h l i g h t s

� Bio-aviation fuel was firstlysynthesized by hydroprocessingcastor oil.� Castor oil was converted into aviation

fuel by the two-step process and aone-pot process respectively.� The degree of hydrodeoxygenation

and hydrocracking could be adjustedby Ni supported on moderate acidiczeolites.

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

Different fuel range alkanes can be synthesized from hydroprocessing castor oil by Ni supported on dif-ferent acidic zeolites, and bio-aviation fuel can be obtained by Ni supported on moderate acidic strengthzeolites.

a r t i c l e i n f o

Article history:Received 3 December 2014Received in revised form 8 February 2015Accepted 9 February 2015Available online 19 February 2015

Keywords:Bio-aviation fuelCastor oilHydroprocessingAcidic zeolitesReaction mechanism

a b s t r a c t

Bio-aviation fuel was firstly synthesized by hydroprocessing castor oil in a continuous-flow fixed-bedmicroreactor with the main objective to obtain the high yield of aviation fuel and determine the elemen-tal compositions of the product phases as well as the reaction mechanism. Highest aviation range alkaneyields (91.6 wt%) were achieved with high isomer/n-alkane ratio (i/n) 4.4–7.2 over Ni supported on acidiczeolites. In addition, different fuel range alkanes can be obtained by adjusting the degree of hydrodeoxy-genation (HDO) and hydrocracking. And the observations are rationalized by a set of reaction pathwaysfor the various product phases.

� 2015 Elsevier Ltd. All rights reserved.

1. Introduction

One of the world’s most serious problems is the depletion ofpetroleum-based resources because of increased industrializationand motorization. It has been reported that the transport sector

uses 40% of the primary energy consumed in the world (Mohanet al., 2006). Furthermore, the combustion of transportation fuels,especially in the aviation industry, contributes to the greenhouseeffect due to the carbon emission. Therefore, it is necessary to findrenewable, sustainable and efficient energy sources with loweremissions. Biofuels from vegetable oils show promising potentialin the manufacture of liquid fuels. Hydrocarbons has been pro-duced by HDO technology (Hancsók, 2014; Kumar et al., 2010;

94 S. Liu et al. / Bioresource Technology 183 (2015) 93–100

Verma et al., 2011; Warner et al., 2014; Mu, 2014; Bu et al., 2012;Sukrutha, 2014; Ren, 2014; Mascarelli, 2009), and there are somereports in the literature on using vegetable oils, such as palm, soy-bean, sunflower and coconut (Bykova et al., 2012). Castor is a kindof widely cultivated, inexpensive, environmentally friendly andindustrial oil plant, meanwhile, castor oil, as an important and spe-cial feed, has long been known as the medicinal oil and an ingredi-ent in the oleochemical industry (Mutlu, 2010; Horvath et al.,2008; Brown et al., 2010; Huang et al., 2014; Elliott, 2007; Antalet al., 1991; Ambekar, 1957), which has not been studied for bio-aviation fuel by HDO technology so far.

Aviation turbine fuels are a complex hydrocarbon mixture con-sisting of different classes, such as paraffin (C8AC15), naphtheneand aromatics. Bio-aviation fuels are mainly composed of alkanesthat are used as mixing components in aviation turbine fuels, withthe largest proportion being less than 50% according to ASTMD7566. Therefore, high C8AC15 selectivity and the degree of iso-merization become the key points for meeting the mixing stan-dards of ASTM D7566. The properties of biofuels, such as freezingpoint, flash point and viscosity can be influenced by the compo-nents and the degree of isomerization.

Vegetable oils are composed of triglycerides, long-chain carbon(16–18) fatty acids; therefore, diesel range carbon chains (C16AC18)can be produced by the reaction pathways including dehydration,decarbonylation and decarboxylation (Bezergianni andDimitriadis, 2013; Peng et al., 2012; Li et al., 2013; Berenblyumet al., 2012). In this context, there have been numerous studieson biodiesel production through HDO technology. Šimácek(2009) studied the hydroprocessing of rapeseed oil over NiAMo/a-lumina hydrorefining catalysts at various temperatures (260–340 �C) under a pressure of 7 MPa in a laboratory flow reactor toproduce biodiesel fuels. The obtained organic liquid product (most-ly C17 and C18) does not contain basically any intermediates norany remaining raw material at a sufficiently high reaction tem-perature (above about 310 �C). Sotelo-Boy (2010) investigated thehydrocracking of rapeseed oil over three different types of bifunc-tional catalysts: Pt/HY, Pt/H–ZSM-5 and sulfided NiMo/c-Al2O3. Pt/zeolite-supported catalysts were found to give a strong crackingactivity, producing gasoline range alkanes for the most part dueto the strong acid sites. However, the less acidic sulfided NiMo/c-Al2O3 produced more C17 and C18 (diesel range). Therefore, theselectivity of HDO production is influenced by the acid strength.However, C8AC15 hydrocarbons and moderate isomerization selec-tivity are ideally suited for jet fuels also because of the stringentinternational specifications. Thus, adjusting the degree of HDO

Fig. 1. Hydroprocessing castor oil by Ni-based b

and hydrocracking plays a vital role in production of bio-aviationfuel.

