Catalytic upgrading of biomass pyrolysis vapours using

faujasite zeolite catalysts

T.S. Nguyen
a
, M. Zabeti
a
, L. Lefferts
a
, G. Brem
b
, K. Seshan
a,
*
a
Catalytic Processes and Materials, Faculty of Science and Technology, University of Twente, PO Box 217, 7500 AE Enschede,
The Netherlands
b
Laboratory of Thermal Engineering, Faculty of Engineering Technology, University of Twente, PO Box 217, 7500 AE Enschede,
The Netherlands
a r t i c l e i n f o
Article history:
Received 25 April 2012
Received in revised form
16 October 2012
Accepted 31 October 2012
Available online 23 December 2012
Keywords:
Bio-oil
Catalytic
Faujasite
Fuel
Pinewood
Zeolite
a b s t r a c t
Bio-oil produced via fast pyrolysis of biomass has the potential to be processed in a FCC
(uid catalytic cracking) unit to generate liquid fuel. However, this oil requires a signicant
upgrade to become an acceptable feedstock for renery plants due to its high oxygen
content. One promising route to improve the quality of bio-oil is to pyrolyse the parent
biomass in the presence of a catalyst. This work investigates the inuence of faujasite
catalysts on the pyrolysis of pinewood. Pyrolysis process with Na-faujasite, Na
0.2
H
0.8
-fau-
jasite, and H-faujasite (Na-FAU, Na
0.2
H
0.8
-FAU, and H-FAU) were carried out in a xed-bed
reactor at 500

C. It is shown that, in the same condition, catalytic upgrading of pyrolysis
vapour is superior to in-situ catalytic pyrolysis of biomass when it comes to quality of bio-
oil. The yields of coke, gas and water increase while that of organic phase decreases
proportional with the concentration of protons in catalysts. Compared to the other two
catalysts, Na
0.2
H
0.8
-FAU removes the most oxygen from bio-oil, reduces amount of acids
and aldehydes/ketones which result in a higher energy-contained and more stable oil with
less corrosive property. However, the biggest contribution to the oxygen removal is via the
formation of reaction water, which is not an optimum path. This leaves space for future
development.
2012 Elsevier Ltd. All rights reserved.
1. Introduction
Declining fossil resources, combined with increased demand
for petroleum by emerging economies, as well as political and
environmental concerns about fossil fuels are driving our
society to search for new sources of fuels. Biomass is a prom-
ising sustainable source of green fuels [1]. One option for the
production of green fuels is to use biomass-derived feed-
stocks/fuel precursors in existing petroleum renery plants.
Petroleum reneries are already built and use of this existing
infrastructure for the production of biofuels therefore requires
little capital investment [2]. These feedstocks/fuel precursors
are generated using a number of techniques involving gasi-
cation and liquefaction [3]. Among these techniques, fast
pyrolysis which involves short vapour residence times (less
than 2 s), rapid heating rates (up to 10
4
C s
1
) and moderate
temperatures (<550

C) has drawn much interest since it has
been shown to favour the formation of easy-to-handle liquid
product (bio-oil) with a higher energy density (19 MJ kg
1
) than
the original biomass (15 MJ kg
1
) [4e7]. The basic conceptual
* Corresponding author. Tel.: 31 53 4893254; fax: 31 53 4894683.
E-mail addresses: t.s.nguyen@utwente.nl (T.S. Nguyen), m.zabeti@utwente.nl (M. Zabeti), l.lefferts@utwente.nl (L. Lefferts), g.brem@
utwente.nl (G. Brem), k.seshan@utwente.nl, i.v.babich@tnw.utwente.nl (K. Seshan).
Available online at www.sciencedirect.com
ht t p: / / www. el sevi er. com/ l ocat e/ bi ombi oe
b i oma s s a nd b i oe ne r g y 4 8 ( 2 0 1 3 ) 1 0 0 e1 1 0
0961-9534/$ e see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.biombioe.2012.10.024
scheme of the biomass pyrolysis coupled with fuels produc-
tion is shown in Fig. 1. Biomass pyrolysis results in three
fractions, bio-oil, gas and char. In the proposed scheme, solid
heat carrier (sand) and catalyst transport char to a regenerator
where it is combusted and catalyst regenerated. Heat from
regenerator is integrated to the endothermic pyrolysis process
similar to FCC scheme in a renery.
The biggest technical challenge for such a green renery is
that the bio-oil consists of variety of oxygenates, e.g., organic
acids, aldehydes, alcohols and ketones. This results in an
acidic, corrosive, viscous and unstable oil with a relatively
low-energy density compared to that of fossil fuels [5].
Chemical methods for improving the properties and applica-
bility of bio-oil mainly include catalytic de-oxygenation [8,9],
hydrogenation/hydrodeoxygenation (HDO) [10e12] and
esterication [13,14]. De-oxygenation is an easier and prom-
ising route [15,16] as it does not require use of other chemicals,
for e.g., alternative HDO process requires hydrogen, which is
expensive and not easily available.
In general, de-oxygenation of biomass results in elimina-
tion of oxygen as carbon oxides, water and, depending on the
extent of oxygen removal, a mixture of oxygen containing
organic, aliphatic/aromatic, acids aldehydes, alcohols, etc.
Complete de-oxygenation (Eq. (1)) results in an aromatic hy-
drocarbon mixture (H/C w 1e1.2) due to the low hydrogen
content of the starting biomass (H/C w1.3) [17].
C
6
H
8
O
4
/4:6CH
1:2
1:4CO
2
1:2H
2
O (1)
In the case of complete de-oxygenation (Eq. (1)), the organic
yield is about 42%(mass fraction), and this corresponds to 50%
energy recovery from the biomass feedstock. Incomplete de-
oxygenation, i.e, retaining some of the oxygen in the organic
fraction, will help enhance the liquid yields, however, this is
an option only if the resulting product has properties that are
compatible with fossil fuel/fuel additives (see Fig. 1, co-pro-
cessing section).
The inuence of H-MFI (ZSM-5) zeolite catalyst on the
pyrolysis of biomass based on rice husks in a xed-bed system
has been investigated [18]. It was found that, in the presence of
ZSM-5, oxygen in the feedstock was removed mostly as H
2
O at
lower temperatures (<500

