Recent Advances in Hydrotreating of Pyrolysis Bio-Oil

8/9/2019 Recent Advances in Hydrotreating of Pyrolysis Bio-Oil

1/24

Recent Advances in Hydrotreating of Pyrolysis Bio-Oil and ItsOxygen-Containing Model CompoundsHuamin Wang, Jonathan Male, and Yong Wang*,,

Pacic Northwest National Laboratory, P.O. Box 999, 902 Battelle Boulevard, Richland, Washington 99352, United StatesVoiland School of Chemical Engineering and Bioengineering, Washington State University, Pullman, Washington 99163, UnitedStates

ABSTRACT: Considerable worldwide interest exists indiscovering renewable energy sources that can substitute forfossil fuels. Lignocellulosic biomass, the most abundant andinexpensive renewable feedstock on the planet, has a greatpotential for sustainable production of fuels, chemicals, andcarbon-based materials. Fast pyrolysis integrated with hydro-

treating, one of the simplest, most cost-eff

ective, and mostefficient processes to convert lignocellulosic biomass to liquidhydrocarbon fuels for transportation, has attracted signicantattention in recent decades. However, effective hydrotreatingof pyrolysis bio-oil presents a daunting challenge to thecommercialization of biomass conversion via pyrolysis-hydro-treating. Specically, the development of active, selective, andstable hydrotreating catalysts is problematic due to the poor quality of current pyrolysis bio-oil feedstock (i.e., high oxygencontent, molecular complexity, coking propensity, and corrosiveness). Signicant research has been conducted to address thepractical issues and provide fundamental understanding of hydrotreating and hydrodeoxygenation (HDO) of bio-oils and theiroxygen-containing model compounds, including phenolics, furans, and carboxylic acids. A wide range of catalysts have beenstudied, including conventional Mo-based sulde catalysts and noble metal catalysts. Noble metal catalysts have been the primaryfocus of recent research because of their excellent catalytic performances and because they do not require the use ofenvironmentally unfriendly sulfur. Recently, the reaction mechanisms of the HDO of model compounds on noble metal catalysts

and their effi

cacy for hydrotreating or stabilization of bio-oil have been reported. This review provides a survey of relevantliterature, published over the past decade, reporting advances in the understanding of the HDO chemistry of bio-oils and theirmodel compounds, mainly on noble metal catalysts.

KEYWORDS: biomass, lignocellulose, pyrolysis bio-oil, hydrodeoxygenation, catalysts, noble metal catalysts, model compounds

1. INTRODUCTION

Most activities in modern society depend heavily on fossilresources (e.g., oil, natural gas, and coal). In 2010, fossil fuelsaccounted for 87% of energy consumed worldwide and 87 and92% of energy consumed in the United States and China,respectively.1 Increased utilization of fossil fuels is aninternation a l concern due to dwindling foss i l- fuelresources,24 environmental consequences of CO2 emissionfrom fossil fuels,5 and economic and political problemsresulting from uneven distribution of fossil-fuel resources.Signicant research has been conducted worldwide to discoveralternative energy sources, which should be renewable andcarbon-neutral and have the potential to replace fossil fuels inthe current energy production and conversion system.6

Attractive resources include solar, wind, hydroelectric, geo-thermal, and biomass. Of these, biomass is the only sustainableresource for the production of fuels, chemicals, and carbon-

based materials, especially for liquid hydrocarbon fuels fortransportation.6,7 Currently, oil is the major resource (94%)used in the transportation energy sector; in 2009, it accounted

for 69% of oil consumption and 29% of total energyconsumption in the United States.8 New strategies must bedeveloped for the efficient and large-scale production of fuelsfrom biomass sources that can be used in current energysystem.

Lignocellulosic biomass (e.g., woods, grass, energy crop, andagricultural waste) is currently the most inexpensive andabundant source of plant biomass.6,9 It is desirable to convert

the whole energy in the lignocellulosic biomass to trans-portation fuels using existing infrastructure.10 Three primaryroutes have attracted the most attention:6,7,1113 gasication tosynthesis gas (CO and H2) followed by FischerTropschsynthesis to liquid hydrocarbons (biomass to liquids,BTL);6,7,14,15 pyrolysis or liquefaction to bio-oils followed bycatalytic upgrading to hydrocarbon fuels;6,7,1621 and pretreat-ment-hydrolysis to aqueous sugars followed by aqueous-phase

Received: January 29, 2013Revised: April 5, 2013Published: April 9, 2013

Review

pubs.acs.org/acscatalysis

2013 American Chemical Society 1047 dx.doi.org/10.1021/cs400069z| ACS Catal. 2013, 3, 10471070
http://localhost/var/www/apps/conversion/tmp/scratch_7/pubs.acs.org/acscatalysishttp://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-000.jpg&w=239&h=127http://localhost/var/www/apps/conversion/tmp/scratch_7/pubs.acs.org/acscatalysis


2/24

catalyt ic processing or fermentation to hydrocarbonfuels.6,7,22,23 The rst two routes can utilize whole lignocellu-losic biomass, but the third route can only utilize the celluloseand hemicellulose portion of lignocellulosic biomass. Amongthese routes, pyrolysis integrated with upgrading is the simplestand most cost-effective,24 and, thus, the resulting bio-oils(pyrolysis oils)have been identied as the cheapest renewableliquid fuels.18 Pyrolysis is the thermal decomposition oflignocellulose in the absence of oxygen at temperatures

between 648 and 800 K. During pyrolysis, a number ofreactions occur (e.g., depolymerization, dehydration, and CC

bond cleavage), leading to the formation of acomplex mixtureof >200 oxygenated compounds.18,19,21,25 Advanced fastpyrolysis can produce bio-oils in high yields (up to 80 wt %of dry feed) with up to 70% of the energy sto red in the biomassfeedstocks retained in the nal bio-oils.6,7 Catalytic fastpyrolysis, wherein fast pyrolysis is integrated with a catalysisprocess to upgrade the pyrolysis vapor with similar operationconditions, can lead to a higher quality bio-oilbut at theexpense of product yield.2636 Another variation, hydro-pyrolysis, is fast pyrolysisconducted in the presence of reactivegases such as H2.

37,38 Both catalytic fast pyrolysis andhydropyrolysis suffer higher operational complexity; however,

both have the potential to produce hydrocarbons directly frombiomass or produce higher quality bio-oils more amenable forthe subsequent upgrading process. Recently, several excellentreviews on pyrolysis have been published.18,19,21,39 In addition,a number of fast pyrolysis technologies have been commer-cialized,21,39 including a small portable pyrolysis reactor beingcommerciallydeveloped for the densication of biomass closeto its location.13,18

Although bio-oil can be produced in a simple and efficientway, its properties result in multiple signicant problems duringits utilization as transportation fuels in standard equipmentsuch as combustion boilers, engines, and turbines constructedfor use of oil-derived fuels.18 Primarily, the high oxygen content

of bio-oils, usually 20 to 50 wt %, leads to different physical andchemical properties and combustion behaviors between bio-oilsand petroleum fuels.40,41 Specically, the low heating value,poor stability, poor volatility, high viscosity, and corrosivenessof bio-oils limit their utilization as transportation fuels.18,40

Catalytic fast pyrolysis and hydropyroysis can produce bio-oilwith improved quality; however, similar issues still exist,although to a lesser extent. Therefore, extensive oxygenremoval is required for upgrading bio-oils to liquid trans-portation fuels with similar properties to petroleum fuels (e.g.,high energy density, high stability, and high volatility).

Several approaches have been studied to upgrade bio-oils(see Figure1) including hydrotreating,7,17,20,42 zeolite upgrad-ing,7,4348 and aqueous-phase processing.49 Further, additional

technologies (e.g., condensation and fractionation) have been

used to make bio-oil feedstock more amenable to theseupgrading processes. These processes have also been combinedto achieve better hydrocarbon yields and deeper oxygenremoval.50 Hydrotreating is the most common route toupgrading bio-oil, involving oxygen removal through hydro-deoxygenation (HDO) to form hydrocarbons and H2O withsaturation of double bonds or aromatic rings by hydro-genation.16,41 Hydrotreating takes place between 400 and 773K under high-pressure hydrogen in thepresence of supportedtransition metal or their sulde catalysts.17,42 Zeolite upgrading,a route similar to the catalytic cracking in petroleum rening,treats bio-oils using cracking catalysts (normally acidic zeolites)at atmospheric pressure without the requirement of hydrogenat between 623 and 773 K.7,4348 In zeolite upgrading, oxygenin the oxygenated compounds of bio-oils is mainly removed asCO, CO2, and H2O and the products obtained are mostlyaromatic and aliphatic hydrocarbons via a number of reactions(e.g., dehydration, cracking, and aromatization).7,45 However,this method suffers from low hydrocarbon yield because of thehigh yield to coke and dealumination of zeolite due to the waterin bio-oils, which consequently deactivates the catalyst. Anotherapproach, aqueous-phase processing, can treat some fraction of

bio-oil to form hydrocarbons and hydrogen. Specically, bio-oils are rst separated by the addition of water and then theaqueous fraction is treated in an aqueous-phase reformingprocess to produce hydrogen49 or in an aqueous-phasedehydration/hydrogenation process to produce alkanes.49,50

As noted previously, in recent decades bio-oil hydrotreatinghas attracted the most attention and is the most commonmethod to remove oxygen from bio-oils. Hydrotreating is oneof the key processes in modern oil rening; it involveshydrodesulfurization (HDS), hydrodenitrogenation (HDN),and HDO to remove the sulfur, nitrogen, and oxygenheteroatoms, respectively, often accompanying hydrogenation(saturation) of olens and aromatics in petroleum feed-stocks.51,42 Less attention has been paid to HDO as compared

to HDS during petroleum upgrading research because the lowoxygen content in oils (


3/24

compounds of bio-oils. Acids, in either aqueous form or solidform, have been combined with metals to achieve faster HDO

by bifunctional catalysis. Signicant challenges still exist in therational design of more effective HDO catalysts and processesfor upgrading bio-oil in a simpler process with higher reactionrates, lower H2requirement, higher carbon yield, and improvedcatalyst life.

