See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/267990372

Characterisation of Oil Palm Empty Fruit Bunches for Fuel Application

Article  in  Journal of Physical Therapy Science · January 2011

CITATIONS READS

107 2,194

3 authors, including:

Nurhayati Abdullah
Universiti Sains Malaysia, George Town, Malaysia
54 PUBLICATIONS   1,248 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Characteristics of graphene produce from biochar View project

Bioenergy and Biofuels study View project

All content following this page was uploaded by Nurhayati Abdullah on 04 June 2015.

The user has requested enhancement of the downloaded file.

Journal of Physical Science, Vol. 22(1), 1–24, 2011

Characterisation of Oil Palm Empty Fruit Bunches for

Fuel Application

N. Abdullah1*, F. Sulaiman1 and H. Gerhauser2

1
School of Physics, Universiti Sains Malaysia, 11800 USM Pulau Pinang, Malaysia
2
Biomass, Coal and the Environment Unit, Energy Research Centre of the Netherlands
(ECN), 1755 ZG Petten, The Netherlands

*Corresponding author: nurhaya@usm.my

Abstract: This study was an attempt to produce bio-oil from empty fruit bunches (EFB)
of oil palm waste using fast pyrolysis technology. A 150 g/h fluidised bed bench scale fast
pyrolysis unit operating at atmospheric pressure was used to obtain the pyrolysis liquid.
A comparison of the elemental composition of unwashed and washed feedstock was made
in this study. With the five methods of treatment being considered, elements such as Al, P,
Cl, Ti, Fe and Cu were removed during the washing. However, Na, S and K decreased
with the reduction of the ash content of the feedstock. The properties of the liquid product
were analysed and compared with wood derived bio-oil and petroleum fuels. The liquids
produced had high acid content, with a High Heating Value (HHV) of about 50% of
conventional petroleum fuel. The char content was in the range of 0.2–2.0%. The
composition and particle size distribution of the washed and unwashed feedstock were
determined, and the thermal degradation behaviour was analysed by thermogravimetric
analysis (TGA). The product yield for washed EFB was similar to that of low ash wood,
while the product yield for unwashed EFB was much closer to that of higher ash
feedstock. The pyrolysis liquids derived from unwashed EFB were found to be in both an
aqueous and an organic phase, which presents challenges for their commercial
application as a fuel. Some possibilities of upgrading were also discussed in this work.

Keywords: Empty fruit bunches, fast pyrolysis, bio-oil, ash, washed feedstock, TGA

1. INTRODUCTION

Palm oil production is on a steeply rising path. The empty fruit bunches
(EFB) are waste by-products that are being investigated for further uses.
Currently, palm oil mills typically use the shell and the drier part of the fibre
product stream, rather than EFB, to fuel their boilers, as the raw EFB contain
nearly 60% water.1

Fast pyrolysis represents a potential route to upgrade the EFB waste to

value added fuels and renewable chemicals. For example, the pyrolysis of woody
feedstock at temperatures around 500ºC, together with short vapour residence
times, will produce bio-oil yields of around 70% and char and gas yields of

© Penerbit Universiti Sains Malaysia, 2011

Characterisation of Oil Palm EFB 2

around 15% each.2–4 Bio-oil is a high-density oxygenated liquid, which can be

burned in diesel engines, turbines or boilers, though further work is still required
to evaluate the long-term reliability.5 It is also used for the production of
speciality chemicals, such as flavourings, which are the main products.
Renewable resins and slow release fertilisers are other potential applications,
which have been the subject of research.6 At this stage, fast pyrolysis is a novel
and relatively untested technology. There are several pilot plants in North
America and Europe, but there is no consistent track record yet outside of the
manufacture of flavourings. Of special significance to the work reported here is
that a fast pyrolysis pilot plant processing EFB has recently started production in
Malaysia.7

Di Blasi et al.8 investigated the pyrolysis of unwashed and washed straw.

A bench scale system was used to investigate the pyrolysis of these samples.
They found that the liquid yields increased with the decrease in char yields for
washed straw. For example, at a temperature of 500°C, the liquid yield is 43 mf
wt% for unwashed feedstock compared to 58 mf wt% for washed feedstock,
whereas the char yield is 28 mf wt% for unwashed feedstock compared to 23 mf
wt% for washed feedstock. Mullen et al.9 carried out fast pyrolysis experiments
on corn cobs and corn stover (stalks, leaves and husks), using a pilot scale
fluidised bed reactor. They found that the bio-oil yield was about 60% from the
corn crop residue feedstocks. The specific gravity of these bio-oils was greater
than 1.0 and the High Heating Value (HHV) was about 20 MJ/kg, thus, giving
these bio-oils 20–32 times the energy density of the biomass feedstocks on a
volume basis. They concluded that the transportation of bio-oil was more cost
effective than transportation of biomass.

The work on water washing has received only limited attention by

researchers. Jenkins et al.10 studied ash reduction in rice straw and wheat straw in
order to reduce slagging and fouling in furnaces and other thermal conversion
systems. Alkali in the ash of crop biomass fuels, mainly potassium and sodium,
creates serious slagging and fouling in conventional boilers. All components,
except silica and calcium, were significantly reduced, although only a small
proportion of the ash for rice straw was removed. Piyali Das et al.11 investigated
vacuum pyrolysis of deashed sugarcane bagasse. The objective of their work was
to understand the change in the quantity and quality of the oil fraction obtained
from the pyrolysis of the deashed biomass. They believed that the presence of
alkali metals in ash catalyses the polymerisation reactions in the oils and thereby
increases the oil viscosity.12,13 In the case of water leaching, it was found that the
extractives were washed out, which reduced their levels from 25.8 mf wt% to
13.1 mf wt%.
Journal of Physical Science, Vol. 22(1), 1–24, 2011 3

Scott et al.14 published a preliminary study of the removal rates of

alkaline cations present in poplar wood (potassium and calcium mainly) by
washing with hot water and mild acid. They found that hot water washing alone
is able to remove a major amount of the alkaline cations from wood as the results
indicated that over 90% of the potassium and over 65% of the calcium was
readily washed out by hot water. A mild acid can readily remove most of the
alkaline cations from biomass and it is more significant at a temperature of 60°C.