C) and as CO/CO
2
at higher temper-
atures (>550

C). At the higher temperatures, the bio-oils ob-
tained from catalytic experiments contain signicantly higher
amounts of single ring and polycyclic aromatic hydrocarbons
(PAH) because of deeper de-oxygenation. The evaluation of
different commercial catalyst, including H-MFI, FCC catalysts,
transitional metals (Fe/Cr) and aluminas (g, a) for biomass
pyrolysis has been carried out in a xed bed reactor [19]. It was
concluded fromthis study that H-MFI is a promising catalyst for
the selective production of aromatic hydrocarbons while tran-
sitional metal catalysts lead to the selective production of
phenol and light phenolics from biomass feedstocks. De-
oxygenation of glycerol or sorbitol, as model compounds rep-
resenting bio-oil, has been reported over several catalysts (H-
MFI, g-Al
2
O
3
, H-FAU(USY), SiCandcommercial FCCcatalyst) [20]
and it was concluded that, at high de-oxygenation levels, ach-
ieved with H-FAU (USY) catalyst, aromatic hydrocarbons (fossil
fuel compatible) and coke were the major products.
Alkali and alkali earth metals e.g., Na, K, Ca have attracted
attention as promising catalysts for upgrading of biomass
vapours in recent years [21e24]. This is also because they are
naturally present in the biomass. For instance, Na

found to
be effective for the conversion of biomass by improving the
bio-oil energy content [4,25]. In the case of holo-celluloses,
degradation is suggested to occur via depolymerization to
give compounds such as levoglucosan, or sugars which
subsequently undergo ring opening degradation to give
aliphatic acids, aldehydes, etc. It is suggested that choice of
the type of alkali metal, its content and reaction temperature
can inuence yields of levoglucosan [24]. Levoglucosan can be
further converted to aromatic components via oxygen
removal as H
2
O [26]. Role for alkali metals on the decompo-
sition of lignin is less discussed since in general these metals
are less active to lignin conversion than to holo-cellulose
conversion. The nature of working of alkali cations is not
well established [16], but basicity of alkali and alkaline earth
metals have been suggested to catalyse de-oxygenation,
especially de-carboxylation [27,28].
In this work, inuence of catalysts on the conversion of
lignocellulosic biomass to various organic components is
investigated with a view to establishing a catalytic pyrolysis
process for a green renery. Faujasite catalyst, which is the
active ingredient in uid catalytic cracking (FCC) catalyst, was
employed because of its price, availability, mechanical/
thermal stability, its ability to crack CeC, CeO bonds and
regenerability from coke. H-FAU was further modied by Na

Fig. 1 e Conceptual scheme for green fuels fromlignocellulosic biomass. (For interpretation of the references to colour in this
gure legend, the reader is referred to the web version of this article.)
b i o ma s s a nd b i o e ne r gy 4 8 ( 2 0 1 3 ) 1 0 0 e1 1 0 101
ion exchange to probe role for alkali in pyrolytic de-
oxygenation. Advantages of using catalyst over the thermal
pyrolysis (in absence of the catalysts) is discussed.
2. Experimental
2.1. Biomass
Woodchips fromCanadianwhite pine (Pinus strobus) were used
as the feedstock in this study (ThoroughBed, purchased from
Long Beach Shavings Co.). Prior to all experiments, the chips
were ground by milling and sieved to particle sizes of
0.3e0.6mm. Characteristicsof thepinewoodis giveninTable1.
2.2. Catalyst
The catalysts used in our study were obtained from Zeolyst
(Na-FAU, Na
0.2
H
0.8
-FAU) and Albermale (H-FAU). The cata-
lysts, which were originally in the form of ne powder, were
pelletized and sieved to particle sizes of 0.3e0.6 mm. Before
each catalytic experiment, the catalysts were dried at 500

C
for 60 min in the pyrolysis set up (Fig. 2). This allows removal
of free water presented in the micropores of the zeolites, and
hence, a better estimation of the mass balance of the experi-
ments. Details of the catalysts are listed in Table 2.
X-ray diffraction (XRD) patterns of the catalysts were
collected over the range 2q 5e60

on a Bruker D2 Phaser XRD

device using Cu Ka1 radiation source.
2.3. Pyrolysis experiment
The experimental set-up used in this study is shown sche-
matically in Fig. 2. In all experiments, 2 g of biomass sample
was packed inside a tubular quartz reactor. The reactor
consists of two compartments: the biomass and the catalyst
bed, of which the inner diameter/length are 10/350 mmand 7/
250 mm, respectively. Before each experiment, the whole
systemwas purged by a owof 100 cm
3
min
1
of Ar for 30 min
to guarantee an inert atmosphere during the pyrolysis. After
the ushing period, the ow of Ar was reduced to
70 cm
3
min
1
and kept constant during the experiment. Ex-
periments were conducted in one of the two modes: mixing
mode or post-treatment mode. In mixing mode, the biomass
and catalyst were mixed and packed in the same compart-
ment #5 (biomass bed). In this mode, the electrical furnace #7
was not used. Biomass-catalyst mixture was heated to the
desired temperature using the IR furnace #3. Using the IR
furnace, it is possible to raise the temperature to 500

C in less
than 10 s. In the post-treatment mode, the biomass and
catalyst were kept in compartment #5 and #8 (biomass and
catalyst bed), respectively. In this experiment, catalyst in
position #8 was brought to the required temperature and
maintained there. Pyrolysis experiment begins by activating
the IR furnace and the vapours formed in compartment #5
travel throughthe catalyst bed. Temperature was measured at
points marked T in the gure using thermocouples. For all the
experiments (in both mixing and post-treatment modes),
biomass to catalyst mass ratio of 10 was used. Lower ratios
(more catalyst) were not applied because they resulted in bio-
oil with very high water contents (>90 mass fraction %) and
makes bio-oil analyses susceptible to severe errors. All the
tests were conducted at atmospheric pressure at 500

