Vous êtes sur la page 1sur 44

Accepted Manuscript

Title: Pyrolysis of poplar, cellulose and lignin: Effects of


acidity and alkalinity of the metal oxide catalysts

Authors: Chenting Zhang, Xun Hu, Hongyu Guo, Tao Wei,


Dehua Dong, Guangzhi Hu, Song Hu, Jun Xiang, Qing Liu, Yi
Wang

PII: S0165-2370(17)30728-3
DOI: https://doi.org/10.1016/j.jaap.2018.08.009
Reference: JAAP 4392

To appear in: J. Anal. Appl. Pyrolysis

Received date: 26-4-2018


Revised date: 14-8-2018
Accepted date: 15-8-2018

Please cite this article as: Zhang C, Hu X, Guo H, Wei T, Dong D, Hu G, Hu S,


Xiang J, Liu Q, Wang Y, Pyrolysis of poplar, cellulose and lignin: Effects of acidity
and alkalinity of the metal oxide catalysts, Journal of Analytical and Applied Pyrolysis
(2018), https://doi.org/10.1016/j.jaap.2018.08.009

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Pyrolysis of poplar, cellulose and lignin: Effects of acidity and alkalinity of the
metal oxide catalysts

Chenting Zhanga, Xun Hua,*, Hongyu Guoa, Tao Weia, Dehua Donga, Guangzhi
Hub, Song Huc, Jun Xiangc, Qing Liud, Yi Wangc,*

T
IP
a
School of Material Science and Engineering, University of Jinan, Jinan, 250022, P. R.

R
China.

SC
b
Key Laboratory of Chemistry of Plant Resources in Arid Regions, State Key
Laboratory Basis of Xinjiang Indigenous Medicinal Plants Resource Utilization,

U
Xinjiang Technical Institute of Physics and Chemistry, Chinese Academy of Sciences,
N
Urumqi 830011, China.
A
c
School of Energy and Power Engineering, Huazhong University of Science and
M

Technology, Wuhan, 430074, P. R. China.


ED

d
Key Laboratory of Low Carbon Energy and Chemical Engineering, College of

Chemical and Environmental Engineering, Shandong University of Science and


PT

Technology, Qingdao 266590, Shandong, China


E

*Corresponding author. Tel. /fax: +86–531–89736201; E-mail: Xun.Hu@outlook.com


CC

(X. Hu); alenwang@hust.edu.cn (Y. Wang)


A

1
Research highlights

 Less gases while more tar formed over acidic oxides than that over basic oxides.

 Over CaO, the production of phenolic compounds was suppressed.

 Coking was more significant over the basic oxides than that over the acidic

oxides.

T
 The tar from cellulose contributed more to coking than that from lignin.

IP
 Cellulose–derivatives and lignin–derivatives affect coke structures in varied

R
ways.

SC
Abstract

U
This study investigated the effects of the acidity/alkalinity of seven oxides
(Al2O3, SiO2, ZnO, K2O, MgO, CaO, La2O3) on the pyrolysis of poplar, cellulose and
N
lignin. The results showed that the basic supports such as CaO and MgO promoted the
A
formation of gaseous products. CO, CO2, and CH4 were the main gaseous products in
M

the pyrolysis of poplar, cellulose and lignin, and CO was formed first, followed by
CO2 and CH4. Some H2 was also formed from the dehydrogenation reactions over
ED

CaO with cellulose as the feedstock. The acidic oxides promoted the tar formation,
while the basic oxides suppressed tar formation. The oxides like CaO could
PT

remarkably suppress the production of phenolic compounds. The coke formation over
the basic oxides were also much more significant than that over the acidic oxides, and
E

the tar from cellulose contributed more towards coking. The heating of the coke in
CC

inert gas released CO2, CO, H2, CH4 via probably decarboxylation/decarbonylation,
dehydrogenation and etc. The coke from the pyrolysis of lignin was much more stable.
A

CaO and La2O3 reacted with the CO2 produced in pyrolysis and form the carbonates,
while MgO could not. The TPO–MS characterization showed that the coke species
were multiple types over CaO, and a single type over MgO. The cellulose–derivatives
and the lignin–derivatives have distinct effects on the structural configuration of the
coke.
2
Keywords: Catalytic pyrolysis; Cellulose and lignin; Oxide catalysts;
Acidity/alkalinity; Properties of the coke.

1. Introduction
Biomass is a sustainable raw material for the manufacture of chemicals and fuels
[1–3]. Pyrolysis is a technology developed to convert biomass into bio-oil and biochar
via heating at 400–600°C in oxygen-deficit atmosphere [4–6]. Bio-oil has a very

T
complex compositions, containing hundreds of chemical substances and a substantial

IP
amount of water [7,8]. The complexity of bio-oil leads to the poor thermal stability
and the low heating value of bio-oil [9–11]. Furthermore, although bio-oil is a pool of

R
hundreds of chemicals, extraction of certain chemicals or a family of chemicals is

SC
difficult, due to the complex composition of bio-oil [12,13]. During pyrolysis of
biomass, there are many ways for the session of the chemical bonds in cellulose,

U
hemicellulose and lignin in biomass, leading to the formation of large number of
N
chemicals [14–16]. The involvement of a catalyst in pyrolysis could remarkably
A
modify the reaction routes of the organics, regulating the composition of the products
M

[17,18]. Many catalysts have been developed or tested for the catalytic pyrolysis of
biomass, cellulose, hemicellulose and lignin, such as molecular sieves–based catalysts
ED

including Al–MCM–41 materials, HZSM–5 catalyst, ZSM–5 [19–23]. In addition,


many oxides supported metal catalysts have also been evaluated in pyrolysis of
PT

biomass [24–27]. The metallic species could also reform the conversion of the
organics and modified the reaction network in pyrolysis of biomass. The oxides in the
E

metal–based catalysts play the role of supporting and dispersing the metal species,
CC

which might also involve in pyrolysis reactions [28–30]. The molecular sieves-based
catalysts have some acidity/alkalinity, which significantly affect pyrolysis process
A

[31]. Similarly, metal oxides also have some acidity/alkalinity, which might also
impact the pyrolysis process.
In this study, seven metal oxides including Al2O3, SiO2, ZnO, K2O, MgO, CaO,
La2O3 were used as the catalysts for catalytic pyrolysis of biomass, cellulose or lignin.
The effects of acidity/alkalinity of the supports on the distribution of pyrolysis

3
products and the formation of coke were investigated. Our results showed that the
oxides with weak acidity could promote the degradation of the organics and form a
bio-oil with higher yields. The supports with alkalinity promoted the coking of the
phenolic compounds and the cellulose derived organics, especially the CaO oxides. In
addition, the cellulose derivatives have more contribution towards coke formation
than the lignin derivatives.

T
2. Materials and methods

IP
2.1 Materials

R
Poplar wood, cellulose and alkali lignin were used as the feedstock for pyrolysis.

SC
The raw materials were milled sieved to 100 to 200 mesh, and dried at 105°C for 12 h
before the pyrolysis experiments.

U
The metal oxides were prepared by a citric acid combustion method, procedures
N
of which was similar to that reported in the literatures [32,33]. Briefly, 5 g of nitrate
A
salts and a certain amount of citric acid (the molar ratio of nitrate to citric acid = 2 ×
valence of metal ion /3.) were dissolved in water and the pH of the mixed solution
M

was adjusted to 7 (For the calcium nitrate solution, the pH was adjusted until
ED

precipitation had just took place.). The mixture was then heated to 80°C and
maintained for 4 h to form a gel. After that, the gel was then dried at 150°C for 2 h
before calcined in a muffle furnace at 600°C for another 2 h with a ramping rate of
PT

2 °C/min. After the combustion, the metal oxide catalysts were used directly for the
pyrolysis of poplar wood, cellulose and lignin.
E

2.2 Catalytic pyrolysis experiments


CC

The pyrolysis or catalytic pyrolysis of biomass was conducted in a fixed–bed


reactor at 500°C with a 30 min of reaction time. 4.5g feedstock was placed in a quartz
A

boat, and the quartz boat was wrapped with tinfoil. The opening of the tinfoil is 4 cm
in length and 2 cm in width. The opening was covered with a nickel mesh (200 mesh)
and the catalyst was placed on top of the tinfoil. Before the experiments, the air inside
the reactor was purged with nitrogen with the flow rate of 140 ml/min for 30 min. The
quartz boat together with the feedstock and the catalysts was located in the inlet of the
4
reactor, which was outside of the furnace. After heating the reactor to 500°C, the
quartz boat was pushed to the middle of the reactor and the pyrolysis experiments
started and maintained for 30 min in nitrogen stream. After about 68 seconds, the
quartz boat containing the feedstock was heated to 500°C. The organic vapor was
collected with four condensers filled with chloroform/methanol (volumetric ratio: 4 :
1). The condensers were connected in tandem and immersed in icy water. The change
of the abundance of gaseous products (CO, CO2, CH4, H2) versus the pyrolysis time

T
was monitored with in situ infrared gas detectors.

IP
After finishing the experiments, the reactor was cooled down and disconnected.

R
The tar retained in the reaction tube was rinsed and collected with

SC
chloroform/methanol, and mixed together with the trapped liquids from the cold traps.
The procedures for calculation of the tar yield generally followed the ones reported in

U
the literature [34–36]. Briefly, 5 g of the tar solution was loaded in a small pan
N
(diameter: 6 cm), which was then placed in an oven at 35°C and hold for 4 h to a
A
constant weight to measure the weight residual. Reduplicated tests were performed to
ensure the accuracy. The error for the reproducibility of tar yield determination was
M

about ± 0.3 wt%. The residual left was in fact the heavier component of the bio-oil
ED

produced.
2.3 Characterizations
The composition of the organic elements (C、H、S、N) in the sample was
PT

measured with EuroEA3000–Single instrument. The samples were burned in the oven
heated at 950°C in an O2/He flow of 100ml min–1. The gases formed were separated
E

and quantified with detectors.