The palm oil industry is currently expanding rapidly and produces large
amounts of poorly utilised waste biomass. Therefore, an investigation of fast
pyrolysis technology to convert oil palm waste, such as EFB to liquids, is
worthwhile. Results for the production of bio-oil derived from unwashed EFB
were presented elsewhere,15 and it was found that in all cases, the liquid product
separated into two phases, which presented difficulties for fuel applications. The
aim of this work is to study fast pyrolysis of EFB in a fluidised bed reactor with a
nominal capacity of 150 g/h to produce bio-oil as a potential liquid fuel. Another
aim of this work is to study the effect of ash reduction on the yield and maximum
ash level to produce a homogenous bio-oil. The results are compared to those
from pyrolysis of various feedstocks with different ash contents.

2. MATERIALS AND METHODS

2.1 Feedstock Preparation

The EFB used in the experiments were supplied by the Malaysian Palm
Oil Board (MPOB). Samples received were relatively dry, having less than 10 mf
wt% moisture, and were in the form of whole bunches. Particle size reduction
was required to allow fast pyrolysis of the EFB on the available 150 g/h system.
The bunches were first manually chopped into smaller pieces that could be fed
into a shredder. Then, a Fritsch grinder with a screen size of 500 µm was used to
reduce the feedstock size to less than 500 µm. After extensive feeding trials, it
was found that only particles between 250–355 µm were easily fed. Both the size
fractions below and above this range frequently led to blockage of the available
feeder.

To reduce the ash content of the EFB by varying amounts, a range of

simple water washing techniques was employed. The treatments used for this
work are briefly described in Table 1. Notably, freshly sterilised empty fruit
bunches typically have a water content of 60% on a wet basis. Water washing, as
described in Abdullah,15 therefore does not have to bear a large additional cost
burden from drying, as drying is required in any case.
Characterisation of Oil Palm EFB 4

Table 1: Ash content of washed EFB according to washing method.

Ash content, mf wt% Water washing method
Soak 100 g of feedstock of size 250–355 µm for 24 hours at ambient
1.03
temperature in 7 litres of distilled water.
Soak 100 g of feedstock of size 2–3 cm for 20 minutes at ambient
2.14 temperature in 5 litres of distilled water (mentioned as method 3 in
Section 3.1.1 and shown in Table 3)
Soak 100 g of feedstock of size 2–3 cm for 10 minutes at ambient
3.05 temperature in 5 litres of distilled water (mentioned as method 2 in
Section 3.1.1 and shown in Table 3)
By manual agitation of 100 g of feedstock of size 2–3 cm for 1 minute
3.68 at ambient temperature in 5 litres of distilled water (mentioned as
method 1 in Section 3.1.1 and shown in Table 3)
5.43 Unwashed feedstock that was not subjected to any washing.

2.1.1 Surface structure and EDX analysis

Scanning Electron Microscope (SEM) analysis and Electron Dispersive

X-ray Spectroscopy (EDX) analysis were carried out on unwashed and washed
feedstock for a range of ash content in order to study their surface structure and
elemental composition. These analyses were performed at Aston Microscopy of
Engineering with the assistance of the expert operator, who used a Cambridge S-
90 Scanning Electron Microscope with digital computer imaging.

2.1.2 Thermogravimetric analysis of unwashed EFB and washed EFB

The thermal characteristics of washed and unwashed EFB were analysed

with a computerised Perkin-Elmer Pyris 1 thermogravimetric analyser.
Thermogravimetric analysis (TGA) was performed under 100 ml/min nitrogen
with a heating rate of 10 C/min.

2.2 Experimental Setup

A fluidised bed bench scale fast pyrolysis unit with a nominal capacity of
150 g/h available at University of Aston was employed for nearly all runs. This
reactor was operating at atmospheric pressure. Figure 1 shows a flow chart of the
fluidised bed pyrolysis process. The reactor was a stainless steel cylinder and the
heating medium in the reactor was inert sand. The fluidising gas was nitrogen,
which was preheated prior to entering the base of the reactor.
Journal of Physical Science, Vol. 22(1), 1–24, 2011 5

EFB

Grinder

GAS sampled to GC Sieve Feeder

Reactor

Liquid Cyclone Char

Condenser
collector collector

LIQUID CHAR

Figure 1: Flow chart of a fluidised bed pyrolysis process.

The char was carried out of the reactor body by the fluidising gas flow,
which is known as ‘blow-through’ mode.16 The char was then separated from the
product stream in a cyclone. The condensable vapours were collected in the
liquid products collection system, which consisted of two cooled condensers, an
electrostatic precipitator (EP) and a cotton wool filter. The incondensable gases
left the system through a gas meter and were then sampled by gas
chromatography (GC) to assess the quantity and type of gas produced.

To investigate the quality and characteristics of pyrolysis liquid of this

work, the pyrolysis experiment was carried out at a temperature of about 500°C
and at a residence time of 1.1 s on the washed feedstock with a size range of
355–600 µm and ash content of 1.06 mf wt% using a 1 kg/h rig. Even though
most of the pyrolysis liquids from washed feedstock were homogenous, the
liquids collected were not sufficient for other analyses.

The operation of the 1 kg/h system is similar to the 150 g/h system,
which also employed nitrogen as the fluidising gas and sand as the bed material.
However, the reaction products were passed through two cyclones where the char
was removed. The remaining vapours were quenched with cooled recycle isopar,
a proprietary isoparaffin supplied by Exxon Mobil Corporation. To remove
aerosols not captured by the quench column, the gas was then passed through an
EP, which was very effective at capturing aerosols. The liquids collected by the
Characterisation of Oil Palm EFB 6

quench column and EP drain into a common storage tank, which also contained
the isopar that is lighter than bio-oil. Bio-oil sank to the bottom of the tank,
whereas isopar accumulated at the top, where the pump inlet for the quench
column recycle pump was located.

2.2.1 Char content

A filtration method was used to determine the char content in pyrolysis

liquids.17 To enable the pyrolysis liquids to pass through the filter, the pyrolysis
liquids need to be diluted with a suitable solvent. Approximately 1 g of pyrolysis
liquids was weighed and then diluted in about 10 g of methanol before it was
filtered. The filter was then placed in an oven at 105 C for a minimum of 6 hours
and then cooled in a desiccator for 1 hour before weighing. The char content was
then calculated, and an average of three samples was taken. All measurements of
weight were taken up to four decimal places with an accuracy of ±0.001 g.