In this work, we focused on the effects of nonsulfided nickel-based bifunctional catalysts on HDO and hydrocracking. In general,transition metals are mainly responsible for the hydrogenation ofunsaturated triglyceride and the acidic zeolite contributes to thedegree of hydrocracking. More importantly, moderate cracking,adjusted by the acidity of catalysts, can convert C16AC18 straightchain alkanes into C8AC15 or directly obtain C8AC15 hydrocarbonswith high isomerization selectivity from castor oil using a one-potprocess (as shown in Fig. 1 and Table 1). Bio-aviation fuel can beattained through a two-step process and a one-pot process. Thetwo-step process focuses on the hydrocracking of C16AC19 toC8AC15, which mainly studies the addition of Ag into Ni/SAPO-11to adjust the degree of hydrocracking. However, the C8AC15 canbe produced via the one-pot process over (3-aminopropyl)-tri-ethoxysilane (APTES) modified MCM-41/USY composite-supportedNi. In this composite system, MCM-41 and USY are covalentlybound together by the addition of APTES. The strong acid site ofUSY can be moderately covered by adjusting the ratio of MCM-41 and USY being supported by NH3–TPD to lessen cracking, andthe addition of APTES contributes to the homogeneous mixing ofMCM-41 and USY. Furthermore, the mesoporous–microporous sys-tem improves the mass transfer ability of triglyceride moleculeleading to excellent catalytic efficiency.

2. Methods

2.1. Materials

The castor oil was commercially available from Tianjin GuangfuTechnology Co. Ltd. MCM-41 (SBET: 842 m2/g), SAPO-11 (SBET:240.3 m2/g), USY (SBET: 655 m2/g), ZSM-5 (SBET: 262.5 m2/g) andH-Beta (SBET: 650 m2/g) were purchased from the catalyst plantof Nankai University.

2.2. Catalyst preparation and characterization

2.2.1. Catalyst preparationWe applied the decomposition of hypophosphite precursors

method to synthesize Ni2P/SAPO-11 as we previously reported(Guan et al., 2009), and metal phosphides would not be easilyinfluenced by AlAO of SAPO-11 in the case of high temperature cal-cination. Ni2P/SAPO-11 was prepared with P:Ni mole ratio of 1.5.

ifunctional catalysts with variable acidity.

Table 1Hydrocoversion of castor oil over Ni supported on different acidic strength zeolites.

Entry Catalyst Feed T (�C) Yield %/lighterproducts

Yield %/C5AC7 Yield %/C8AC15 Yield %/C16AC19 Conv. (%) i/n ratio (C8AC15)

1 (a) 25% Ni/SAPO-11 Castor oil 300 0.2 1.2 3.2 95.4 99 0.12 (b) 25% Ni2P/SAPO-11 Castor oil 300 0.3 4.9 2.5 92.3 99 0.13 25% NiAg/SAPO-11

(molar Ni/Ag: 44)ProductI 360 0.3 3.8 91.6 4.3 98 6.8

4 25% NiAg/SAPO-11(molar Ni/Ag: 44)

ProductII 360 0.3 5.6 88.9 5.2 98 7.2

5 25% Ni/H-Beta Castor oil 300 0.8 97.7 1.5 0 99 06 25% Ni/ZSM-5 (Si/Al: 38) Castor oil 300 1.2 96.8 2.0 0 99 07 25% Ni/ZSM-5 (Si/Al: 50) Castor oil 300 0.9 93.4 3.2 2.3 99 0.58 25% Ni/ZSM-5 (Si/Al: 100) Castor oil 300 0.7 58.8 25.3 15.2 99 2.49 25% Ni/USY Castor oil 300 0.6 79.4 19.2 0.8 99 5.310 25% Ni/USY–APTES–

MCM-41 (APTES: 7.5%)Castor oil 300 0.5 13.8 80.3 5.4 99 4.4

i/n ratio (C8AC15) = ratio of isomers (i) and normal (n) alkanes. 3 MPa, WHSV = 2 h�1.ProductI: C16AC19 (hydroprocessing castor oil by 25% Ni/SAPO-11).ProductII: C16AC19 (hydroprocessing castor oil by 25% Ni2P/SAPO-1).Lighter products: C1AC4, CO, CO2.