CC

GCMS–QP2020 (Shimadzu) was used for analyzing the light organics in bio-oil.
The operating conditions were listed as follows: carrier gas: helium; initial oven
A

temperature: 35°C (maintained for 2.5 min), ramping from 35 to 250°C (ramping rate:
10°C/min), maintained at 250°C for 5 min; injection volume: 0.5 μL; solvent delay:
1.25 min; column: DB–Wax. The compounds in bio-oil were identified with the
standard spectrum in the NIST library. It was difficult to find the standards for all the
compounds in bio-oil, and thus the normalized peak areas were used for comparison
5
of the relative abundance of some major products.
Diffuse Reflectance Infrared Fourier transform spectroscopy (DRIFTS) was used
to explore the change of the functionalities of the tar versus the increase of
temperature by using Nicolet IS50 instrument with a modified Harrick Praying Mantis
DRIFT cell instrument. 0.2 g of tar was used for the test. The temperature rose from
room temperature to 700°C at a rate of 20°C/min in the vacuum environment. The
FT-IR spectrum was recorded in a time interval of every 7 seconds.

T
High performance liquid chromatography (HPLC, LC–20AD/T LPGE KIT) was

IP
used to quantify some sugars in bio-oil using a SUGAR SH1011 column. The

R
temperature of the column was 50°C and the flow phase velocity is 0.6 mL/min. The

SC
quantification was achieved by using external standards.
Thermogravimetric analysis (TG) of the tar or the coke on/in the used catalysts

U
were performed in HCT–1 (Henven) instrument connected with an in situ MS detector.
N
The sample was loaded in the crucial of the TG and was purged with nitrogen at room
A
temperature for 20 min to remove the air inside the furnace. The tar or the spent
catalyst was then heated to 900°C from room temperature at a ramping rate of
M

20°C/min in a N2 flow of 80 ml min–1. The abundance of CO2, CO, H2, CH4 and C2H4
ED

in the gaseous products was analyzed with an in–situ mass spectrometer (Pfeiffer, MS
GSD 320).
UV–fluorescence spectra of the bio-oil were measured with a Shimadzu
PT

(RF–6000) fluorescence spectrometer. Prior to the characterization, the bio-oil was


diluted to 500 ppm in methanol, and then the sample was placed in a nominal 10 mm
E

light pass length quartz cell for test. The emission spectrum was 230 nm to 580 nm
CC

and the excitation spectrum was 250 to 600 nm with a scan speed of 600 nm/min.
The ultraviolet spectrophotometer of bio-oil was recorded by using a UV–800S
A

spectrometer (Metash Corp). All samples were diluted to 500 ppm in methanol before
measuring. In addition, the methanol solution was used as the blank or background.
Fourier Transform infrared spectroscopy (FT–IR) IS50 infrared spectrum
instrument was used to characterize the functional groups of the samples. The bio-oil
was mixed with KBr with a mass ratio of 1 : 200 and was thoroughly grounded. The
6
sample was then placed in a vacuum oven set at 35°C and dried for 4 h. Finally, the
dried sample was compressed and tested.
Temperature programmed oxidation (TPO) was conducted by using the
chemisorption instrument of PCA–1200 instrument to explore the nature of the coke
species deposited on the spent catalyst. 20 mg powder catalyst was loaded in the
quartz reactor and maintained at 120°C for 1 h in order to thoroughly remove the
absorption of moisture from the catalysts in argon atmosphere. After that, the sample

T
was heated to the 900°C under a 5 vol% H2/Ar flow of 20 ml min−1. The gaseous

IP
products formed versus the reaction time was measured by an in situ MS, which was

R
named as TPO–MS.

SC
Temperature programmed desorption (TPD) was used to characterize the
acidity/alkalinity of oxide by using the instrument of PCA–1200 instrument. The 50

U
mg powder sample were loaded and heated to 200°C to remove the impurities on
N
surface of the catalysts. The samples were then cooled to room temperature (25°C)
and the absorption of in CO2 (CO2–TPD) or in NH3 (NH3–TPD) flow (30 ml/min) for
A

1 hour. After that, the samples were purged with He (30 ml/min) until the baseline
M

was leveled. The temperature rose from room temperature to 800°C at a ramping rate
ED

of 20°C/minute.
Powder X–ray diffraction (XRD) patterns of the catalysts were scanned at 10º
min–1 in the Bragg angle (2Ɵ) range of 10–80º by a Cu–K α monochromatic X–ray
PT

source with Bruker D8–Advance working at 40 kV and 100 mA.


E

3. Results and discussion


CC

3.1 NH3/CO2–TPD patterns of the oxide catalysts


Figure S1 showed the NH3/CO2–TPD patterns of the oxide catalysts. In the
A

NH3–TPD pattern (Figure S1a), Al2O3 had two NH3-desorption peaks at 465 and
670°C, indicating that alumina had medium acidic sites and strong acidic sites. and its
intensity was higher than that of the other two catalysts. The desorption peak of
silicon oxide turned up at both 320 and 605°C, which were hardly discernable,
accounting for the weak acidity. Zinc oxide mainly had one desorption peak at 720°C,
7
suggesting the presence of the strong acidic sites. CO2–TPD of ZnO did not show
obvious desorption peaks of CO2. ZnO used here not neutral but had some acidity.
The CO2 desorption peak of potassium oxide appeared at around 170°C and at the
high temperature regions. K2O had both the weak and the strong basic sites. In
comparison, strong CO2 desorption peaks appeared at ca. 350°C for MgO while at
720°C for CaO. MgO and CaO had medium basic sites and strong basic sites,
respectively. La2O3 is also a basic oxides, but the amount of the strong basic sites was

T
much less than that of CaO. The distinct acidity/alkalinity of the oxides were closely

IP
related to their catalytic properties, which is discussed in the following sections.

R
SC
3.2 Yields of the main products in the pyrolysis experiments
Figure 1 exhibited the yields of the gaseous products (mainly CO, CO2, H2 and

U
CH4) and the tar yields for pyrolysis of poplar, cellulose and lignin over the various
metal oxides. The pyrolysis of cellulose and lignin were performed only over CaO
N
and MgO, due to the more serious coke formation over these two oxides, which will
A
be discussed later. Besides, since the catalysts did not have direct contact with the
M

feedstock, the yield of char was the same in all the experiments. The char yields from
the pyrolysis of poplar, cellulose and lignin were shown in Figure 1d.
ED

The metal oxides affected the yields of the gaseous products in different ways.
Over the acidic oxides (from Al2O3 to ZnO), the yields of total gas was lower than
PT

that of the blank experiments while the total gas yields over the basic oxides were
close to that of blank experiment (Figure 1a). Nevertheless, the tar yields over Al2O3,
E

SiO2 and ZnO were relatively higher. These oxides could suppress the further
CC

decomposition of the volatiles to gaseous products and retained them in tar. The basic
oxides did not have such an effect.
A

Such a phenomenon was also observed in pyrolysis of cellulose over MgO and
CaO. The yields of the gaseous products were also relatively higher over MgO and
CaO while the yields of tar were relatively lower. The acidic sites in the oxides could
facilitate the dehydration reaction of organic molecules, which could also promote the
cleavage of the macromolecules into small ones, thereby increasing the tar yield. In
8
comparison, the alkaline center in the basic oxides could effectively promote the
fracture of the C–H, C–O or C–C bonds, accelerating the formation of the gaseous
products and correspondingly leading to lower tar yields.
For the pyrolysis of cellulose and lignin in the absence/presence of a catalyst.
The total gas yields and the tar yields varied significantly. The total gases yields for
the pyrolysis of cellulose ranged from 22–25%, while that from lignin ranged from
18–22%. The tar yields from the pyrolysis of cellulose were also much higher than

T
that of lignin (ca. 25% for cellulose versus ca. 5% for lignin). Cellulose is comprise of

IP
a linear chain of several hundred to many thousands of β(1→4) linked D–glucose

R
units [37–39]. The breakdown of the C–O linkage bonds produces sugar monomers or

SC
oligomers. The retro-aldol condensation of the sugars produces the smaller organics
like hydroxyl aldehyde, hydroxyl acetone and etc. The cleavage of the C–O bonds

U
generated the gaseous products like CO, CO2, CH4 or some alkanes. Cellulose is more
N
prone to pyrolysis, producing the lower yields of the residual char (23.1%). Cleavage
A
of the side chains of lignin produces gaseous products, while the cleavage of the
linkage between the benzene ring in lignin are relatively difficult, leading to a low
M

yields of tar and a much higher yields of char. The basic benzene ring units in lignin is
ED

associated with ether bonds (C–O bond) or C–C bonds [40–42]. These chemical
bonds directly connect to benzene ring or located in the local vicinity of benzene ring,
which could have σ–π conjugation for C–C bonds or n–π conjugation for C–O bonds
PT

with the π bonds in benzene ring, making the dissociation more difficult. The kinetics
for releasing the gaseous products versus the pyrolysis reaction time were further
E

investigated with on–line infrared gas detectors.


CC

3.3 Analysis of the gaseous products


A

The abundance of the gaseous products (H2, CO, CH4, and CO2) versus the
pyrolysis time was given in Figure S2 and 2, respectively. In general, the abundance
of the CO, CO2, and CO2 were higher in the blank experiments than those over the
oxide catalysts, which was in lined with the total gas yields in Figure 1. The gaseous
products were formed not only during the pyrolysis (the mass transfer of the gases
9
from the inner structure to the outer surface of biomass particles), but also during the
travel of the volatiles inside the reactor. The oxides impacted the structures of the
volatiles during their pass–through the catalyst bed, affecting their further degradation
to the gases. H2 was not detected during the pyrolysis process, except that over CaO.
The other three gases were detected in all the experiments, and the starting time for
their formation followed the orders: CO < CO2 < CH4.
CO was probably formed from the decarbonylation, or more probably from the

T
dissociation of the HC*OH radical formed from the cleavages of the C–C bonds in the

IP
sugar structures. The dehydrogenation of HC*OH radical would be cracked to form

R
two H*, the combination of which would produce H2. This, however, did not take

SC
place at least over the oxides except CaO. The H in HC*OH radical is bonded with
oxygen or carbon atom. The bond energies are different and the synchronous

U
dissociation of the C–H and O–H would be difficult to take place and the released H*
N
would quickly be captured by other radical such as CH* or CH3* to form CH4. CO2 is
A
unlikely formed from the decarboxylation reaction, but might be formed from the
combination of the CO* radical with the O* radical. The formation CO needs the least
M

involvement of the other radicals and thus could be form first, followed by CO2 and
ED

CH4.
The abundance of the gaseous products over the different oxides were not
significantly different, except that over CaO. Our further study showed that CO2
PT

reacted with CaO, which will be discussed later. H2 was formed over CaO oxides,
while H2 was also formed for the pyrolysis of cellulose over CaO. CaO probably
E

promote the hydrogenation reactions of the glucoside unit of cellulose, resulting in H2


CC

formation.
Compared with the gases formation over lignin, the abundance of the total gases
A

were higher over cellulose than over lignin, as is demonstrated in Figure 2. During the
pyrolysis, cellulose has a higher contribution than lignin towards the gaseous products
formation. In addition, the amount of CO formed was much higher over cellulose than
over lignin. In converse, more CH4 was formed in the pyrolysis of lignin. The basic
units of lignin have more aliphatic carbon chain or methyl group, while in cellulose
10
there are more C–O bonding structures, causing the distinct distribution of the
products. The distinct pyrolysis behaviors of the poplar, cellulose and lignin not only
led to the different distribution of the gaseous products, but also significantly affected
elemental composition of the char and tar products.