2.2.2 Viscosity

Viscosity describes a fluid's internal resistance to flow and can be

thought of as a measure of fluid friction. For fuel, the viscosity is important
because of its effect on pumping and injecting.18 Viscosity measurements were
carried out using capillary viscometers at a fixed temperature of 25 C. Only
about 4 ml of sample was required for the test using this type of viscometer. The
viscosity was measured by determining the time taken for a fixed volume of
pyrolysis liquids to flow between two marked points, and the viscosity was
calculated using the following equation:

V = Ct (1)

where,

V is the viscosity in centipoises, C is the viscometer constant and is not affected

by the temperature, and t is the time in seconds it takes for the liquid to fall from
the first timing mark to the second mark.

2.2.3 Stability

The Viscosity Index (VI) measures the relative change in viscosity with
time and is used to provide a measure of stability. VI was calculated using the
following equation:

VI = (µf - µi) / µi (2)

Journal of Physical Science, Vol. 22(1), 1–24, 2011 7

where,

µi is the initial viscosity, the viscosity of the fresh pyrolysis liquid tested at 25°C
after 24 hours from which it was produced, and µf is the final viscosity, the
viscosity of the pyrolysis liquid stored for 24 hours at 80°C, which is now
referred to as an aged pyrolysis liquid, as recommended by the Aston Viscosity
Index.

The viscosity of two pyrolysis liquids were re-tested at 25°C as well. A

lower VI number indicates greater stability of the pyrolysis liquid. A VI of zero
would be a completely stable liquid.

2.2.4 Fourier transform infra-red spectroscopy

Fourier transform infra-red (FTIR) is a fast scanning, high resolution,

high signal-to-noise ratio, and non-destructive analytical technique that is used to
ascertain the presence of specific functional groups in pyrolysis liquids. The
PerkinElmer Spectrum 100 Series FT-IR spectrometers were used in this work.
Analysis of FTIR spectra at wave numbers 4000 to 200 cm–1 of pyrolysis liquids
provides a quick and simple qualitative technique that uses the standard IR
spectra to identify the functional group(s) of the components of the derived
pyrolysis liquids.

2.2.5 Lignin content

The pyrolytic lignins were obtained from pyrolysis liquids by dropwise

addition of the pyrolysis liquids to distilled water19 under vigorous stirring (>
5000 rpm) using an ultra turrax. The ratio of pyrolysis liquids to water used in
this study was 1:50. After slowly adding the pyrolysis liquids to distilled water,
pyrolytic lignin precipitated. After precipitation, the liquid smoke (water soluble
part of pyrolysis liquids) was dissolved in a large amount of distilled water.
Subsequently, the solid was filtered and thoroughly washed until the filtrate
reached a neutral pH. Finally, the wet solids (pyrolytic lignins) were dried in a
vacuum at ambient temperature for 24 hours.

2.2.6 Gel permeation chromatography

The molar mass and molecular weight of pyrolytic lignins were

determined using a gel permeation chromatography (GPC) assay. The Waters 717
Plus Autosampler GPC system was used in this study. The chromatography was
set to 40°C and equipped with columns which have pore sizes of 50 and 100 Å
and lengths of 300 and 600 mm. The pyrolytic lignin sample was diluted in THF
(tetrahydrofuran, Aldrich 99.5% / HPLC grade) to make a 5 mg/ml solution. This
Characterisation of Oil Palm EFB 8

sample solution was filtered through filter paper with a 0.2 µm pore size to
remove insoluble impurities, and toluene was added into the sample solution to
act as a flow marker. About 100 µl of sample solution was injected into a mobile
phase to allow the sample to elute through the columns.

A series of polystyrene samples with molar masses ranging from

161 g/mol to 72200 g/mol were used as calibration standards. The Cirrus GPC
software was used to determine the experimental data process with regards to
molar mass and analysis. The number average molar mass (Mn), weight average
molar mass (Mw), molecular weight at highest peak (Mp) and polydispersity index
(PD = Mw/Mn) were calculated from refractive index (RI) signals using the
software. Detectors consist of a differential refractive index and a photometer in
the UV range of 280 nm.

3. RESULTS AND DISCUSSION

A dark brown, sticky mass stuck to the walls of the EP, although the
pyrolysis experiment was carried out using a 1 kg/h rig. The amount of dark
brown, sticky mass on the walls of the EP and liquids collected in the tank were
about 22.8% and 74%, respectively, of total pyrolysis liquids produced from the
pyrolysis experiment. Therefore, in the future, a system that is capable of
collecting the liquid in one single collecting point is recommended, by
implementing a flushing system in the EP for the purpose of analysing the liquid
quality as shown in Figure 2. The liquid product collected in the tank was
analysed for its properties as a potential fuel in comparison to petroleum fuel or
other applications. However, only a few analyses were carried out for the organic
mass that stuck to the walls of the EP, as it was very viscous.

3.1 Properties of Feedstock

The key properties of EFB, both measured for this research and from the
literature, are given in Table 2. The high ash and potassium values are
noteworthy because ash, and potassium in particular, leads to reduced liquid
yields in fast pyrolysis.5 The carbon and hydrogen content are comparable to
woody biomass, as is the measured heating value. The lowest HHV in the
literature20 could be due to confusion between values quoted on a dry basis as
opposed to a wet basis, a problem apparent elsewhere in the literature. For
example, a value of 10 MJ/kg is quoted for dry oil palm matter,21 which is clearly
too low for a ligno-cellulosic biomass on a dry basis, or values for wet FFB were
used for dry FFB.22
Journal of Physical Science, Vol. 22(1), 1–24, 2011 9
Characterisation of Oil Palm EFB 10

Table 2: Properties of EFB (mf wt%).