S. Liu et al. / Bioresource Technology 183 (2015) 93–100 95

The catalysts Ni/SAPO-11, Ni/H-Beta, Ni/ZSM-5 (Si/Al: 38, 50and 100) and Ni/USY were prepared by wet impregnation of aque-ous solutions of Ni(NO3)2 on the support. A series of different Niloadings (10%, 15%, 20%, 25% and 30%) were investigated. NiAAg/SAPO-11 was also prepared by co-impregnation of aqueous solu-tions of Ni(NO3)2 and AgNO3 on the support. The best catalyticactivity of NiAAg was at a mole ratio of 44.

Ni/MCM-41–APTES–USY was prepared by impregnation ofaqueous solutions of Ni(NO3)2 on MCM-41–APTES–USY. The sup-port was prepared as follows: APTES was added to a three-neckedflask containing ethanol, USY and MCM-41. After ultrasonic disper-sion for 30 min, the solution was refluxed for 10 h. Then, the mix-ture was filtered using a membrane with a pore size of 0.45 lmand washed thoroughly three times with ethanol and deionizedwater. This was followed by drying at 80 �C in vacuo for 12 h andthen calcined at 550 �C in air for 3 h to form MCM-41–APTES–USY.

The catalyst was pelleted, crushed and sieved with 20–40 mesh.The catalyst (3.00 g) was diluted with silica sand to a volume of5.0 mL in the reactor.

2.2.2. Catalyst characterizationPowder X-ray diffraction (XRD) was performed on a Bruker D8

focus diffractometer with Cu Karadiation at 40 kV and 40 mA. X-ray photoelectron spectra (XPS) were recorded using a Kratos AxisUltra DLD spectrometer employing a monochromated Al Ka X-raysource (hm = 1486.6 eV), hybrid (magnetic/electrostatic) optics,and a multichannel plate and delay line detector. All XPS spectrawere recorded using an aperture slot of 300 � 700 lm survey spec-tra were recorded with a pass energy of 80 eV, and high-resolutionspectra with a pass energy of 40 eV. To subtract the surface charg-ing effect, the C1 speak has been fixed at a binding energy of284.6 eV. NMR spectra were recorded at 25 �C using an Avance500 MHz spectrometer (Bruker Biospin, Billerica, MA) usingdeuterated dimethyl sulfoxide as solvent. Temperature pro-grammed desorption (TPD) experiments were performed with aMicromeritics AutoChem II 2920. All samples were pretreated inN2 (25 mL/min) at 325 �C for 2 h. The desorption step was per-formed from 100 �C to 700 �C at a heating rate of 10 �C/min. Trans-mission electron microscopy (TEM) images were acquired using aJEOL-2010 FEF high resolution transmission electron microscope.Nitrogen adsorption desorption isotherms of samples at 77 K weremeasured using a BEL-MINI adsorption analyzer. Samples weredegassed at 473 K for 3 h before measurement. The surface areawas calculated using a multi-point Brunauer–Emmett–Teller(BET) model. The pore size distribution was obtained from the des-

orption isotherms using the Barret–Joyner–Halenda (BJH) model,and the total pore volume was estimated at a relative pressure of0.99, assuming full surface saturation with nitrogen. The liquidproducts were analyzed using an Agilent 7890 FID–GC equippedwith an HP-5 capillary column. The following temperature pro-gram was used: initial temperature 50 �C for 10 min, heating10 �C min�1 to 100 �C with a dwell time of 10 min, and then10 �C min�1 to 200 �C for 10 min. The identification was furtherconfirmed using GC–MS analysis. The gas-phase products were col-lected once under every reaction condition and analyzed using astandard three-column GC setup with flame ionization and ther-mal conductivity detectors.

2.2.3. Catalytic activity measurementsCastor oil (raw material) was diluted with an inert solvent,

cyclohexane, (mass ratio 1.5:1) before feed introduction. TheHDO reaction was carried out in a continuous-flow fixed-bedmicroreactor. The catalyst was pelleted, crushed and sieved with20–40 mesh. Samples of reaction products were collected at 1 hintervals after a stabilization period of 8 h.