Over the past 30 years, a wider range of studies have beenreported in the literature regarding fundamental and practicalassessments of hydrotreating bio-oils and their oxygen-containing model compounds over a variety of catalysts. Areview by Furimsky focused on the chemistry (mechanisms andkinetics) of the HDO of representative compounds (mainlyfurans and phenols) on Mo-based sulde catalysts and the roleof HDO in hydrotreating of traditional oils.42 A review byElliott in 2007 provided a very detailed summary of researchefforts on processing actual bio-oil products.17 Other recentreviews also discussed hydroprocessing of actual bio-oils,20,41

and more general reviews have highlighted the importance ofthe HDO of bio-oils.6,7

This review focuses on literature from the past decaderelevant to furthering the understanding of HDO chemistry of

bio-oils and their model compounds on mainly non-suldemetal catalysts. HDO on sulde catalyst will be disseminatedfor the purpose of comparison, and the most recent advances inHDO of actual bio-oil will be summarized.

2. BIO-OIL AND ITS MAJOR OXYGEN-CONTAININGCOMPONENTS

Bio-oils from fast pyrolysis of biomass are multicomponentmixtures of a large number of oxygenated compounds derivedfrom lignocellulosic biomass feedstocks (e.g., woods, grasses,energy corps, agricultural wastes, and forest wastes). Thestructured portion of lignocellulosic biomass is composed ofcellulose (28 to 55%), hemicellulose (17 to 35%), and lignin(17 to 35%), which are oxygen-containing organic polymers.The weight percentage of these three major components variesin different biomass species, as shown in Table1for pinewood,poplarwood, and switchgrass.7,53 Other minor biomasscomponents include organic extractives and inorganic minerals.

The structure of cellulose, hemicellulose, and lignin areshown in Figure2. Cellulose is a high-molecular-weight linearpolymer of -D-anhydroglucopyranose units (AGUs, 10,000

AGUs for cellulose chain in wood), and its basic repeat unitconsists of two AGUs, called a cellobiose unit, as shown inFigure 2A.54 Cellulose degrades at 573 K to producelevoglucosan and other anhydrocellulose.55 Hemicellulose isderived from several sugars in addition to glucose, especially

xylose but also mannose, galactose, and arabinose (Figure2B),all of which are highly substituted with acetic acid. Hemi-cellulose consists of shorter and branched chains (200 sugar

units). Decomposition of hemicellulose at between 473 and533 K leads to acetic acid and anhydrocellulose.19 Lignin, anamorphous, highly branched, and substituted polymer, consistsof an irregular array of variously hydroxyl- and methoxyl-substituted phenylpropane units with relatively hydrophobicand aromatic nature. Three representative units, p-coumaryl,coniferyl, and sinapyl, are shown in Figure2C. Lignin structureand properties are different for different feedstocks. Lignindecomposes at between 553 and 773 K to form oligomers andmonomers of polysubstituted phenols via the cleavage of etherand CC linkages.19,56

Lignocellulosic biomass decomposes during fast pyrolysis toproduce a wide range of products including vapors, aerosols,and charcoal-like char. Cooling and condensation of the vaporsand aerosols results in the formation of a dark brown liquidreferred to as bio-oil. The bio-oil yields from wood are in therange of 72 to 80 wt %, depending on the feedstock and processused.19,21 Other products include noncondensable gases (10 to20 wt %) and solid char (15 to 25 wt %). 19,21 Advanced fast

pyrolysis to achieve high liquid yield requires careful pretreat-ment of biomass (drying and milling), high heating rates, highheat transfer rate, well-controlled temperature (698 to 773 K),short residence time (0.5 to 5 s), and rapid cooling andquenching of pyrolysis vapor.19,21,57 The reaction during fastpyrolysis involves rapid depolymerization and fragmentation ofthe cellulose, hemicellulose, and lignin to their primarydecomposed molecules, as mentioned above, and some

volatiles. Subsequent reactions including isomerization, dehy-dration, repolymerization/condensation, decomposition, andCC bond cleavage occur during pyrolysis and rapidquenching, leading to a very complicated mixture of >200organic compounds. This product is not thermodynamically

Table 1. Typical Lignocellulose Content of Some BiomassSpecies

lignocellulose content (wt %)

biomass species cellulose hemicellulose lignin

pinewooda 4650 1922 2129

poplarwoodb 4046 1723 2128

switchgrassb 2837 2329 1720aBased on data reported in ref7. bBased on the data reported in ref53.

Figure 2.Structure of different lignocellulosic biomass fractions: (A)cellulose, (B) hemicellulose, (C) lignin.

ACS Catalysis Review

dx.doi.org/10.1021/cs400069z| ACS Catal. 2013, 3, 104710701049
http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-002.png&w=204&h=290


4/24

equilibrated, and therefore its chemical composition tends tochange during storage.

Table2 lists the major compounds in typical bio-oils fromwood, based on the detailed analysis reported in variousliteratures.5862 Major groups of compounds include water,simple oxygenates (e.g., acids, esters, alcohols, ketones,aldehydes), miscellaneous oxygenates, sugars, furans (e.g.,alkylated furan, furfural, hydroxymethyl furfural), phenolics(e.g., phenols, guaiacols, syringols), and high-molecular-weightcompounds. Miscellaneous oxygenates, sugars, and furans areprimarily derived from the cellulose and hemicellulose fractionsof biomass, whereas phenolics are derived from the lignincomponent of biomass. Simple oxygenates (e.g., acids, esters,

alcohols, ketones, and aldehydes) likely form from thedecomposition of the miscellaneous oxygenates, sugars, andfurans. High-molecular-weight compounds are primarilyoligomers of phenolic compounds with molecular weightsranging from several hundred to as much as 5,000 or more.

The complicated components in bio-oils pose signicantchallenges for their further upgrading by hydrotreating.Traditional hydrotreating of petroleum-derived fuels (e.g.,HDS to remove sulfur) is predominantly performed to removesulfur from ve-membered thiophenic heterocyclics (e.g.,thiophene, benzothiophene, dibenzothiophene, and theiralkylated compounds; including 20 compounds with about1.0 to 1.8 wt % sulfur) in feedstocks.51 Here, bio-oil has >200

oxygen-containing compounds, representing nearly all kinds ofoxygenated organics and oxygen function groups, which canundergo intertwining interactions, making it extremely difficultto understand HDO catalysis reaction pathways, mechanisms,kinetics, and propertyreactivity correlation of hydrotreatingcatalysts. In addition, extensive oxygen removal from bio-oiland some signicant problems of bio-oil (e.g., high watercontent, acidity, and chemical instability) bring signicantchallenges for bio-oil hydrotreating catalysts and processes.Table 3 compares the typical compositions and physicalproperties of wood pyrolysis bio-oil and conventional fueloil.18,40,42,61,62 The presence of oxygen in almost every bio-oilcomponent and, thus, the overall high oxygen content of bio-oil

constitute the primary difference between bio-oil andpetroleum-derived oils and are the cause of difficulties in bio-oil hydrotreating. Polymerization and condensation of bio-oilsoccur over time and can be accelerated by increasingtemperature because of interactions of different components,especiallyhighly reactive species (e.g., furfurals, guaiacols, andphenols),52,63 leading to a high propensity of coke formationduring hydrotreating and the consequent catalyst deactivation,and even plugging of catalyst bed.64 Water, initially present in

bio-oil with a high content (1730 wt %) and also formedduring hydrotreating, inhibits hydrotreating reactions bycompetitive adsorption on active sites and causes catalystdeactivation by modication of the catalyst structure.63,65 Other

Table 2. Major Compounds in Typical Wood Pyrolysis Bio-Oilsa

aBased on results in refs 5862. bR, R= H or CH3.




5/24

issues include deposition of inorganic species (e.g., alkali) oncatalyst17 and leaching of catalyst by bio-oil because of bio-oilacidity and corrosivity, which lead to the deactivation ofhydrotreating catalysts. Catalytic fast pyrolysis and hydro-pyrolysis can produce bio-oil with improved quality (i.e., morefavorable oxygen content, hydrocarbon content, acidity, andstability),3339 which is expected to alleviate challengesassociated with downstream hydrotreating processes. However,even using catalytic fast pyrolysis and hydropyrolysis, asubstantial amount of oxygen, reactive species, and water inthe bio-oil leads to similar hydrotreating issues as described,although to a lesser extent. Rigorous data is lacking todemonstrate the advantages of hydrotreating bio-oil produced

by catalytic fast pyrolysis or hydropyrolysis over bio-oilproduced by conventional fast pyrolysis.

Improved hydrotreating processes and catalysts are requiredto efficiently upgrade bio-oils to transportation fuels. Extensive

research has been conducted over the past decade, andsignicant advancement has been achieved in the fundamentalunderstanding of the chemistry of reactions taking place during

bio-oil hydrotreating and in the development of catalysts andprocesses to upgrade actual bio-oils and their modelcompounds.17,42 Next, we will summarize the catalysts usedin the HDO of bio-oil and provide a detailed review of theHDO chemistry of bio-oil model compounds.

3. HDO CATALYSTS

Various catalysts with different active phases, promoters, andsupports have been studied in HDO of bio-oils and theiroxygen-containing model compounds. Table4 summarizes thecatalysts in six groups: Mo-based suldes, noble metals, base

metals, metal phosphides, other metal catalysts, and bifunc-tional catalysts. Metals are present as zerovalent metal, sulde,oxide, and others and in either mono- or bimetallic form.Supports include carbon, silica, alumina, zirconia, titania,amorphous silicaalumina, and various zeolites. Secondfunction of bifunctional catalysts includes aqueous and solidacids.

Molybdenum sulde, normally promoted by cobalt or nickeland supported on porous supports (e.g., -Al2O3), has been

widely used in modern hydrotreating processes. The reactionnetwork, mechanism, and site requirement of HDS and HDNreactions on Mo-based sulde catalysts have been studiedextensively and are well-understood.51 Mo-based sulde

catalysts are also active for HDO and, therefore, are mostcommonly used for bio-oil hydrotreating.17 Elliott et al.17,52

reported that extensive oxygen removal from bio-oil (from 45to


6/24

Early studies have employed Mo-based sulde with Al2O3asa support, which have shown appropriate pore structure anddispersion of active phase, and promotion effects for hydro-treating reactions. However, Al2O3is known to be responsiblefor coke formation due to its acidity and is unstable in thepresence of large amounts of water.63,76 To avoid theseproblems, neutral supports with better water tolerance (e.g.,activated carbon) have been widely usedin recent studies forhydrotreating of actual bio-oils52,6769 and model com-pounds.77 In addition, TiO2 and ZrO2 have been used assupports with a better activity than Al2O3.

78 Further,mesoporous materials (e.g., MCM-41)79 and zeolites (e.g.,HBeta,80 HY,81 and HZSM-582) have been utilized to prepare

bifunctional catalysts or catalysts with enhanced diffusionproperties. However, intensive investigation is still neededregarding the stability of TiO2-, ZrO2-, and zeolite-supportedcatalysts during hydrotreating of actual bio-oil.