Component / Literature References Measured Method
Property values
Cellulose 59.7 [24] na na
Hemicellulose 22.1 [24] na na
Lignin 18.1 [24] na na
Elemental analysis
Carbon 48.9 [25] 49.07 Combustion
Hydrogen 6.3 [26] 6.48 analysis
Nitrogen 0.7 [26] 0.7
Sulphur 0.2 [26] <0.10
Oxygen 36.7 [26] 38.29 By difference
K 2.24 [27] 2.00 Spectrometry
K2O 3.08–3.65 [28] na na
Proximate analysis
Moisture na na 7.95 ASTM E871
Volatiles 75.7 [25] 83.86 ASTM E872
Ash 4.3 [20] 5.36 NREL LAP005
Fixed carbon 17 [26] 10.78 By difference
HHV (MJ/kg) 19.0 [25] 19.35 Bomb
calorimeter
LHV (MJ/kg) 17.2 [27] na na
Notes: na - not available

The distribution of feed particle size after grinding is shown in Figure 3.

The ash content of each size fraction was determined using National Renewable
Energy Laboratory (NREL) Standard Analytical Method LAP005,23 and the mass
average of the size fractions of 5.39% compares well, within the accuracy of the
measurements, with the original sample sent by MPOB, which had an ash content
of 5.36%.

Figure 3: Particle size distribution of EFB.

Journal of Physical Science, Vol. 22(1), 1–24, 2011 11

3.1.1 Surface structure and EDX analysis

We found that the surface of the washed feedstock was clean, and there
was no dirt or particular matter on the surface. However, the unwashed feedstock
of sizes ranging from 355–500 µm could not be easily fed into the feeder because
the particles tended to stick to each other easily, thus, making the feeding process
for the 150 g/h reactor difficult.

Table 3 shows the results of the elemental analysis of unwashed

feedstock and washed feedstock over a range of ash content and leachate, using
Electron Dispersive X-ray Spectroscopy (EDX). It shows that hydrogen cannot
be detected by this method. Elements such as Al, P, Cl, Ti, Fe and Cu were
removed during the washing. However, only Na, S and K were decreased with
the reduction of the ash content of the feedstock. The treatment of manual
agitation on large feedstock at ambient temperature for 1 minute and soaking the
large feedstock at ambient temperature for 10 minutes are referred to as method 1
and method 2, respectively. We found that method 2 was more effective than
method 1 in removing potassium, sodium and sulphur. The treatment of stirring
the large feedstock at 90°C for 2 hours, referred to as method 4, was more
effective in removing potassium, sodium and sulphur by 70.4%, 100% and
87.9%, respectively. However, soaking the large feedstock at ambient
temperature for 20 minutes, referred as method 3, was also effective at removing
potassium, sodium and sulphur by 67.2%, 100% and 37.9%, respectively. The
treatment of soaking the large feedstock at ambient temperature for 1 minute
refers to method 5. The leachate is the washed water that was collected after the
washing treatment. An EDX analysis on the leachate of method 5 showed that it
has a high potassium content of about 82.53 wt%, which was expected because
the ash content of the leachate was even higher. It was expected that the leachate
contain insoluble inorganics, such as soil and insoluble organics.

Table 3: Elemental composition of unwashed and washed feedstock.

Element Unwashed Method 1 Method 2 Method 3 Method 4 Method 5
EFB (leachate)
(wt%)
C 64.3 65.8 69 73.4 72.4 0.03
O 19.8 20.2 18.9 17.8 19.0 3.79
Na 0.33 0.41 0 0 0 0.63
Mg 0.19 0.25 0.29 0.35 0 1.68
Al 0.16 0.1 0.51 0.03 0 0
Si 1.95 1.87 1.34 2 1.72 0.1
P 1.06 0.43 0.8 1.39 1.71 5.39
S 0.48 0.55 0.39 0.36 0.07 1.54
Cl 1.37 0.37 0.5 0.19 0.36 3.48
(continued)
Characterisation of Oil Palm EFB 12

Table 3: (continued)
Element Unwashed Method 1 Method 2 Method 3 Method 4 Method 5
EFB (leachate)
(wt%)
K 8.61 9.1 7.68 2.82 2.55 82.5
Ca 1.43 1.47 1.76 1.56 1.53 1.06
Ti 0.13 0 0.06 0 0.29 0.03
Cr 0.15 0.25 0 0.55 0 0
Mn 0 0 0.06 0 0.05 0.02
Fe 0.29 0 0.48 0.08 0 0
Ni 0 0.26 0 0 0.59 0
Cu 0.31 0 0.28 0 0.23 0.04
Zn 0.12 0 0 0 0.66 0.06
Ash 5.43 3.68 3.05 2.14 1.15 34.1
content
(mf wt%)

3.1.2 Thermogravimetric analysis of unwashed EFB and washed EFB

The thermal degradation characteristics of unwashed and washed

feedstock are displayed in Figure 4 and Figure 5 by thermogravimetry (TG) and
differential thermogravimetry curves (DTG), respectively. The washing
procedure employed for the washed sample gave an ash content of 1.15 mf wt%
after two hours of stirring 100 g of EFB in 7 litres of distilled water at 90°C. The
ash content of the unwashed sample was 5.43 mf wt%. The thermal behaviour of
the two samples substantially differed. The unwashed sample showed a small
DTG peak around 200°C, which might be indicative of extractives.29 No such
peak was apparent for the washed sample. A possible explanation is that the high
washing temperature could have led to the loss of extractives during washing.

100

80
% dry weight

60 Washed EFB Unwashed EFB

40

20

0
100 200 300 400 500 600
Temperature °C
Figure 4: Thermogravimetric analysis of washed and unwashed EFB.
Journal of Physical Science, Vol. 22(1), 1–24, 2011 13

0.3

mg/min weight loss 0.25

0.2 Unwashed EFB

0.15
0.1 Washed EFB

0.05
0
100 200 300 400 500 600
Temperature °C

Figure 5: Differential thermogravimetric analysis of washed and unwashed EFB.

The unwashed sample showed a single peak at 355°C, and the washed
sample had its greatest weight loss at 383°C. Furthermore, there was a hump at
around 320°C for the washed sample, which was not visible in the DTG curve of
the unwashed sample. Similar trends have been reported by other authors for low
ash and high ash biomass feedstock.30–32 It is generally assumed that the
temperature shift is due to the catalytic effect of the ash and that this shift leads to
a single DTG peak for both cellulose and hemicellulose for high ash biomass.
However, the hump apparent in the DTG curve for low ash biomass is indicative
of hemicellulose and the actual peak is due to the decomposition of cellulose.