C. This
was also the temperature of the catalyst bed in experiments in
post-treat mode. The average temperature ramp and vapour
residence time inside the quartz tube (#2) was 40

C s
1
and
4 s, respectively. The average vapour residence time here was
calculated based on the Ar ow rate.
Products of pyrolysis reactions were pushed by the Ar ow
into two consecutive condensers immersed in isopropanol/
liquid nitrogen mixture, at around - 40

C to make sure that
only non-condensable gases escaped for collection in a gas
bag. Duration of experiments was set at 10 min to ensure
collection of all products. This should not be mistaken with
the pyrolysis reaction time (in the order of a few seconds).
After each experiment the condensers were rinsed with
known amounts of isopropanol to collect all the liquid prod-
ucts for latter analysis. The weight of the liquid products was
considered equal to the gain in the weight of the two
Table 1 e Main characteristics of the Canadian pinewood.
Proximate analysis (mass fraction %) Ultimate analysis (mass fraction %)
Fixed carbon Volatile Ash Moisture HHV
a
(MJ kg
1
) C H O
b
N
14.51 78.40 2.59 4.50 16.01 48.34 5.81 45.46 0.40
a Higher heating value.
b By difference.
Fig. 2 e Schematic of the packed-bed system for fast
pyrolysis of biomass.
b i oma s s a nd b i oe ne r g y 4 8 ( 2 0 1 3 ) 1 0 0 e1 1 0 102
condensers after experiments. The change in the weight of
the quartz reactor before and after pyrolysis is equal to the
amount of volatiles which has left the reactor during this
process (weight of liquid weight of gas). By applying the
mass conservation law (weight of biomass weight of
solid weight of liquid weight of gas), it is possible to
calculate the amount of solid products. Non-condensable
gases collected in the gas bag were injected to a micro GC
(using PPQ and MS 5A columns) to analyse the composition
(based on volume fraction). The total volume of gas which
ows to the bag was measured using a gas ow metre (Fig. 2).
This, together with composition analysed by the micro GC,
provide the exact volume of each gas in the mixture. The ideal
gas law was employed to calculate the molar amount of each
gas and close the mass balance.
2.4. Bio-oil analysis
In order to acquire the liquid bio-oil, the condensers were
washed with a known amount of 2-propanol. The mass ratio
between 2-propanol and the bio-oil was limited at around 1:2
to prevent errors in sequential analyses. All the analyses of
bio-oils were reported on a water- and solvent-free basis. The
water content of bio-oil (mass fraction) was determined using
Karl Fischer titration (using Metrohm 787 KF titrator). The
energy content of bio-oil was determined by a bomb calo-
rimeter unit (C2000 Kikawerke). A Flash 2000 (Interscience)
instrument was employed to analyse the elemental compo-
sition (C, H, and O mass fraction) of bio-oil and char. In this
analysis, the samples were completely oxidized in O
2
and the
[C], [H] mass fraction were calculated directly based on the
amount of CO
2
and H
2
O formed, and the [O] mass fraction was
obtained by difference.
The total acid number (TAN) of bio-oil was determined
according to ASTM standard D664 using the titrator 785 DMP
Titrino (Metrohm). In the measurement, the potential of the
bio-oil/solvent solution was plotted against the amount of
KOH (1 mol L
1
) used to titrate. The end point was determined
as the point with the highest slope.
The amount of coke formed on the catalysts was deter-
mined by Thermogravimetric analysis (TGA) (Mettler Toledo,
TGA/SDTA851e) in air. The procedure for the coke determi-
nation was the following: the samples were heated fromroom
temperature to 110

C and maintained there for 30 min to
remove all the physically absorbed water, which was followed
by a ramp to 800

C. The analyses were carried out in the ow
of 50 cm
3
min
1
air and using the heating rate of 10

C min
1
.
The weight distribution of bio-oils was analysed in the
Agilent 1200 series gel permeation chromatography (GPC)
equipped with a refractive index detector.
Thecompositionof the gas mixture obtainedfrompyrolysis
was measured ofine using a micro GC (Varian e CP4500)
consistedof two columns anda TCDdetector. The rst column
(MS 5A) was set at 70

C and used for determination of H
2
, CO
and CH
4
. The second column (PPQ) was set at 80

C for the
separation of CO
2
and the rest of the HeCs. Both columns used
Ar as the carrier gas. Finally, an Agilent 7890AGCcoupled with
5975C MS was used to determine the chemical composition of
the bio-oil. The GC was equipped with a capillary column
(Varian VF-1701ms 60 m 0.25 mm 250 mm) using He as the
carrier gas at a constant owrate of 1 cm
3
min
1
. The injection
port of the GC was set at 250

C and a split ratio of 50:1 was
employed. The initial oven temperature of 40

C was held for
5 min. The oven was programmed from 40 to 280

C at
5

Cmin
1
andthe maximumtemperature was maintainedfor
30 min. Data were collected in the full-scanmode between m/z
25e500 with a solvent delay of 8 min. The compounds were
identiedby comparing their mass spectra to the mass spectra
of the built-in NIST 2008 library, interpreting the observed
fragmentation and evaluating the retention times in compar-
ison with standard compounds.
3. Results and discussions
First, results of thermal pyrolysis experiments in the absence
of catalysts are discussed. It is intended to use these results to
establish the proper choice of laboratory experiments and
determine the possibility to carry out catalytic pyrolysis
experiments.
3.1. Thermal pyrolysis experiments
Yields of products in the thermal pyrolysis experiments were,
based on mass fraction, 20.4% of fast pyrolysis char, 10.3% of
gas and 61.2% of liquid which adds up to a total mass balance
of 92%. Liquid product contained 36.9% (in mass, based on
initial weight of biomass) of organic fractionand 24.3%yield of
water. The relatively high liquid yield and low gas and solid
yield shows resemblance to those obtained fromfast pyrolysis
in bigger scale systems [29e32].
Obtained solid char (i.e., excluding minerals/ash), con-
tained carbon (78.7 mass fraction %), hydrogen (2.8%) and
oxygen (18.5%). Energy content of this char was measured to
be 26 MJ kg
1
. This, combined with the yield of 20.4%, provides
more than enough energy (by burning the char in air) to drive
the pyrolysis reaction without the necessity of external heat,
whichmakes the whole process autothermal (Fig. 1) via proper
heat integration.
Carbon monoxide and carbon dioxide are major compo-
nents in pyrolysis gas (total >90% selectivity) with yields of
4.66 and 4.75% (mass fraction compared to biomass), respec-
tively. There are smaller amounts of (C
1
eC
3
) hydrocarbons
and hydrogen (yields of 0.83 and 0.02%, respectively).
Table 2 e Basic properties of faujasite catalysts.
Catalyst SiO
2
/Al
2
O
3
molar ratio Nominal cation form Na
2
O (mass fraction %) Surface area (m
2
g
1
)
Na-FAU 5.1 Na

