Screening acidic zeolites for catalytic fast pyrolysis of biomass and its components

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Journal of Analytical and Applied Pyrolysis 92 (2011) 224–232

Contents lists available at ScienceDirect

Journal of Analytical and Applied Pyrolysis

journa l homepage: www.e lsev ier .com/ locate / jaap

Screening acidic zeolites for catalytic fast pyrolysis of biomass and itscomponents�

David J. Mihalcik, Charles A. Mullen, Akwasi A. Boateng ∗

Eastern Regional Research Center, Agricultural Research Service, U. S. Department of Agriculture, 600 E. Mermaid Lane, Wyndmoor, PA 19038, United States

a r t i c l e i n f o

Article history:Received 6 April 2011Accepted 6 June 2011Available online 14 June 2011

Keywords:Catalytic fast pyrolysisZeolitePyrolysis oil

a b s t r a c t

Zeolites have been shown to effectively promote cracking reactions during pyrolysis resulting inhighly deoxygenated and hydrocarbon-rich compounds and stable pyrolysis oil product. Py/GC–MS wasemployed to study the catalytic fast pyrolysis of lignocellulosic biomass samples comprising oak, corncob, corn stover, and switchgrass, as well as the fractional components of biomass, i.e., cellulose, hemicel-lulose, and lignin. Quantitative values of condensable vapors and relative compositions of the pyrolyticproducts including non-condensable gases (NCG’s) and solid residues are presented to show how reactionproducts are affected by catalyst choice. While all catalysts decreased the oxygen-containing productsin the condensable vapors, H-ZSM-5 was most effective at producing aromatic hydrocarbons from thepyrolytic vapors. We demonstrated how the Si/Al ratio of the catalysts plays a role in the deoxygenationof the vapors towards the pathway to aromatic hydrocarbons.

Published by Elsevier B.V.

1. Introduction

Owing to its vast availability and potential application as aliquid fuel, the use of biomass for renewable energy productionhas received increasingly more attention over the last 40 years[1].Among the various conversion technologies being studied, fastpyrolysis has received significant attention because it is feedstockneutral [2] and has the potential to be easily incorporated intoa distributed process model [3]. This notwithstanding the majordrawback towards commercialization of fast pyrolysis oil (bio-oil)is its instability caused by high oxygen content and acidity, therebyrendering it unsuitable for incorporation into existing petroleum-based infrastructure “as is” [2,4]. Upgrading technologies that favorpathways towards reducing the oxygen content are therefore nec-essary [1]. Among these technologies, incorporation of catalystsinto the pyrolysis reaction, in situ, to reduce the more reactiveoxygenated compounds appears to be most pragmatic [5–8]. Ofthe various catalysts studied to date, [9–19] zeolites have receivedmuch attention due to their vast availability, relatively low cost, andfacile tunability [20–32]. Zeolite catalysts have been shown to beeffective in the selective deoxygenation of pyrolytic vapors, result-ing in the formation of aromatics and effectively increasing the C/O

� Mention of trade names or commercial products in this publication is solely forthe purpose of providing specific information and does not imply recommendationor endorsement by the U.S. Department of Agriculture. USDA is an equal opportunityprovider and employer.

∗ Corresponding author. Tel.: +1 215 233 6493; fax: +1 215 233 6406.E-mail address: [email protected] (A.A. Boateng).

ratio [33]. Zeolites with higher acidity, or lower Si/Al ratios, havebeen shown to be effective in promoting cracking reactions dur-ing the incipient pyrolysis reactions [34–37]. However, there are asmany catalysts as there are types of biomass available so screen-ing of various catalysts/biomass combinations is useful. Py/GC–MSserves as a useful tool to understanding pathways leading to sta-ble pyrolysis liquids so promising catalysts/biomass candidates canbe down-selected for larger experiments. In this study, a series ofzeolitic catalysts, chosen based on their acidity and framework,and several types of biomass, including their pure components,were screened using Py/GC–MS to determine critical degradationpathways of in situ catalytic pyrolysis.

The objective of the study was to use Py/GC–MS to screen ligno-cellulosic and component biomass samples against different typesof commercially available zeolite catalysts to determine relativeabilities to deoxygenate pyrolytic condensable components andproduce aromatic hydrocarbons.

2. Materials and methods

2.1. Pyrolysis GC-MS

Pyrolysis experiments were conducted using a Frontier LabDouble-Shot micro pyrolyzer PY-2020iD equipped with theFrontier Lab Auto-Shot Sampler AS-1020E coupled to a gas chro-matograph, Shimadzu GC-2010. Pyrolysis products were detectedusing a Shimadzu GCMS-QP2010S mass spectrometer (MS). Themicro pyrolyzer comprised an interface heater (100–400 ◦C) anda pyrolysis furnace (40–800 ◦C). In this setup, approximately 1 mg

0165-2370/$ – see front matter. Published by Elsevier B.V.doi:10.1016/j.jaap.2011.06.001

Author's personal copy

D.J. Mihalcik et al. / Journal of Analytical and Applied Pyrolysis 92 (2011) 224–232 225

of biomass is placed at the bottom of a deactivated stainless steelcup, followed by the addition of approximately 5 mg of catalyst(∼1:5/w:w, biomass to catalyst ratio). Among the limitations ofthis small scale system is that there is no actual determinationof the deactivation of the catalyst, which necessitates the largeexcess of catalyst. The biomass/catalyst sample is not initiallymixed and is placed in the Auto-Shot Sampler (48 sample capac-ity). The sample is gravity fed into the inert atmosphere of thepre-heated pyrolysis furnace (550 ◦C), where the sample is sub-jected to pyrolysis conditions for 18 s. The helium carrier gas isalso used to purge air in the sample prior to pyrolysis and to con-vey the pyrolysis gas through the pyrolysis reactor, a quartz tube,then to the GC–MS. The fraction of the pyrolysis vapors that arenot detected by GC can be analyzed by LC–MS, NMR, etc., whenproduced on larger scale systems where all condensation productsare collected. To analyze the high molecular weight compoundseluted in the pyrolysis vapor, i.e., pyrolysis liquid products thatcould be collected by condensation, GC analyses were performedon a DB-1701 60 m × 0.25 mm, 0.25 �m film thickness. The ovenwas programmed to hold at 45 ◦C for 4 min and ramped at 3 ◦C/minto 280 ◦C after which it was held at this temperature for 20 min.The injector temperature was kept at 250 ◦C with the injector splitratio set to 30:1 and helium flow rate was maintained at 1 mL/min.A separate experiment was conducted using the same procedureas above for analysis of non-condensable pyrolysis gases but usinga 100:1 split ratio. In this case, GC analysis was performed usinga fused silica capillary column, CP-PoraBOND Q, 25 m × 0.25 mm(Varian, Palo Alto, CA) with the following program: 3 min at 35 ◦Cthen ramped at 5 ◦C/min up to 150 ◦C followed by 10 ◦C/min upto 250 ◦C and held for 45 min for a total run time of 81 min. Thepermanent gas yields were quantified with calibration curves pro-duced using a standard gas mixture comprising CO, CO2, CH4, C2H4,C2H6, C3H8, and C4H10 in helium (custom-mixed by Scott Spe-cialty Gases, Plumsteadville, PA). Hydrogen was not included inthe NCG analysis because it was not able to be detected usingthe GC–MS setup, and thus, the total product distribution of theNCG’s is not presented. Each run was performed in triplicate andrelevant compounds were quantitatively analyzed. To quantify thesolid residue remaining after pyrolysis we carried out manual inser-tions of stainless steel cups containing biomass/catalyst mixture ina similar 1:5/w:w, ratio, into the pyrolyzer under a flow of heliumand subjected it to temperatures of 550 ◦C for 18 s and subse-quent weight loss analysis was performed on an analytical scale.The yield wt%’s for a total of 15 condensable pyrolysis products,7 non-condensable gases, and residual solids, were quantified foreach feedstock/catalyst combination. The wt% of the majority of thecondensable pyrolysis products was then calculated by differenceto allow for total % composition to be determined for comparisonpurposes.