The total weight loss between 100–450°C is 77.4% for the washed
sample and only 65.8% for the unwashed sample. There are two explanations.
Firstly, the ash itself will largely not volatilise and therefore contribute to the char
yield. Secondly, the ash is also believed to contribute to greater char production
through its catalytic activity.

A number of other washed samples with intermediate ash contents have

been subjected to TGA analysis. The resulting curves vary smoothly between the
two extremes as presented in Figure 4 and Figure 5. The total weight loss
between 100–450°C and temperature of maximal weight loss are listed in
Figure 6.
Characterisation of Oil Palm EFB 14

Figure 6: Weight loss in TGA between the temperatures of 100–450 C.

3.2 Properties of Liquid Product

Table 4 shows the characteristics of pyrolysis liquid from this work in

comparison with pyrolysis liquid derived from olive bagasse33 and petroleum
fuel.34 The HHV was calculated from the data of the elemental analysis using
Channiwala and Parikh’s formula35 as shown below:

HHV = 0.3491C + 1.1783H + 0.1005S – 0.1034O – 0.0151N – 0.0211A (3)

where HHV stands for high heating value on a dry basis and the unit is MJ/kg.

Most of the characteristics of the pyrolysis liquid are in Table 4, and the
results of liquid product analysis discussed in following subtopics.

Table 4: Comparison of characteristics of pyrolysis liquid and petroleum fuel.

Pyrolysis liquid
Pyrolysis liquid Pyrolysis liquid Petroleum
Type of liquids collected in EP’s
collected in tank of olive bagasse fuel
wall
Reference this work this work [33] [34]
Water content,
6.66 21.68 – 0.1
mf wt%
Elemental analysis (mf wt%)
C 56.47 41.86 66.9 85.2
H 7.85 7.82 9.2 11.1
O 35.46 33.94 21.9 1.0
N 0.22 0.1 2.0 0.3
(continued)
Journal of Physical Science, Vol. 22(1), 1–24, 2011 15

Table 4: (continued)
Pyrolysis liquid
Pyrolysis liquid Pyrolysis liquid Petroleum
Type of liquids collected in EP’s
collected in tank of olive bagasse fuel
wall
Reference this work this work [33] [34]
S 0 0 0 2.3
HHV, MJ/kg
(Channiwala and 25.29 20.32 31.9 42.94
Parikh’s formula)
Density, kg/m3 1548 1206 1070 940
Viscosity,
not determined 46.31 51 180
cP@25°C
pH 2.33 2.7 – –
GPC analysis
(g/mol)
Mp 447 222 – –
Mn 382 357 – –
Mw 562 564 – –
Polydispersity 1.47 1.57 – –

3.2.1 pH value

All the liquids produced from unwashed feedstock and washed feedstock
gave a pH of between 2 and 3, which indicated higher acid contents that arose
from organic acids, such as acetic and formic acids. It showed that treating the
feedstock by water washing did not increase the pH value of liquids.
Furthermore, mild steel is not suitable for handling or storage as it could have a
reaction with the pyrolysis liquid. Polypropylene piping, for example, has been
used to overcome this problem.

3.2.2 Density

As expected, the density of liquids stuck on the walls of the EP are more
viscous and have much higher density than the density of liquids collected in the
tank. The densities of liquids collected in the tank and on the walls of the EP are
1206.1 kg/m3 and 1548 kg/m3, respectively. The densities of petroleum and a
typical pyrolysis liquid are 940 kg/m3 and 1230 kg/m3, respectively.34 Pyrolysis
liquid density is higher compared to the density of petroleum, which means that
the liquid has about 42% of the energy content of fuel oil on a weight basis, but
61% on a volumetric basis.

3.2.3 Heating value

Typically, pyrolysis liquid has a heating value of about 40–50% of

conventional petroleum fuel. The HHV of the liquid collected in the tank and on
Characterisation of Oil Palm EFB 16

the walls of the EP had an average value of 21.49 MJ/kg, which is about 50% of
conventional petroleum fuel.

3.2.4 Char content

Char is a coarse powder with particle sizes ranging from less than a
micron to several milimetres, which have similar size and shape characteristics as
the feedstock. Overall, the char content of all liquids from unwashed and washed
feedstock is in the range of 0.2–2.0%. According to Peiyan,36 very small particle
size char was more easily blown into the collection system. Therefore, higher
temperatures gave high oil char contents.

3.2.5 Viscosity

Because the liquid stuck on the walls of the EP was too viscous, the
viscosity was measured on the liquid collected in the tank only. The viscosities of
the fresh liquid and old liquid at 25°C were 46.31 cp and 67.58 cp, respectively.
The fresh liquid is the liquid tested after 24 hours from which it was produced
and old liquid is the liquid that is stored for 24 hours at 80°C. We expected that
the viscosity on the whole liquid of this run would be higher than 46.31 cp if the
viscosity of the liquid stuck on the walls of the EP was taken into account. As
shown in Table 4, the viscosities of petroleum and pyrolysis liquid of olive
bagasse at 25ºC is 180 cP and 51 cP, respectively. Therefore, this result shows
that the density of petroleum is much higher compared to the viscosity of
pyrolysis liquid derived from olive bagasse or pyrolysis liquid from this study.

3.2.6 Stability

The stability of the pyrolysis liquid was examined by monitoring the

changes in viscosity, density and pH. The density and pH of the pyrolysis liquid
remained constant but the viscosity changed with time. The oil’s stability was
measured by calculating the viscosity index using Equation 2. The Aston
Viscosity Index (AVI) number for pyrolysis liquid produced from the pyrolysis
process on woody biomass at 500°C is about 2.2.37 The AVI for the liquid in this
study is 0.46. Thus, this liquid is quite stable as the AVI number is low.
According to Salter,36 the lignin content of the pyrolysis liquid greatly influenced
the stability of the liquid because the more lignin in the pyrolysis liquid, the more
stable the liquid is, hence, the lower AVI.