13 900
Na
0.2
H
0.8
-FAU 5.1 H

/Na

2.8 730
H-FAU 7 H

0.2 650
b i o ma s s a nd b i o e ne r gy 4 8 ( 2 0 1 3 ) 1 0 0 e1 1 0 103
Table 3 lists the differences in the major properties of the
obtained liquid (bio-oil) with those of the original biomass and
a representative fossil fuel-oil. Compared to the parent
biomass, bio-oil contains almost the same amount of
hydrogen(w5.8 mass fraction %), slightly more carbon (52.11%
of bio-oil compared to 48.74% of biomass) and less oxygen
(42.34% of bio-oil compared to 45.6% of biomass).
However, this difference is very small compared to the
difference betweenbio-oil and fuel oil. Compared to bio-oil (H/
C 1.3), fuel oil has much higher carbon and hydrogen content
(85.30% and 11.47%, respectively, H/C 1.6) and much less
oxygen (1.05%). Higher [C], [H] content of fuel oil gives it higher
energy content. Table 3 also summarizes the problems when
working with bio-oil. Presence of high [O] content in bio-oil
results in a corrosive, acidic, viscous, unstable and particu-
lates containing oil with a relatively low-energy density
compared to that of fossil fuels.
To have a better understanding of the undesired properties
of bio-oil, and the factors that cause them, GC/MS was used to
analyse the chemical composition of bio-oil.
Since bio-oil is a complicated mixture of hundreds of
different organic compounds, it is reasonable to classify
them into different groups based on their chemical func-
tionality. Such a classication of the components of bio oil is
shown in Fig. 3. In Fig. 3, the vertical axis shows the relative
proportions of different groups in the bio-oil calculated by
summing the total ion chromatogram (TIC) area percentages
of all the compounds belonging to these groups. This
method, commonly used by other authors [33e35], generates
semi-quantitative results. It can be used to compare the
concentrations of a certain component in bio-oil obtained
under different conditions. Fig. 4 shows the components of
biomass and the possible products that can be formed from
them by pyrolysis. Bio-oil contains a variety of carboxylic
acids e.g., formic, acetic acids, mainly formed by the
decomposition of cellulose and hemicellulose fractions of
biomass [16]. These are the main components that cause
acidity of the bio oil [36]. Phenols which form by the
decomposition of lignin also contribute to acidity but to
a much lesser extent [36]. Carbonyl compounds such as
aldehydes and ketones have tendency to undergo conden-
sation reactions causing viscosity increase make the oil
unstable [16]. Hydrocarbons and furans (formed from hol-
ocelluloses) are the required components due to their high-
energy content [16]. In order to improve the properties of bio-
oil, it is thus essential to remove/modify the problematic
components during de-oxygenation. Essentially de-
oxygenation of harmful components can be achieved by
converting them to products such as CO
2
, CO or H
2
O. In
order to illustrate how effective each route is, in an energy
content point of view, glucose (C
6
H
12
O
6
) was chosen as
a model compound. From glucose, we can either remove one
or more CO
2
or CO or H
2
O molecules and obtain corre-
sponding products (Fig. 5). The higher heating values (HHV)
of these products can be estimated using Dulong formula,
a widely-used model to estimate the HHV of a compound
based on its elemental composition ([C], [H], [O]%, mass
fraction) [37e39]. The higher heating values and oxygen
contents of the products were then compared with those of
the original glucose. These comparisons are represented by
the energy enhancement (%) and the degree of de-oxygena-
tion (%), respectively (Fig. 5). It is obvious from Fig. 5 that
selective de-oxygenation to CO
2
is the most benecial. This
is because per carbon atom lost, two oxygen atoms are
removed and thus carbon loss is lower. For the same reason,
de-oxygenation as CO is less efcient. Formation of water
lowers hydrogen content of biomass, and is thus the least
preferred.
Incorporation of a catalyst in the pyrolysis process must
have the ability to selectively deoxygenate unwanted pyrol-
ysis product oxygenates and form CO
2
. Results of catalytic
pyrolysis are discussed next.
3.2. Post-treatment vs. mixing catalytic pyrolysis modes
Catalytic pyrolysis experiments were carried out with three
faujasite catalysts, details of which are given in Table 2. Since
both catalysts and biomass are solids, there are two ways to
introduce the catalysts into the pyrolysis system, namely: (i)
mixing mode and (ii) post-treatment mode. The aim of
applying catalysts in our work is to reduce the oxygen content
of biomass to get a higher quality bio-oil. It is therefore
Fig. 3 e Distribution of different chemical groups in
non-catalytic bio-oil.
Table 3 e Comparison of main characteristics of bio-oil with those of biomass and fuel-oil.
Materials C
a
H
a
O
a
Energy density (MJ kg
1
) pH Water content
a
Biomass 48.74 5.80 45.46 16 e 5.6
Bio-oil 52.11 5.70 42.34 19 2.4 30.5
Fuel-oil 85.30 11.47 1.05 40 5.7 0.1
a *Data are in mass fraction %.
b i oma s s a nd b i oe ne r g y 4 8 ( 2 0 1 3 ) 1 0 0 e1 1 0 104
reasonable to use degree of de-oxygenation to compare the
efciency of the two modes. The formula for the oxygen
removal degree is shown below.
Oxygen removal degree