2.2.4. The blank test for HDO and hydroisomerization reaction usingsilica

HDO reaction of castor oil: silica was used in the HDO reactionof castor oil under the condition of 300 �C, 3 MPa, andWHSV = 2 h�1, with a H2 flow rate of 160 mL min�1 at atmosphericpressure. The liquid products do not have any changes compared tothe raw material (detected by gas chromatography).

Hydroisomerization reaction of C17AC18 (HDO products of cas-tor oil): silica was used in the Hydroisomerization reaction ofC17AC18 under the condition of 360 �C, 3 MPa, and WHSV = 2 h�1,with a H2 flow rate of 160 mL min�1 at atmospheric pressure.According to the analysis of gas chromatography, the liquid prod-ucts do not have any changes compared to the raw material(detected by gas chromatography).

3. Results and discussion

3.1. Synthesis of Ni2P/SAPO-11 and USY–APTES–MCM-41

Ni2P/SAPO-11 can successfully be synthesized by decomposi-tion method of hypophosphite precursors according to the XRDpatterns (a), TEM images and XPS spectra (b) of Ni 2p for Ni2P/SAPO-11 (see Fig. S1). Peaks arising from the Ni2P phase and theSAPO-11 phase are visible (Fig. S1(a)). And it can be seen that the

96 S. Liu et al. / Bioresource Technology 183 (2015) 93–100

Ni2P particles disperse evenly in SAPO-11. The XPS spectra of Ni 2pof Ni2P/SAPO-11 and Ni/SAPO-11 are shown in Fig. S1(b). The bind-ing energy of Ni 2p for Ni/SAPO-11 is lower than that of Ni2P/SAPO-11. The higher binding energy indicates a lower electron density,which is influenced by the formation of the Ni2P phase (Saidiet al., 2014; Ardiyanti et al., 2012; Barta et al., 2010; Parizottoet al., 2007; Jeong and Kang, 2010; Shi et al., 2014). Fig. S2 showsthe nitrogen adsorption–desorption isotherms of USY, MCM-41,USY–MCM-41 (1:1) and USY–APTES–MCM-41 (USY: MCM-41 = 1:1 and the amount of APTES accounts for 5%). MCM-41 exhi-bits representative type IV curves with sharp capillary condensa-tion steps at relative pressures of 0.4–0.7, which is typical for themesoporous structures, and USY shows the typical microporestructure. In contrast, USY–MCM-41 and USY–APTES–MCM-41exhibits a combination of mesoporous and microporous structuresfrom the adsorption–desorption isotherms. The capillary conden-sation steps of USY–APTES–MCM-41 become gentler than thoseof USY–MCM-41, which indicates better mixing in the ATPESsystem.

3.2. The distribution of acidic sites for the Ni-based bifunctionalcatalysts

The distribution of acidic sites of the bifunctional catalysts isshown in Table 2. In short, the desorption temperature of Ni/SAPO-11, Ni2P/SAPO-11 and NiAAg/SAPO-11 are found to be 93,71 and 80 �C, respectively (entries 1–3), being ascribed to the weakacid sites. While the desorption temperatures of Ni/H-Beta, Ni/ZSM-5 (38, 50 and 100) and Ni/USY are in the range of 456–500 �C which attributes to the strong acid sites (entries 5–8) (Shiet al., 2014). However, for the mesoporous–microporous system,the addition of MCM-41 can cover the strong acid sites and makethe adsorption peak shifting, to some extent, to a lower tem-perature (entries 10–12). The strong acid peak disappears withan increase APTES, leading to the increases of the amount of weakacid. Consequently, fuels with different alkane ranges can beobtained using bifunctional catalysts with differing acidity levels.

3.3. The two-step process for production of aviation fuel byhydroprocessing castor oil

The main products from hydroprocessing castor oil are given inFig. S3. The presence of the gaseous products CO, CO2, CH4 and

Table 2The acidic site distribution of different acid strength and amount Ni-based bifunctional ca

Entry Catalyst Amount of acid sites (mmol NH3 g�1)

T < 200 �C 200

Amount of acidsites

Desorptiontemperature (�C)

Amsite

1 Ni/SAPO-11 1.4 93 –2 Ni2P/SAPO-11 0.5 71 –3 NiAg/SAPO-11 2.4 80 –4 Ni/H-Beta 1.6 103 –5 Ni/ZSM-5 (Si/Al: 38) 1.7 166 0.16 Ni/ZSM-5 (Si/Al: 50) 1.2 159 0.27 Ni/ZSM-5 (Si/Al: 100) 1.4 171 0.38 Ni/USY 2.7 73 and 1909 Ni/USY–MCM-41 (1:1) 2.2 72 0.210 Ni/USY–APTES (5%)–MCM-

41 (1:1)3.2 67 0.1

11 Ni/USY–APTES (7.5%)–MCM-41 (1:1)

3.0 68 and 170 0.0

12 Ni/USY–APTES (10%)–MCM-41 (1:1)

2.8 64 and 173

C3H8 (GC analysis) is attributable to decarbonylation, decarboxyla-tion and hydrocracking.