3.4 Analysis of char and tar


The elemental composition of the feedstock, char and tar produced from

T
pyrolysis of poplar, cellulose and lignin differed significantly, as indicated in Table 1.

IP
Cellulose has a lower content of carbon while a higher content of hydrogen, which

R
was determined by the basic sugar unit in the structure. The alkali lignin has a high

SC
content of carbon but also sulfur. Sulfur was introduced during the separation of lignin
from cellulose, while after the pyrolysis at 500°C, the sulfur was released into the tar.

U
Interestingly, in the char derived from lignin, the oxygen content was much higher
N
than that in the char of cellulose. The ether bonds or the oxygen–containing
A
functionalities did not decompose much, which was retained in the char structure. In
comparison, the pyrolysis of cellulose proceeded to a significant extent, releasing
M

majority of the oxygen–containing functionalities in form of gases or volatiles in tar.


ED

The elemental composition of the tar depended on the oxide catalysts used. The
tar from the blank experiment and that with alumina as catalyst had a comparable
compositions, while the others generally have a C/H molar ratio around 0.5. More
PT

hydrogen was retained in the tar produced over the oxide catalysts. The use of CaO or
MgO as the catalyst did not change much of the elemental composition of the tar
E

produced from pyrolysis of cellulose. For the pyrolysis of lignin, MgO as the catalyst
CC

led to the formation of the tar with a lower carbon content while that with CaO was
higher. This result indicated that there were more dehydrogenation or dehydration
A

reactions taking place over CaO catalyst. Moreover, as mentioned above, the sulfur in
the alkali lignin feedstock was transferred to the tar. The sulfur content in the tar
produced from pyrolysis of lignin over CaO was relatively much lower. CaO probably
absorbed the sulfur species, while the further characterization of the used CaO did not
find the presence of sulfur. With CaO as the catalyst, sulfur probably existed in the
11
light volatiles, which was not contained in the tar as the light components was not
included in the tar analyzed here.

3.5 Analysis of the bio-oil trapped in the condensers


3.5.1 Characterization with GC–MS
The bio-oil produced from pyrolysis of poplar wood was composed of a complex
mixture of ketones, cyclopentanones, phenols, aldehydes, furans, acids and various

T
aromatics. The compounds in bio-oils were identified and the normalized abundance

IP
of the typical compounds were shown in Figure 3, S3 and S4, respectively. The oxide

R
catalysts impacted the formation of the organic compounds in varied ways. Acetic

SC
acid, formed mainly from the degradation of hemicellulose, is a main component of
bio-oil [43]. The oxide catalysts excepted La2O3 all promoted the formation of acetic

U
acid. Nevertheless, the production of hydroxyl aldehyde, formed from retrograde
N
aldol condensation of sugars [44,45], was promoted over the acidic supports while
A
was suppressed over the basic supports, especially MgO and CaO. For the production
of ketones (Figure 3c) and cyclopentones (Figure 3d), the acidic supports generally
M

facilitated their productions, while the basic supports also promoted their production,
ED

but the promotional effect was not so noticeable. Similar phenomenon was observed
for the formation of furans such as furfural (Figure 3e). All the oxides promoted the
production of phenol (Figure 3f). Nevertheless, the production of other phenolics such
PT

as guaiacol or 4–allyl–2,6–dimethoxyphenol were remarkably suppressed over K2O


and CaO catalysts, especially over CaO. The results showed that the acidic supports
E

were benefited for the formation of the sugar–derivatives such as acetic acid and
CC

hydroxyl aldehyde, and the basic supports especially CaO could suppress the
formation of the phenolics, leading to the lower tar yield.
A

The pyrolysis of cellulose and lignin over the basic supports, MgO and CaO,
were also investigated to further confirm their effects on evolution of the organics.
The results shown in Figure S3 showed that the basic oxides could suppress the
formation of hydroxyl aldehyde and cyclopentones, especially over CaO.
Nevertheless, the basic supports promoted the production of carboxylic acids (Figure
12
S3b), furans like furfural (Figure S3d) and HMF (Figure S3c) to some extent. The
results here were similar to that in pyrolysis of the poplar wood (Figure 3). However,
for the main anhydrate sugar, levoglucosan, the oxide catalysts had negligible effects
(Figure S3h). For the pyrolysis of lignin, CaO could suppress the formation of the
phenolics, which was also similar to that of pyrolysis of poplar wood.

3.5.2 Characterization with UV–fluorescence spectroscopy

T
The heavy phenolics, which could not be analyzed with GC–MS, were

IP
characterized with UV–fluorescence spectroscopy. Synchronous spectra of various

R
bio-oils were given in Figure S5. The peak position and peak value of the

SC
fluorescence spectrum correspond to the size and concentration of the aromatic rings
[46,47]. Over the different oxides, the profiles of the spectra for the pyrolysis products

U
of poplar were similar. Nevertheless, the peak intensity over ZnO, SiO2 and Al2O3
N
were relatively higher, indicating that more phenolics were formed, which was similar
A
to that of GC–MS results. The basic oxides also could promote the production of the
heavier phenolics, but to a milder extent. This was different from that of the GC–MS
M

results, where the production of the phenolics except phenol were suppressed over
ED

CaO.
Interestingly, pyrolysis of cellulose also produced the bio-oil showing abundant
fluorescence diffractions, although the intensity was lower than that of poplar. The
PT

appearance of the fluorescence was due to the existence of the π–conjugated


structures, which was probably the phenolics. The basic oxides had little effects on the
E

formation of the π–conjugated structures from the pyrolysis of cellulose. Similar


CC

phenomenon was observed for pyrolysis of lignin over the basic supports. From the
GC–MS and the UV fluorescence results, it showed that the basic oxides mainly
A

impacted the formation of the light phenolics while had little effect on the formation
of heavier phenolics.

3.5.3 Characterization with UV–Vis spectrophotometers


The UV–Vis spectra of the various bio-oil were recorded in Figure S5. The
13
bio-oil of lignin had two absorption peaks in the UV–Vis spectrum, which located at
205 and 280 nm, respectively. The main peak at 205 nm was the characteristic E2
absorption strip that belongs to benzene ring [48]. The weak peak near 280 nm was
the B strip undergone a redshift. The spectra of the bio-oils were similar, indicating
the similar abundance of the phenolics. For the spectrum of the bio-oil produced from
poplar, the peak intensity especially at 205 nm decreased, due to the lower content of
aromatic compounds. In addition, the peak intensity was also affected by the

T
conjugated π bonds in the furans like furfural, furfuryl alcohol, HMF, leading to the

IP
increase of the peak intensity of 268 nm (Figure S5e). The bio-oils formed over the

R
oxide catalysts had higher peak intensities than that from the blank experiment at 268

SC
nm, especially over Al2O3, SiO2 and ZnO, which was in line with the GC–MS results
where the formation of the furans were promoted (Figure 3e). According to the

U
fluorescence spectrum for the bio-oil produced from cellulose, the content of
N
phenolics was relatively low. However, in the ultraviolet spectrum, the intensified
A
peak at 205 nm and at 268 nm indicated that the bio-oil formed from the pyrolysis of
cellulose contained a large number of compounds with conjugated structures and the
M

presence of presence of MgO and CaO promoted the formation of the conjugated
ED

structures in the bio-oil.

3.6 Tar characterization


PT

3.6.1 Characterization of the tar with FT–IR


The FT–IR spectra for the tar produced from the poplar, cellulose and lignin
E

were presented in Figure 4. For the tar from pyrolysis of poplar, there was a large
CC

peak at 3400 cm–1, which was the typical absorption of hydroxyl group. The
corresponding substance may be acids, alcohols, phenolics or water in tar. In the range
A

of 2980 to 2850 cm–1, there were three peaks corresponding to the stretching vibration
of the C–H bond of alkane chain. The peak at 1705 cm–1 were assigned to the
stretching vibration of ketonic functional group. The absorption bands in the region of
1600–1500 cm–1 demonstrated the presence of aromatic substances. The stretching
vibration between 1750 and 1390 cm–1 is an indication of the presence of alkane
14
fragments. The higher the relative peak of 1390 cm–1, the longer the alkane chain in
the tar. The stretching vibration of C–O at 1260 and 1050 cm–1 indicated that tar
might also contain ester and ether functionalities. The absorption at 1110 cm–1
corresponded to the stretching vibration of –OH of alcohols. The absorption band in
the fingerprint area could belong to the deformation vibrations of aromatic rings [49].
The profile for the tar from pyrolysis of cellulose were similar to that of poplar tar.
The deformation vibrations of aromatic rings was also observed, indicating the

T
presence of aromatics in the tar produced from cellulose, which have been confirmed

IP
from the UV–Vis spectra and UV fluorescence spectra (Figure S5). For the tar

R
produced from lignin, the characteristic absorption of the functionalities were similar

SC
to that of poplar tar but had different intensities. The distribution of the functionalities
in the poplar tar and in the lignin tar were similar.