O%
biooil
Yield
biooil
=O%
biomass

100
In this formula [O]% represents the oxygen content (mass
fraction) of a specic material. Oxygen removal degree can be
used to judge the amount of oxygen that has been removed
from the biomass and retained in the bio-oil. In non-catalytic
experiment, the bio-oil contains 21% less oxygen than which
was originally present in biomass. By using Na
0.2
H
0.8
-FAU
catalyst in mixing mode, it is possible to remove 23% oxygen
from biomass, which is slightly better compared to non-
catalytic experiment. The same catalyst in post-treatment
mode, however, shows much higher activity in terms of
oxygen removal, which is 31%.
In post-treatment mode, the biomass vapours always have
to travel through the catalyst bed and contact the catalytic
sites at the required temperature. In the mixing mode the
situation is different. The problems are (i) solidesolid contact
between the catalyst and the biomass (particle size:
300e600 mm) is not ideal, and (ii) since catalyst and biomass
are subjected to the temperature ramp simultaneously,
biomass vapours that already form at lower temperatures
(around 300

C) pass the catalyst before it has reached pyrol-
ysis temperature (500

C). For these reasons, not surprisingly,
post-treatment mode is superior to mixing mode in terms of
oxygen removal, and hence this mode is chosen for further
catalytic tests.
Fig. 4 e Origin of components in bio-oil related to chemical functionalities of holocellulose and lignin.
b i o ma s s a nd b i o e ne r gy 4 8 ( 2 0 1 3 ) 1 0 0 e1 1 0 105
3.3. Catalytic experiments
3.3.1. Products distribution
In this section the catalytic inuences on the yields of pyrol-
ysis products (gas, liquid and solid) and on the composition of
solid and gas products are discussed. The catalytic effects on
the composition of liquid phase are discussed separately in
Section 3.3.2. Fig. 6 shows the yields of pyrolysis products
obtained from catalytic and non-catalytic (thermal) experi-
ments. For all the experiments the mass balances are more
than 90%. One can see from Fig. 6 that the solid and gas yields
increase while that of liquid decreases in presence of the
catalysts. H-FAU gave the lowest liquid yield (46.3 mass frac-
tion %, based on initial weight of biomass), highest solid yield
(28.6%) and high gas yield (16.3%).
In Fig. 6, solid yields shown are the sum of solid char left in
the pyrolysis compartment (>90 mass fraction %of total solid)
plus the coke/char formed over the catalyst. Coke/char
formed on the catalyst is maximum for H-FAU catalyst with
the yield of 1.8% (mass fraction, based on initial weight of
biomass) compared to that of 1.6% on Na
0.2
H
0.8
-FAU and 1.3%
on Na-FAU. Among the catalysts, H-FAU also has the highest
concentration of acid sites (H

) and the highest cracking

activity of CeC, CeH and CeO bonds and it is not surprising
that it produces the maximum amount of coke/char.
Comparable experiment with inert material (quartz) in place
of catalyst did not produce any char/coke. H-FAUalso gave the
maximumamount of water (yield of 22 mass fraction %, based
on initial weight of biomass) as expected, since dehydration is
strongly catalysed by acids.
For all the catalysts, elemental compositions and higher
heating value (HHV) of the char and coke were analysed. All
the results are based on ash-free calculations, which means
that the compositions and heat values were only attributed to
the organic fraction of the solid. Elemental composition of
char formed in the pyrolysis chamber was, as expected,
similar for all experiments (78.7% carbon, 2.8% hydrogen and
18.5% oxygen, mass fraction). Surprisingly, compositions of
coke/char formed on the catalysts were also similar, only the
amount of coke varied and was maximum for H-FAU catalyst.
This coke/char contains 87.5% carbon, 6% oxygen and 6.5%
hydrogen, mass fraction. As can be seen, the oxygen content
of coke/char formed on the catalyst is much lower than that of
char in the pyrolysis chamber, but it still contains 6% of
oxygen (mass fraction). This implies that it is not the classical
hydrocarbon coke that is formed on the catalyst but oxygen-
containing oligomeric species. Moreover, it can be seen that
this coke/char is aromatic in nature (H/C w1). Both acidic H