Fig. 2 shows NH3–TPD curves of the two-step process catalysts.The first step focuses on the HDO of castor oil over the weak acidicstrength catalyst, such as Ni/SAPO-11, Ni2P/SAPO-11 and Ni/MCM-41 which can produce high yields of C17AC18, but the mole ratio ofC17/C18 is different as shown in Fig. 2(a). The higher C17/C18 moleratio of Ni/SAPO-11 indicates the greater importance of the decar-bonylation pathway, while the lower C17/C18 mole ratio of Ni2P/SAPO-11 and Ni/MCM-41 indicates the greater importance of thedehydration pathway. Ni2P/SAPO-11 and Ni/MCM-41 possess moreamounts of weak acid than Ni/SAPO-11 (see NH3–TPD results) as aresult of increasing dehydration trend, which leads to theincreased molar ratio of C17/C18 is. In addition, for Ni2P/SAPO-11,there are two reasons for greater amounts of weak acid. Firstly,the large particle diameter of Ni2P may cover the acidity sites ofSAPO-11, as shown in Fig. S1(a). Secondly, PAOH formation, replac-ing AlAOH, results in the peak moving to the low temperature area,as shown in Fig. 2(b). In summary, high C17AC18 yields (90–96 wt%)with different molar ratio were achieved by Ni supported on differ-ent acidic strength zeolites. Thus the high yield of aviation rangealkanes could be produced from C17AC18 products by the secondprocess which mainly focuses on the degree of cracking and iso-merization for catalysts.

C8AC15 alkanes with high isomerization selectivity wasobtained by hydrocracking and the isomerization of the first step’sproduct (C17AC18) on a series of catalysts with different Ni/Ag moleratios supported on SAPO-11 (Ni + Ag = 25%), as summarized inTable S1. We has used Ni/SAPO-11 and Pt/SAPO-11, which are con-sidered as effective hydroisomerization catalysts to hydrocrack C17

and C18, and found that Ni/SAPO-11 and Pt/SAPO-11 catalysts had astrong cracking activity leading to a lower yield in the jet-fuelrange and high yields of C5AC7 (Table S1 entries 1 and 2). However,the addition of Ag can restrain cracking as a result of the highyields of C8AC15 and low yields of C5AC7 (Table S1). Therefore,the level of cracking is also depended on the acid strength andamount of the catalysts (see Table 2). In addition, the NH3 adsorp-tion peak of NiAAg/SAPO-11 is lower than those of Ni/SAPO-11 andPt/SAPO-11. With the increasing Ni/Ag molar ratio, the yields ofC8AC15 and the isomerization selectivity increased. However,when the Ni/Ag molar ratio is higher than 44, the yields ofC8AC15 begin to decrease (Table S1 entries 3–8), which indicatedthat excessive Ag can cover parts of the acidity on SAPO-11 leadingto lower yields of C8AC15 and isomerization selectivity.

talyst.

�C < T < 400 �C T > 400 �C

ount of acids

Desorptiontemperature (�C)

Amount of acidsites

Desorptiontemperature (�C)

– – -– – –– – –– 0.5 456

260 0.6 469220 0.4 420389 0.2 –

– 0.1 468216 0.05 501205 – –

8 210 – –

– –

Fig. 2. NH3–TPD curves of Ni/SAPO-11, Ni/MCM-41 and Ni2P/SAPO-11 (a) as well as the conversion of C17 and C18 from hydrotreatment of castor oil, NH3–TPD curves ofNiAAg/SAPO-11, Pt/SAPO-11 and Ni/SAPO-11 (b).

Fig. 3. The product distribution for hydroprocessing castor oil (a) and bio-jet fuel range alkane yields and isomerization selectivity (b) on Ni/USY, Ni/MCM-41–USY, Ni/MCM-41–APTES (5%)–USY, Ni/MCM-41–APTES (7.5%)–USY and Ni/MCM-41–APTES (10%)–USY.