3.2.7 Functional group

The pyrolysis liquid was analysed using FTIR spectroscopy to obtain a

qualitative impression of the basic functional groups present in the oil. The
Journal of Physical Science, Vol. 22(1), 1–24, 2011 17

measured absorption frequency spectrum is shown in Table 5. The C-H bend

vibration between 1500 and 1450 cm–1 indicates the presence of alkanes. The
absorption peak between 1750 and 1625 cm–1 represents the C=O stretching
vibration and is suggestive of the presence of carboxylic acids, ketones and
aldehydes. The absorbance peaks between 1675 and 1600 cm–1 represent C=C
stretching vibrations. The peaks between 1300 and 1000 cm–1 are due to the
presence of phenols and alcohols, and absorption peaks between 900 and
650 cm–1 indicate the presence of single, polycyclic or substituted aromatic
groups.

Table 5: FTIR functional group composition of pyrolysis liquid.

Frequency range (cm–1) Group Class of compound
3000 – 2800 C-H stretching alkanes
1750 – 1625 C=O stretching aldehydes, carboxylic acids, ketones
1675 – 1600 C=C stretching alkenes
1500 – 1450 C-H bending alkanes
1300 – 1000 C-O stretching alcohol
900 – 650 O-H bending phenol aromatic compounds

3.2.8 Pyrolytic lignin content

Lignin is one of the major wood components, and pyrolytic lignin

represents the water insoluble part of the pyrolysis liquids. Pyrolytic lignin can be
obtained by precipitating pyrolysis liquid in an excess of water. In this work,
pyrolysis liquid collected in the tank derived from washed EFB was precipitated.
The content of pyrolytic lignin in the pyrolysis liquid collected in the tank is
12.42% w/w, similar to the pyrolytic lignin yield from bio-oil derived from hard
wood.37 The pyrolytic lignin has a relatively high calorific value and has been
proposed as a phenol formaldehyde resin replacement.39 For this application, it
would have to be separated from the remainder of the bio-oil by differential
condensation and collection and presents an energy efficient option.

3.2.9 Molecular weight

The pyrolytic lignin of liquid collected in the tank was prepared by

precipitating this liquid in water, and the molar masses of the lignin were
determined using GPC. We found that the average molar mass (Mw), the average
molar number (Mn) and the molar mass at maximum peak (Mp) of the pyrolytic
lignin were 886 g/mol, 588 g/mol and 660 g/mol, respectively. The results of
molar masses for both liquid fractions (EP fraction and tank fraction) are
mentioned in Table 4. The molecular weight is an important indicator of the
degree of polymerisation and that the peak molecular weight of the liquid
collected in the tank is much lower than that of the pyrolytic lignin fraction.
Characterisation of Oil Palm EFB 18

3.2.10 Feedstock comparison

A comparison of EFB with results from the literature for other biomass
feedstock is presented in Figure 7 and Table 6. The product yield for washed
EFB is similar to that of a low ash wood, such as poplar, whereas the product
yield for unwashed EFB is much closer to that obtained for higher ash feedstock,
such as wheat straw.40,41

Table 6: Comparison of EFB with other biomass feedstock.

Feedstock Unwashed Washed Poplar Corn Wheat
EFB EFB aspen stover straw
Reference this work this work [36] [37] [37]
Cellulose [mf wt%] 59.7 – 42.3 31.0 32.4
Hemicellulose [mf wt%] 22.1 – 31.0 43.0 41.8
Lignin [mf wt%] 18.1 – 16.2 13.0 16.7
Moisture [mf wt%] 7.48–8.96 6.04–6.54 5.0 6.5 9.0
Ash [mf wt%] 5.29 1.03 0.39 11.0 4.6
Yields [mf wt%]
Char 24.52 10.76 11 42.2 22
Gas 22.31 14.70 14 13.7 19
Organics 34.71 61.34 58.9 27.3 37
Total Liquids 49.8 72.36 74.1 43.1 54

unwashed EFB
65 washed EFB
60 poplar-aspen
corn stover
Organics yield [mf wt%]

55
wheat straw
50
45
40
35
30
25
20
400 450 500 550 600 650
Temperature °C
Figure 7: Organics yield as a function of temperature for a number of feedstock.
Journal of Physical Science, Vol. 22(1), 1–24, 2011 19

3.2.11 Physical and chemical properties of products

The pyrolysis liquids produced separated into two phases; a phase

predominated by tarry organic compounds and an aqueous phase. The relative
shares of the total liquid product yield of the two phases are approximately 60%
for the former and 40% for the latter. A comparison of key properties for the two
phases with those of wood derived bio-oil, light fuel oil and heavy fuel oil is
given in Table 7, which also contains the ultimate analysis of the char product.
The value for sulphur was not determined, as there is very little sulphur in the
EFB itself. Because of its high water content, the higher heating value of the
aqueous phase was not measured.

The viscosity of the aqueous phase is close to that of water, whereas the
organic phase hardly flows at all. At room temperature, the organic phase clearly
has a viscosity above 10,000 cP, the limit for the measuring equipment available
to the authors of this work.

The empirical formula of the organics in the organic phase is CH1.51O0.14.

By comparison, the empirical formula for carbohydrates is CH2O, that of phenol
is CHO0.17 and longer chain straight alkanes approach an empirical formula of
CH2. It is possible that the organic phase contains a small amount of palm oil, as
the organics in the organic phase have a significantly higher hydrogen to carbon
ratio than is the case for wood-derived slow pyrolysis tars.42

Table 7: Characteristics of pyrolysis oil compared to petroleum fuel.

EFB Wood
Light Heavy
derived
fuel oil fuel oil
bio-oil
Organics Aqueous
Char
phase phase
C 69.35 13.83 71.43 32–48 86.0 85.6
H 9.61 11.47 1.80 7–8.5 13.6 10.3
N 0.74 0.14 0.63 <0.4 0.2 0.6
O 20.02 74.56 8.72 44–60 0 0.6
S – – – <0.05 <0.18 2.5
Moisture content 7.9 64.01 – 20–30 0.025 0.1
HHV (MJ/kg) 36.06 – – – – –
LHV (MJ/kg) – – – 13–18 40.3 40.7

The phase separated liquid product would represent a challenging fuel for
boilers and engines, because of the high viscosity of the organic phase and the
high water content of the aqueous phase. These could be overcome by upgrading.
The addition of polar solvents such as methanol or ethanol represents one of the
easiest routes,5 and it was established by the authors of this study that the two
phases both readily dissolve in methanol to give a homogeneous single-phase
Characterisation of Oil Palm EFB 20

product with a low viscosity. Several researchers43–45 have studied the addition of
solvent to reduce viscosity and aging rates because the bio-oil composition
changes during aging and mutual solubility of the components changes to make
phase separation more likely. Further research would be required to establish how
much methanol would have to be added at a minimum to obtain a single-phase
liquid. The addition of ethanol could also reduce corrosivity and allow the
removal of the water through low temperature vacuum distillation,46 which is
otherwise difficult because of the thermal instability of pyrolysis liquids.