2.2. Biomass/catalyst combinations

The biomass samples included woody biomass, energy crops,and agricultural residues. Additionally, primary biomass compo-nents, i.e., cellulose (Sigma–Aldrich), hemicellulose (xylan frombeechwood, Sigma–Aldrich), and lignin (“Asian”, from GranitResearch and Development SA) and a 50/50 mixture of celluloseand lignin (ETEK lignin) were screened (Table 1).

Catalyst used included the activated forms of several commer-cially available zeolites including: H-Mordenite, H-ZSM-5, H-Y,H-Beta, and H-Ferrierite, in a five factorial experimental design.The selected catalysts were obtained from Zeolyst International andencompassed different frameworks (Table 2) and varying aciditieswith Si/Al ratio ranging from 5.1 to 360. All catalysts were activatedto their protonated form in a muffle furnace at 400 ◦C for 5 h.

O

OH OH

HOO

O

OH

OHO

O OO

H

OH

OH

OCH3

OCH3OH

OCH3

OH

1 2 3 4 5

6 7 8 9 10

11 12 13 14 15

Acetic Acid Furfural Furfural Alcohol Acetol Levoglucosan

Guaiacol Syringol Phenol Benzene Toluene

Ethyl Benzene p-Xylene o-Xylene Naphthalene 1-MethylNaphthalene

Fig. 1. Typical oxygenated compounds associated with fast pyrolysis of biomass(1–8). Products of catalytic fast pyrolysis using zeolites are typically aromatic hydro-carbons (9–15).

A series of 15 condensable pyrolysis products were groupedunder oxygenated and aromatic compounds and were evaluatedby analysis of variance to determine the de-oxygenation effectsof catalyst and feedstock interactions, as well as catalyst abil-ity to produce aromatic hydrocarbons. All the response variableswere considered with all “0” values replaced by a random numberbetween 0 and 0.03 (0.03 is limit of detection) where the effectsand interactions deemed to be significant (p < 0.05) were furtherexamined by performing mean separations using the BonferroniLSD technique. Use of 0.03 is an attempt to keep the large num-ber of “0” responses from biasing the estimate of variation. Eachresponse variable has the mean separations presented in two formsof data comparison. The first type allows for the direct compari-son of yield observed for a specific pyrolytic product from a singlecatalyst across all feedstocks. The second type facilitates the com-parison of the effect of all catalysts on the wt% production of singlepyrolytic product for each individual feedstock.

3. Results

3.1. Non-catalytic pyrolysis

Fast pyrolysis initiates decomposition of the biomass to pro-duce a complex mixture of products. The product mixture canvary depending on the amounts of each component foundin the feedstock [39,40]. The efficiency of the catalysts wasmonitored by restricting the quantitative analysis to fifteen com-pounds commonly associated with the catalytic/non-catalyticpyrolysis of biomass, eight oxygenated compounds and sevenaromatic hydrocarbons (Fig. 1). Also, several non-condensablegases (NCGs), with the exception of hydrogen, and condens-able components of the pyrolysis reactions were measured usingPy/GC–MS.

Table S1 presents the statistical analysis and yields of all feed-stocks under non-catalytic pyrolysis conditions. They are consistentwith previously reported data and will not be discussed in depthhere, rather, Table S1 is provided as a baseline for comparison withcatalytic pyrolysis screening. Consistent with previously reportedresults, [41] ∼61.50 wt% levoglucosan was produced following non-catalytic pyrolysis of cellulose. Where the feedstock comprised anequal mixture of cellulose and lignin in 50:50 ratio (ETEK lignin) thelevoglucosan yield reduced to 12.61 wt% similar to that previouslyreported by Mullen and Boateng [42]. The non-catalytic pyroly-

Author's personal copy

226 D.J. Mihalcik et al. / Journal of Analytical and Applied Pyrolysis 92 (2011) 224–232

Table 1Composition assumed for selected lignocellulosic feedstock materials [38] and approximate amounts of cellulose, hemi-cellulose, and lignin associated with tested feedstocksand feedstock components.

Feedstock Cellulose fraction Hemi-cellulose fraction Lignin fraction

Energy crops 36.6 16.1 21.9Crop residues 38.0 32.0 17.0Woody biomass 43.7 28.3 24.3

Cellulose 100 0 0Hemi-cellulose (xylan) 0 100 0Pure lignin (Asian) 0 0 100ETEK lignin 50 0 50Oak 47 20 27Corn cob 30 38 3Corn stover 48 29 6Switchgrass 32 28 16

sis products of hemicellulose included increased levels of acetolat 7.06 wt%, the highest amount observed for all feedstocks tested.The non-catalytic pyrolysis of pure lignin, Asian lignin in this case,yielded relatively large amounts of guaiacol (0.49 wt%) and syringol(1.01 wt%). Additionally, all feedstock components produced traceamounts of aromatic hydrocarbons.

Non-catalytic pyrolysis of lignocellulosic biomass samples pro-duced about 6–8 wt% of acetic acid which is significantly more thanis produced by biomass components, 1.8–3.1 wt%. Corn cob hadincreased levels of phenol (0.25 wt%), similar to levels producedby pure lignin (0.20 wt%). Consistent with biomass components,low values were observed for all aromatic hydrocarbons followingnon-catalytic pyrolysis of lignocellulosic biomass.

3.2. Catalytic pyrolysis over zeolites

3.2.1. H-Ferrierite and H-MordeniteH-Ferrierite and H-Mordenite zeolites have similar acidities

(Si/Al ratio = 20), surface areas and pore sizes with each having anorthorhombic structure and respective 8–10, and 12-memberedrings. The products of pyrolysis of the cellulosic biomass andbiomass primary compounds over these zeolite catalysts are pre-sented in Tables S2 and S3 to discern the effect of catalyst for eachfeedstock. For both component and lignocellulosic biomass, the useof H-Ferrierite (20) gave similar yields for most oxygenates butthere were statistical differences in the yields of furfural, guaiacoland syringol. Oxygenate yields varied less when H-Mordenite (20)was employed, but statistical differences were more prevalent inthe case of furfural, guaiacol, syringol, and phenol. Very little vari-ation was observed for either catalyst in terms of the quantifiedaromatic hydrocarbon products among both primary compoundsand the biomass used as feedstocks, except to say that they weredifferent in comparison with non-catalytic pyrolysis.

A rather significant reduction in acetic acid was observed fol-lowing the pyrolysis of lignocellulosic biomass over these catalysts,compared to the component feedstocks as is shown in Table S4.Compared to non-catalytic pyrolysis, increase in the production ofaromatic hydrocarbons was detected with H-Mordenite (20), moreso than with H-Ferrierite (20), across all feedstocks. However, theincreased detection of aromatic hydrocarbons in either case waslimited to amounts under 2 wt%.