S. Liu et al. / Bioresource Technology 183 (2015) 93–100 97

3.4. The one-pot process for production of aviation fuel byhydroprocessing castor oil

Knowing that the bio-aviation fuel range alkanes could not beobtained using acidic bifunctional catalysts that were too strongor too weak, the Ni/MCM-41–APTES–USY was synthesized for

preparing bio-aviation fuel range alkanes to adjust the acidstrength of the bifunctional catalysts. The catalytic process of25% Ni/MCM-41–APTES–USY and 25% Ni/MCM-41–USY are shownin Fig. S4. When triglyceride molecules are adsorbed on USY, whichmakes more cracking resulting in the production of C5AC10. Whentriglyceride molecules are adsorbed on MCM-41, C17AC19 can be

Scheme 1. The conversion of castor oil with H2 catalyzed by nickel-based bifunctional catalyst.

98 S. Liu et al. / Bioresource Technology 183 (2015) 93–100

obtained. Therefore, Ni/MCM-41–USY produced less bio-aviationfuel range alkanes. However, this problem is solved by the additionof APTES, which covalently bound USY and MCM-41 based on theexperimental results (Fig. 3). Fig. 3(a) shows the product distribu-tion from the deoxygenation of castor oil over Ni/USY, Ni/MCM-41–USY and Ni/MCM-41–APTES–USY. We can clearly see that theaddition of MCM-41 could cover the strong acid sites (proved byNH3–TPD) to obtain more yield of bio-aviation fuel range alkanes;however, C17AC19 also account for 30%. This is because of theuneven system for MCM-41–USY, as shown in Fig. S3. ForFig. 3(b), Ni/MCM-41–APTES (7.5 wt%)–USY could yield 80.3%C8AC15 with high isomerization selectivity (i/n 4.4) under the reac-tion conditions (300 �C, 3 MPa, 2 h�1). In addition, the meso-porous–microporous system improves the mass transfer abilityof triglyceride molecule in the pore channel of zeolite leading toimproved catalytic efficiency. With increases in ATPES, the yieldof aviation range alkanes began to decrease presumably becausethe excess APTES blocks the zeolite pores. On the other hand, withincreasing temperatures, the degree of cracking and isomerization

selectivity increased. In summary, adjustments in the acid strengthof zeolite played an important role in the transformation of triglyc-erides for preparing bio-aviation fuel range alkanes.

3.5. The HDO reaction mechanism for castor oil

Scheme 1 shows the reaction process for the HDO of castor oilover the nickel-based bifunctional catalysts. In the triglyceridesof ricinoleic acid, there are three types of functional group, theester group, unsaturated double bonds and hydroxyl group. Gas-phase and liquid-phase analyses revealed the formation of CO,CO2 and C3H8, and C5AC19 alkanes and H2O, respectively. Firstly,based on GC–MS analysis, the component of gas-phase and liq-uid-phase production suggest four pathways for hydroprocessingcastor oil: dehydration (route C), decarbonylation, decarboxylation(route A), the hydrogenation of unsaturated double bond (route B)and hydrocracking, which dominates the distribution of products.Transition metals are mainly responsible for the hydrogenationof unsaturated triglyceride and the acidic zeolite contributes to

Fig. 4. 1H (400 MHz) spectrums of the products from the hydroprocessing of castor oil at different distillation phases (I); and 13C (400 MHz) spectra of castor oil and productsfrom the hydroprocessing of castor oil at different distillation phases (II), 1H (400 MHz) spectrums of castor oil (III); the 13C (400 MHz) spectrums of castor oil (IV); (a, a1:castor, b, b1: HDO product, c, c1: 150 �C distillation HDO product, d, d1: 150–200 �C distillation HDO product, e, e1: 200–250 �C distillation HDO product, and f, f1: 250–300 �Cdistillation HDO product).

S. Liu et al. / Bioresource Technology 183 (2015) 93–100 99

the degree of hydrocracking. Therefore, different fuel range alkanescan be obtained by Ni supported on different acidic strength zeo-lites. The 1H (400 MHz) spectrum of castor oil clearly shows theresonances of protons including ester group (2), unsaturated dou-ble bond (9–10) and hydroxyl group (12) as shown in Fig. 4(III)(Sheldon, 2012). In comparison with the different distillation phaseproduction of hydroprocessing castor oil, the resonances of thesespecies disappeared; whereas, resonances of protons of cyclohex-ane (castor oil feed comprising 60% castor oil and 40% cyclohexane)and C5AC19 alkanes appeared, as shown in Fig. 4(I). The 13C spectraof castor oil is depicted in Fig. 4(IV), which reveals the same resultsas the 1H spectra. On the basis of the above results, HDO, decar-bonylation, decarboxylation and hydrocracking occur during thehydroprocessing of castor oil to produce alkanes and that no car-boxyl group, ester group, unsaturated double bond or hydroxylgroup were not found in the production of bio-aviation fuel rangealkanes.