4. HDO OF MODEL COMPOUNDS OF BIO-OIL:REACTION NETWORK, MECHANISM, AND KINETICS

The rational design of more effective HDO catalysts requiresadvancement in the fundamental understanding of themechanism, kinetics, and site requirements for HDO reactions.Most studies for elucidation of HDO mechanism and kineticshave been conducted using model compounds includingphenolics (phenols, anisole, and guaiacol), furans, andcarboxylic acids, which are the major components in bio-oil.Reaction routes of HDO of model compounds have beeninvestigated in recent studies, whereas kinetics and siterequirements have received little attention. Both continuous-ow and batch reactor have been used for the HDO of modelcompounds. It is notable that the industrial process for bio-oilhydrotreating prefers continuous-ow reactor and thatchallenges exist in comparing data between continuous owand batch reactors.

Cleavage of CO bonds must occur for the removal of

oxygen from oxygen-containingcompounds. Some CO bondstrengths are listed in Table 5.42 CO bond strengths of the

OH group or RO group (aliphatic-oxygen) attached to anaromatic carbon (ArOH, phenols; or ArOR, aryl ethers) areabout 83 kJ/mol greater than that attached to the aliphatic

carbon (ROH, alcohols; or ROR, aliphatic ethers),respectively. This suggests that cleavage of CO bondsattached to phenols and aryl ethers is more difficult than thatof CO bonds attached to alcohols and aliphatic ethers.Therefore, removal of oxygen from phenols or aryl ethers will

be enhanced byrst hydrogenation of the aromatic ring to thecorresponding cycloalkane by converting of ArO bond to RO bond. This also implies that a hydrogenationdeoxygenationroute would be preferred in HDO of phenols or aryl ethers oncatalysts with good hydrogenation activity (e.g., noble metalcatalysts). Deoxygenation of alcohols (OH group attached tothe aliphatic carbon) could occur by hydrogenolysis catalyzed

by a metal catalyst or, alternately, by dehydration catalyzed by

an acid under milder conditions. Recent results showed that anacid function greatly promotes the deoxygenation activity ofmetal catalysts.

The sequence of CO bond strength is also consistent withreported reactivity sequence of oxygenated groups on Mo-

based sulde catalysts under hydrotreating conditions. Grangeet al.83 compared the reactivity of oxygen-containing groups/compounds by showing the isoreactivity, which is thetemperature at which the conversion rates reach a signicantidentical value on a sulded CoMo/Al2O3 catalyst (see Table6). In addition, Grange et al. estimated that unsaturated double

bonds (olens), aliphatic alcohols, and ethers would react ateven lower temperature than the ketonic group. Therefore, theysuggested a low-temperature hydrotreatment would undergohydrogenation of unsaturated double bonds (olens, aldehydes,ketones) to stabilize the bio-oils, whereas deep HDO wouldrequire a higher temperature for the elimination of the phenolicand furanic oxygen-containing compounds. Similar reactivityranking of oxygenated groups was also developed by Elliott17

on the basis of the results from the literature. As shown inFigure 3, hydrogenation of olens, aldehydes, and ketones

occurs at low temperatures. Alcohols are reacted from 423 to473 K by catalytic hydrogenation but also by thermaldehydration to form olens, while carboxylic acids, phenolicethers, and phenols are reacted at from 573 to 623 K.

4.1. HDO of Phenols. Because functionalized phenols arethe major components of bio-oils obtained from the ligninfraction of lignocellulosic biomass, compounds includingphenol (and alkylated phenols) are the main model compoundsin HDO studies. Studies on the HDO of phenol and cresol onsulded CoMo catalyst established the basic reaction schemes

Table 5. Bond Dissociation Energies42

bond dissociation energy (kJ/mol)

ROR 339

ROH 385

ArOR 422

ArOH 468

Table 6. RelativeReactivities of Oxygen-Containing Groupsand Compounds83

activation energy (kJ/mol) temp of isoreactivity (K)

ketone 50.2 476

carboxylic acid 108.9 556

methoxy phenol 113.0 574

4-methylphenol 140.7 613

2-ethylphenol 149.9 640

Figure 3.Reactivity scale of oxygenated groups under hydrotreatment

conditions. Redrawn with permission from ref 17. Copyright 2007American Chemical Society.




7/24

involving deoxygenation and hydrogenation reactions.42,8488

As shown in Figure 4, primary reactions of 4-methylphenol

include two parallel routes: direct deoxygenation (COcleavage) to toluene (DDO route) and hydrogenation to 4-methyl-cyclohexanol (HYD route).8488 Subsequent hydro-genation of toluene or deoxygenation of 4-methylcyclohexanoloccurs to produce the nal product: methylcyclohexane. Kineticstudies showed a preference of the DDO route over the HYDroute during the HDO of phenol and 3- or 4-methylphenol onCoMo sulde catalysts.84,87 Figure 5 summarizes the rate

constants of the two paths of the HDO of phenol and itsmethylated compounds measured in a continuous-ow reactorat 573 K and 2.85 MPaof hydrogen pressure over a suldedCoMo/Al2O3 catalyst.

87 The dependencies of adsorption andreaction rates upon methyl-group substitution are because ofthe effects on the electrostatic potential and orbitals rather thansteric effects.87 Massoth et al.86,87 proposed that the DDO sites

consist of a vacancy site with Mo (or Co) exposure and thatadjacent HYD sites have S or SH saturated sites, consistent

with the results that the DDO route was more inhibited thanthe HYD route by H2S, which could block vacancy site bycompletion adsorption. Romero et al.88 proposed that HYDsites could be either the multiple vacancies or the fully suldedmetallic edges (so-called brim sites) located on the basal planeof the MoS2slabs, which does not contain vacancies. These are

very similar to the proposed site requirements of HDS reactionof thiophene-derived molecules on Mo-based sulde catalysts.51

The DDO and HYD reactions of phenols require different sitesand different binding phenols on catalyst surfaces. As shown inFigure6, Romero et al.88 proposed vertical 1adsorption of 2-

ethylphenol through oxygen on vacancy sites for the DDOroute and at5adsorption of 2-ethylphenol through aromaticring and oxygen for the HYD route on multiple vacancies onMoS2 surfaces. The at adsorption of phenol on fully suldedmetallic edges could be expected for the HYD route.

Reports on both vapor-phase and aqueous-phase HDO ofphenol on non-sulded catalysts (e.g., Pd, Pt, and Ni) indicatedthat the HYDroute isfavored in the conversion of phenol onthese catalysts.74,75,81,8997 The vapor-phase hydrogenation ofphenol over Ptand Pd supported on alumina was studied by

Talukdar et al.

89

in a continuous-

ow reactor at 1 MPa and473548 K. The selectivity to the HYD products, cyclo-hexanone and cyclohexanol, was up to 99% at conversion rangeof 40 to 97%. Pt catalysts favored the production ofcyclohexanol while Pd catalysts favored cyclohexanoneproduction. In addition, Pt catalysts showed twice as muchconversion as Pd catalysts. Supports also affected gas-phasehydrogenation of phenols on Pd metal. Neri et al.90 showedthat a higher selectivity to cyclohexanone was found on Pd/MgO compared to Pd/Al2O3 at atmospheric pressure and at393 to 573 K, which was hypothesized to be due to the twodifferent forms of adsorbed phenol at the interface between thesupport and palladium metal particles.However, contradictoryresults were reported by Mahata et al.92 They showed that Pd/

Al2O3was selective for cyclohexanone production whereas Pd/MgO produced cyclohexanone along with cyclohexanol as aminor product for hydrogenation of phenol at 503 K underambient pressure in a continuous-ow reactor. In addition,

Al2O3-based catalysts showed initial deactivation, while MgO-based catalysts showed signicantly improved catalyst life.92

Wan et al.98 studied the HDO of cresol using a batch reactor at573 K and 8.3 MPa. HDO of cresol occurred via both HYD andDDO routes to form toluene and methylcyclohexane when

water was used as solvent. The occurrence of the DDO route isbecause of low H2 availability by the mass-transfer limitationswith water as the solvent. When supercritical heptane was usedas solvent and the mass-transfer limitations were eliminated, themajor products were methylcyclohexanol and methylcyclohex-ane via the HYD route. The proposed reaction mechanism, as

shown in Figure 7, showed both vertical and coplanaradsorption of cresol on catalyst surface, in which the verticaladsorption via oxygen favored the DDO route to toluene andthe coplanar surface adsorption of the aromatic ring favored theHYD route to ring-saturated products. The proposedadsorption modes of cresol on noble metal catalysts resembledthose for phenol HDO on MoS2surfaces.

Studies on aqueous-phase HDO of phenol using Pd oncarbon and Raney Ni catalyst in a batch reactor indicated thatthese catalysts favorphenol hydrogenation to cyclohexanol at340 to 423 K.74,9396 Phenol does not undergo directhydrogenolysis (DDO) to benzene on Pd/C at 353 K and 5MPa of H2.

74,95,96 The hydrogenation product of phenol,

Figure 4.Reaction network of HDO of 4-methylphenol.

Figure 5.DDO rate constant (k1, A) and HYD rate constant (k2, B)of phenol (P), methylphenols (MP), and dimethylphenols (DMP) onsulde CoMo catalyst. Reprinted with permission from ref 87.Copyright 2006 American Chemical Society.

Figure 6.Proposed binding modes of 2-ethylphenol on MoS2catalyst

surface: (A) vertical 1adsorption of through oxygen and (B)

at 5adsorption through aromatic ring and oxygen. Redrawn withpermission from ref88. Copyright 2010 Elsevier.


http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-007.png&w=218&h=59http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-006.jpg&w=161&h=143http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-005.png&w=173&h=101


8/24

cyclohexanol, can only undergo oxygen removal by dehydrationto cyclohexene catalyzed by acids such as H3PO4 at >473K.74,95,96As shown in Figure8, the aqueous-phase dehydrationof the intermediate cyclohexanol to cyclohexene is the rate-determining step in the overall phenol HDOreaction on Pd/Cand H3PO4 at 473 K and 5.0 MPa of H 2.