U
N
3.6.2 Characterization of the tar with TG-MS
A
The TG–MS analysis of the various tars in nitrogen atmosphere was given in
Figure 5. The weight loss for the tars obtained from pyrolysis of poplar, lignin, and
M

cellulose were 79.2, 77.4 and 92.4%, respectively. The components in the cellulose tar
ED

were more volatile than that in poplar tar or lignin tar. The MS results showed the
change of the gases abundance versus the reaction time. CO was firstly formed and
the formation of CO2 followed. From the GC–MS and the FT–IR characterization, it
PT

was known that the tar contained the aldehydes or ketones with the carbonyl group
and carboxylic acids withcarboxylic group. The decarbonylation and decarboxylation
E

probably led to the formation of CO and CO2. The cleavage of methyl group at the
CC

higher temperature probably led to the formation of CH4. Furthermore, hydrogen was
also formed at the temperature above 650°C, which was probably formed from the
A

dehydrogenation reactions to form the conjugated system or the ring structures such
as aromatic rings.
The poplar tar produced over CaO had a different pattern for the formation of
gases. Ethylene only formed above 650°C and no hydrogen was formed. The
components bearing the methyl group was different from that in the blank
15
experiments and the dehydrogenation reactions were not preferred. The cellulose tar
showed a similar pattern to that of poplar tar for release of the gaseous products
during the TG–MS characterization, indicating that the formation of CO, CO2 or H2
from the poplar tar were mainly due to the cellulose derived components. The release
of the gases during the heating of lignin tar, as shown in Figure 5d, confirmed this.
The lignin tar during the heating mainly experienced dehydrogenation reaction to
form hydrogen and possibly the decarboxylation reaction to produce CO2. Our

T
previous study showed that there are a significant portion of heavy carboxylic acids,

IP
probably derived from the pyrolysis of lignin, exist in bio-oil [50], the

R
decarboxylation of which would produce CO2.

SC
3.6.3 Characterization of the tar with DRIFTS

U
Figure 6 showed the DRIFTS spectra of the tar samples at different temperature.
N
For the tar from poplar (Figure 6a), the peak position of the DRIFTS spectrum was
A
basically consistent with that of the FT–IR spectrum (Figure 4), but the intensity was
relatively lower at low temperature. The peaks at ca. 3400, 1705, 1260, 1050 cm-1
M

indicated the presence of –OH, C=O and C–O, intensity of which was gradually
ED

weakens with the increase of temperature. This was consistent with the inference in
thermal analysis. At the same time, the peaks belong to CO and CO2 were appeared at
the 2230 cm-1 and 2350 cm-1 respectively. When the temperature rose from 400 to
PT

500°C, the vibration peaks of alkanes in the range of 2980 to 2850 cm-1 were
significantly weakened. At the 3070 cm-1, a new small peak was generated, which
E

represented the generation of methyl group. Methyl group is a precursor for the
CC

formation methane. For the functional groups of aromatic hydrocarbons, the


characteristic absorption peaks enhanced with the increase of temperature. It could be
A

related to the dehydrogenation reactions, resulting in the formation of aromatic


hydrocarbons. In the tar obtained from pyrolysis poplar over CaO catalyst, the
absorption band of the C–H bonds decreased while correspondingly the absorption of
aromatics in the fingerprint area intensified (Figure 6c), indicating occurrence of the
aromatization process to form ring structures. For the tar produced from pyrolysis of
16
cellulose, although aromatic structures were also formed when the temperature
reached to 700°C (Figure 6e), the C–H bonds and the C=O bonds in the aliphatic
structures were quite stable. Aromatization of the compounds in the lignin derived tar
was also observed (Figure 6g). The C=C, C–H and C–O bonds were the main
functionalities in the tar produced from pyrolysis of lignin, and the latter two types of
bonding were quite stable during the heating. The more detailed structural
configuration of the tar produced from the poplar, cellulose and lignin needs further

T
investigation.

R IP
3.7 Characterization of coke on the catalysts

SC
3.7.1 Elemental analysis of the coke
The pyrolysis of poplar, cellulose and lignin over the oxide catalysts might form

U
coke, which was further analyzed by using an elemental analyzer. Elemental contents
N
(N, C, H, S) of the spent catalysts were shown in Table 2. Compared with the basic
A
oxides, the Al2O3, SiO2 and ZnO had lower carbon contents, while the coke formation
over the basic oxides was significant, especially over CaO, whether in the pyrolysis of
M

poplar wood, cellulose or lignin. The C/H molar ratio in the coke in pyrolysis of
ED

poplar over the CaO was ca. 2.4, indicating that the coke on the calcium oxide had a
higher degree of aromatization than the coke on the other catalysts, and the aromatic
ring structure formed was larger. The C/H molar ratio in the coke over MgO was ca.
PT

0.48, indicating that the coke was partially aromatic and partially aliphatic. The
structure of the coke seemed quite different over MgO or CaO, which was further
E

characterized.
CC

3.7.2 TG–MS and DTG analysis of the coke


A

The TG/DTG characterization of the coke formed on the oxide catalysts in air
atmosphere was presented in the Figure 7. The weight loss over the Al2O3, SiO2, ZnO
and K2O was not significant, indicating the negligible coke formation, which was in
line with the EA characterization (Table 2). In comparison, the coke formation was
serious over MgO, CaO and La2O3, especially over CaO. Over the MgO catalyst. The
17
weight loss initiated earlier for the coke formed from cellulose than the coke formed
from lignin, indicating that the coke from cellulose was more reactive than the coke
from lignin. More importantly, more coke was formed from the pyrolysis of cellulose
than that of lignin. The derivatives from the pyrolysis of cellulose thus had more
contribution towards coke formation. Similarly, the coke formed from the pyrolysis of
cellulose over CaO also contributed more to the coking. In addition, the spent CaO
catalyst started to lose weight at the temperature above 200°C, which was probably

T
due to the release of some volatiles and the oxidation of organic species on surface of

IP
the catalysts. Nevertheless, there was a significant weight decrease at ca. 700°C. In

R
addition, for the coke in pyrolysis of lignin over CaO, as shown in Figure 7e, the

SC
weight loss at ca. 700°C was also the main peak. It seemed that the coke formed over
CaO was quite inert and could only be combusted at the high temperatures. To further

U
understand the properties of the coke, the spent catalysts were characterized
N
subsequently with TG–MS in N2 atmosphere.
A
In N2 atmosphere, the coke species pyrolysed, resulting in the release of gases
such as CO and CO2. The carbon oxides was probably formed from the degradation of
M

the functionalities of the coke species [51,52]. In addition, there was hydrogen
ED

formation over the spent CaO catalysts (Figure 8f). Dehydrogenation reactions took
place at the elevated temperature, forming π–bonds or π–conjugated structures [53].
This also indicated the coke species over CaO was different from others. The release
PT

of hydrogen was also observed during the pyrolysis of the coke formed from cellulose
pyrolysis over either MgO or CaO catalysts, while was insignificant in the pyrolysis
E

of lignin. The coke was probably more aliphatic from the pyrolysis of cellulose. The
CC

CO2 signals were more intensified over the basic oxides and was the most significant
over the spent CaO catalyst in pyrolysis of poplar wood. Nevertheless, the CO2
A

signals were also significant over the spent CaO catalyst whether in pyrolysis of
cellulose or lignin (Figure 9b and 9d). CO2 might be formed from pyrolysis of the
coke species, which might also formed from the CO2 absorbed on surface of the
catalysts. CO2 was one of the major gaseous product in the pyrolysis experiments,
which might react with the basic oxides. To confirm this, XRD characterization was
18
subsequently performed.

3.7.3 XRD analysis of the coke


Figure 10 showed the characteristic diffraction of the oxide catalyst before the
pyrolysis experiment and spent catalysts. Before the pyrolysis, it was found that the
nominal “K2O”, in fact, did not exist in the oxide form, but in the K2CO3 form.
During the calcination of the catalyst precursor in air, the K2O formed reacted with

T
the CO2 in air, forming K2CO3 that was actually the catalyst for the pyrolysis

IP
experiments. In addition to K2CO3, La2O3 and CaO also reacted with the CO2

R
produced during the pyrolysis, formed La2CO5 and CaCO3, respectively. MgO did not

SC
react with CO2 during the pyrolysis. La2CO5 and CaCO3 could decompose at the high
temperature such as above 700°C, resulting in the significant weight loss and the

U
release of significant amount of CO2, as evidenced by both the TG and the TG–MS
N
characterizations (Figure 7 and 8). From the results here it demonstrated that the
A
significant weight loss during the TG characterization was not totally due to the
combustion of coke. The results here also explained the very high C/H molar ratio in
M

the coke. Some of the carbon was from that in CaCO3, not from the coke species.
ED

The results here brought the doubt about the coke formation over CaO and La2O3
catalysts. Was the weight loss due to the decomposition of the carbonates or really
from the coke species? To clarify this question, TPO–MS was subsequently
PT

performed.
E

3.7.4 TPO–MS analysis of the coke


CC

The results for the TPO characterization of the used catalysts after the pyrolysis
tests were given in Figure S6 and S7. The results showed that the oxygen
A

consumption peak over Al2O3, SiO2, ZnO and K2O was relatively small and the peak
mainly located at ca. 500°C. The oxygen consumption peak over the spent MgO was
more remarkable, the gaseous products formed include CO and CO2. The incomplete
oxidation of the coke species resulted in the formation of CO. For the spent CaO,
there are three main oxidation peaks for CaO, locating at 370, 470 and 730°C,
19
respectively. It has been reported that coke could be classified into the ones with
amorphous structures with poor thermal stability, the carbon fiber structures, and the
highly developed carbon fiber structures [54,55]. The coke prone to oxidation at low
temperatures probably belong to an amorphous structure. The TG–MS
characterization showed that coke could release hydrogen during the heating in N2,
which probably belonged to the coke with amorphous structure with aliphatic
fragments. CaCO3 decompose at ca. 700°C, releasing CO2. Nevertheless, there was a

T
significant oxygen consumption peak and also the peak for CO formation, indicating

IP
that there was coke species combusted at this temperature. The significant weight loss

R
in the TG characterizations (Figure 7) was partly due to the combustion of the coke

SC
species. The TPO–MS results for the spent MgO catalysts were in line with that in the
TG characterizations, where the coke generating in cellulose pyrolysis was more

U
significant. For the CaO catalyst, the oxygen consumption peak in cellulose pyrolysis
N
included one main peak at ca. 400°C and another one at ca. 700°C. The oxygen
A
consumption peak for the catalyst in pyrolysis of lignin mainly located at 700°C,
indicating that the different type of coke species formed in pyrolysis of cellulose and
M

lignin.
ED

3.8 Effects of acidity/alkalinity of oxide catalysts on distribution of pyrolysis


products
PT

The TPD results showed in Figure S1 showed that the oxides had different
acidity or alkalinity. The acidity or alkalinity of the oxides did remarkably affect the
E

distribution of some of the products, as detailed in Figure 11. As shown in Figure 11a,
CC

the acidity of the oxides clearly affected the yields of tar. The tar yields over the acidic
oxide such as Al2O3 was much higher than that over the basic oxides such as CaO or
A