3.6. Physicochemical properties of HDO product from hydroprocessingcastor oil

Aviation turbine fuels and missile fuels are generally mixturesof hydrocarbons derived from the distillation of petroleum crudeoil which the hydrocarbons selectivity is mainly between the gaso-line fraction and the diesel fraction (C8AC15). According to ASTMD7566 (ASTM D7566-09, 2009), aviation biofuels can be used bymixing traditional aviation fuels, provided the mixing proportionis less than 50%. Fig. S5 shows the distribution of hydrocarbonsfrom Jet fuel No. 3, the HDO product from castor oil and the

mixture of HDO product and Jet fuel-No. 3 (1:1). The selectivityof the above–three production is similar to, as a result, HDOproduction from the single-pot and two-step processes meet allof the basic jet-fuel mixing requirements, as shown in Table S2.It is normal that the density of the HDO product is a little lowbecause of the absence of cycloalkanes, which can be overcomeby mixing traditional aviation fuels. In addition, Table S3 shows acomparison of the physicochemical properties of castor oil, theHDO product and Jet fuel-No. 3, which shows that there is no oxy-gen in the HDO product, as investigated by elemental analysis.

4. Conclusions

Renewable bio-aviation fuel was produced by hydroprocessingcastor oil using bifunctional catalysts with differing acid strength.By employing the two-step and one-pot processes that wedesigned, bifunctional catalysts with moderate acidity could pro-mote the conversion of castor oil or C16AC19 alkanes into highyields of bio-aviation fuel with high isomerization selectivity. TheHDO mechanism of castor oil was investigated by GC–MS andNMR analyses. Meanwhile, the HDO product we prepared wascompared with Jet fuel No. 3, which showed that the HDO productmet the basic jet fuel mixing requirements.

Acknowledgements

This work was financially supported by the NSFC (21376123,U1403293), MOE (IRT-13R30 and 113016A), and the Research Fundfor 111 Project (B12015).

100 S. Liu et al. / Bioresource Technology 183 (2015) 93–100

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.biortech.2015.02.056.

References

Ambekar, V.R., 1957. Detoxification of castor cake. Indian J. Dairy Sci. 10, 107–122.Antal, M.J., Leesomboon, T., Mok, 1991. Mechanism of formation of 2-furaldehyde

from d-xylose. Carbohydr. Res. 217, 71–85.Ardiyanti, A.R., Khromova, S.A., Venderbosch, R.H., 2012. Catalytic hydrotreatment

of fast-pyrolysis oil using non-sulfided bimetallic Ni–Cu catalysts on a d-Al2O3

support. Appl. Catal. B 117–118, 105–117.ASTM D7566-14a Standard Specification for Aviation Turbine Fuel Containing

Synthesized Hydrocarbons [S], ASTM International, West Conshohocken, PA,2014.

Barta, K., Matson, T.D., Fettig, M.L., Scott, S.L., Iretskii, A.V., Ford, P.C., 2010. Catalyticdisassembly of an organosolv lignin via hydrogen transfer from supercriticalmethanol. Green Chem. 12, 1640–1647.

Berenblyum, A.S., Shamsiev, R.S., Podoplelova, T.A., Danyushevsky, V.Y., 2012. Theinfluence of metal and carrier natures on the effectiveness of catalysts of thedeoxygenation of fatty acids into hydrocarbons. Russ. J. Phys. Chem. 86, 1199–1203.

Bezergianni, S., Dimitriadis, A., 2013. Comparison between different types ofrenewable diesel. Renew. Sustainable Energy Rev. 21, 110–116.

Brown, T., Duan, P., Savage, P.E., 2010. Hydrothermal liquefaction and gasification ofNannochloropsis sp. Energy Fuels 24, 3639–3646.

Bu, Q., Lei, H., Zacher, A.H., 2012. A review of catalytic hydrodeoxygenation oflignin-derived phenols from biomass pyrolysis. Bioresour. Technol. 124, 470–477.

Bykova, M.V., Ermakov, D.Y., Kaichev, V.V., Bulavchenko, O.A., Saraev, A.A., Lebedev,M.Y., Yakovlev, V.A., 2012. Ni-based sol–gel catalysts as promising systems forcrude bio-oil upgrading: guaiacol hydrodeoxygenation study. Appl. Catal. B113–114, 296–307.