74 Consequently, abifunctional catalyst system, including a metal function and an

acid function, was developed to catalyze phenol HDO with amuch improved oxygen removal rate under mild conditions.The reaction pathway followed the sequence of phenolhydrogenation on metal surface to form cyclohexanol, whichis dehydrated on acidic sites to form cyclohexene. Variousaqueous Brnsted acids (e.g., H3PO4, CH3COOH, and Naonsolution), together with Pd/C, Ru/C, Rh/C, or metalnanoparticles, were used as bifunctional catalysts for aqueous-phase HDO of phenol to produce cyclohexane via ahydrogenationdehydration route.74,95,96,99 In addition, varioussolid acids were used as either a support or a cocatalyst toachieve bif unctional catalysts for phenol and cresolHDO.75,81,96,97,100103 Zhao et al.75,96 reported that a nearly100% yield of cycloalkane was achieved for aqueous-phase

HDO of phenol at 473 K and 4.0

5.0 MPa of H2using Pd/Cor Raney Ni with SiO2-supported Naon or HZSM-5 as theBrnsted solid acid. Zhao et al.101 later investigated the kineticsof aqueous phenol HDO on Ni/HZSM-5 catalysts and showedthat phenol hydrogenation was the rate-determining step,

which was in contrast to the results on Pd/C with H3PO4shown above, indicating that the balance of activity andquantity of metal and acid siteswas critical for the performanceof the catalysts. Foster et al.103 reported that the increase ofnumber and strength of acid sites on the -Al2O3 support, bymodication with base or acid treatments, increased the rate ofHDO on Pt/-Al2O3at 533 K and 0.05 MPa.

Zhao et al.97 used a large-pore molecular sieve HBEA-supported Pd catalyst97 or Pd/C with H/La-BEA100 for theconversion of phenol and substituted phenols to form desired

bicycloalkanes via HDO and hydroalkylation reaction at 473 to523 K. The proposed mechanism involves the parallel reactionsoccurring after the hydrogenation of phenol to cyclohexanol:cyclohexanol dehydration to form cyclohexene or addition ofcyclohexanol or cyclohexene to phenol by alkylation, both on

Brnsted acid sites. The preference of the two reactions wasinuenced by variations of solid acid, steric constraints,temperature, metal sites, and metalreactant ratios. A highlyselective aqueous-phase hydroalkylation and deoxygenation ofsubstituted phenols was achieved over an HBEA-supported Pdcatalyst (metalacid ratio: 1:22 mol/mol) for the productionofC12C18 bicycloalkanes with yields up to 80%. Hong et al.81

used zeolite-supported Pt catalyst for HDO of aqueous phenolin a continuous-ow reactor at 473 to 523 K and 4 MPa of H2.The selectivity of HYD route was more than 98%, and the acidfunctions of zeolite led to high activity and selectivity tomonocyclics and the production of useful bicyclics. Acidfunction of the bifunctional catalyst apparently could catalyze

both the dehydration reaction to accelerate phenol HDO tocyclohexane and the hydroalkylation reaction to produce useful

bicycloalkanes.4.2. HDO of Anisole, Guaiacol, and Phenolic Dimers.

The HDO of anisole was conducted on an Al2O3 supportedCoMo sulde catalyst at between473 and 573 K and 1.5 MPaof H2in a continuous-ow reactor.

104As shown in Figure9, the

primary reaction of anisole included demethylation to formmethane and phenol and transalkylation to produce methyl-phenol and dimethylphenol. The further conversion of phenolsfollows either HYD or DDO routes. A methyl transfer reactionis desirable in the HDO reaction because it prevents the loss ofcarbon in methoxyl (OCH3), one of the major functionalgroups in lignin-derived phenolics. Li et al.72 studied the HDOof anisole over silica-supported Ni

2P, MoP, and NiMoP

catalysts at 573 K and 1.5 MPa. Only phenol, benzene, andcyclohexane were detected as main products, indicating that theHDO of anisole under these conditions mainly proceeds viademethylation to phenol followed by DDO or HYD routes ofphenol to hydrocarbon. HDO activities decreased in the

Figure 7.Proposed reaction mechanism for HDO of 4-methylphenolon supported Pt, Pd, and Ru metal surface. Redrawn with permissionfrom ref98. Copyright 2010 Springer.

Figure 8.Reaction pathway for aqueous-phase HDO of phenol on dual-functional catalysts of Pd/C and H 3PO4at 473 K. The turnover frequencyfor each step is also presented. Redrawn with permission from ref74. Copyright 2010 Elsevier.

Figure 9. Reaction network of HDO of anisole on sulde CoMocatalyst. Redrawn with permission from ref 104. Copyright 2000Elsevier.


http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-010.png&w=203&h=180http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-008.png&w=239&h=102


9/24

sequence of Ni2P/SiO2> NiMoP/SiO2 > MoP/SiO2, and theNi phosphide-containing catalysts showed much higheractivities than a conventional NiMo/-Al2O3catalyst.

The HDO of anisoleon a Pt/Al2O3 catalyst was conductedby Runnebaum et al.105 at 573 K and 0.14 MPa of H2 in acontinuous-ow reactor, which showed the major primaryreactions of demethylation to phenol, methyl transfer reactionto methyl phenols, and deoxygenation to benzene andmethanol. As shown in Figure 10A a study of anisoleHDOon a SiO2-supported Pt catalyst conducted by Zhu et al.

80 at673 K and 0.1 MPa of H2in a continuous-ow reactor indicatedthat demethylation of anisole to form phenol and methane wasthe primary reaction. No direct hydrogenation of anisole or theproduced phenol was observed. The decomposed methyl group

did not remain on the surface nor was it transferred to anothermolecule, rather it was rapidly hydrogenated to form methane.

An isomerization catalyst, Pt/H-Beta, was chosen as abifunctional catalyst to provide acid sites to catalyze the alkyltransfer reaction. As shown in Figure10B, a signicant amountof toluene, xylenes, and C9 isomers were produced at high W/F(the ratio of catalyst mass to organic feed ow rate), indicatingthat acidic function (H-Beta) catalyzed the methyl transferreaction (transalkylation) from methoxyl to the aromatic ring.Bifunctional catalysts showed higher rates for the formation ofaromatics, lower hydrogen consumption, and a signicantreduction in carbon losses compared to catalysts with only ametal function.80 In addition, bifunctional Pt/H-Betaexhibitedlower rate of deactivation and less coke deposition.80 Zhao et

Figure 10. Product yield against W/F(the ratio of catalyst mass to organic feed ow rate) and the corresponding reaction network of HDO ofanisole on (A) Pt/SiO2 and (B) Pt/H-Beta catalyst at 673 K and 0.1 MPa (Ben, benzene; Tol, toluene; Ph, phenol; Xy, xylene isomers; MA,methylanisole; DMA, dimethylanisole isomers; Cr, cresol isomers; Xol, xylenol isomers). Redrawn with permission from ref80. Copyright 2011Elsevier.

Figure 11. Reaction scheme for guaiacol conversion on alumina-supported CoMo sulde catalyst at 573 K under 4 MPa of H2. Redrawn withpermission from ref108. Copyright 2009 Elsevier.


http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-012.png&w=304&h=135http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-011.jpg&w=473&h=295


10/24

al.74 tested the HDO of anisole at 423 K and 5 MPa of H2onthe bifunctional catalyst system consisting of Pd/C with H3PO4in a batch reactor. The two primary reactions were found to behydrogenation of anisole to methoxycyclohexane by metalfunction and hydrolysis of anisole to phenol by acid function.The appearance of hydrogenation of anisole as the primaryreaction was because of the high pressure of H2used.

Guaiacol and substituted guaiacols have attracted muchattention among studies of HDO of bio-oil model compounds

because they have the representative functional groups inphenolics and are relatively stable under the temperature ofpyrolysis.42 Early studies using mainly supported Mo-basedsulde catalysts showed that the basic reaction scheme ofguaiacol HDO involves a consecutive pathway from guaiacol tocatechol by demethylation and then to phenol by deoxygena-tion.42,76,106,107 However, some recent studies also suggestedthe additional routes including direct demethoxylation ofguaiacol to phenol and methyl transfer reaction to methyl-catachols during HDO of guaiacol on Mo-based suldecatalysts.71,78,108,109 Figure11shows a general reaction schemereported by Bui et al.108 for guaiacol conversion on alumina-supported CoMo sulde catalyst at 573 K under 4 MPa of H2.In addition, heavier products (coke precursors) might beformed during the reaction, especially with the acidic catalyticsystems.

Zhao et al.73 conducted HDO of guaiacol on 573 K andatmospheric pressure on SiO2-supported transition metalphosphides including Ni2P, Fe2P, MoP, Co2P, and WP in acontinuous-ow reactor. The major products for the mostactive phosphides were benzene and phenol (and a smallamount of anisole), with formation of reaction intermediates(e.g., catechol and cresol) at short contact times. This indicates

that HDO of guaiacol on metal phosphide follows similarroutes on metal suldes including demethylationdeoxygena-tion, demethoxylation, and methyl transfer reaction. Theobservation of anisole as a product also indicated that thedirect deoxygenation of guaiacol to remove OH group occurs asa new route of guaiacol HDO. The activity for HDO of guaiacolfollows the order Ni2P > Co2P > Fe2P, WP, MoP. The 8.6 wt %Ni2P/SiO2 catalysts are less active then a commercial 5 wt %Pd/Al2O3 catalyst, but much more active than a commercialsulde CoMo/Al2O3.

Elliott and Hart77 reported on HDO of guaiacol on carbon-supported Ru and Pd catalysts at 423 to 523 K and 4.0 MPa ofhydrogen in a batch reactor. As shown in Figure 12,

hydrogenated guaiacol, 2-methoxycycohexanol, is the majorproduct, indicating a completely different reaction route thanon Mo-based sulde catalysts. The hydrogenation route using

the Ru catalyst proceeded via methoxycyclohexanol tocyclohexanediols at low temperatures and continued to formcyclohexanol at higher temperatures (Figure 12A). At 523 Kand above, gasication reactions dominated. The Pd catalyst ledrst to methoxycyclohexanone at 423 K and then methox-

ycyclohexanol at 473 K with some cyclohexanediol. At 523 K,as shown in Figure 12B, the product slate was shifted towardcyclohexanol and cyclohexane (without gasication), but at 573K, the product slate was shifted strongly to cyclohexane with alarge amount of methanol byproduct. Gutierrez et al.110 studiedHDO of guaiacol on ZrO2-supported mono- and bimetallicnoble metal (i.e., Rh, Pd, Pt) catalysts using a batch reactor at373 and 573 K and 8.0 MPa of H2. At low temperature (373K), hydrogenated oxygen-containing compounds were pro-

duced as major products and methyl transfer reaction occurred.The performance of the noble metal catalysts, especially theRh-containing catalysts, was similar to or better than that oftheconventional sulded CoMo/Al2O3 catalyst. Bykova et al.