sites and sodium ions Na

seem to catalyse this oligomer

formationsince both H-FAUand Na-FAUresulted in coke/char
formation.
The gases formed during catalytic experiments are mainly
composed of COand CO
2
, the rest being smaller hydrocarbons
(C
1
eC
3
) and hydrogen. Details are shown in Table 4. As can be
seen, the yields of CO and CO
2
increased in the presence of
catalysts, especially in the cases of Na
0.2
H
0.8
-FAU and H-FAU.
Using Na
0.2
H
0.8
-FAU, the yield of CO and CO
2
increased w1.7
times compared to the non-catalytic experiment. Na
0.2
H
0.8
-
FAU also shows the maximum amount of hydrocarbon and
hydrogen compared to all experiments. Consequence of these
results on the properties of bio oil is discussed in the next
section.
3.3.2. Catalytic effects on the composition of bio-oil
This section discusses the change in bio-oil composition, as
analysed by GC/MS, for the various catalytic experiments and
the correlations between the composition and properties,
namely: acidity/corrosiveness, stability, and energy density of
bio-oils.
3.3.2.1. Acidity. Pyrolysis of wood results in the formation of
carboxylic acids and other acidic components which make the
bio-oil corrosive and negatively affect its applicability as
Fig. 5 e Efciency of different oxygen removal routes,
calculated using Dulong formula**HHV (MJ/kg) [ (0.3388*
[C]) D (1.442*([H] L [O]/8)).
Fig. 6 e Product yields in catalytic and non-catalytic
experiments.
Table 4 e Composition of non-condensable gas yields;
data are given in mass fraction %, based on initial weight
of biomass.
Thermal Na-Y Na
0.2
H
0.8
-Y H-Y
CO 4.66 4.89 8.01 7.25
CO
2
4.75 4.37 7.55 6.03
HC 0.83 1.10 3.66 2.91
H
2
0.02 0.03 0.18 0.09
b i oma s s a nd b i oe ne r g y 4 8 ( 2 0 1 3 ) 1 0 0 e1 1 0 106
a fuel. It was shown that among the various components
present in bio-oil, carboxylic acids contribute to the acidity of
bio-oil the most (60e70%) [36]. Other components which
contribute to the acidity of bio-oil are phenols (5e10%) and
sugars (20%). For this reason, the acidity of bio-oil can be
correlated to the amount of carboxylic acids determined by
GC/MS. Validity of this correlation is shown in Fig. 8 by plot-
ting the acidity of several bio-oil samples determined by
potassium hydroxide titration (TAN) versus the corresponding
carboxylic acid concentrations determined by GC/MS.
Fig. 7 shows the inuence of catalysts on the amounts of
acids, phenols, and sugars formed. All the catalysts reduced
the amount of carboxylic acids compared to thermal experi-
ment with H-FAU having the most inuence. It is noteworthy
to recall that the reduction in carboxylic acids is correlated to
the increase in yield of CO
2
in gas stream (Table 4). Since the
amount of acids and the acidity of bio-oils is well correlated, it
is expected that H-FAU is a good catalyst to reduce the acidity
of bio-oil. In the meantime, the amount of phenols increased
drastically in the presence of catalysts (Fig. 7). Phenolic
components are less troublesome in terms of causing acidity
and, together with their high-energy content, are more
desired than carboxylic acids. This increase also indicates the
ability of FAU catalysts to decompose more lignin-derived
vapour.
3.3.2.2. Energy density. Hydrocarbons are the most valuable
fuel components from pyrolysis of biomass due to their high
heating value. The amount of hydrocarbons formed in bio-oil
using each catalyst is compared in Fig. 7. As can be seen from
the gure, there is only a slight increase in hydrocarbon
concentration when Na-FAU is used compared to thermal
pyrolysis. However, when H-FAUand especially Na
0.2
H
0.8
-FAU
were used, the area percentage of hydrocarbons in bio-oil
increased signicantly from 0.5% (thermal) to 3.4% (H-FAU)
and 9.5% (Na
0.2
H
0.8
-FAU). During catalytic de-oxygenation of
biomass, hydrocarbons are proposed to be formed from
sugars through several reactions including dehydration of
monomer sugars to dehydrated products, i.e. furans, and
subsequent decarbonylation, decarboxylation and oligomeri-
zation of dehydrated products to aromatics [16]. Another
possible pathway for hydrocarbon formation is decarboxyl-
ation of organic acids to CO
2
and hydrocarbons. Because the
hydrogen content of biomass is very low, in the absence of
external source of hydrogen most of the hydrocarbons in bio-
oil can be found as aromatics. Moreover, aromatic hydrocar-
bons can also be produced from phenolic compounds by
scission of all oxygenate substituents fromphenol ring [28,35].
It can be concluded that in the presence of Na
0.2
H
0.8
-FAU
catalyst hydrocarbons are most likely produced from de-
oxygenation of phenols and sugars because the amount of
phenols and sugars obtained using this catalyst was the
lowest (Fig. 7).
3.3.2.3. Stability. Ageing process of bio-oil results in an
increase in molecular weight and viscosity of bio-oil due to
the presence of reactive components, i.e. aldehydes and
ketones. A number of possible reactions responsible for the
instability of bio-oil was proposed [40], including reactions
of (i) organic acids with alcohols forming ester and water, (ii)
aldehydes and alcohols forming hemiacetals or acetals and
water, (iii) aldehydes forming oligomers and resins, (iv)
aldehydes and phenols forming resins and water, (v) and
aldol type of condensation reaction between aldehydes and
ketones. Reactions ii, iv and v consume aldehydes, phenols
and form water and the fact that the amount of aldehydes
and phenols in the old oil decreased and water increased
compared to fresh oil supports this (Fig. 9). In order to have
an overview of how the molecular weight distribution of the
compounds in bio-oil has shifted during ageing process, two
bio-oil samples were analysed using GPC technique: one is
a sample of a fresh bio-oil and the other is the bio-oil which
was produced in the same conditions but has been kept at
laboratory conditions for 1 year. The results are shown in
Fig. 9. The GPC chromatogram can be divided into 2 regions:
low-molecular weight (M < 400 g mol
1
) and high-molecular
weight (M > 400 g mol
1
). In the low-molecular weight re-
gion, the curve corresponding to the old oil has similar
shape compared to that of the fresh oil but with lower in-
tensity. The opposite occurs with the high-molecular
weight region (similar shape, higher intensity). From these
results, it can be concluded that during storage process, the
Fig. 7 e Distribution of chemical groups in bio-oils obtained
from catalytic and non-catalytic experiments.
Fig. 8 e Correlation of TAN with total amount of acids in
bio-oil; measurement error of TAN analysis calculated to
be 1.47% based on 95% condence level.
b i o ma s s a nd b i o e ne r gy 4 8 ( 2 0 1 3 ) 1 0 0 e1 1 0 107
low-molecular weight compounds in the fresh oil has reac-
ted with each other to form heavier molecules and increase
the viscosity of bio-oil. Fig. 7 compares the effect of each
catalyst on the formation of aldehydes/ketones (carbonyls)
in bio-oil. It is obvious from the gure that all the catalysts
reduce the amount of carbonyls in bio-oil compared to
thermal reaction with H-FAU and Na
0.2
H
0.8
-FAU having the
most inuence. It can be concluded that the higher the
proton concentration of the catalyst is, the more carbonyls
are removed from bio-oil and the higher stability of bio-oils
is expected.
Highest de-oxygenation extent obtained with Na
0.2
H
0.8
-
FAU, accompanied by the highest amount of CO
2
formation, as
expected, also results in a bio oil with higher energy content
(24 MJ kg
1
, compared to 22 MJ kg
1
for Na-FAU and H-FAU).
Na
0.2
H
0.8
-FAU is a promising catalyst because it (i) decreases
the acid content of the bio-oil, (ii) removes unstable compo-
nents such as aldehydes and ketones, (iii) and by increasing
Fig. 9 e Molecular weight distributionof freshandoldbio-oils.
Fig. 10 e (a) XRD analyses. (b) Comparison of activities of fresh and regenerated catalysts; tested after one pyrolysis cycle.
b i oma s s a nd b i oe ne r g y 4 8 ( 2 0 1 3 ) 1 0 0 e1 1 0 108
the phenols and hydrocarbon contents it not only improves
the energy density of bio-oil but also makes it easier to blend
with crude [41].
3.4. Catalyst regeneration/recycle
In our proposed scheme, the catalyst/coke is burnt to generate
heat for the process and recycled back to the reactor (Fig. 1)
similar to the FCC process. In order to investigate the regen-
erability of the catalysts, pyrolysis experiments were con-
ducted with fresh and regenerated Na
0.2
H
0.8
-FAU catalyst.
Regeneration by coke/char combustion was probed in
a TGA experiment which showed that complete removal of
coke/char is achieved by 600