Elliott, D.C., 2007. Historical developments in hydroprocessing bio-oils. EnergyFuels 21, 1792–1815.

Guan, Q.X., Li, W., Zhang, M.H., 2009. Alternative synthesis of bulk and supportednickel phosphide from the thermal decomposition of hypophosphites. J. Catal.263, 1–3.

Hancsók, J., 2014. Sustainable production of bioparaffins in a crude oil refinery.Clean Technol. Environ. Policy 16, 1445–1454.

Horvath, I.T., Mehdi, H., Fabos, V., Boda, L., Mika, L.T., 2008. C-valerolactone – asustainable liquid for energy and carbon-based chemicals. Green Chem. 10,238–242.

Huang, Y.-B., Chen, M.-Y., Yan, L., Guo, Q.-X., Fu, Y., 2014. Nickel–tungsten carbidecatalysts for the production of 2,5-dimethylfuran from biomass-derivedmolecules. ChemSusChem 7, 1068–1072.

Jeong, H., Kang, M., 2010. Hydrogen production from butane steam reforming overNi/Ag loaded MgAl2O4 catalyst. Appl. Catal. B 95, 446–455.

Kumar, R., Rana, B.S., Tiwari, R., 2010. Hydroprocessing of jatropha oil and itsmixtures with gas oil. Green Chem. 12, 2232–2239.

Li, G., Li, N., Yang, J., Wang, A., Wang, X., Cong, Y., Zhang, T., 2013. Synthesis ofrenewable diesel with the 2-methylfuran, butanal and acetone derived fromlignocelluloses. Bioresour. Technol. 134, 66–72.

Mascarelli, A.L., 2009. Gold rush for algae. Nature 461, 460–461.Mohan, D., Pittman, C.U., Steele, P.H., 2006. Pyrolysis of wood/biomass for bio-oil: a

critical review. Energy Fuels 20, 848–889.Mu, W., 2014. Noble metal catalyzed aqueous phase hydrogenation and

hydrodeoxygenation of lignin-derived pyrolysis oil and related modelcompounds. Bioresour. Technol. 173, 6–10.

Mutlu, H., 2010. Castor oil as a renewable resource for the chemical industry. Eur. J.Lipid Sci. Technol. 112, 10–30.

Parizotto, N.V., Roch, K.O., Damyanova, S., 2007. Alumina-supported Ni catalystsmodified with silver for the steam reforming of methane: effect of Ag on thecontrol of coke formation. Appl. Catal. A 330, 12–22.

Peng, B., Yao, Y., Zhao, C., Lercher, J.A., 2012. Towards quantitative conversion ofmicroalgae oil to diesel-range alkanes with bifunctional catalysts. Angew.Chem. 124, 2114–2117.

Ren, X., 2014. Thermogravimetric investigation on the degradation properties andcombustion performance of bio-oils. Bioresour. Technol. 152, 267–274.

Saidi, M., Samimi, F., Karimipourfard, D., Nimmanwudipong, T., Gates, B.C.,Rahimpour, M.R., 2014. Upgrading of lignin-derived bio-oils by catalytichydrodeoxygenation. Energy Environ. Sci. 7, 103–129.

Sheldon, R.A., 2012. Fundamentals of green chemistry: efficiency in reaction design.Chem. Soc. Rev. 41, 1437–1451.

Shi, H., Chen, J.X., Yang, Y., Tian, S.S., 2014. Catalytic deoxygenation of methyllaurate as a model compound to hydrocarbons on nickel phosphide catalysts:remarkable support effect. Fuel Process. Technol. 118, 161–170.

Šimácek, P., 2009. Hydroprocessed rapeseed oil as a source of hydrocarbon-basedbiodiesel. Fuel 88, 456–460.

Sotelo-Boy, R., 2010. Renewable diesel production from the hydrotreating ofrapeseed oil with Pt/zeolite and NiMo/Al2O3 catalysts. Ind. Eng. Chem. Res. 50,2791–2799.

Sukrutha, S., 2014. Harnessing indigenous plant seed oil for the production of bio-fuel by an oleaginous fungus, cunninghamella blakesleeana–JSK2, isolated fromtropical soil. Appl. Biochem. Biotechnol. 172, 1027–1035.

Verma, D., Kumar, R., Rana, B.S., 2011. Aviation fuel production from lipids by asingle-step route using hierarchical mesoporous zeolites. Energy Environ. Sci. 4,1667–1671.

Warner, G., Hansen, T., Riisager, A., 2014. Depolymerization of organosolv ligninusing doped porous metal oxides in supercritical methanol. Bioresour. Technol.161, 78–83.