3.2.2. H-YH-Y zeolite is described as having a cubic structure with a

pore system composed of 12-membered rings. Of all the zeolitecatalysts tested, H-Y has the highest acidity (Si/Al = 5.1), largestsurface area (925 m2/g), and largest average pore size (0.74 nm).The quantitative and statistical analyses of products formed overH-Y (5.1) zeolite are presented in Table S5. Upon pyrolysis of allfeedstocks over the catalyst, furfural alcohol and levoglucosan wereproduced in similar amounts and levels of phenol were similar forlignocellulosic biomass. Otherwise, few statistically relevant trendswere observed in terms of oxygenate production. Statistically, nosignificant differences in aromatic hydrocarbon compounds wereencountered among all feedstocks, primary or lignocellulosic.

Compared with non-catalytic pyrolysis, the H-Y (5.1) catalystinfluenced pyrolysis of biomass and its primary products. It effec-tively reduced the production of acetic acid from cellulose, 1.89to 0.015 wt%, and decreased acetic acid formation from all ligno-cellulosic feedstocks. The presence of acetic acid was significantlyreduced for corn cob from around 8.19 to 3.23 wt%. Likewise, ace-tol levels decreased from 5.83 wt% non-catalytically for corn cobto 1.33 wt% with H-Y (5.1). The catalyst was also effective at thereduction of levoglucosan. Again, the increased observation of sub-stituted furans, suggests the switch from a depolymerization to afragmentation mechanism in the presence of H-Y (5.1) catalyst.There was not a substantial increase in the amounts of aromatichydrocarbons produced, roughly < 2 wt% increase, a value similarto that of the H-Ferrierite (20) catalyst.

3.2.3. H-ZSM-5The H-ZSM-5 zeolite consists of a MFI orthorhombic crystal

structure and is characterized as having a 3D network of intercon-nected pores arranged in straight 10-membered ring pore system5.2 × 5.7 A channels connected by sinusoidal 5.3 × 5.6 A channels.A Si/Al ratio of 23, combined with a potentially ideal pore size andstructure, leads to an increased production of aromatic hydrocar-bons from each feedstock and feedstock components tested. Thequantitative and statistical analyses of products formed over H-ZSM-5 (23) zeolite are presented in Table 3. Compared to previouscatalysts described, H-ZSM-5 (23) reduced nearly all oxygenatedcompounds to statistically similar levels regardless of feedstock.Conversely, there were larger statistical differences for the produc-

Table 2Characteristics of zeolites tested.

Catalyst (product #) SiO2/Al2O3 ratio Pore dimensions (A) Average pore size (nm) Description

Ferrierite (cp914c) 20 4.2 × 5.4, 3.5 × 4.8 0.39–0.51 8,10-Membered rings (FER) OrthorhombicMordenite (cbv21a) 20 6.5 × 7.0, 3.4 × 4.8, 2.6 × 5.7 0.42–0.58 12-Membered rings (MOR) OrthorhombicY (cbv300) 5.1 7.4 × 7.4 0.74 12-Membered rings (FAU) CubicZSM-5 (cbv2314, cbv5524, cbv28014) 23, 50, 280 5.1 × 5.5, 5.3 × 5.6 0.52–0.55 10-Membered rings (MFI) OrthorhombicBeta (cp814e, cp814c, cp811c) 25, 38, 360 6.6 × 6.7 0.61–0.62 10, 12-Membered rings (BEA) Tetragonal

Author's personal copy

D.J. Mihalcik et al. / Journal of Analytical and Applied Pyrolysis 92 (2011) 224–232 227

Table 3Quantitative and statistical analysis of the wt% of pyrolytic products (1–15) observed following catalytic pyrolysis of component and lignocellulosic feedstocks over H-ZSM-5(23); relative values of NCG’s, solid residue, and total calculated condensable vapors using H-ZSM-5 series.a

Cellulose Hemi-cellulose Lignin ETEK Lignin Oak Corn cob Corn stover Switchgrass

Acetic acid 2.25A 2.02A 0.0091B 1.72A 2.31A 3.07A 1.72A 2.17A

Furfural 0.76A 0.0079B 0.022B 0.011B 0.016B 0.013B 0.014B 0.020B

Furfural Alcohol 0.019A 0.013A 0.018A 0.010A 0.023A 0.021A 0.017A 0.010A

Acetol 0.12A 0.12A 0.010A 0.12A 0.93A 0.31A 0.013A 0.020A

Levoglucosan 0.31A 0.022A 0.019A 0.019A 0.48A 0.0094A 0.025A 0.019A

Guaiacol 0.019B 0.010B 0.12AB 0.15A 0.099AB 0.10AB 0.084AB 0.093AB

Syringol 0.013C 0.023C 0.22A 0.018C 0.12B 0.048C 0.037C 0.016C

Phenol 0.14ABC 0.10C 0.15ABC 0.22A 0.16ABC 0.19AB 0.17ABC 0.11BC

Benzene 3.06A 1.48C 1.39C 2.96A 2.93A 3.12A 3.17A 1.86B

Toluene 3.79A 2.99BC 2.68CD 2.24D 2.28D 2.23D 2.17D 3.36AB

Ethyl benzene 1.77B 0.26CD 0.21D 0.30CD 0.37CD 0.41C 0.37CD 2.10A

p-xylene 2.15E 2.59D 2.46DE 3.17C 4.12A 3.49BC 3.65B 2.56D

o-xylene 0.54CD 0.59BC 0.40D 0.63ABC 0.74AB 0.77A 0.77A 0.55CD

Naphthalene 1.44A 0.50C 0.75B 1.33A 1.37A 1.33A 1.33A 0.89B

Methyl Naphthalene 0.85C 0.60D 0.70CD 1.19AB 1.28A 1.16AB 1.12B 0.81C

H-ZSM-5 (23)NCG’s 29.4 20.4 15.5 26.9 26.5 28.1 40.4 28.9Solid residue 8.8 12.2 19.8 11.9 13.9 23.4 28.2 19.4Condensables 61.8 67.4 64.7 61.2 59.6 48.5 31.4 51.7

H-ZSM-5 (50)NCG’s 32.3 24.7 20.9 23.9 34.9 33.3 31.7 31.1Solid residue 3.6 18.7 11.4 29.0 20.5 15.3 27.7 15.9Condensables 64.1 56.6 67.7 47.1 44.6 51.4 40.6 53.0

H-ZSM-5 (280)NCG’s 16.6 17.1 14.7 21.9 21.3 25.3 26.3 28.1Solid residue 4.3 20.9 33.2 10.7 8.9 14.0 22.8 22.5Condensables 79.1 62.0 52.1 67.4 69.8 60.7 50.9 49.4

a Any two means in the same row with no letter in common are significantly different (p < 0.05) by the Bonferroni least significant difference (LSD) technique. The lettersA–F refer to the highest estimates to the least.

tion of aromatic hydrocarbon compounds, more so from feedstockcomponents than for lignocellulosic biomass samples.