Thermochemical upgrading is another possibility, either via gasification

and Fischer-Tropsch, where it may be advantageous to gasify a bio-oil char slurry
rather than the biomass itself,47 or via catalytic steam reforming of the aqueous
phase to obtain the hydrogen for hydrogenation of the organic phase.48

The organic phase might also be directly used in engines, turbines and
boilers, if it is first pre-heated to reduce its viscosity, though its thermal
instability might limit the temperature it can be brought up to. The aqueous phase
could potentially be co-fired to reduce thermal NOx and enable efficient burning
of its dissolved dry matter. As mentioned in the introduction, it might also be
possible to obtain useful chemicals, such as phenolic compounds for resin
making, as a commercial by-product.

Due to the low liquid yield, not all of the char and gas would be required
for internal process heat. Some could be used to dry the very wet fresh EFB.
Burning the char, which contains nearly all the minerals, would allow the
recovery of the ash, which is useful as a fertiliser.21 The gas is high in carbon
dioxide and has a higher heating value of only approximately 4 MJ/kg. It may
potentially be utilised for carbon sequestration enhanced oil recovery schemes.49

Because the fresh EFB is very wet, washing of the EFB prior to fast
pyrolysis represents another interesting avenue to enhance the value of the fuel
products obtained, as drying is required already and would therefore add a little
extra expense. Simple soaking in water at room temperature can remove most of
the ash, with small biomass losses, giving a feedstock with similar yield
characteristics as low ash woody biomass.15

4. CONCLUSION

A study of fast pyrolysis on unwashed and washed Empty Fruit Bunches

(EFB) was performed using a 150 g/h fluidised bed reactor. The liquids produced
from both the unwashed and washed EFB had high acid content and a density of
a typical pyrolysis liquid, with a High Heating Value (HHV) of about 50% of the
Journal of Physical Science, Vol. 22(1), 1–24, 2011 21

conventional petroleum fuel. The char content of all liquids from unwashed and
washed feedstock was in the range of 0.2–2.0%. The viscosity of the liquid had
an overall value higher than 46.31 cp. The liquid had low Aston Viscosity Index
(AVI) and therefore was considered stable. The measured absorption frequency
spectrum was shown in Table 5. The content of pyrolytic lignin in the pyrolysis
liquid collected was similar to pyrolytic lignin yielded from bio-oil derived from
hard wood. The results of molar masses for both liquids were presented in Table
4. By feedstock comparison, the product yield for washed EFB was similar to
that of a low ash wood, whereas the product yield for unwashed EFB was much
closer to that of higher ash feedstock. However, the pyrolysis liquids that were
produced separated into two phases; a phase predominated by tarry organic
compounds (60%) and an aqueous phase (40%).

5. ACKNOWLEDGEMENTS

The authors would like to thank Universiti Sains Malaysia for fully
funding the work described in this publication. Most of the experimental work
was performed at the University of Aston in Birmingham, while Nurhayati
Abdullah was simultaneously a PhD student of the University of Aston in
Birmingham and a full-time employee of USM.

6. REFERENCES

1. Husain, Z., Zainal, Z. A. & Abdullah, M. Z. (2003). Analysis of biomass-

residue-based cogeneration system in palm oil mills. Biomass Bioenergy,
24, 117–124.
2. Pattiya, A., Titiloye, J. O. & Bridgwater, A. V. (2008). Fast pyrolysis of
cassava rhizome in the presence of catalysts. J. Anal. Appl. Pyrolysis,
81(1), 72–79.
3. Heo, H. S., Park, H. J., Park, Y. K., Ryu, C., Suh, D. J. & Suh, Y. W.
(2010). Bio-oil production from fast pyrolysis of waste furniture sawdust
in a fluidized bed. Bioresour. Technol., 101, 91–96.
4. Park, H. J., Heo, H. S., Park, Y. J., Yim, J. H., Jeon, J. K. & Park, J.
(2010). Clean bio-oil production from fast pyrolysis of sewage sludge:
Effects of reaction conditions and metal oxide catalysts. Bioresour.
Technol., 101, 83–85.
5. Chiaramonti, D., Oasmaa, A. & Solantausta, Y. (2007). Power generation
using fast pyrolysis liquids from biomass. Renewable and Sustainable
Energy Rev., 11(6), 1056–1086.
6. Bridgwater, A. V., Meier, D. & Radlein, D. (1999). An overview of fast
pyrolysis of biomass. Org. Geochem., 30(12), 1479–1493.
Characterisation of Oil Palm EFB 22