111

studied the guaiacol HDO on Ni-based solgel catalysts at 593K and 17 MPa in a batch reactor. Major products werehydrogenated intermediates and cyclohexane, indicating apreference ofhydrogenation route.

Lin et al.71 demonstrated the different reaction routesinvolved in HDO of guaiacol between noble metal and Mo-

based sulde catalyst by comparing reactions on Rh-basedcatalysts and sulde CoMo and NiMo catalysts at 573 to 673 Kand 5.0 MPa of H2 in a batch reactor. Sulded CoMo andNiMo produced methoxybenzene, methylphenol, phenol,

benzene, cyclohexene, and cyclohexane, consistent with theirproposed reaction network for guaiacol HDO on suldecatalysts, which assembles the general reaction scheme onsulde catalyst as discussed above. However, Rh-based catalystsproduced 2-methoxycyclohexanol, 2-methoxycyclohexanone, 1-methoxycyclohexane, cyclohexanol, cyclohexanone, and cyclo-hexane, suggesting that the mechanism of guaiacol HDO byRh-based catalysts involved two consecutive steps: hydro-genation of the guaiacol benzene ring, followed by demethox-

ylation and dehydroxylation of oxygenates, as shown in Figure13. They concluded that the Rh-based catalyst exhibited the

best HDO activity with the preference to saturate benzenerings, while conventional sulded CoMo and NiMo catalysts

Figure 12.Reactant and product distribution of guaiacol HDO at 4.2 MPa of H2on (A) Ru/C at 473 K and (B) Pd/C at 523 K. Adapted withpermission from ref77. Copyright 2009 American Chemical Society.




11/24

were much less active compared to noble metals on bothhydrogenation and deoxygenation reaction.

Zhao et al.74,95 tested the aqueous-phase HDO of guaiacol ona bifunctional catalyst system (Pd/C and H3PO4) at 423 K and5 MPa of H2in a batch reactor. The primary product was theintermediately hydrogenated 2-methoxycyclohexanone (selec-tivity of 100% at t= 0) by the metal-catalyzed hydrogenation ofthe aromatic ring. As shown in Figure14, the proposed reaction

pathway includes the acid-catalyzed hydrolysis of the methylgroup of hydrogenated intermediate to form 1,2-cyclo-

hexanediol, followed by acid-catalyzed dehydration to cyclo-hexenol. Another bifunctional catalyst system (Pd/C andHZSM-5) showed 100% conversion of guaiacol to cyclohexaneat 473 K and 5 MPa of H2in a batch reactor.

75

In addition, HDO of guaiacol was investigated at a low H2pressure (


12/24

aryl ether) was quantitatively converted to phenol and alkylatedaromatic at a low conversion and to cyclohexane and alkylatedcyclohexane as nal products (Table7).75 The mechanism forthe HDO of-O-4 and -O-4 model compounds started withthe metal-catalyzed hydrogenolysis of the ether to phenol andarenes, followed by the hydrogenationdehydration of thephenols and the hydrogenation of the arenes on the metal/acid

bifunctional catalyst. The 4-O-5 dimers were quantitativelyconverted to cyclohexane; however, they remained unchangedusing HZSM-5 only, indicating that both metal and HZSM-5are essential for the cleavage of the arylaryl ether bond. TheCC linkages in 5-5,-1, and-were preserved, whereas thesubstituted hydroxyl and ketone groups were selectivelyremoved, leading to C12, C14, and C16 bicycloalkanes,respectively. Strassberger et al.117 studied the HDO of-O-4lignin-type dimers on supported Cu catalysts at 423 K and 2.5MPa in a batch reactor, which showed both -O-4 cleavageproducts (phenolics) and HDO products (aromatics).

4.3. HDO of Furans and Furfurals. HDO of furans,including furan, methylfuran, dimethylfuran, benzofuran, anddibenzofuran, have attracted the most attention among thestudies of the HDO of fuel oil using supported Mo-basedsulde catalysts. The products from the HDO of furan,methylfuran, and dimethylfuran are mainly hydrocarbonsincluding isomers of alkene, alkane, and small amounts ofalkadiene.42Alkene and alkane arose from hydrogenated furans

(HYD route) and/or hydrogenation of alkadiene that formedfrom direct deoxygenation of furans (DDO route). Furan can

be completely hydrogenated to tetrahydrofuran under typicalhydroprocessing conditions, and the HDO of tetrahydrofuran

was much faster than that of the furan, indicating thattetrahydrofuran is an important intermediate during furanHDO.118,119 Detailed studies on the vapor-phase HDO of

benzofuran on sulde CoMo catalyst at 6.5 MPa of H2and 533to 583 K have shown that the reaction followed the sequencesof hydrogenation to 2,3-dihydrobenzofuran, CO cleavage too-ethylphenol, second CO cleavage to ethylbenzene, andhydrogenation to ethylcyclohexene and ethylcyclohaxane. CC

bond hydrogenolysis occurred to produce small amounts of

phenol, toluene, and benzene. A study by Bunch et al.120 onvapor-phase HDO of benzofuran on reduced NiMo catalyst at393 to 633 K and 2.0 to 5.0 MPa found the furtherhydrogenation of 2,3-dihydrobenzofuran to octahydrobenzofur-an. HDO of benzofuran appeared to occur exclusively via

hydrogenation (HYD) route. However, it is possible thathydrogenation of styrene, the primary product of directdeoxygenation (DDO), is fast and therefore is not detectedin the products. HDO of dibenzofuran occurred via both DDOroute to biphenyl and HYD route to hexahydrodibenzofuran onsulde NiMo and CoMo catalysts.121 Hydrogenation of

biphenyl and deoxygenation of hexahydrodibenzofuran oc-curred subsequently. The formed hydrocarbons were furtherconverted to the single-ring products.

Few studies have investigated the HDO of furan on noblemetal catalysts. Furan hydrogenation was carried out by Klieweret al.122 over Pt nanoparticles with 10 Torr of furan and 100Torr of H2 at 353 to 413 K. Dihydrofuran, tetrahydrofuran,

Figure 16. Demonstration of rst-order kinetics of (A) overallconversion and (B and C) conversion to each primary product ofguaiacol HDO on Pt/-Al2O3at 573 K and 0.14 MPa of H 2. The termxi represents the conversion to product i. Conversion was varied bychanging the catalyst mass. Redrawn with permission from ref113.

Copyright 2011 American Chemical Society.

Table 7. Aqueous-Phase HDO of Phenolic Dimers onPd/Cand HZSM-5 Catalysts at 473 K and 5.0 MPa of H2

75,a

aThe conversion is 100% for all feeds.


http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-018.png&w=208&h=442http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-017.jpg&w=189&h=179


13/24

butanol, and propylene were all detectable products. As shownin Figure17A, the conversion of butanol to propylene increases

as the temperature increases from 353 to 413 K. A reactionnetwork was proposed, as shown in Figure17B. Spectroscopyanalysis suggested that the furan ring adsorbed on Pt surfaces

via at adsorption and tetrahydrofuran and oxametallacycle

intermediate bound on metal surfaces vertically under reactionconditions. Study of HDO of tetrahydrofuran on supported Ptcatalysts at 423 to 573 K by Kreuzer and Kramer123 showedthat the primary reaction was CO bond cleavage to open theve-membered ring. The primary product butanol underwent asecond CO bond cleavage to form butane and H2O or CC

bond cleavage via butanol to form propane and CO.

Liu et al.124 studied the HDO of benzofuran over silicaalumina-supported Pt, Pd, and alloyed PtPd catalysts at 553 Kand 3.0 MPa. Only the hydrogenationdeoxygenation (HYD)

route was found for the conversion of benzofuran, which wasrst hydrogenanted to 2,3-dihydrobenzofuran and then to

octahydrobenzofuran, followed by deoxygenation to cyclo-hexanes, as shown in Figure 18. Bimetallic PtPd catalystsshowed higher activities in hydrogenation and oxygen removalthan their monometallic counterparts did. Bowker et al.125

reported that the Ru phosphide catalysts (Ru2P/SiO2and RuP/SiO2) exhibited furan HDO activities similar to or higher than aRu/SiO2 catalyst, and the phosphide catalysts favored C4hydrocarbon products (butene and butane, CO cleavage)

while the Ru metal catalyst produced primarily C3 hydro-carbons (propene and propane, CC cleavage). These Ru-

based catalysts were much more active than a commercialCoMo sulde catalyst.

Furfural is another important component commonly foundin bio-oil. Due to its high reactivity, this compound needs to becatalytically hydrogenated or hydrodeoxygenated to improve

bio-oil stability. Early studies on furfural have focused on thehydrogenation of furfural to furfuryl alcohol,2-methylfuran, andtetrahydrofuran using supported Cu,126 Pt,127 NiB alloy,128

and CuCr129 catalysts. Figure19shows a reaction pathway for

furfural hydrogenation proposed by Zheng et al.129 for gas-phase reaction at 543 K and atmospheric pressure on Cu/Zn/

Al/Ca/Na catalysts. Primary reaction of furfural involveshydrogenation of CO to furfural alcohol and hydrogenolysisof CC bond to furan. Furan reacts further via HYD or DDOto tetrahydrofuran or butanal and butanol, respectively, asdiscussed above. Furfural alcohol produces 2-methylfuran viaCO bond cleavage, furan via CC bond cleavage, andtetrahydrofurfuryl alcohol via hydrogenation.

Elliott and Hart77 studied HDO of furfural on carbon-supported Ru and Pd catalysts at 423 to 523 K and 4.0 MPa of

hydrogen in a batch reactor. As shown in Figure 20A, at 473 K,the major product of the HDO of furfural on Ru catalyst was

Figure 17. HDO of furan over Pt nanoparticles. (A) productselectivity as a function of reaction temperature over 3 nm Ptnanoparticles and (B) reaction network. Redrawn with permissionfrom ref122. Copyright 2010 American Chemical Society.

Figure 18.Reaction network for HDO of benzofuran over silicaalumina-supported Pt, Pd, and PtPd catalysts at 553 K and 3.0 MPa. Redrawnwith permission from ref124. Copyright 2012 American Chemical Society.