As mentioned above, H-ZSM-5 (23) reduced the production ofoxygenates to similar levels, regardless of feedstock. Like othercatalysts, H-ZSM-5 (23) drastically reduced the amount of levoglu-cosan found in the pyrolysis products of cellulose to levels below0.31 wt% compared to non-catalytic results (Table S4). Although,substituted furans were not observed as with previous catalystsmost likely due to the increased cracking ability of H-ZSM-5 (23).Upon pyrolysis of hemicellulose over H-ZSM-5 (23) levels of ace-tol dropped from 7.06 wt% for non-catalytic pyrolysis, to 0.12 wt%.Pyrolysis of lignin over H-ZSM-5 (23) resulted in reduction of aceticacid to trace amounts while syringol was reduced from 1.01 to0.22 wt%.

H-ZSM-5 (23) increased the production of aromatic hydrocar-bons in every case when compared to non-catalytic pyrolysis andwas the most successful of any catalyst tested (Table S4). Thepresence of H-ZSM-5 (23) also increased the likelihood of the for-mation of substituted benzenes and naphthalenes, especially whenpyrolyzing cellulose. Furthermore, when pyrolyzing componentfeedstocks over H-ZSM-5 (23), toluene was produced in the great-est abundance and is highest for cellulose at 3.79 wt%. For H-ZSM-5(23) over lignocellulosic biomass, with the exception of switch-grass, p-xylene was the most abundant hydrocarbon produced, andwas highest for oak (4.12 wt%).

Because H-ZSM-5 (23) was effective at reducing the oxygen con-tent of pyrolytic vapors and increasing the production of aromatichydrocarbons, a series of H-ZSM-5 catalysts with Si/Al ratios of 50and 280, were compared. As Fig. 2 shows, with the exception ofcorn stover, the abundance of oxygenates was lowest when usingthe H-ZSM-5 (50) catalyst, while the presence of oxygenates washighest with H-ZSM-5 (280). On the other hand, total aromatics for-mation from H-ZSM-5 (23) was more than three times the amount

observed for the less acidic catalysts (Fig. 3) with p-xylene beingproduced in the largest amounts in nearly every case. This trendindicates that the relative acidities of the zeolites play a larger rolein the formation of aromatics than in the processing of oxygenates.

3.2.4. H-BetaH-Beta zeolite has an acidity number of (25) and is composed

of a tetragonal crystal structure with straight 12-membered ringchannels (7.6 × 6.4 A) with crossed 10-membered ring channels(5.5 × 6.5 A). The quantitative and statistical analyses of productsformed over H-Beta (25) zeolite are presented in Table 4. Statis-tically, H-Beta (25) was equally effective at reducing oxygenatedcompounds from pyrolysis of component and lignocellulosic feed-stocks. Statistically, larger differences were observed for theproduction of aromatic hydrocarbons from biomass componentswhen compared to lignocellulosic biomass samples. Yields of aro-matic hydrocarbons were significantly larger than all catalystsexcept for H-ZSM-5 (23).

0123456789

10

wt %

Fee

dsto

ck

H-ZSM-5 (23) H-ZSM-5 (50) H-ZSM-5 (280)

Fig. 2. Comparison of the wt% of oxygenated compounds 1–8 produced by H-ZSM-5zeolites.

Author's personal copy

228 D.J. Mihalcik et al. / Journal of Analytical and Applied Pyrolysis 92 (2011) 224–232

Table 4Quantitative and statistical analysis of the wt% of pyrolytic products (1–15) observed following catalytic pyrolysis of component and lignocellulosic feedstocks over H-Beta(25); relative values of NCG’s, solid residue, and total calculated condensable vapors using H-Beta series.a

Cellulose Hemi-Cellulose Lignin ETEK lignin Oak Corn cob Corn stover Switchgrass

Acetic acid 2.01A 1.92A 0.017B 1.75A 0.0082B 1.76A 1.90A 2.23A

Furfural 0.68A 0.018B 0.018B 0.016B 0.025B 0.016B 0.013B 0.23B

Furfural Alcohol 0.020A 0.0080A 0.021A 0.018A 0.017A 0.021A 0.012A 0.016A

Acetol 0.35A 0.58A 0.0046A 0.14A 0.0088A 0.12A 0.016A 0.11A

Levoglucosan 0.013A 0.012A 0.016A 0.018A 0.14A 0.0093A 0.018A 0.021A

Guaiacol 0.018A 0.021A 0.13A 0.11A 0.019A 0.080A 0.077A 0.073A

Syringol 0.019A 0.0087A 0.063A 0.023A 0.014A 0.010A 0.016A 0.016A

Phenol 0.10A 0.083A 0.16A 0.12A 0.087A 0.11A 0.11A 0.10A

Benzene 1.12BCD 0.84D 1.33ABC 1.48AB 1.44AB 1.51A 1.43AB 0.96CD

Toluene 1.47BC 1.27C 1.83AB 1.95AB 2.05A 2.17A 2.21A 1.44BC

Ethyl benzene 0.62A 0.62A 0.17B 0.20B 0.20B 0.25B 0.26B 0.66A

p-xylene 0.73C 0.73C 0.96ABC 0.97ABC 1.10A 1.09AB 1.21A 0.78BC

o-xylene 0.32A 0.31A 0.39A 0.38A 0.43A 0.43A 0.47A 0.33A

Naphthalene 0.40CD 0.38D 0.70A 0.53BCD 0.60AB 0.63AB 0.56ABC 0.48BCD

Methyl Naphthalene 0.18C 0.25BC 0.43A 0.30ABC 0.35AB 0.39AB 0.37AB 0.15C

H-Beta (25)NCG’s 30.9 24.8 18.2 26.2 28.4 31.9 30.3 27.8Solid residue 5.4 22.4 42.9 23.1 29.5 20.9 23.1 28.8Condensables 63.7 52.8 38.9 50.7 42.1 47.2 46.6 43.4

H-Beta (38)NCG’s 31.9 23.8 18.6 24.1 30.7 33.3 26.1 27.8Solid residue 11.6 12.4 44.6 24.7 40.1 34.3 32.9 29.0Condensables 56.5 63.8 36.8 51.2 29.2 32.4 41.0 43.2

H-Beta (360)NCG’s 20.4 20.1 18.9 23.9 27.3 27.7 25.5 24.9Solid residue 7.2 12.8 11.6 18.4 18.5 27.4 17.5 26.5Condensables 72.4 67.1 69.5 57.7 54.2 44.9 57.0 48.6

a Any two means in the same row with no letter in common are significantly different (p < 0.05) by the Bonferroni least significant difference (LSD) technique. The lettersA–F refer to the highest estimates to the least.

For all feedstocks, levoglucosan levels were brought below0.026 wt%. The lignocellulosic biomass samples showed a markeddecrease in the levels of acetic acid and furfural compared to non-catalytic pyrolysis. As is the case with H-ZSM-5 (23), there wasan increased presence of substituted benzenes and naphthalenes,while aromatic hydrocarbon production increased in all instanceswith biomass components but the amount produced were less thanthose achieved with H-ZSM-5 (23). Toluene was the most abundanthydrocarbon produced for all lignocellulosic feedstocks as well aslignin and ETEK lignin (Table S4).

In Fig. 4, the reduction in oxygenates was investigated usingan H-Beta zeolite series with Si/Al = 25, 38, 360. The acidity differ-ence has little effect on ability to reduce oxygenates with regard tocellulose and hemicellulose feedstocks. However, a more dramaticeffect between H-Beta catalysts was observed for lignin, oak, corncob, and corn stover. As shown in Fig. 5, H-Beta (25) was effective atproducing aromatics at levels of 4.5–6.5 wt% for all feedstocks. Useof H-Beta (38) decreased aromatics production to levels ranging

02468

1012141618

wt

% F

eeds

tock

H-ZSM-5 (23) H-ZSM-5 (50) H-ZSM-5 (280)

Fig. 3. Comparison of the wt% of aromatic hydrocarbons 9–15 produced by H-ZSM-5zeolites.