C. Catalyst after a single pyrol-
ysis experiment was regenerated at 600

C in owing air
(30 cm
3
min
1
) for 5 h. XRD analyses of the two samples: fresh
and regenerated are shown in Fig. 10a. No appreciable differ-
ences were observed. Results of catalytic pyrolysis experi-
ments for fresh and regenerated catalyst are shown together
in Fig. 10b. As can be seen, the fresh and regenerated catalysts
have almost the same performance in pyrolysis reaction. It
can be concluded that the catalyst is regenerable. However,
this conclusion is only valid for the catalyst which was
regenerated after one cycle, while in industry (e.g. in FCC
process) catalysts were required to be functional after many
cycles. Testing catalytic performance after multiple cycles is,
nevertheless, out of scope of this paper.
4. Conclusions
Catalytic upgrading of biomass pyrolysis vapours was carried
out in a lab-scale x-bed reactor at 500

C. Three zeolite
catalysts with different H

and Na

concentrations including
H-FAU, Na
0.2
H
0.8
-FAU and Na-FAU were tested in the reaction.
The results showed that the amount of H

and Na

of the
catalyst plays an important role in the product yields and
product distribution. The higher the concentration of H

of
the catalyst is, the lower the liquid yield, and the higher the
solid and gas yields are obtained. It was shown that two of
major problems with bio-oil, namely the corrosiveness and
instability, were mainly caused by carbonyl and carboxylic
acid compounds, respectively. The best catalyst candidate is
Na
0.2
H
0.8
-FAU, which reduced the most acid and carbonyl
compounds while boosted the amount of the desirable
phenolic and hydrocarbon compounds compared to non-
catalytic experiment and to the other two catalysts.
Na
0.2
H
0.8
-FAU also removed the most oxygen as CO
2
, resulting
in an oil with lowest oxygen content (38 mass fraction %) and
highest energy content (24 MJ kg
1
). It was shown possible to
regenerate the spent catalyst without changing its crystalline
structure and catalytic performance.
Acknowledgements
Financial support from STW Project No. 07972 is gratefully
acknowledged.
r e f e r e n c e s
[1] Lynd LR, Cushman JH, Nichols RJ, Wyman CE. Fuel ethanol
from cellulosic biomass. Science 1991;251(4999):1318e23.
[2] Huber GW, OConnor P, Corma A. Processing biomass in
conventional oil reneries: production of high quality diesel
by hydrotreating vegetable oils in heavy vacuum oil
mixtures. Appl Catal A: Gen 2007;329(0):120e9.
[3] Bridgwater AV, Peacocke GVC. Fast pyrolysis processes for
biomass. Renew Sust Energ Rev 2000;4(1):1e73.
[4] Babich IV, van der Hulst M, Lefferts L, Moulijn JA, OConnor P,
Seshan K. Catalytic pyrolysis of microalgae to high-quality
liquid bio-fuels. Biomass Bioenerg 2011;35(7):3199e207.
[5] Bridgwater AV, Meier D, Radlein D. An overview of fast
pyrolysis of biomass. Org Geochem 1999;30(12):1479e93.
[6] OConnor P, Stamires D, Daamen S, inventors; Kior, Inc.,
assignee. Process for the conversion of biomass to liquid
fuels and specialty chemicals. United States patent US
2012190062; 2012 Jul 26.
[7] OConnor P, Corma A, Huber GW, Sauvanaud L, inventors;
Kior, Inc., assignee. Fluid catalytic cracking of oxygenated
compounds. United States patent US 2012198756; 2012 Aug 9.
[8] Priecel P, Kubicka D, Capek L, Bastl Z, Rysanek P. The role of
Ni species in the deoxygenation of rapeseed oil over NiMo-
alumina catalysts. Appl Catal A: Gen 2011;397(1e2):127e37.
[9] Stefanidis SD, Kalogiannis KG, Iliopoulou EF, Lappas AA,
Pilavachi PA. In-situ upgrading of biomass pyrolysis vapors:
catalyst screening on a xed bed reactor. Bioresour Technol
2011;102(17):8261e7.
[10] Busetto L, Fabbri D, Mazzoni R, Salmi M, Torri C, Zanotti V.
Application of the Shvo catalyst in homogeneous
hydrogenation of bio-oil obtained from pyrolysis of white
poplar: new mild upgrading conditions. Fuel 2011;90(3):
1197e207.
[11] Ferna ndez MB, Sa nchez M, Jhon F, Tonetto GM, Damiani DE.
Hydrogenation of sunower oil over different palladium
supported catalysts: activity and selectivity. Chem Eng J
2009;155(3):941e9.
[12] Fisk CA, Morgan T, Ji YY, Crocker M, Crofcheck C, Lewis SA.
Bio-oil upgrading over platinum catalysts using in situ
generated hydrogen. Appl Catal A: Gen 2009;358(2):150e6.
[13] Lohitharn N, Shanks BH. Upgrading of bio-oil: effect of light
aldehydes on acetic acid removal via esterication. Catal
Commun 2009;11(2):96e9.
[14] Mahfud FH, Melian-Cabrera I, Manurung R, Heeres HJ.
Biomass to fuels e upgrading of ash pyrolysis oil by reactive
distillation using a high boiling alcohol and acid catalysts.
Process Saf Environ 2007;85(B5):466e72.