Figure 19. Reaction pathway proposed for furfural HDO (VL:

valerolactone). Redrawn with permission from ref 129. Copyright2006 Elsevier.


http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-022.png&w=212&h=153http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-021.png&w=361&h=72http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-020.png&w=135&h=33http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-019.png&w=198&h=252


14/24

tetrahydrofurfuryl alcohol (THFMeOH), a completely hydro-genated furfural. Other prominent products included -

valerolactone (GVL), methyltetrahydrofuran (mTHF), penta-nediol, and cyclopentanol. At higher temperatures (i.e., 523 and573 K), tetrahydrofurfuryl alcohol quickly converted tomethyltetrahydrofuran and tetrahydrofuran. As shown inFigures 20B and 20C, at 473 K and on Pd catalyst, majorproducts were cyclopentanone and cyclopentanol. Methylte-

trahydrofuran, tetrahydrofurfuryl alcohol, and -valerolactonewere other prominent products. At higher temperature,methyltetrahydrofuran, -valerolactone, and pentanol were themajor products at the end of the test. Elliott and Hartconcluded that a route from furfural to cyclopentanone andpentanols was an important mechanistic route. At lowtemperatures, furfural hydrogenation to tetrahydrofurfurylalcohol appeared to be the primary route for both Ru and Pdcatalysts. At 523 K and above, the pathway through -

valerolactone to 1,4-pentanediol and methyltetrahydrofuranwas more important. These authors also performed a singleuncatalyzed test at 523 K and observed a solid, polymericmaterial formed from furfural, indicating that polymerizationand even charring of furfural might occur simultaneously duringHDO reaction.

Sitthisa et al.130132 studied the HDO of furfural on SiO2-supported Pd, Cu, Ni, and PdCu catalysts at 503 to 563 K and0.1 MPa of H2in a continuous-ow reactor. Table8 compares

the activity and product selectivities on the three monometalliccatalysts. The conversion of furfural on Cu/SiO2yields mainlyfurfuryl alcohol from hydrogenation of the carbonyl, with onlysmall amounts of 2-methylfuran, obtained from a subsequentcleavage of the CO bond in furfuryl alcohol. A LangmuirHinshelwood model was used to t the kinetic data and providethe parameters of physical signicance. The rate constant forthe hydrogenation of furfural was signicantly higher than thatfor HDO of furfuryl alcohol to produce 2-methylfuran. Theheat of adsorption for furfural (12.3 kcal/mol) was similar tothat of adsorption of water (12.4 kcal/mol), but higher than

those for furfuryl alcohol (6.9 kcal/mol) and 2-methylfuran(3.7 kcal/mol). On the Pd catalyst, furfural conversion wasdescribed as two parallel routes: (i) decarbonylation to furanthat subsequently hydrogenates to tetrahydrofuran and (ii)hydrogenation to furfuryl alcohol that subsequently hydro-genates to tetrahydrofurfuryl alcohol, as shown in Figure21forconversion and product distribution as a function ofW/Fandthe corresponding reaction network. Decarbonylation was thedominant reaction even at low W/F. The proposed reactionmechanism (see Figure 21B) includes two parallel reactionsrequiring different intermediates. The preferential formation ofan acyl intermediate at higher temperatures, which can readilydecompose into CO and hydrocarbons, led to a higher

Figure 20. (A) Cyclic ether pathway product distribution of furfuralconversion over Ru/C at 473 K. (B) Cyclic ether pathway productdistribution of furfural conversion over Pd/C at 473 K. (C) Cyclic

ketone pathway product distribution from furfural conversion overPd/C at 473 K. (GVL, -valerolactone; GBL, -butyrolactone; THF,tetrahydrofuran.) Adapted with permission from ref 77. Copyright2011 American Chemical Society.

Table 8. Activity and Product Distribution of HDO ofFurfural over SiO2-Supported Cu, Pd, and Ni Catalysts at573 K and 0.1 MPa131

10% Cu/SiO2

1% Pd/SiO2

5% Ni/SiO2

TOF (s1) 1.3 265.8 76.5

conversion (%) 69 69 72

hydrogenation (%)

furfuryl alcohol 98 14 25

2-methylfuran 2 3

tetrahydrofurfurylalcohol

5 4

decarbonylation (%)

furan 60 43

tetrahydrofuran 20

ring-opening (%)

butanal 12

butanol 3butane 10




15/24

decarbonylation/hydrogenation ratio. The PdCu alloys, with

a lower extent of electron back-donation to the *

system ofthe aldehydes, had less preference of formation of acylintermediates of furfural. Therefore, the decarbonylation rate

was reduced on PdCu catalysts, but the hydrogenation ratewas increased. The Ni catalyst had a product distributionsimilar to that of the Pd catalyst. Furan was not as abundant

because it further reacted to form ring-opening products due tothe interaction of the ring with the surface, which was evenstronger than on the Pd catalyst. Zhang et al.133 conducted astudy on the furfural decarbonylation reaction on K-doped Pd/

Al2O3at 453 to 533 K and ambient pressure in a continuous-ow reactor. The doping of K not only promoted the furfuraldecarbonylation but also suppressed the hydrogenation sidereaction.

Zhao et al.82 studied HDO of furfural and methylfufural inwater on a Ni/HZSM-5 catalyst at 523 K and 5 MPa of H2in abatch reactor. Two parallel reactions, intermolecular dehy-dration of furfural to tetrahydropyran and HDO of furfural topentane, competed to ultimately produce 64% pentane and36% tetrahydropyran. The HDO of furfural occurred byhydrogenation of the furan ring, followed by hydrolysis of theC5 ring, and subsequent alcohol dehydration/hydrogenation toform the straight-chain alkane.

4.4. HDO of Carboxylic Acids. Carboxylic acids, such asacetic acid, formic acid, and propanoic acid, are found in largeamounts in bio-oils and contribute to the acidic and corrosivenature of bio-oil. Therefore, conversion of carboxylic acid

during hydrotreating of bio-oil is critical to produce suitabletransportation fuels. Early studies on acetic acid conversionhave focused on the hydrogenation of acetic acid toacetaldehyde using oxidecatalysts (e.g., titania, iron oxide, tinoxide, chromium oxide).134138 The addition of Pt to thecatalyst enhanced selectivity and activity.134 Detailed studies onreaction mechanism and kinetics showed that the reduction ofacetic acid on oxide-supported Pt catalysts started via a reaction

between adsorbed H atoms from the Pt surface and an acylspecies on oxide. This pathway was the major route for theproduction of acetaldehyde, which could be further hydro-genated to ethanol.138 Pt was used as active sites to activatehydrogen.

Reaction on pure Pt was studied by testing acetic acid

conversion on Pt supported on SiO2, an inert support, via atemperature-program reaction at 0.12 MPa of H2and from 293to 723 K. Acetic acid was converted at 593 K, and the mainreaction products were methane, water, CO, andCO2, obtainedfromthe total decomposition of acetic acid.134 Gursahani etal.139 studied reaction pathways for the catalytic conversions ofacetic acid over a Pt/SiO2catalyst at temperatures from 500 to600 K and showed that the stoichiometric reaction created anequimolar mixture of CO and CH4: CH3COOH + H2 CH4+ C O + H2O. A small amount of CO2, a decarboxylationproduct of acetic acid, was observed at temperatures >673 K.

Elliott and Hart studied HDO of acetic acid on Pd/C andRu/C catalystsat 473 to 573 K and 4.0 MPa of hydrogen in a

batch reactor.77 Acetic acid was hydrogenated to ethanol with

moderate yields on Pd/C at 573 K, with a small amount ofethyl acetate as a secondary condensation product. Acetic acidwas not effectively hydrogenated to ethanol on Ru catalyst,rather it decomposed to methane and carbon dioxide attemperatures where there was signicant activity. Olcay et al.140

studied aqueous-phase hydrogenation of acetic acid overtransition metal catalysts, including carbon-supported Ru, Rh,Pt, Ir, and Pd, Raney Ni, and Raney Cu at 5.17 MPa of H2andat 383 to 683 K. The turnover rate of acetic acid conversiondecrease in the order Ru > Rh Pt > Pd Ir > Ni > Cu. TheRu/C catalyst showed 70 to 80% selectivity for ethanol, thehydrogenation product, at temperatures below 448 K and 83%selectivity for methane,the decomposition product, at 498 K.Pallassana and Neurock141 carried out density functional theory

calculations to examine alternative mechanisms for thehydrogenolysis of acetic acid to ethanol over Pd. The overallreaction energy results indicated that the most favorable pathfor acetic acid hydrogenolysis involved the formation of anacetyl intermediate, followed by its hydrogenation toacetaldehyde. The acetaldehyde was subsequently hydro-genated to form ethanol.

Gas-phase HDO of propionic acid on Pd/SiO2and Pt/SiO2catalysts at 523 to 673 K under atmospheric pressure led toalmost 100% selectivity to ethane, suggesting that hydro-genolysis of CC bond played an essential role.142 However,Cu/SiO2 catalyst primarily converted propionic acid topropanal and 1-propanol by CO hydrogenation. Turnover

Figure 21. Conversion and product distribution (A) from furfural over 0.5% Pd/SiO2 as a function of W/Fat 503 K and 0.1 MPa of H2 andproposed reaction mechanism (B). (THF, tetrahydrofuran; FOL, furfuryl alchol; HFOL, tetrahydrofurfuryl alcohol.) Redrawn with permission fromref132.Copyright 2011 Elsevier.


http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-024.jpg&w=405&h=173


16/24

rates of propionic acid conversion followed the order Pd > Pt >Cu. Improved activities were observed when bulk acidic salt(Cs2.5H0.5PW12O40, CsPW) was used as a support to achievemetalacid bifunctional catalysis. Hydrogenation of propionicacid was observed on Pd/CsPW and Pt/CsPW catalysts at lowtemperature, in contract with SiO2-supported catalysts,suggesting that propionic acid hydrogenation took place onthe CsPW polyoxometalate, possibly by a Marsvan Krevelenmechanism. Aqueous-phase HDO of propanoic acid over Ru/ZrO2 catalysts at 463 K and 6.4 MPa showed that the CC

bond cleavage to methane and ethane was the dominantreaction compared to CO hydrogenation to propanol andpropane at high temperatures,143 indicating a preferentialformation of propanoyl intermediate, which decomposed intoCO and hydrocarbons. Addition of Mo to Ru catalyst inhibitedthe CC bond cleavage reaction of propanoic acid; however,the conversion decreased due to the formation RuMoOxandthe stability and variety of propanoyl intermediate species onthe RuMo surface.