La2O3. The acidic oxides could help to promote polymerisation and cracking reactions.
Polymerisation of the organics especially to coke would diminish tar production. The
cracking reaction, in converse, could convert more solid tar to the condensable
volatiles (the tar) or the heavier organics to the lighter ones. The results here showed
that the cracking reactions dominated, as more tar was formed over the acidic oxides.
20
In comparison, the tar yields over the basic oxides was lower. The basic oxides during
the pyrolysis could suppress the formation of some aldehydes, ketones and phenolics,
probably via catalysing their condensation reactions. Further to this, the basic oxides
also promoted the formation of more gaseous products (Figure 1a) and more coke
(Figure 7a). These factors together resulted in the formation of less tar over the basic
oxides.
Figure 11b herein showed the weight loss in the TG characterisation of the spent

T
catalyst in air stream. The weight loss was due to the coke deposits or the CO2

IP
released from the spent catalysts. For the K2O catalyst, it actually existed in the form

R
of K2CO3, as evidenced by the XRD characterisation of the catalysts (Figure 10). The

SC
coke deposits over K2CO3 was insignificant but significant over the other basic oxides
when compared with the acidic oxides. Over the basic oxide catalysts, the formation

U
of tar was lower, which was probably due to the polymerisation reactions. The
N
formation of the aldehydes like hydroxyl aldehyde, the ketones like hydroxyl acetone,
A
the carboxylic acids like acetic acid and the phenolics like phenol all decreased. These
compounds or their precursors probably polymerised over the basic oxides, leading to
M

the diminished production of tar in the products and more coke deposits. Over the
ED

acidic oxides, the formation of the coke deposits was insignificant. Although acidic
oxides could catalyze polymerisation reactions, they also could catalyze cracking
reactions, which facilitate the production of more tar while less coke deposits.
PT

Figure 11c shows the abundance of the typical carboxylic acids, acetic acid, the
typical sugars, levoglucsoan, and the typical phenolics such as phenol and 3-methyl
E

phenol in bio-oil produced over the oxide catalysts. The production of acetic acid was
CC

suppressed over the basic oxides, especially over La2O3 and CaO. Acetic acid in
bio-oil was mainly originated from degradation of hemicellulose in biomass. Over the
A

basic oxides, it was probably the cracking of hemicellulose or the precursors of acetic
acid was suppressed. Much more coke deposit was observed over the basic oxides.
The precursor(s) of acetic acid was probably polymerized, leading to the diminished
production of acetic acid. In comparison, the production of levoglucosan was not
suppressed over the basic oxides, formation of which was even enhanced to some
21
extent. Obviously, the formation of levoglucosan via cracking of the
1,4-linked β-D-glucan units in cellulose was not affected much by the basic oxides.
The alkalinity of the oxide catalysts also affected the formation of some of the
phenolics such as phenol and 3-methyl phenol. The production of these two phenolic
compounds were suppressed over the basic oxides. Phenol or 3-methyl phenol was
not the primary compounds from pyrolysis of the lignin in biomass. It seemed the
further degradation of the heavier phenolics were suppressed to some extent over the

T
basic oxides. Nevertheless, the effects of acidity/alkalinity of the oxide catalysts on

IP
the formation of the phenolic compounds was rather complicated. As shown in Figure

R
3g and 3h, the production of other phenolics over the oxides catalysts did not follow

SC
some specific trend. During the pyrolysis of lignin over MgO or CaO, the formation
of the typical phenolics was suppressed, especially over CaO (Figure S4). Clearly, the

U
formation of some of the phenolics were not favoured over the basic oxides,
N
especially over CaO.
A
During the ex-situ catalytic pyrolysis of biomass, the initial pyrolysis of the
components of biomass was not affected by the catalysts. The organic volatiles
M

produced from the primary pyrolysis would then pass through the catalyst bed. The
ED

contact of the organic vapours with the catalysts would modify the composition of the
products. Nevertheless, the transformation of the organic volatiles continued after
leaving the catalyst bed, especially at the elevated temperature. This would lead to the
PT

complicated reaction routes for the conversion of the organic volatiles and
consequently the composition of bio-oil.
E

In comparison to that of the phenolics, clear trends for formation of the


CC

aldehydes like hydroxyl aldehyde, the furans like furfural and the ketones like
hydroxyl acetone were observed. Their production was suppressed over the basic
A

oxides. Before contacting the oxide catalysts (in the pure pyrolysis), the amount of the
aldehydes/ketones were the same. The aldehydes/ketones were clearly transformed
during the contact with the oxide catalysts, especially over the basic oxide catalysts.
Aldehydes/ketones have carbonyl functionality and/or activated ɑ-H. It is well known
that base could catalyze the polymerisation of aldehydes/ketones via the Aldol
22
condensation reaction. Some aldehydes/ketones were probably polymerised over the
basic oxides. The coke deposits were also observed on surface of the spent basic
oxides catalysts, which might be related to condensation of the aldehydes/ketones.
The specific reaction routes for the conversion of the aldehydes/ketones over the basic
oxides could be understood by further model compound studies.

4. Conclusions

T
In summary, different kinds of oxide catalysts have distinct effects on the

IP
distribution of the products from the pyrolysis of poplar, cellulose and lignin. The

R
acidic oxides such as Al2O3, SiO2 and ZnO promoted the formation of tar, while the

SC
basic supports such as CaO and MgO promote the formation of more gaseous
products, which correspondingly suppress the tar formation to some extent. CO, CO2,

U
and CH4 were the main gaseous products in the pyrolysis or catalytic pyrolysis of
N
poplar, cellulose and lignin. The starting time for their formation followed the
A
sequence: CO < CO2 < CH4. Only over CaO there was some H2 formed with cellulose
as the feedstock. For the gaseous products the pyrolysis of cellulose yields more CO
M

while the pyrolysis of lignin yields more CH4, due to their distinct configuration. In
ED

addition, the amount of tar formed from cellulose during the pyrolysis or catalytic
pyrolysis was five times more than that of lignin. The pyrolysis of cellulose also
forms the π–conjugated structures that are different from that of lignin.
PT

The acidic oxide catalysts generally promoted the formation of carboxylic acids,
aldehydes/ketones, furans and phenolics, while the basic oxides like CaO could
E

remarkably suppress the formation of the phenolics. The coke formation over the
CC

basic oxides were also much more significant than that over the acidic oxides. The
derivatives from the pyrolysis of cellulose contributed more towards the coke formed.
A

The structure of the coke formed from cellulose contained long aliphatic chains. The
heating of the coke in inert gas was accompanied with the release of a substantial
amount of CO2, CO, H2, CH4 via probably decarboxylation/decarbonylation,
dehydrogenation and etc. The coke from the pyrolysis of lignin was much more stable.
It was also found that CaO and La2O3 could react with the CO2 produced in the
23
pyrolysis process and form the carbonates, while MgO could not. The TPO–MS
characterization showed that the coke species were multiple types over CaO, while
that over MgO was a single type. The properties of the coke and the effects of the
cellulose–derivatives and the lignin–derivatives on the structural configuration of the
coke needs further investigation. The results from this work also indicated that the
acidity/alkalinity of the catalysts significantly impacted the pyrolysis process, which
needs to be taken into consideration during the design of the catalysts for pyrolysis of

T
biomass.

R IP
Acknowledgements

SC
This work was supported by the Strategic International Scientific and Technological
Innovation Cooperation Special Funds of National Key R&D Program of China (No.

U
2016YFE0204000), the Program for Taishan Scholars of Shandong Province
N
Government, the Recruitment Program of Global Young Experts (Thousand Youth
A
Talents Plan), Natural Science Fund of Shandong Province (ZR2017BB002) and the
key research and development program of Shandong Province (2018GSF116014).
M
ED

References
[1] S. Wang, G. Dai, H. Yang, Z. Luo, Lignocellulosic biomass pyroly mechanism: a
state–of–the–art review, Prog. Energy Combust. Sci. 163 (2017) 33–86.
PT

[2] C. Tian, B. Li, Z. Liu, Y. Zhang, H. Lu, Hydrothermal liquefaction for algal
biorefinery: a critical review. Renew. Sust. Energ. Rev. 38 (2014) 933–950.
E

[3] L. Zhang, G. Hu, S. Hu, J. Xiang, X. Hu, Y. Wang, D. Geng, Hydrogenation of


CC

fourteen biomass-derived phenolics in water and in methanol: their distinct


reaction behaviours. Sustainable Energy & Fuels. 2.4 (2018) 751–758.
A

[4] Z. Ma, D. Chen, J.Gu, B. Bao, Q. Zhang, Determination of pyrolysis


characteristics and kinetics of palm kernel shell using TGA–FTIR and
model–free integral methods. Energ. Convers. Manage. 89 (2015) 251–259.
[5] L. Qiang, X. Yang, C. Dong, Z. Zhang, X. Zhang, and X. Zhu, Influence of
pyrolysis temperature and time on the cellulose fast pyrolysis products:
24
Analytical Py–GC/MS study. J. Anal. Appl. Pyrol. 92.2 (2011) 430–8.
[6] D. Mourant, C. Lievens, R. Gunawan, Y. Wang, X. Hu, L. Wu, SS. A Syed–Hassan,
C. Li, Effects of temperature on the yields and properties of bio-oil from the fast
pyrolysis of mallee bark. Fuel. 108 (2013) 400–408.
[7] S. Wang, X. Guo, T. Liang, Y. Zhou, Z. Luo, Mechanism research on cellulose
pyrolysis by Py–GC/MS and subsequent density functional theory studies.
Bioresource Technol. 104 (2012) 722–728.

T
[8] S. Xiong, J. Zhuo, B. Zhang, Q. Yao. Effect of moisture content on the

IP
characterization of products from the pyrolysis of sewage sludge. J. Anal. Appl.

R
Pyrol. 104 (2013) 632–639.