0

0.5

1

1.5

2

2.5

3

3.5

4

wt %

Fee

dsto

ck

H-Beta (25) H-Beta (38) H-Beta (360)

Fig. 4. Comparison of the wt% of oxygenated compounds 1–8 produced by H-Betazeolites.

from 1.5 to 2.5 wt%. Lastly, the use of H-Beta (360) was limited inits ability to produce aromatics where levels were around 1.0 wt%.

0

1

2

3

4

5

6

7

8

wt %

Fee

dsto

ck

H-Beta (25) H-Beta (38) H-Beta (360)

Fig. 5. Comparison of the wt% of aromatic hydrocarbons 9–15 produced by H-Betazeolites.

Author's personal copy

D.J. Mihalcik et al. / Journal of Analytical and Applied Pyrolysis 92 (2011) 224–232 229

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

No Catalyst H-Ferrierite (20)

H-Mordenite (20)

H-ZSM-5 (23) H-Beta (25)

% C

ompo

si�

on

NCGs

Aroma�cs 9-15

Other Condensibles

Solids

Fig. 6. Average total compositional wt% of pyrolytic condensables, NCG’s, and solidresidue over lignocellulosic feedstocks for each catalyst.

3.3. Total compositional analysis of catalytic pyrolysis products

3.3.1. Comparison of zeolitesFig. 6 depicts the average compositional analyses for lignocellu-

losic biomass feedstocks. Non-catalytic pyrolysis produces about70 wt% condensables on average for all feedstocks tested. Thisvalue is less for each of the catalysts and smallest for H-Beta(25) (∼48 wt%). Gas production associated with catalytic and non-catalytic pyrolysis was characterized by CO, CH4, CO2, C2H4, C2H6,C3H6, C3H8, and C4H10. A distinct increase in non-condensable gasproduction is observed for all feedstocks over all catalysts whencompared to non-catalytic pyrolysis which averaged 14 wt%. NCGranged from a low of 19.9 wt% for H-Mordenite (20) to a high of30.9 wt% for H-ZSM-5 (23). These trends are also consistent withcatalytic pyrolysis over zeolite where bio-oil yields are consider-ably reduced and the conversion of the oxygenated species in theoils was largely due to H2O loss at lower catalyst temperatures andCO and CO2 formation at higher catalyst temperatures [5].

Correspondingly, use of H-Beta yielded the larger amount ofcoke (25.5 wt%), which most likely contributed to premature deac-tivation of the catalyst. H-ZSM-5 (23) produced the least residueof all catalysts, about 21.2 wt%, while non-catalytic reactions pro-duced 15.8 wt% of feedstock on average. This indicates that evenwith the large amount of aromatic hydrocarbons produced byH-ZSM-5 (23) (12.6 wt%), instantaneous coking was kept to a min-imum. H-Beta (25) causes a large amount of coking of the pyrolysisvapors. On the other hand, H-Ferrierite (20) and H-Mordenite (20)lacked reactivity towards the pyrolytic vapors presumably due toframework characteristics associated with the pore system of thecatalysts.

3.3.2. Comparison of H-ZSM-5 seriesFig. 7 compares the average compositional analysis of H-ZSM-5

catalysts and their effects on the pyrolytic vapors of lignocellu-losic feedstocks. The total amount of condensables is estimated aseverything not accounted for by the solid and NCGs, more thanlikely, leading to a slight overestimation of the total liquid yield. Thetotal amount of condensable pyrolytic products decreases for eachcatalyst when compared to non-catalytic reactions, a drawbackreported for catalytic pyrolysis of biomass [5,8]. NCG productionis increased for each catalyst run compared to non-catalytic butthere does not appear to be a straightforward trend among the H-ZSM-5 catalyst series. The overall NCG’s production was slightlyhigher for H-ZSM-5 (50) at 32.7 wt% than for H-ZSM-5 (23) 30.9 wt%while total NCG yields from H-ZSM-5 (280) were 25.2 wt%. Thistrend indicates the relative amount of aromatics formation cannotbe directly related to formation of gases. Solid formation is fairlyconsistent among the three H-ZSM-5 and all are slightly increasedcompared to non-catalytic. This consistency indicates that for theH-ZSM-5 catalysts, coking is not dependent upon relative acidi-

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

No Catalyst

H-ZSM-5 (280)

H-ZSM-5 (50)

H-ZSM-5 (23)

% C

ompo

si�

on

NCGs

Aroma�cs 9-15

Other Condensables

Solid

Fig. 7. Average total compositional wt% of pyrolytic condensables, NCG’s, and solidresidue over lignocellulosic feedstocks when using H-ZSM-5 catalysts.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

No Catalyst

H-Beta (360)

H-Beta (38)

H-Beta (25)

% C

ompo

si�

on

NCGs

Aroma�cs 9-15

Other Condensables

Solid

Fig. 8. Average total compositional wt% of pyrolytic condensables, NCG’s, and solidresidue over lignocellulosic feedstocks when using H-Beta catalysts.

ties of the catalysts or on the amounts of aromatic hydrocarbonsproduced during the 18 s pyrolysis time period.

3.3.3. Comparison of H-Beta seriesFig. 8 compares the average compositional analysis of the H-Beta

catalyst series compared to non-catalytic runs. The average solidproduction for the series in the order of increasing acidity is 22.4,34.0, and 25.5 wt% of biomass sample. The large amount of solidsformed with H-Beta (38) leads to the least amount of condensablesbeing formed, 36.4 wt%, of all Beta catalysts. The amount of solidresidues recovered increased in all cases and was most significantfor oak where solid amounts increased from 16.2 to 29.5 wt%. Therelatively small increase in solid residue when comparing H-Beta(360) to non-catalytic values may be attributed to the catalysts rel-ative inactivity. NCG formation tends to increase with increasedcatalytic activity, or increasing acidity within the series.

4. Discussion

4.1. Catalyst chemistry

In general, the acid catalyzed formation of aromatic hydrocar-bons begins with donation of a proton to incoming substrates suchas hydrocarbons at the acidic sites of the catalyst. This protonationleads to carbocation formation within the hydrocarbon which thenreacts further by way of �-hydrogen elimination to form olefinsand the sequence continues. The olefins are transformed into aro-matics through oligomerization, cyclization, and hydrogen transferreactions also performed on the catalyst acid sites [43–45].

Zeolite catalyzed fast pyrolysis of biomass carbohydrates includ-ing glucose, cellobiose, cellulose, and xylitol as feedstocks in apyroprobe was described by Huber, et al. [46] The general pathwaythey established was that cellulose undergoes initial dehydration

Author's personal copy

230 D.J. Mihalcik et al. / Journal of Analytical and Applied Pyrolysis 92 (2011) 224–232

reactions to form anhydro-sugars which then interact with theactive sites of the catalyst to form dehydrated products via acid cat-alyzed dehydration. The dehydrated products further undergo acidcatalyzed oligomerization, decarboxylation, and decarbonylation,or cracking reactions, to form C2-C6 olefins which then combineto yield aromatics. The deoxygenation occurs by the productionof CO, CO2, and H2O. Consistent with this pathway, we observeddecreases in the production of oxygenates in general when zeolitecatalysts were applied in the pyrolysis of cellulose and hemicellu-loses. Lignin pyrolysis over the catalysts studied followed pathwaysimilar to that described by Mullen and Boateng where similarcatalysts specifically H-ZSM-5 and the mixed metal oxide catalystCoO/MoO3 were applied [42]. This resulted in the production of aro-matic products which further undergo cracking reactions to olefinsleading to production of deoxygenated aromatics.