4.5. HDO of Other Oxygenates. Wildschut et al.144

conducted HDO of D-glucose, D-cellobiose, and D-sorbitol, therepresentative model components for the carbohydrate fractionin bio-oil at 523 K and 10 MPa of H2 using Ru on carboncatalysts in water. Two parallel reaction routes were reported: ahydrogenation route leading to smaller polyols and gaseoushydrocarbons (e.g., methane and ethane) and a thermalnoncatalyzed route leading to insoluble humins (char). Theauthors suggested that at least part of the gas-phasecomponents and solids formed upon hydrotreating of bio-oilsarose from thecarbohydrate fraction in the oil.

Sitthisa et al.132 studied gas-phase HDO of 2-methylpentanalon SiO2-supported Pd and PdCu catalysts at 398 K and 0.1MPa of H2. On the Pd catalyst, the primary reactions of 2-methylpentanal were decarbonylation to pentane and hydro-genation to 2-methylpentanol. Further, etherication occurred

between 2-methylpentanal and formed 2-methylpentanol to

produce ether. The reaction network is shown in Figure 22.Upon addition of Cu, both the overall activity and thedecarbonylation selectivity decreased and the selectivity tohydrogenation and etherication increased.

Li and Huber145 investigated the HDO of 1- and 2-butanol,intermediates of HDO of sorbitol, on Pt/SiO2Al2O3at 518 Kand 2.93 MPa in a continuous-ow reactor. Butane, propane,and CO2were the nal products for 1-butanol, whereas butane

was the only product for 2-butanol. CC bond cleavage wasmore preferable than CO bond cleavage for 1-butanol underthe reaction conditions. As shown in Figure 23, 1-butanol cango through a continuous dehydrogenation and decarbonylationreaction to produce propane and CO2. 2-Butanol cannotundergo the CC bond cleavage by decarbonylation becauseits dehydrogenation product is a ketone and not an aldehyde.

4.6. Mutual Inuences of Oxygen-Containing Com-pounds during Simultaneous Reaction. As discussed insection2, bio-oils are complex mixtures of oxygen-containingcompounds. Because of the competitive adsorption ofmolecules in these mixtures and subsequent inhibiting andpoisoning effects, signicant differences exist in reaction ratesand selectivities measured from single-model compounds andfrom the same compound in the various mixtures. Such effectshave been well recognized in studies of the mutual inuence ofS- and N-containing molecules during hydrotreating of

petroleum-based oil.

146

Almost all studies of HDO of modelcompounds were conducted on a single compound. The fewstudies that investigated the inhibition effect primarily focusedon the effects of H2S and water on HDO reactions over Mo-

based sulde catalysts. H2S has been found to strongly affectthe activity and selectivity of the Mo-based sulde cata-lysts.84,104,147150 The effect of H2S depends on the oxygenatedmolecule used, the experimental conditions, and the type ofcatalyst (Ni(Co)Mo-based catalysts). For instance, thepresence of H2S strongly decreased the phenol and anisoleHDO activity of the sulded CoMo/Al2O3 catalyst at 473 to573 K, and the ratio of the HDO reaction pathways dependedon H2S concentration.

104 The presence of H2S suppressed theDDO route to aromatics; however, at moderate H2Sconcentrations, the HYD route to alicyclics remained thesame as in the absence of H2S.104 A study by Bouvier et al.147on HDO of 2-ethylphenol on sulded Mo, NiMo, and CoMocatalysts at 613 K and 7 MPa found that H2S, needed tomaintain the suldation level of the catalysts, has promoting orinhibiting effects depending on the catalyst tested and thedeoxygenation pathway considered. Over the three catalysts,H2S was found to slightly promote the HYDroute and stronglyinhibit the DDO route. Romeroa et al.150 reported that H2Spromoted the HYD route and inhibited the DDO route of 2-ethylphenol HDO over a NiMoP/Al2O3 catalyst. In general,

water caused a decrease of the catalytic activity for HDO due tocompetitive adsorption and modication of catalyst struc-ture.63,65,151,152 However, few studies have reported the effectof water on HDO activity of noble metal catalysts.

In addition, the mutual inuence of HDO and HDS or HDNon Mo-based sulde has been studied for simultaneousreactions during hydrotreating of petroleum-based oils153156

and coprocessing of bio-oils and petroleum-based oils.108,157

For example, Philippe et al.156 reported that both guaiacol andphenol inhibit the HDS of sulfur compounds (dibenthiopheneand dimethyldibenthiophene) on a sulded CoMoP/Al2O3catalyst at 613 K and 4.0 MPa. In addition, they reportedthat, according to a LangmuirHinshelwood model, guaiacolhas a stronger inhibiting effect than phenol due to competitiveadsorption between the oxygen- and sulfur-containing com-pounds on the catalyst surface. Bui et al.108 observed a decreasein HDS performance of sulded CoMo/Al2O3 catalyst at low

Figure 22. Reaction network of the HDO of 2-methylpentanal(MPAL). Redrawn with permission from ref 132. Copyright 2011Elsevier.

Figure 23.Major reaction pathways for the HDO of 1-butanol and 2-butanol over Pt/SiO2Al2O3 catalyst at 518 K. Redrawn withpermission from ref145. Copyright 2010 Elsevier.


http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-026.jpg&w=200&h=67http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-025.png&w=140&h=54


17/24

temperature and high contact time during HDS processing of astraight run gas oil with coprocessing of guaiacol. This decreasein HDS was due to the formation of intermediate phenols,

which compete with sulfur-containing molecules for adsorptionon active sites. At a higher temperature, complete HDO ofguaiacol was observed and HDS could proceed without furtherinhibition.

Few studies have focused on the mutual inuences ofoxygen-containing compounds during their simultaneous HDOreaction in the various mixtures. However, Romero et al.150

studied the competitive effects between furanic compounds(benzofuran and 2,3-dihydrobenzofuran) and a phenoliccompound (2-propylphenol) over a sulded NiMoP/Al2O3catalyst at 613 K and 7 MPa in a xed-bed reactor. Benzofuranand/or 2,3-dihydrobenzofuran strongly inhibited the trans-formation of 2-propylphenol into deoxygenated compounds,

whereas 2-propylphenol hardly affected the conversion ofbenzofuran and 2,3-dihydrobenzofuran, indicating a strongerbinding of benzofuran and 2,3-dihydrobenzofuran on activesites of catalysts than 2-propylphenol. Ferrari158 found that amutual competition for the HDO existed between guaiacol andethyldecanoate. In addition, guaiacol and ethyldecanoate

inhibited the conversion of 4-methylacetophenone; however,the reciprocal effect was less intense. In a brief abstract, Wan etal.159 reported that results from binary combination of aceticacid and p-cresol on Ru/C catalyst indicated that thehydrogenation of acetic acid was suppressed by the presenceofp-cresol. In contrast, the presence of acetic acid promotedthe HDO ofp-cresol by dehydration reaction, leading to highselectivity to methylcyclohexane. Apparently, additional studiesof the HDO of bio-oil model compounds that use feedcompositions more relevant to actual bio-oils are needed.

5. HYDROTREATING OF ACTUAL BIO-OIL FEEDS

Recently, Elliott17 provided a very detailed summary of researchefforts on the hydrotreating of actual bio-oil products. Inaddition, Choudhary20 published a report in 2011 that discussesthe hydroprocessing of actual bio-oils. This section focuses onrecent advances in the hydrotreating of actual pyrolysis bio-oilfeed, primarily on non-sulded metal catalysts.

5.1. One-Stage Hydrotreating of Bio-Oil with NobleMetal Catalyst. Elliott et al.160 tested Ru/C catalyst forhydrotreating of white wood pyrolysis bio-oil and bagassepyrolysis bio-oil at 454 to 513 K, 13.2 to 14.3 MPa of H2, and0.22 to 0.67 liquid hourly space velocity (LHSV) in acontinuous-ow reactor. Table 9 compares the feed andproduct analysis results for bio-oil hydrotreating. For white

wood oil, a deoxygenation of 31 to 70% and a product yield of

0.54 to 0.79 g/g feed (dry basis) were reported. For bagasse oil,a deoxygenation of 32 to 46% and product yield of 0.64 to 0.81g/g feed (dry basis) were reported. Oxygen content decreasedfrom 42 to 20%. Two separate phases, an aqueous phaseand a tar phase, formed in the product oil because of the changein the water solubility of the component in the product oils.The products were less hydrophilic by removal of the carbonyl,olenic, and aromatic characteristics. Signicant loss of catalystactivity was observed during the experiments, probably becauseof coke formation and contaminants (i.e., sulfur and iron) inthe bio-oil.

Wildschut et al.67 investigated a variety of supported noblemetal catalysts (Ru/C, Ru/TiO2, Ru/Al2O3, Pt/C, and Pd/C)for hydrotreating of beech wood pyrolysis bio-oil at temper-atures of 523 and 623 K and pressures of 10 and 20 MPa in a

batch reactor. In addition, typical hydrotreatment catalysts(sulded NiMo/Al2O3 and CoMo/Al2O3) were tested forcomparison. Under mild HDO conditions (523 K and 10MPa), two liquid phases (water and oil) and char wereproduced with mass balance varying between 77 and 96 wt %.Figure24A shows that oil yields on a dry basis ranged between21 and 55 wt % with oxygen contents between 18 and 27 wt %.Both the yields and the levels of deoxygenation were higher forthe noble metal catalysts than for the classical hydrotreatmentcatalysts. For a goal of a high oil yield combined with lowoxygen content, Pd/C was believed to be the best choice formild HDO followed by Ru/TiO2. Under deep HDO conditions(623 K and 20 MPa), two oil phases were obtained with thenoble metal catalysts. Mass balance ranged between 97 and 100

wt %. Figure24B shows that the combined oil yields on a drybasis ranged between 25 and 65 wt % with oxygen contentsbetween 6 and 11 wt %, much lower than after mild HDO.Figure25uses a van Krevelen plot to summarize the effects ofcatalyst and process conditions. Deep reduction of oxygencontent seems possible only under severe conditions. Based onoil yields, deoxygenation levels, and extents of hydrogen

consumption, Ru/C was suggested to be the most promisingcatalyst for further testing. Pd/C was considered to have thepotential to provide higher oil yields than Ru/C, but withhigher product oxygen content and hydrogen consumption.The deep HDO upgraded pyrolysis oil by Ru/C had lowerorganic acid, aldehyde, ketone, and ether content than the feed,

but higher phenolic, aromatic, and alkane content. Wildschutconducted additional studies focused on using a Ru catalyst for

bio-oil hydrotreating and concluded that the highest oil yield(65 wt %) at deep HDO condition was obtained for 4 hreaction time at 623 K and 20.0 MPa.68,69 A longer reactiontime led to a signicant decrease of the oil yield due to theformation of gas products. Catalyst recycling experimentsindicated a severe deactivation in hydrogenation upon

recycling. Characterization of the Ru/C catalyst before andafter reaction under deep HDO conditions showed signicantcoke deposition and a decrease in pore volume and metaldispersion.