SC
[9] Q. Cai, J. Xu, S. Zhang. Upgrading of Bio-oil Aqueous Fraction by Dual–Stage
Hydrotreating–Cocracking with Methanol. ACS Sustain. Chem. Eng. 5.7 (2017)
6329–6342.
U
N
[10] J. Chen, Q. Cai, L. Lu, F. Leng, S. Wang, Upgrading of the acid–rich fraction of
A
bio-oil by catalytic hydrogenation–esterification. ACS Sustain. Chem. Eng. 5.1
(2017) 1073–1081
M

[11] L. Busetto, D. Fabbri, R. Mazzoni, M. Salmi, C. Torri, V. Zanotti, Application of


ED

the Shvo catalyst in homogeneous hydrogenation of bio-oil obtained from


pyrolysis of white poplar: new mild upgrading conditions. Fuel. 90.3 (2011)
1197–1207.
PT

[12] S. Wang, Y. Gu, Q. Liu, Y.Yao, Z. Guo, Z. Luo, K. Cen, Separation of bio-oil by
molecular distillation. Fuel Process. Technol. 90.5 (2009) 738–745.
E

[13] JS. Kim, Production, separation and applications of phenolic–rich bio-oil–a


CC

review. Bioresource Technol. 178 (2015) 90–8.


[14] S. Wang, X. Guo, K. Wang, Z. Luo, Influence of the interaction of components
A

on the pyrolysis behavior of biomass. J. Anal. Appl. Pyrol. 91.1 (2011) 183–189.
[15] T. He, Y. Zhang, Y. Zhu, W. Wen, Y. Pan, J. Wu, J. Wu, Pyrolysis mechanism
study of lignin model compounds by synchrotron vacuum ultraviolet
photoionization mass spectrometry. Energ. Fuel. 30.3 (2016) 2204–8.
[16] S. Wang, B. Ru, H. Lin, W. Sun, Z. Luo, Pyrolysis behaviors of four lignin
25
polymers isolated from the same pine wood. Bioresource Technol. 182 (2015)
120–7.
[17] X. Hu, R. Gunawan, D. Mourant, C. Lievens, X. Li, Shu Zhang, Weerawut
Chaiwat, Chun–Zhu Li, Acid–catalysed reactions between methanol and the
bio-oil from the fast pyrolysis of mallee bark. Fuel. 97 (2012) 512–522.
[18] X. Zhang, L. Sun, L. Chen, X. Xie, B. Zhao, H. Si, G Meng, Comparison of
catalytic upgrading of biomass fast pyrolysis vapors over CaO and Fe (III)/CaO

T
catalysts. J. Anal. Appl. Pyrol. 108 (2014) 35–40.

IP
[19] A.Sanna, TP. Vispute, GW. Huber. Hydrodeoxygenation of the aqueous fraction

R
of bio-oil with Ru/C and Pt/C catalysts. Appl. Catal. B–Environ. 165 (2015)

SC
446–56.
[20] X. Hu, C. Lievens, D. Mourant, Y. Wang, L. Wu, R. Gunawan, Y. Song, C. Li.

U
Investigation of deactivation mechanisms of a solid acid catalyst during
N
esterification of the bio-oils from mallee biomass. Appl. Energy. 11 (2013)
A
94–103.
[21] S. Wang, Q. Cai, J. Chen, L. Zhang, L. Zhu, Z. Luo, Co-cracking of bio-oil
M

model compound mixtures and ethanol over different metal oxide-modified


ED

HZSM-5 catalysts. Fuel. 160 (2015) 534-543.


[22] EF. Iliopoulou, EV. Antonakou, SA. Karakoulia, IA. Vasalos, AA. Lappas, KS.
Triantafyllidis. Catalytic conversion of biomass pyrolysis products by
PT

mesoporous materials: effect of steam stability and acidity of Al–MCM–41


catalysts. Chem. Eng. J. 134.1–3 (2007) 51–7.
E

[23] P. Pan, C. Hu, W. Yang, Y. Li, L. Dong, L. Zhu, D. Tong, R. Qing, Y. Fan, The
CC

direct pyrolysis and catalytic pyrolysis of Nannochloropsis sp. residue for


renewable bio-oils. Bioresource Technol. 101.12 (2011) 4593–9.
A

[24] EF. Iliopoulou, SD. Stefanidi, KG Kalogiannis, A. Delimitis, AA. Lappas, KS.
Triantafyllidis, Catalytic upgrading of biomass pyrolysis vapors using transition
metal–modified ZSM–5 zeolite. Appl. Catal. B–Environ. 127 (2012) 281–90.
[25] H. Yuan, S. Xing, Huhetaoli, T. Lu, Y. Chen, Influences of copper on the
pyrolysis process of demineralized wood dust through thermogravimetric and
26
Py–GC/MS analysis. J. Anal. Appl. Pyrol. 112(2015)325–332.
[26] C. Liu, H. Wang, AM. Karim, J. Sun, Y. Wang, Catalytic fast pyrolysis of
lignocellulosic biomass. Chem. Soc. Rev. 43.22 (2014) 7594–623.
[27] S. Eibner, F. Broust, J. Blin, A Julbe, Catalytic effect of metal nitrate salts during
pyrolysis of impregnated biomass. J. Anal. Appl. Pyrol. 113 (2015) 143–52.
[28] Z Zhang, X Hu, J Li, G Gao, D Dong, R Westerhof, S Hu, J Xiang, Y Wang,
Steam reforming of acetic acid over Ni/Al2O3 catalysts: Correlation of nickel

T
loading with properties and catalytic behaviors of the catalysts. Fuel. 217 (2018)

IP
389–403.

R
[29] C. Liu, H. Wang, AM. Karim, J. Sun, Y. Wang, Catalytic fast pyrolysis of

SC
lignocellulosic biomass. Chem. Soc. Rev. 43.22 (2014) 7594–7623.
[30] Z Yu, X Hu, P Jia, Z Zhang, D Dong, G Hu, S Hu, Y Wang, J Xiang, Steam

U
reforming of acetic acid over nickel-based catalysts: The intrinsic effects of
N
nickel precursors on behaviors of nickel catalysts. Appl. Catal. B–Environ. 2018,
A
DOI: 10.1016/j.apcatb.2018.06.020.
[31] P. Adisak. Bio-oil production via fast pyrolysis of biomass residues from cassava
M

plants in a fluidised–bed reactor. Bioresource Technol. 102.2 (2011) 1959–1967.


ED

[32] AK. Zak, ME. Abrishami, WA. Majid, R. Yousefi, SM. Hosseini, Effects of
annealing temperature on some structural and optical properties of ZnO
nanoparticles prepared by a modified sol–gel combustion method. Ceram. Int.
PT

37.1 (2011) 393–398.


[33] MY. Nassar, IS. Ahmed, I. Samir, A novel synthetic route for magnesium
E

aluminate (MgAl2O4) nanoparticles using sol–gel auto combustion method and


CC

their photocatalytic properties. Spectrochim. Acta, Part A. 131 (2014) 329–334.


[34] Y Song,Y Wang, X Hu, S Hu, J Xiang, L Zhang, S Zhang, Z Min, CZ Li. Effects
A

of volatile–char interactions on in situ destruction of nascent tar during the


pyrolysis and gasification of biomass. Part I. Roles of nascent char. Fuel. 122
(2014) 60–66.
[35] Y. Wang, X. Li, D. Mouran, R. Gunawan, S. Zhang, CZ. Li, Formation of
aromatic structures during the pyrolysis of bio-oil. Energ. Fuel. 26 (2012)
27
241–247
[36] Z. Min, M. Asadullah, P. Yimsiri, S. Zhang, H, Wu, CZ. Li. Catalytic reforming
of tar during gasification. Part I. Steam reforming of biomass tar using ilmenite
as a catalyst. Fuel. 90 (2011) 1847–1854.
[37] S. Wang, B. Ru, G. Dai, W. Sun, K. Qiu, J. Zhou, Pyrolysis mechanism study of
minimally damaged hemicellulose polymers isolated from agricultural waste
straw samples. Bioresource Technol. 190 (2015) 211–218.

T
[38] S. Xin, H. Yang, Y. Chen, M. Yang, L Chen, X. Wang, H. Chen. Chemical

IP
structure evolution of char during the pyrolysis of cellulose. J. Anal. Appl. Pyrol.

R
116 (2015) 263–71.

SC
[39] S. Wang, B. Ru, H. Lin, Z. Luo, Degradation mechanism of monosaccharides and
xylan under pyrolytic conditions with theoretic modeling on the energy profiles.
Bioresource Technol. 143 (2013) 378–383.
U
N
[40] S. Wang, K. Wang, Q. Liu, Y. Gu, Z. Luo, K. Cen, T. Fransson, Comparison of
A
the pyrolysis behavior of lignins from different tree species. Biotechnol. Adv.
27.5 (2009) 562–567
M

[41] J. Yu, N. Paterson, J. Blamey, M. Millan, Cellulose, xylan and lignin interactions
ED

during pyrolysis of lignocellulosic biomass. Fuel. 191 (2017) 140–9.


[42] S. Wang, B. Ru, G. Dai, Z. Shi, J. Zhou, Z. Luo, M. Ni, K. Cen, Mechanism
study on the pyrolysis of a synthetic β-O-4 dimer as lignin model compound. P.
PT

Combust. Inst. 36.2 (2017) 2225–2233.


[43] X. Hu, Y. Wang, D. Mourant, R. Gunawan, C. Lievens, W. Chaiwat, M.
E

Gholizadeh, L. Wu, X. Li, C.-Z. Li, Polymerization upon heating up of bio-oil: A


CC

model compound study, AIChE J. 59 (2013) 888–900.


[44] X. Hu, L. Wu, Y. Wang, D. Mourant, C. Lievens, R. Gunawan, CZ. Li, Mediating
A

acid-catalyzed conversion of levoglucosan into platform chemicals with various


solvents. Green Chem. 14.11 (2012) 3087–3098.
[45] X. Li, R. Gunawan, C. Lievens, Y. Wang, D, Mourant, S. Wang, H. Wu, M.
Garcia-Perez, CZ. Li, Simultaneous catalytic esterification of carboxylic acids
and acetalisation of aldehydes in a fast pyrolysis bio-oil from mallee biomass.
28
Fuel. 90 (2011) 2530–2537.
[46] S. Jiang, X Hu, D Xia, CZ Li. Formation of aromatic ring structures during the
thermal treatment of mallee wood cylinders at low temperature. Appl. Energ. 183
(2016) 542–551.
[47] Y. Song, Y. Wang, X. Hu, S. Hu, J. Xiang, L. Zhang, S Zhang, Z. Min, CZ. Li.
Effects of volatile–char interactions on in situ destruction of nascent tar during
the pyrolysis and gasification of biomass. Part I. Roles of nascent char. Fuel. 122

T
(2014) 60–6.