Furthermore, upon pyrolysis of biomass components over Mor-denite and Ferrierite catalyst, significant coke depositions werenot observed. For lignocellulosic biomass samples, there was anincrease in solid residue formation in each biomass/catalyst com-bination with the exception of the combination of corn cob andH-Ferrierite. Because corn cob contains very low levels of ligninthis data suggest that lignin concentration is a contributing fac-tor in the increased solid formations. Therefore, interactions oflignin in cellulosic biomass with H-Ferrierite are different thanthose interactions in pure lignin leading to increased coke for-mation in the former. Another potential factor is the presence ofnon-methoxylated lignins (H-lignin) which are more highly con-centrated in herbaceous species (i.e. corn stover and switchgrass)than hardwoods (i.e. oak).

The entire set of catalysts exhibit similar reactivity trendstowards deoxygenation of pyrolytic vapors, regardless of the feed-stock and despite their different frameworks. This trend suggeststhat they are acting on initial pyrolytic degradation products ratherthan the carbohydrates or lignin components. The pyrolytic degra-dation products are then deoxygenated over the zeolite catalyststo form the aromatic hydrocarbons. Adjaye and Bakhshi proposedthat the light oxygenated organics are converted to olefins whichthen aromatize (Scheme 1) [47].

There are three types of pathway mechanisms for the deoxy-genation of the oxygenated organics; dehydration to form water,decarbonylation to form CO, and decarboxylation to form CO2.For products of cellulose pyrolysis, the increase in CO2 observedover the H-ZSM-5 suggests that both dehydration and decarboxyla-tion mechanisms are active. Further, both Mordenite and Ferrieritegreatly reduced the production of levoglucosan from cellulose, witha qualitative, and co-concurrent, increase in production of substi-tuted furans when compared to non-catalytic reactions. This trendindicates the catalytic pyrolysis of cellulose most likely proceedsthrough fragmentation, rather than depolymerization, as is the casewith non-catalytic pyrolysis, but does not indicate that the con-densable fraction of the pyrolysis products was deoxygenated. Thepresence of these catalysts promotes pathways similar to thosecaused by biomass ash, i.e., and the presence of alkali metalsreported to shift degradation mechanisms towards fragmentation[48]. Levoglucosan is deoxygenated in two steps, one dehydra-tion to form furans and similar compounds and then conversion ofthe furans to aromatic hydrocarbons via further dehydration anddecarboxylation (Scheme 2). A possible mechanism for the con-version of furans is acid catalyzed ring opening to form aldehydes(Scheme 3) which can be converted to aromatics by mechanismssimilar to that proposed by Resasco for conversion of propanal[49].

For hemicellulose the major net reaction observed betweennon-catalytic pyrolysis and catalytic pyrolysis over HZSM-5 wasconversion of acetol (a ketone) to aromatics, with an increasedproduction of carbon monoxide. This suggests that the conversion

leading to deoxygenation of ketones proceeds through decar-bonylation as was previously suggested by Adjaye & Bakhshi[50].

4.2. Effects of catalyst acidity and framework

The higher acidity of the catalyst should theoretically increasethe propensity of the catalyst to promote cracking reactions, butin the case of H-Y (5.1) zeolite, leads to the relative instability ofthe catalyst framework. This instability led to catalyst frameworkcollapse and consequently, full compositional analysis of pyrolyticproducts when using H-Y (5.1) was not performed. The low Si/Alratio of this catalyst implies a relatively high acidity, where theAl-O bond is susceptible to hydrolysis reactions which lead to thebreaking of Si-O-Al bonds and removal of the aluminum from thetetrahedral lattice position further leading to lattice defects andultimately framework destabilization and eventual collapse [51].This mechanism could potentially lead to unreliable solid recoveryvalues and as a result, wt% of condensables could not be calcu-lated. This susceptibility to destabilization may also help to explainthe lack of aromatics production despite the relatively favorableacidity.

Regardless of relative acidity, the pore structure of the indi-vidual catalyst plays a very important role. For instance, the H-Y(5.1) catalyst may be described as a cubic structure and is com-posed of straight channels measuring 0.74 nm in diameter definedby a 12-member oxygen ring (0.74 nm); pores in perpendicular x, y,and z planes. The potential problem with this type of pore systemmay be two-fold. Firstly, the size of the channels is relatively largecompared to other zeolites tested. This large size may potentiallyresult in far less contact with incoming pyrolytic substrates andtherefore offers significantly less potential for cracking reactionsto occur. Secondly, the large and straight channels also contributeto the ineffective contact ability but also shape selectivity becomesa non-issue. Essentially, the pyrolytic products may be channelingthrough the catalyst relatively freely.

The pore size and framework of the catalyst tend to affectthe product composition. For instance, the shape selectivity oftendisplayed by H-ZSM-5 catalysts [20] was observed in these experi-ments. The H-ZSM-5 (23) seems to offer shape selectivity in terms offavorably cracking pyrolytic vapors to para rather than ortho-xylenefor all feedstocks tested. o-Xylene has a molecular dimension ofapproximately 0.55 nm which is slightly larger than the averagewidth of H-ZSM-5 pores (0.52–0.55 nm). Whereas p-xylene has awidth closer to that of benzene at 0.51 nm, so it may be moreeffectively produced and transported within the pore system of theH-ZSM-5 catalyst.

Trends indicate that when considering the reduction of oxy-genates, the H-Beta series and especially H-Beta (25) seemto be more feedstock specific when compared to the H-ZSM-5 series. For instance, even with the highest acidity, H-Beta(25) was actually the least effective for deoxygenation of cel-lulose pyrolytic vapors and behaved similarly to the less acidicH-Beta (38) for hemicellulose, pure lignin, and ETEK lignin.Interestingly, for oak, corn cob, and corn stover, a dramaticdifference is observed between the most acidic, and moreeffective H-Beta (25), and its less acidic counterparts. This indi-cates that when using the H-Beta catalyst, the differing Si/Alratios play a larger role than for lignocellulosic biomass feed-stocks than for the component feedstocks. This discrepancybetween feedstocks did not pertain to the H-ZSM-5 catalystseries.

The feedstock specificity observed for the H-Beta series for theproduction of oxygenates was not reciprocated in the production ofaromatics. While production of aromatics from H-ZSM-5 plateaus,production of aromatics using the H-Beta catalyst series is clearly

Author's personal copy

D.J. Mihalcik et al. / Journal of Analytical and Applied Pyrolysis 92 (2011) 224–232 231

Scheme 1. General pathways towards aromatic hydrocarbons from oxygenated pyrolytic substrates.

Scheme 2. Two-step deoxygenation of levoglucosan to form furans and aromatic hydrocarbons via further dehydration and decarboxylation.