Fisk et al.161 conducted liquid phase HDO of a model bio-oil(mixture of ten compounds to reect the composition of typicalpyrolysis oil) over a series of supported Pt catalysts at 623 K ina batch reactor under inert atmosphere using in situ generatedhydrogen. Pt/Al2O3 showed the highest activity for deoxyge-nation with the oxygen content of the model oil decreasingfrom an initial value of 41 to 3 wt % after upgrading with oil

yield of around 30% (dry basis). The major components in thetreated oil were alkyl-substituted benzenes and cyclohexanes,

Table 9. Comparison of Feed and Product Analysis Results

for Bio-Oil Hydrotreating17

bio-oil

whitewood

white wood,hydrogenated bagasse

bagasse,hydrogenated

oil composition, wt%, dry basis

C 48.3 65.674.0 52.0 65.170.7

H 7.4 8.510.1 6.6 7.79.4

O 44.4 16.725.8 41.3 20.327.0

H/C atomic ratio(wet)

1.82 1.491.68 1.51 1.371.59

water content, wt % 30.0 9.715.7 35.0 11.214.0




18/24

and alkyl-substituted phenols were the main residual oxygen-containing compounds. Fisk et al. suggested that the reactionroute for bio-oil upgrading using in situ generated hydrogenproceeded via two steps: light oxygenates underwent reformingto H2 and CO2, and, concomitantly, aromatics underwentHDO in thepresence of H2.

Zhao et al.82 used a bifunctional catalyst (Ni/HZSM-5) toconvert n-hexane-extracted crude bio-oil to produce C5C9paraffins, naphthenes, and aromatics by a hydrogenationhydrolysisdehydrationdehydroaromatization cascade reac-tion in the presence of substantial concentrations of water at523 K and 5 MPa of H2in a batch reactor. The components ofn-hexane-extracted bio-oil mainly include C5C6 substitutedfurans, ketones, and aldehydes and C6C9 substituted phenols.

The resulting gasoline-range hydrocarbons contain less than10% C5C6 paraffins and more than 90% C5C9 naphthenesand C6C9 aromatic molecules. The Ni/HZSM-5 catalyst wasreported to be hydrothermally stable.

5.2. Stabilization of Bio-Oil with Noble MetalCatalysts in a Dual-Stage Hydrotreating Process. Mostpreviously mentioned one-step hydrotreatments of bio-oil onsupported metal catalysts were conducted at the high-temperature range of 523 to 623 K in batch reactors. In thetest using batch reactors, bio-oil could be stabilized byhydrogenation during temperature increase from room temper-ature to reaction temperature, which therefore greatly alleviatedthe char and coke formation. However, the hydrotreating of

pyrolysis bio-oil at high temperature in a continuous-owreactor could result in heavy product char plugging reactor andcatalyst encapsulation by coke-like material, as revealed byElliott et al.17 It is notable that the industrial process for bio-oilhydrotreating prefers continuous-ow ow reactor. Thechemical instability of bio-oil was attributed to unsaturateddouble bonds (e.g., olens, aldehydes, ketones) which mightreact through condensation to form polymerization products.Thus, it is desirable to eliminate these groups via low-temperature hydrogenation before they react to form high-molecular-weight compounds. A two-step process wasdeveloped: a hydrotreating step at temperatures below 573 Kto stabilize bio-oils, followed by the hydrotreating step at moresevere conditions to achieve deep oxygen removal. SuldeCoMo and NiMo, catalysts proven effective in producingstabilized bio-oils, have been the focus of most stabilizationstudies.17,20,162 However, noble metals (e.g., Pd, Pt, or Ru) havealso been used because oftheir excellent hydrogenation activity.

Gagnon and Kaliguine163 investigated the effects of a mildhydrogenating pretreatment using a Ru catalyst on the HDO of

wood-derived vacuum pyrolysis oil in a batch reactor. First-stage studies focused on a Ru/Al2O3 catalyst at 353 to 413 Kand 410 MPa, and second-stage studies focused on a NiO-

WO3/-A12O3catalyst at 623 K and 17 MPa. A temperature of353 K and a pressure of 4 MPa were found to be the optimalconditions for the rst stage. The yield of HDO products wascorrelated with the average molecular weight of the products. Inaddition to aldehyde hydrogenation and polymerization,hydrogenolysis reactions occurred over the Ru catalyst duringthe rst stage. Gagnon and Kaliguine concluded that increasedcontrol of polymerization/coking during the less severe rst-stage operatingconditions resulted in higher HDO conversions.

Elliott et al.52 reported that a Pd on carbon catalyst could beused in a continuous-ow reactor at 583 to 648 K, 0.18 to 1.12

LHSV, and 14 MPa to hydrogenate various bio-oils to partiallyupgraded bio-oils suitable for the next, more severe, hydro-cracking step. Oil yield, product structure, and hydrogenconsumption were not inuenced dramatically by feedstock

variations. Gas yield increased and oil yield decreased astemperature increased from 583 to 648 K; oxygen contentreached a minimum at 613 K. Table 10 shows that oxygencontent for hydrogenated bio-oil from mixed wood was 12.3 wt% (613 K and 0.25 LHSV), which was much lower than that forthe feedstocks (33.7 wt %). Catalyst bed plugging was observed

because of the higher temperature used in this study. The oil-phase products obtained from the rst stage were furtherupgraded by a hydrocracking process using a traditional

Figure 24.Oil yields and oxygen contents (both on a dry basis) of the combined oil phases for (A) the mild HDO of pyrolysis oil (523 K, 10.0 MPa,4 h) and (B) deep HDO of pyrolysis oil (623 K, 20.0 MPa, 4 h) over various catalysts. Redrawn with permission from ref67. Copyright 2009American Chemical Society.

Figure 25. van Krevelen plot for the elemental compositions (drybasis) of the produced oils by mild (523 K, 10.0 MPa, 4 h) and deepHDO (623 K, 20.0 MPa, 4 h) with various catalysts and products byHPTT (high-pressure thermal treatment). Redrawn with permissionfrom ref67. Copyright 2009 American Chemical Society.


http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-028.jpg&w=239&h=141http://pubs.acs.org/action/showImage?doi=10.1021/cs400069z&iName=master.img-027.jpg&w=342&h=126


19/24

hydrocracking sulde catalyst. The tests were performed atlower pressure (10 MPa) and higher temperature (678 K) thanthe hydrotreating of as-produced bio-oils, and stable operation

was observed without reactor plugging, indicating that a

stabilized hydrogenated bio-oil using Pd/C catalyst was suitablefor severe hydrotreating. Table10shows that high oil yield anddeep oxygen removal were achieved after the hydrocrackingprocess. Incorporation of both steps into a nonisothemalreactor system was further explored, and, as shown in Table10,a high yield of hydrocarbon products from the highlyoxygenated bio-oil was demonstrated. A recent report byElliott et al.164 describes a two-stage hydrotreating process forupgrading the bio-oil from softwood biomass in a bench-scalecontinuous-ow xed-bed reactor system. The rst stage wasconducted at 443 K using a carbon-supported suldedruthenium catalyst, and the second stage was operated at 673K using a carbon-supported sulded CoMo catalyst. Using thisprocess, each gram of dry bio-oil feed yielded 0.35 to 0.45 g of

oil product with densities of 0.82 to 0.92 g/mL, oxygencontents of 0.2 to 2.7 wt %, and a total acid number (TAN) of


20/24

hydroxyacetone, furfural, and small amounts of guaiacols. Low-temperature hydrogenation, withaRu/C catalyst at 398 to 448K and 6.9 MPa in a batch reactor49 orat 398 K and 5.2 to 10.0MPa in a continuous-ow reactor,50 was found to stabilizecompounds with high functionalities by converting aldehydes,sugars, and unsaturated aromatics to alcohols, sugar alcohols,and saturated aromatics, respectively. The stabilized aqueousfraction was further treated by aqueous-phase reforming toproduce H2,

49byaqueous-phase dehydration/hydrogenation toproduce alkanes,49 or by zeolite upgrading to producearomatics and olens.50 Stabilization of the aqueous fraction

by hydrogenation followed by zeolite upgrading resulted inmuch higher yields of light alkanes, aromatics, and olenscomparedtothe direct upgrading of aqueous fraction of bio-oil

on zeolite.50

A two-stage stabilization process, involving low-temperature hydrogenation on a Ru/C catalyst at 398 K and10.0 MPa followed by high-temperature hydrogenation on aPt/C catalyst at 523 K and 10.0 MPa, was demonstrated tofurther improve the yields of light alkanes, aromatics, andolens after zeolite upgrading.50 No reactor plugging wasobserved in single-stage or two-stage stabilization during vedays of operation, which implies that the catalysts were stable.50

Li et al.169 studied the aqueous-phase HDO of carbohydratesolutions from hydrolysis of wood by a two-step catalyticprocess using a Ru/C catalyst at 393 K and 6.2 MPa in the rststage and a Pt/zirconium phosphate catalyst at 518 K and 6.2MPa in the second stage. The aqueous carbohydrate solutions

were a mixture of xylose, water-soluble hemicellulose oligomers,

acetic acid, glucose, glucose oligomers, and probably somelignin polymers. The Ru/C catalyst was able to selectivelyhydrogenate xylose into xylitol but could not selectivelyhydrogenate the xylose oligomers. This two-stage process wasable to convert the aqueous carbohydrate streams derived frommaple wood into gasoline-range products with carbon yields ofup to 57% and an estimated octane number of 96.5. Nosignicant catalyst deactivation was observed, indicating thatthe catalysts remained very stable.

de Miguel Mercader et al.70 studied the HDO of aqueousfraction of bio-oil obtained by phase separation after addition of

water. They reported that operation at 493 to 583 K and 19.0MPa on a Ru/C catalyst produced HDO oil suitable for

coprocessing with fossil feeds in a FCC unit. Compared to theoil fraction of bio-oil and the whole bio-oil, the aqueous fractionproduced HDO oils with the highest quality (e.

Date post:	01-Jun-2018
Category:	Documents
Upload:	mauricio-escobar-labra
View:	225 times
Download:	1 times

Recent Advances in Hydrotreating of Pyrolysis Bio-Oil

Documents