IP
[48] R. Lu, GP. Sheng, YY. Hu, P. Zheng, H. Jiang, Y. Tang, HQ. Yu, Fractional

R
characterization of a bio-oil derived from rice husk. Bimass Bioenerg. 35 (2011)

SC
671–678.
[49] R. Yin, R. Liu, Y. Mei, W. Fei, X. Sun, Characterization of bio-oil and bio–char

U
obtained from sweet sorghum bagasse fast pyrolysis with fractional condensers.
N
Fuel. 112 (2013) 96–104.
A
[50] L. Wu, X. Hu, D. Mourant, Y. Wang, C. Kelly, M. Garcia-Perez, M. He, C.-Z. Li,
Quantification of strong and weak acidities in bio-oil via non-aqueous
M

potentiometric titration, Fuel. 115 (2014) 652–657.


ED

[51] B. Zhang, Z. Zhong, P. Chen, R. Ruan, Microwave–assisted catalytic fast


pyrolysis of biomass for bio-oil production using chemical vapor deposition
modified HZSM–5 catalyst. Bioresource Technol. 197 (2015) 79–84.
PT

[52] M. Gholizadeh, R. Gunawan, X. Hu, F. M. Mercader, R. Westerhof, W. Chaitwat,


MM Hasan, D. Mourant, CZ. Li, Effects of temperature on the hydrotreatment
E

behaviour of pyrolysis bio-oil and coke formation in a continuous


CC

hydrotreatment reactor. Fuel Process. Technol. 148 (2016) 175–83.


[53] A.Veses, B. Puértolas, MS. Callén, T. García, Catalytic upgrading of biomass
A

derived pyrolysis vapors over metal–loaded ZSM–5 zeolites: Effect of different


metal cations on the bio-oil final properties. Microporou. Mesoporou. Mat. 209
(2015) 189–96.
[54] X. Zhao, H. Li, J. Zhang, L Shi, D Zhang. Design and synthesis of NiCe@
m–SiO2 yolk–shell framework catalysts with improved coke–and
29
sintering–resistance in dry reforming of methane. Int. J. Hydrogen Energy. 41.4
(2016) 2447–56.
[55] Y. Song, Y. Xu, Y. Suzuki, H. Nakagome, X. Ma, ZG. Zhang. The distribution of
coke formed over a multilayer Mo/HZSM–5 fixed bed in H2 co–fed methane
aromatization at 1073 K: Exploration of the coking pathway. J. Catal. 330 (2015)
261–72.

T
R IP
SC
U
N
A
M
ED
E PT
CC
A

30
Table 1 Elemental analysis of feedstock, char and tar.

Mass% (organic basis)


C/H
N C H S O ratio
Feedstock
Poplar – 43.1 5.4 – 51.5 0.67

T
Cellulose – 40.9 6.4 – 52.7 0.53

IP
Lignin – 48.3 5.1 6.4 46.6 0.79

R
Char

– –

SC
Poplar 77.6 6.3 16.1 1.0

Cellulose – 82.7 5.2 – 12.0 1.3

Lignin – 64.6 4.6 – 30.3 1.2

Tar of poplar U
N
Blank 0.93 58.2 6.3 – 34.6 0.77
A
Al2O3 1.0 59.7 6.4 – 32.9 0.78
M

SiO2 1.3 57.1 9.5 – 32.0 0.50

ZnO 1.5 57.6 9.5 – 31.4 0.50


ED

K2O 1.7 59.0 10.2 – 29.1 0.48


MgO 1.7 53.7 10.4 – 34.2 0.43
CaO 1.9 61.1 10.4 – 26.6 0.49
PT

La2O3 1.6 58.7 9.9 – 29.8 0.50


Tar of cellulose
Blank – 47.5 9.2 – 43.3 0.43
E

MgO – 47.9 8.5 – 43.6 0.47


– –
CC

CaO 48.6 8.9 42.5 0.45


Tar of lignin
Blank 1.2 40.3 7.7 11.7 39.2 0.44
MgO 0.50 38.9 7.9 12.2 40.5 0.41
A

CaO 1.8 44.8 8.0 4.3 41.1 0.47

31
Table 2 Elemental analysis of coke on the catalyst.

Catalysts Mass% (organic basis) C/H ratio


N C H S
Poplar

T
Al2O3 0.05 1.3 0.23 – 0.49
SiO2 – 0.31 0.05 – 0.51

IP
ZnO – 1.7 0.11 – 1.4
K2O – 2.6 0.34 – 0.63

R
MgO – 8.9 1.10 – 0.68

SC
CaO 0.75 21.3 0.75 – 2.4
La2O3 0.18 6.3 0.17 – 3.0
Cellulose
MgO
CaO


10.9
7.9
1.3
0.71 U –

0.69
0.93
N
Lignin
MgO – 6.0 0.76 – 0.66
A
CaO – 13.1 0.65 – 1.7
M
ED
E PT
CC
A

32
(a)
35 (b) 35
Gas
Gas
30 Tar
Tar
30
25
25

Yield (wt%)
Yield (wt%)

20
20
15
15
10
10
5
5
0
0

T
Blank Al2O3 SiO2 ZnO K 2O MgO CaO La2O3
Blank MgO CaO
Catalyst varieties Catalyst varieties

IP
(c) (d)
35 60
Gas Char
30 Tar
50

R
25
40
Yield (wt%)

SC
20
Yield (wt%)
30
15
20
10

0
10

U
N
0
Blank MgO CaO Poplar Cellulose Lignin
Catalyst varieties Feedstock type
A
Figure 1 The gas and tar yields from pyrolysis of the different feedstocks over the
M

various catalysts. (a) feedstock: poplar; (b) feedstock: cellulose; (c) feedstock: lignin.
(d) char yields.
ED
E PT
CC
A

33
(a) (b)
150 Cellulose 150 Lignin
CO2 CO2
120 CH4 120 CH4
CO CO
Intensity (a.u.)

Intensity (a.u.)
H2 H2
90 90

60 60

30 30

0 0
0 400 800 1200 1600 0 400 800 1200 1600
(c)

T
Time (s) (d) Time (s)
150 Cellulose with MgO 150 Cellulose with CaO

IP
CO2 CO2
CH4 CH4
120 120
Intensity (a.u.)

Intensity (a.u.)
CO CO
H2 H2

R
90 90

60 60

SC
30 30

0 0

(f)
0 400 800
Time (s)
1200 1600
(f)
0
U400 800
Time (s)
1200 1600
N
150 Lignin with MgO 150 Lignin with CaO
CO2
CO2
A
120 120 CH4
CH4
Intensity (a.u.)

Intensity (a.u.)

CO
CO H2
90 H2 90
M

60 60

30 30
ED

0 0

0 400 800 1200 1600 0 400 800 1200 1600


Time (s) Time (s)
PT

Figure 2 Gaseous product distribution versus the time during the catalytic pyrolysis
experiments. Reaction conditions: T = 500°C, N2 = 120 ml/min, P = 1 atm. The “0 (s)”
E

in the “X-axis” meant the gases generated from pyrolysis just reached the gas sensors.
CC
A

34
(a) (b)
2.4x107 Acetic acid Propionaldehyde
Butyrate 2.4x106 2-Hydroxyacetaldehyde
Butanedial

1.6x107
Intensity

Intensity
1.6x106

8.0x106 8.0x105

T
0.0 0.0
Blank Al2O3 SiO2 ZnO K2O MgO CaO La2O3 Blank Al2O3 SiO2 ZnO K2O MgO CaO La2O3

IP
Catalyst varieties Catalyst varieties
(c) (d)
4x106 Butanone Cyclopentanone
Hydroxy-acetone 3.0x106 2-Cyclopentene

R
Hydroxy-2-butanone Cyclopentanedione
3-Methy-2-cyclopentanedione
3x106 2.4x106

SC
Intensity
Intensity

1.8x106
2x106

1.2x106

1x106
6.0x105

U
N
0 0.0
Blank Al2O3 SiO2 ZnO K2O MgO CaO La2O3 Blank Al2O3 SiO2 ZnO K2O MgO CaO La2O3
Catalyst varieties Catalyst varieties
A
(e) (f)
Furyl methyl ketone Phenol
6 1.5x107
5x10 Furfural 4-Methyl-phenol
M

Furfuralcohol 3-Methyl-phenol
5-Methylfurfural
4x106 HMF
1.0x107
Intensity

Intensity

3x106
ED

2x106
5.0x106

1x106
PT

0 0.0
Blank Al2O3 SiO2 ZnO K2O MgO CaO La2O3 Blank Al2O3 SiO2 ZnO K2O MgO CaO La2O3
Catalyst varieties Catalyst varieties
(g) (h)
E

3x106 Guaiacol Eugenol


Iso-eugenol Propylene lilac
Propenylguaiacol 2.4x106
CC

4-Allyl-2,6-dimethoxyphenol

2x106
Intensity

Intensity

1.6x106
A

1x106
8.0x105

0 0.0
Blank Al2O3 SiO2 ZnO K2O MgO CaO La2O3 Blank Al2O3 SiO2 ZnO K2O MgO CaO La2O3
Catalyst varieties Catalyst varieties

Figure 3 A list of compounds found in the bio-oils from the pyrolysis of poplar over
the various oxide catalysts.
35
(a) (b)
0.6 OH Blank 0.6 OH K2O
CC CC
Al2O3 CH3 MgO CH3
0.5 SiO2 CC CO 0.5 CaO CO CC CO
CO
ZnO CHOH La2O3 CHOH

Absorbance (a.u.)
CH
Absorbance (a.u.)

CH
0.4 Aromatics 0.4 Aromatics

0.3 0.3

0.2 0.2

0.1 0.1

0.0 0.0
4000 3500 3000 2500 2000 1500 1000 500 4000 3500 3000 2500 2000 1500 1000 500

T
Wavenumber (cm-1) Wavenumber (cm-1)
(c) (d)
0.6 0.6 Lignin
Cellulose

IP
OH
Blank
0.5 Blank 0.5 MgO CO
OH MgO OH CaO
CO

Absorbance (a.u.)
Absorbance (a.u.)