Scheme 3. Proposed acid catalyzed conversion of aldehydes to aromatics.

dependent upon the overall acidity of the catalyst. For each compo-nent and lignocellulosic biomass feedstock, H-Beta (25) producesmore aromatics than other less acidic H-Beta zeolite catalyst. Thislinear effect more closely describes the effect of decreasing acidstrength on the production of aromatics and also exemplifies thedifferences observed between types of catalyst.

H-Beta was affected by increased amounts of coke regardlessof its relative acidity. Coke is the carbonaceous solid result-ing from the decomposition or condensation of hydrocarbonsand is typically composed of polymerized heavy hydrocarbons.Because of the increased amount of hydrocarbons being pro-duced within the pores of the zeolites during catalytic pyrolysis,there is an increased likelihood of hydrocarbon decomposi-tion or condensation leading to coke formation within thepores. The formation of coke within the zeolite pore systemleads to fouling due to blockage and eventual catalyst deac-tivation. Zeolite catalyst deactivation is of major concern andpresents reactor engineering challenges necessary for catalystregeneration.

5. Conclusion

Eight lignocellulosic biomass and component feedstocks werescreened non-catalytically and catalytically using nine com-mercially available acidic zeolite catalysts through Py-GC/MSexperiments. Quantitative analysis of eight oxygenates and sevenaromatic hydrocarbons common to non-catalytic and catalyticpyrolysis was presented. Relative compositions of pyrolytic prod-ucts: NCG, solid residues, and condensable vapors, were presentedto show how reaction products are affected by catalyst choice.While all catalysts decreased the oxygen containing species in thepyrolytic vapors, the H-ZSM-5 (23) catalyst was the most effectivecatalyst at producing aromatic hydrocarbons from the oxygen-richvapors. The framework and Si/Al ratio of the catalysts plays a majorrole in their ability to effectively deoxygenate the vapors and pro-duce aromatic hydrocarbons.

Acknowledgement

Author Mihalcik thanks the Agricultural Research Service forpartial funding through the Administrator Funded Research Asso-ciate Program. We would like to thank statistician John G. Phillipsof the North Atlantic Area (NAA) for analyzing data by analysis ofvariance to determine the effects and interactions of catalyst andfeedstock.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.jaap.2011.06.001.

References

[1] G.W. Huber, S. Iborra, A. Corma, Synthesis of transportation fuels frombiomass: chemistry, catalysts, and engineering, Chemical Reviews 106 (2006)4044–4098.

[2] D. Mohan, C.U. Pittman Jr., P.H. Steele, Pyrolysis of wood/biomass for bio-oil: acritical review, Energy and Fuels 20 (2006) 848–889.

[3] M.M. Wright, R.C. Brown, A.A. Boateng, Distributed processing of biomass tobio-oil for subsequent production of Fischer-Tropsch liquids, Biofuels Bioprod-ucts Biorefining 2 (2008) 229–238.

[4] S. Schwietzke, M. Ladisch, L. Russo, K. Kwant, T. Makinen, B. Kavalov, K. Mani-atia, R. Zwart, G. Shananan, K. Sipila, P. Grabowwski, B. Telenius, M. White, A.Brown, Report of the IEA Bioenergy Task 41, Project 2, 2008.

[5] P.T. Williams, N. Nugranad, Comparison of products from the pyrolysis andcatalytic pyrolysis of rice husks, Energy 25 (2000) 493–513.

[6] J. Adam, M. Blazso, E. Meszaros, M. Stocker, M.H. Nilsen, A. Bouzga, J.E. Hustad,M. Gronli, G. Oye, Pyrolysis of biomass in the presence of Al-MCM-41 typecatalysts, Fuel 84 (2005) 1494–1502.

[7] M.A. Jackson, D.L. Compton, A.A. Boateng, Screening heterogeneous catalystsfor the pyrolysis of lignin, Journal of Analytical and Applied Pyrolysis 85 (2009)226–230.

[8] H. Zhang, R. Xaio, H. Huang, G. Xiao, Comparison of non-catalytic and catalyticfast pyrolysis of corncob in a fluidized bed reactor, Bioresource Technology 100(2009) 1428–1434.

[9] A. Pattiya, J.O. Titiloye, A.V. Bridgwater, Fast pyrolysis of cassava rhizome inthe presence of catalysts, Journal of Analytical and Applied Pyrolysis 81 (2008)72–79.

[10] E.F. Iliopoulou, E.V. Antonakou, S.A. Karakoulia, I.A. Vasalos, A.A. Lappas, K.S.Triantafyllidis, Catalytic conversion of biomass pyrolysis products by meso-

Author's personal copy

232 D.J. Mihalcik et al. / Journal of Analytical and Applied Pyrolysis 92 (2011) 224–232

porous materials: effect of steam stability and acidity of Al-MCM-41 catalysts,Chemical Engineering Journal 134 (2007) 51–57.

[11] J. Adam, E. Antonakou, A. Lappas, M. Stocker, M.H. Nilsen, A. Bouzga, J.E. Hustad,G. Oye, In situ catalytic upgrading of biomass derived fast pyrolysis vapors in afixed bed reactor using mesoporous materials, Microporous and MesoporousMaterials 96 (2006) 93–101.

[12] A. Pattiya, J.O. Titiloye, A.V. Bridgwater, Evaluation of catalytic pyrolysis ofcassava rhizome by principle component analysis, Fuel 89 (2010) 244–253.

[13] E. Antonakou, A. Lappas, M.H. Nilsen, A. Bouzga, M. Stocker, Evaluation of var-ious types of Al-MCM-41 materials as catalysts in biomass pyrolysis for theproduction of bio-fuels and chemicals, Fuel 85 (2006) 2202–2212.

[14] D. Fabbri, C. Torri, V. Baravelli, Effect of zeolites and nanopowder metal oxideson the distribution of chiral anhydrosugars evolved from pyrolysis of cellulose:an analytical study, Journal of Analytical and Applied Pyrolysis 80 (2007) 24–29.

[15] M.I. Nokkosmaki, E.T. Kuoppala, E.A. Leppamaki, A.O.I. Krause, Catalytic con-version of biomass pyrolysis vapors with zinc oxide, Journal of Analytical andApplied Pyrolysis 55 (2000) 119–131.

[16] Q. Lu, W.-M. Xiong, W.-Z. Li, Q.-X. Guo, X.-F. Zhu, Catalytic pyrolysis of cellulosewith sulfated metal oxides: a promising method for obtaining high yield of lightfuran compounds, Bioresource Technology 100 (2009) 4871–4876.

[17] S.J. Juutilainen, P.A. Simell, A. Outi, I. Krause, Zirconia: selective oxidation cat-alyst for removal of tar and ammonia from biomass gasification gas, AppliedCatalysis B: Environmental 62 (2006) 86–92.

[18] J. Li, B. Xiao, L. Du, Y. Rong, T.D. Liang, Formation of NiO nanowires on the surfaceof nickel strips, Journal of Fuel Chemistry and Technology 36 (2008) 42–47.

[19] J. Li, R. Yan, B. Xiao, D.T. Liang, D.H. Lee, Preparation of nano-NiO particles andevaluation of their catalytic activity in pyrolyzing biomass components, Energyand Fuels 22 (2008) 16–23.

[20] P.B. Weisz, W.O. Haag, P.G. Rodewald, Catalytic production of high-grade fuel(gasoline) from biomass compounds by shape-selective catalysis, Science 206(1979) 57–58.