[15] Czernik S, Bridgwater AV. Overview of applications of
biomass fast pyrolysis oil. Energ Fuel 2004;18(2):590e8.
[16] Huber GW, Iborra S, Corma A. Synthesis of transportation
fuels from biomass: chemistry, catalysts, and engineering.
Chem Rev 2006;106(9):4044e98.
[17] Bridgwater AV. Catalysis in thermal biomass conversion.
Appl Catal A: Gen 1994;116(1e2):5e47.
[18] Williams PT, Nugranad N. Comparison of products from the
pyrolysis and catalytic pyrolysis of rice husks. Energy 2000;
25(6):493e513.
[19] Samolada MC, Papafotica A, Vasalos IA. Catalyst evaluation
for catalytic biomass pyrolysis. Energ Fuel 2000;14(6):1161e7.
[20] Corma A, Huber GW, Sauvanaud L, OConnor P. Processing
biomass-derived oxygenates in the oil renery: catalytic
cracking (FCC) reaction pathways and role of catalyst. J Catal
2007;247(2):307e27.
[21] Fahmi R, Bridgwater AV, Darvell LI, Jones JM, Yates N,
Thain S, et al. The effect of alkali metals on combustion and
b i o ma s s a nd b i o e ne r gy 4 8 ( 2 0 1 3 ) 1 0 0 e1 1 0 109
pyrolysis of Lolium and Festuca grasses, switchgrass and
willow. Fuel 2007;86(10e11):1560e9.
[22] Mullen CA, Boateng AA. Chemical composition of bio-oils
produced by fast pyrolysis of two energy crops. Energ Fuel
2008;22(3):2104e9.
[23] Nowakowski DJ, Woodbridge CR, Jones JM. Phosphorus
catalysis in the pyrolysis behaviour of biomass. J Anal Appl
Pyrol 2008;83(2):197e204.
[24] Patwardhan PR, Satrio JA, Brown RC, Shanks BH. Inuence of
inorganic salts on the primary pyrolysis products of
cellulose. Bioresour Technol 2010;101(12):4646e55.
[25] Zabeti M, Nguyen TS, Lefferts L, Heeres HJ, Seshan K. In situ
catalytic pyrolysis of lignocellulose using alkali-modied
amorphous silica alumina. Bioresour Technol 2012;118(0):
374e81.
[26] Vispute TP, Zhang HY, Sanna A, Xiao R, Huber GW.
Renewable chemical commodity feedstocks from integrated
catalytic processing of pyrolysis oils. Science 2010;330(6008):
1222e7.
[27] Selva M, Fabris M, Perosa A. Decarboxylation of dialkyl
carbonates to dialkyl ethers over alkali metal-exchanged
faujasites. Green Chem 2011;13(4):863e72.
[28] Ding LH, Rahimi P, Hawkins R, Bhatt S, Shi Y. Naphthenic
acid removal from heavy oils on alkaline earth-metal
oxides and ZnO catalysts. Appl Catal A: Gen 2009;371(1e2):
121e30.
[29] Choi HS, Choi YS, Park HC. Fast pyrolysis characteristics of
lignocellulosic biomass with varying reaction conditions.
Renew Energ 2012;42(0):131e5.
[30] Bridgwater AV. Review of fast pyrolysis of biomass and
product upgrading. Biomass Bioenerg 2012;38:68e94.
[31] Zhang HY, Xiao R, Wang DH, He GY, Shao SS, Zhang JB, et al.
Biomass fast pyrolysis in a uidized bed reactor under N(2),
CO(2), CO, CH(4) and H(2) atmospheres. Bioresour Technol
2011;102(5):4258e64.
[32] Pattiya A. Bio-oil production via fast pyrolysis of biomass
residues from cassava plants in a uidised-bed reactor.
Bioresour Technol 2011;102(2):1959e67.
[33] DominguezA, Menendez JA, Inguanzo M, BernardPL, Pis JJ. Gas
chromatographic-mass spectrometricstudyof theoil fractions
produced by microwave-assisted pyrolysis of different sewage
sludges. J Chromatogr A 2003;1012(2):193e206.
[34] Marin N, Collura S, Sharypov VI, Beregovtsova NG,
Baryshnikov SV, Kutnetzov BN, et al. Copyrolysis of wood
biomass and synthetic polymers mixtures. Part II:
characterisation of the liquid phases. J Anal Appl Pyrol 2002;
65(1):41e55.
[35] Goncalves AR, Schuchardt U, Meier D, Faix O. Pyrolysis gas
chromatography of the macromolecular fractions of oxidized
organocell lignins. J Anal Appl Pyrol 1997;40(1):543e51.
[36] Oasmaa A, Elliott DC, Korhonen J. Acidity of biomass fast
pyrolysis bio-oils. Energ Fuel 2010;24:6548e54.
[37] Mason DM, Gandhi KN. Formulas for calculating the caloric
value of coal and coal chars: development, tests, and uses.
Fuel Process Technol 1983;7(1):11e22.
[38] Wanignon Ferdinand F, Van de Steene L, Kamenan Blaise K,
Siaka T. Prediction of pyrolysis oils higher heating value with
gas chromatographyemass spectrometry. Fuel 2012;96(0):
141e5.
[39] Kathiravale S, Yunus MNM, Sopian K, Samsuddin AH,
Rahman RA. Modeling the heating value of Municipal Solid
Waste. Fuel 2003;82(9):1119e25.
[40] Diebold J. A review of the chemical and physical
mechanisms of the storage stability of fast pyrolysis bio-oils.
Golden, Colorado: National Renewable Energy Laboratory;
2000 Jan. p. 60, Report No. SR-570-27613 (Contractor
Thermalchemie, Inc).
[41] Anderson BD, Rytting JH, Higuchi T. Inuence of self-
association on the solubility of phenol in isooctane and
cyclohexane. J Am Chem Soc 1979;101(18):5194e7.
b i oma s s a nd b i oe ne r g y 4 8 ( 2 0 1 3 ) 1 0 0 e1 1 0 110