CaO CH3
CO 0.4
0.4 CH CC
CH CO Aromatics CH CC
CO

R
CO Aromatics
CH
CO
0.3 0.3

SC
0.2 0.2

0.1 0.1

0.0
0.0
4000 3500 3000 2500 2000
Wavenumber (cm )
1500
-1
1000 500 4000

U
3500 3000 2500 2000
Wavenumber (cm-1)
1500 1000 500
N
Figure 4 FT–IR characterization of tar produced over different oxide catalysts or
A
feedstocks: (a) Poplar; (b) Poplar; (c) Cellulose; (d) Lignin.
M
ED
E PT
CC
A

36
(a) 0.20 (b) 0.20
100 100 Poplar with CaO
Poplar
90 TG
TG
90 CO
CO
0.15 80 CO2 0.15
80 CO2

Weight loss (%)


70 H2
Weight loss (%)

Intensity (a.u.)
H2

Intensity (a.u.)
70 CH4
CH4
60 C2H4
60 C2H4 0.10 0.10
50
50 40
40 0.05 30 0.05
30 20
20 10
0.00 0.00
10 0
100 200 300 400 500 600 700 800 100 200 300 400 500 600 700 800
Temperature (C)

T
Temperature (C)
(c) 0.20 (d) 0.20
100 100
Cellulose Lignin

IP
90 TG TG
CO
CO 80 0.15
80 CO2
0.15 CO2

Weight loss (%)

Intensity (a.u.)
H2
Weight loss (%)

70

Intensity (a.u.)
H2
CH4

R
60 CH4 60
0.10 C2H4 0.10
C 2H 4
50
40

SC
40
30 0.05 0.05

20 20

10 0.00
0.00 0
0
100 200 300 400 500
Temperature (C)
600 700 800

U 100 200 300 400 500


Temperature (C)
600 700 800
N
Figure 5 Weight loss of the tar and abundance of the gaseous products during the
A
TG–MS tests in the N2 atmosphere: (a) Poplar; (b) Poplar with CaO; (c) Cellulose (d)
M

Lignin.
ED
E PT
CC
A

37
(a)
Poplar
CC Aromatics
CO2
3
CO2 CC CO
CO CH
CH4 OH
Absorbance (a.u.)

CO
OH CH
700C
2
600C

500C

400C
1 300C

200C
100C
0
4000 3500 3000 2500 2000 1500 1000
Wavenumber (cm-1)

T
(c) 4.5
CO2

IP
Poplar with CaO CC
4.0 Aromatics

CO2 CC CO


3.5 CO CH
OH
Absorbance (a.u.)

CH4
CO

R
3.0 OH CH

700C
2.5
600C

SC
2.0 500C

1.5
400C
1.0 300C

0.5 200C

0.0
4000
100C

3500 3000 2500 2000 1500 1000


U
N
Wavenumber (cm-1)
(e) 3.0
Cellulose CO2
A
Aromatics
CC
2.5
CC CO
CO2 CO
OH
Absorbance (a.u.)

CH4
2.0
M

CO CH
700C OH CH

1.5 600C

500C
1.0 400C
ED

300C
0.5

0.0 200C
100C

-0.5
PT

4000 3500 3000 2500 2000 1500 1000


Wavenumber (cm-1)
(e) 4 CO2
Lignin CC Aromatics
CO
E

CO2 CC CO


CO CH
3 700C
CH4 OH
Absorbance (a.u.)

CO
OH CH
CC

600C
2
500C

400C
1
300C
A

200C
100C
0
4000 3500 3000 2500 2000 1500 1000
Wavenumber (cm-1)

Figure 6 DRIFTS characterizations of the changes of the functionalities of the tarry


products versus temperature: (a) and (b): Poplar; (c) and (d): Poplar with CaO; (e) and
(f): Cellulose, (g) and (h): Lignin.
38
(a) (b)
100 Al2O3
0.024 SiO2
Al2O3
90 ZnO
SiO2 K2O
Weight loss (%)

80 ZnO 0.018 MgO


K2O CaO

DTG
MgO La2O3
70
CaO 0.012
La2O3
60
0.006
50

40 0.000
100 200 300 400 500 600 700 800
200 300 400 500 600 700 800

T
(c) Temperature (C) (d) Temperature (C)

IP
100 Poplar
0.0024 Cellulose
Lignin
MgO
95
Weight loss (%)

Poplar

R
Cellulose DTG 0.0018
Lignin
90

SC
0.0012

85
0.0006

80

100 200 300 400 500


Temperature (C)
600 700 800
0.0000

(f)
100 200
U 300 400 500 600 700 800
N
Temperature (C)
(e)
100 0.012 Poplar
Cellulose
A
CaO 0.010 Lignin
90
Poplar
Weight loss (%)

Cellulose 0.008
M

80
DTG

Lignin
0.006
70
0.004
60
ED

0.002
50
0.000
40
100 200 300 400 500 600 700 800 100 200 300 400 500 600 700 800
Temperature (C)
PT

Temperature (C)

Figure 7 TG and DTG characterization of the catalysts used in the pyrolysis of


different feedstock: (a) and (b) feedstock: Poplar; (c) and (d) catalyst: MgO; (e) and (f)
E

catalyst: CaO.
CC
A

39
(a) 0.6 (b) 0.6
100 100
0.5 0.5
99 Al2O3 SiO2
0.4
Weight loss (%)

0.4

Weight loss (%)

Intensity (a.u.)
TG

Intensity (a.u.)
TG
CO 99 CO
98 0.3
CO2 CO2 0.3
H2 H2
97 CH4 0.2 CH4
98 0.2
C2H4 C2H4
96 0.1 0.1

95 0.0 97
0.0

200 300 400 500 600 700 800 200 300 400 500 600 700 800

T
Temperature (C) Temperature (C)
(c) 0.6 (d) 0.6

IP
100
0.5 0.5
100
ZnO K2O

Intensity (a.u.)
TG 0.4 99 0.4
Weight loss (%)

Weight loss (%)


TG

Intensity (a.u.)
CO

R
CO
CO2
98 0.3 CO2 0.3
H2
98 H2
CH4

SC
0.2 CH4
C 2H4 0.2
C2H4
96 97
0.1 0.1

0.0 96 0.0

(e)
94
200 300 400 500
Temperature (C)
600 700 800

(f) U
200 300 400
Temperature (C)
500 600 700 800
N
0.6 4.0
100
100 3.5
0.5 CaO
A
MgO 90
TG TG 3.0
96 0.4 CO
Weight loss (%)

Weight loss (%)

CO
80
Intensity (a.u.)

CO2 2.5

Intensity (a.u.)
CO2
M

H2
H2 0.3
CH4 2.0
92 CH4 70
C2H4
C2H4 0.2 1.5
60
88 1.0
0.1
ED

50 0.5
84 0.0
0.0
40
200 300 400 500 600 700 800 200 300 400 500 600 700 800
Temperature (C) Temperature (C)
PT

(g) 2.0
100
La2O3
TG 1.6
E

98
CO
Weight loss (%)

CO2
Intensity (a.u.)

96 H2 1.2
CC

CH4
94 C2H4
0.8
92
0.4
A

90
0.0
88
200 300 400 500 600 700 800
Temperature (C)

Figure 8 Weight loss of the spent catalysts used in pyrolysis of poplar and the
released gaseous products during the TG–MS characterisation in N2 atmosphere.
40
(a) 0.6 (b)
100 3.6
100 CaO
0.5
TG
90 3.0
CO
MgO

Weight loss (%)


0.4
Weight loss (%)

Intensity (a.u.)
CO2
96

Intensity (a.u.)
TG 80 2.4
CO H2
CO2 0.3 CH4
H2 70 C 2H 4 1.8
92 CH4 0.2
C2H4 60 1.2
0.1
88 50 0.6
0.0
40 0.0
200 300 400 500 600 700 800 200 300 400 500 600 700 800

Temperature (C) Temperature (C)

T
(b) 0.6 (d)

IP
100 3.6
100
0.5
MgO
3.0

Intensity (a.u.)
TG 90 CaO
0.4

Weight loss (%)


Weight loss (%)

CO

Intensity (a.u.)
TG

R
96 CO2 2.4
CO
H2 0.3 80 CO2
CH4 H2 1.8

SC
C2H4 0.2 CH4
92 70
C2H4 1.2
0.1
60 0.6
88 0.0

200 300 400 500 600


Temperature (C)
700 800
50
200
U 300 400 500 600
Temperature (C)
700 800
0.0
N
A
Figure 9 Weight loss of the spent catalysts used in pyrolysis of cellulose (a and b) and
M

lignin (c, d), and the released gaseous products during the TG–MS characterisation in
N2 atmosphere.
ED
E PT
CC
A

41
(a) (b)
Fresh catalyst Al2O3  ZnO

After the use in pyrolysis 
 
  
 
    

Intensity (a.u.)
Intensity (a.u.)

 SiO2  :K2CO3
 
  
     
     

T
20 40 60 80 20 40 60 80
2 Theta (degree) 2 Theta (degree)
(d)

IP
(c)
:MgO La2O3

La2CO5

R
   

Intensity (a.u.) 


Intensity (a.u.)

 

SC
 

 CaCO3
CaO
 
 
  





  U   



N
20 40 60 80 20 40 60 80
2 Theta (degree) 2 Theta (degree)
A
Figure 10 XRD patterns of the catalysts before and after the use in the pyrolysis
M

experiments.
ED
E PT
CC
A

42
(a) (b)
24 60

50
22

Weight loss in TG (wt%)


Tar yield (wt%)

40

20 30

20
18
10
16
0

T
Al2O3 SiO2 ZnO K2O MgO CaO La2O3 Al2O3 SiO2 ZnO K2O MgO CaO La2O3

IP
(c) (d)

R
Phenol
Hydroxy-acetone
3-Methyl-phenol 2-Hydroxyacetaldehyde

SC
Acetic acid Furfural
Levoglucosan
Intensity

Intensity

U
N
A
Al2O3 SiO2 ZnO K2O MgO CaO La2O3 Al2O3 SiO2 ZnO K2O MgO CaO La2O3
M

Figure 11 (a) The tar and gas yields over different catalysts in the pyrolysis of poplar;
ED

(b) The weight loss of the catalysts in the TG characterisation in air stream; (c and d)
A list of compounds found in the bio-oils from the pyrolysis of poplar over the
different catalysts.
E PT
CC
A

43

Vous aimerez peut-être aussi