[21] M.I. Haniff, L.H. Dao, Deoxygenation of carbohydrates and their isopropylidenederivatives over ZSM-5 zeolite catalysts, Applied Catalysis 39 (1988) 33–47.

[22] M. Olazar, R. Aguado, J. Bilbao, Pyrolysis of sawdust in a conical spouted-bed reactor with a HZSM-5 catalyst, Reactors Kinetics and Catalysis 5 (2000)1025–1033.

[23] P.A. Horne, N. Nugranad, P.T. Williams, Catalytic coprocessing of biomass-derived pyrolysis vapors and methanol, Journal of Analytical and AppliedPyrolysis 34 (1995) 87–108.

[24] R. French, S. Czernik, Catalytic pyrolysis of biomass for biofuels production, FuelProcessing Technology 91 (2010) 25–32.

[25] A.A. Lappas, M.C. Samolada, D.K. Iatridis, S.S. Voutetakis, I.A. Vasaloa, Biomasspyrolysis in circulating fluid bed reactor for the production of fuels and chem-icals, Fuel (2002) 2087–2095.

[26] H.J. Park, J.-I. Dong, J.-K. Jeon, K.-S. Yoo, J.-H. Yim, J.-M. Sohn, Y.-K. Park, Con-version of the pyrolytic vapor of radiate pine over zeolites, Journal of IndustrialChemistry 13 (2007) 182–189.

[27] A. Aho, N. Kumar, K. Eranen, T. Salmi, M. Hupa, D.Y. Murzin, Catalytic pyrolysisof biomass in a fluidized bed reactor: influence of the acidity of H-Beta zeolite,Process Safety and Environmental Protection 85 (2007) 473–480.

[28] A. Aho, N. Kumar, K. Eranen, T. Salmi, M. Hupa, D.Y. Murzin, Catalytic pyrolysisof woody biomass in a fluidized bed reactor: influence of the zeolite structure,Fuel 87 (2008) 2493–2501.

[29] R.T. Carlson, G.A. Tompsett, W.C. Conner, G.W. Huber, Aromatic productionfrom catalytic fast pyrolysis of biomass-derived feedstocks, Topics in Catalysis52 (2009) 241–252.

[30] E.M. Sulman, V.V. Alferov, Y.Y. Kosivtsov, A.I. Sidorov, O.S. Misnikov, A.E.Afanasiev, N. Kumar, D. Kubicka, J. Agullo, T. Salmi, D.Y. Murzin, The devel-opment of the method of low-temperature peat pyrolysis on the basis ofaluminosilicate catalytic system, Chemical Engineering Journal 134 (2007)162–167.

[31] H. Zhang, R. Xiao, D. Wang, Z. Zhong, M. Song, Q. Pan, G. He, Catalytic fastpyrolysis of biomass in a fluidized bed with fresh and spent fluidized catalyticcracking (FCC) catalysts, Energy Fuels 23 (2009) 6199–6206.

[32] H. Li, P. Yu, B. Shen, Biofuel potential production from cottonseed oil: a com-parison of non-catalytic and catalytic pyrolysis on fixed-fluidizing bed reactor,Fuel Processing Technology 90 (2009) 1087–1092.

[33] T.R. Carlson, T.P. Vispute, G.W. Huber, Green gasoline by catalytic fastpyrolysis of solid biomass derived compounds, ChemSusChem 1 (2008)397–400.

[34] A. Atutxa, R. Aguado, A.G. Gayubo, M. Olazar, J. Bilbao, Kinetic description ofthe catalytic pyrolysis of biomass in a conical spouted bed reactor, Energy Fuel19 (2005) 765–774.

[35] A. Aho, N. Kumar, K. Eranen, T. Salmi, M. Hupa, D.Y. Murzin, Fast pyrolysisof biomass for bio-oil with ionic liquid and microwave radiation, I Chemi-cal Engineering Part B Process Safety and Environmental Protection 85 (2007)473–480.

[36] P.A. Horne, P.T. Williams, Upgrading of biomass-derived pyrolytic vapors overzeolite ZSM-5 catalyst: effect of catalyst dilution on product yields, Fuel 75(1996) 1043–1050.

[37] A.A. Lappas, M.C. Samolada, D.K. Latridis, S.S. Voutetakis, I.A. Vasalos, Biomasspyrolysis in a circulating fluid bed reactor for the production of fuels and chem-icals, Fuel 81 (2002) 2087–2095.

[38] M. Kaylen, D.L. Van Dyne, Y.S. Choi, M. Blasé, Economic feasibility of produc-ing ethanol from lignocellulosic feedstocks, Bioresource Technology 72 (2000)19–32.

[39] A. Dermibas, Mechanisms of liquefaction and pyrolysis reactions of biomass,Energy Conversion and Management 41 (2000) 633–646.

[40] R.J. Evans, T.A. Milne, Molecular characterization of the pyrolysis of biomass,Energy and Fuels 1 (1987) 123–137.

[41] F. Schafizadeh, Introduction to pyrolysis of biomass, Journal of Analytical andApplied Pyrolysis 3 (1982) 283–305.

[42] C.A. Mullen, A.A. Boateng, Characterization of water insoluble solids isolatedfrom various biomass fast pyrolysis oils, Fuel Processing Technology 91 (2010)1446–1458.

[43] D.B. Lukyanov, N.S. Gnep, M. Guisnet, Kinetic modeling of ethane and propenearomatization over HZSM-5 and GaHZSM-5, Industrial and Engineering Chem-istry Research 33 (1994) 223–234.

[44] Y. Song, X. Zhu, S. Xie, Q. Wang, L. Xu, The effect of acidity on olefin aroma-tization over potassium modified ZSM-5 catalysts, Catalysis Letters 97 (2004)31–36.

[45] V.R. Choudhary, D. Panjala, S. Banerjee, Aromatization of propene and n-butene over H-galloaluminosilicate (ZSM-5 type) zeolite, Applied Catalysis A231 (2002) 243–251.

[46] T.R. Carlson, G.A. Tompsett, W.C. Conner, G.W. Huber, Aromatic productionfrom catalytic fast pyrolysis of biomass-derived feedstocks, Topics in Catalysis52 (2009) 241–252.

[47] J.D. Adjaye, N.N. Bakhshi, Production of hydrocarbons by catalytic upgrading ofa fast pyrolysis bio-oil.Part II: comparative catalyst performance and reactionpathways, Fuel Processing Technology 45 (1995) 185–202.

[48] R. Fahmi, A.V. Bridgwater, L.I. Darvell, J.M. Jones, N. Yates, S. Thain, I.S. Donnison,The effect of alkali metals on combustion and pyrolysis of Lolium and Festucagrasses, switchgrass, and willow, Fuel 86 (2007) 1560–1569.

[49] T.Q Hoang, X. Zhu, T. Sooknoi, D.E. Resasco, R.G. Mallinson, A comparison ofthe reactivities of propanal and propylene on HZSM-5, Journal of Catalysis 271(2010) 201–208.

[50] J.D. Adjaye, N.N. Bakhshi, Catalytic conversion of a biomass-derived oil to fuelsand chemicals I: model compound studies and reaction pathways, Biomass andBioenergy 8 (1995) 131–149.

[51] L. Moscou, R. Mone, Structure and catalytic properties of thermally andhydrothermally treated zeolites: acid strength distribution of REX and REY,Journal of Catalysis 30 (1973) 417–422.