7. Genting Group unveils Malaysia’s first commercially produced bio-oil

using breakthrough technology. (2005). Press release by Genting Group,
21st August.
8. Di Blasi, C., Branca, C. & D’Errico, G. (2000). Degradation
characteristics of straw and washed straw. Thermochim. Acta, 364, 133–
142.
9. Mullen, C. A., Boateng, A. A., Goldberg, N. M., Lima, I. M., Laird, D.
A. & Hicks, K. B. (2010). Bio-oil and bio-char production from corn
cobs and stover by fast pyrolysis. Biomass Bioenergy, 34, 67–74.
10. Jenkins, B. M., Bakker, R. R. & Wei, J. B. (1996). On the properties of
washed straw. Biomass Bioenergy, 10(4), 177–200.
11. Das, P., Ganesh, A. & Wangikar, P. (2004). Influence of pretreatment for
deashing of sugarcane bagasse on pyrolysis products. Biomass
Bioenergy, 27, 445–457.
12. Bridgwater, A. V. & Peacocke, G. V. C. (1995). Biomass fast pyrolysis.
Proceedings of the Second Biomass Conference of the Americas. Energy
Environment, Agriculture, and Industry, NREL/cp-2008098, DE
95009230, NREL, Golden CO.
13. Diebold, J. P. (2002). A review of the chemical and physical mechanisms
of the storage stability of fast pyrolysis bio-oils. In Fast pyrolysis of
biomass: A handbook, ed. Bridgwater, A. V. London: CPL Press.
14. Scott, D. S., Paterson, L., Piskorz, J. & Radlein, D. (2000). Pretreatment
of poplar wood for fast pyrolysis: Rate of cation removal. J. Anal. Appl.
Pyrolysis, 57, 169–176.
15. Abdullah, N. (2005). An assessment of pyrolysis for processing empty
fruit bunches. PhD thesis, Aston University, Birmingham, UK.
16. Scott, D. S. & Piskorz, J. (1984). The continuous flash pyrolysis of
biomass. Can. J. of Chem. Eng., 62, 404–412.
17. ASTM. (1998). D7579 - 09 Standard test method for pyrolysis solid
content in pyrolysis liquids by filtration of solids in methanol. (1998).
Annual book of ASTM Standard, vol. 11–05.
18. Dyroff, G. V. (1993). Manual on significance of tests for petroleum
products, 6th ed. Philadelphia: ASTM.
19. Scholze, B., Hanser, C. & Meier, D. (2001). Characterization of the water
insoluble fraction from fast pyrolysis liquids (pyrolytic lignin): Part II.
GPC, carbonyl groups, and 13C-NMR. J. Anal. Appl. Pyrolysis, 58, 387–
400.
20. H-Kittikun, A., Prasertsan, P., Srisuwan, G. & Krause, A. (2000).
Environmental management for palm oil mill. Internet Conference on
Material Flow Analysis of Integrated Bio-Systems 2000, March–October.
21. Yusoff, S. (2006). Renewable energy from palm oil – innovation on
effective utilization of waste. J. Cleaner Prod., 14, 87–93.
Journal of Physical Science, Vol. 22(1), 1–24, 2011 23

22. Husin, M., Ramli, R., Mokhtar, A., Hassan, W. H. W., Hassan K, Mamat,
R. (2002). Research and development of oil palm biomass utilization in
wood-based industries. Palm Oil Developments, 36, 1–5.
23. Ehrman, T. (1994). Standard Method for Ash in Biomass. Standard
Analytical Method LAP005, National Renewable Energy Laboratory,
Golden, Co., 41–46.
24. Ani, F. N. (2003). Personal communication, Faculty of Mechanical
Engineering, Universiti Teknologi Malaysia.
25. Yan, R., Yang, H., Chin, T., Liang, T. D., Chen, H. & Zheng, C. (2005).
Influence of temperature on the distribution of gaseous products from
pyrolyzing palm oil wastes. Combustion and Flame, 142(1–2), 24–32.
26. Hussain, Z., Zainac, Z. & Abdullah, Z. (2002). Briquetting of palm fibre
and shell from the processing of palm nuts to palm oil. Biomass
Bioenergy, 22(6), 505–509.
27. Menon, N. R., Rahman, Z. A. & Bakar, N. A. (2003). Empty fruit
bunches evaluation: Mulch in plantation vs. fuel for electricity
generation. Oil Palm Industry Econ. J., 3(2), 15–20.
28. Saletes, S., Caliman, J. & Raham, D. (2004). Study of mineral nutrient
losses from oil palm empty fruit bunches. J. Oil Palm Research, 16(1),
11–21.
29. Raveendran, K., Ganesh, A. & Khilar, K. (1995). Influence of mineral
matter on biomass pyrolysis characteristics. Fuel, 74(12), 1812–1822.
30. Yang, H., Yan, R., Chin, T., David, T. L., Chen, H. & Zheng, C. (2004).
Thermogravimetric analysis - fourier transform infrared analysis of palm
oil waste pyrolysis. Energy Fuels, 18, 1814–1821.
31. Raveendran, K., Ganesh, A. & Khilar, K. (1996). Pyrolysis
characteristics of biomass and biomass components. Fuel, 75(8), 987–
998.
32. Jensen, A., Dam-Johansen, K., Wojtowicz, M. A. & Serio, M. A. (1998).
TG-FTIR study of the influence of potassium chloride on wheat straw
pyrolysis. Energy Fuels, 12(5), 929–938.
33. Sensöz, S., Demiral, I. & Gerçel, H. F. (2005). Olive bagasse (Olea
europea L.) pyrolysis. Bioresour. Technol., 97(3), 429–436.
34. Maggi, R. & Elliott, D. C. (1997). Upgrading overview. In Developments
in thermochemical biomass conversion, ed. Bridgwater, A. V. &
Boocock, D. G. B., 575–588. London: Blackie Academic & Professional.
35. Channiwala, S. A. & Parikh, P. P. (2001). A unified correlation for
estimating HHV of solid, liquid and gaseous fuels. Fuel, 81, 1051–1063.
36. Liu, P. (2002). Improvement of bio-oil stability in wood pyrolysis. PhD
thesis, Aston University, Birmingham, UK.
37. Salter, E. H. (2001). Catalytic pyrolysis of biomass for improved liquid
fuel quality. PhD thesis, Aston University, Birmingham, UK.
 Characterisation of Oil Palm EFB 24

38. Effendi, A., Gerhauser, H., Bayerbach, R., Meier, D. & Bridgwater, A.
V. (2005). Characterisation of pyrolytic lignin: molar masses by GPC
and MALDI TOF/MS. Institute for Wood Chemistry and Chemical
Technology of Wood, Hamburg, Germany (unpublished result).
39. Effendi, A., Gerhauser, H. & Bridgwater, A. V. (2008). Production of
renewable phenolic resins by thermochemical conversion of biomass: A
review. Renewable Sustainable Energy Rev., 12(8), 2092–2116.
40. Scott, D. S. & Piskorz, J. (1982). The flash pyrolysis of aspen-poplar
wood. Can. J. Chem. Eng., 60, 666–674.
41. Scott, D. S., Piskorz, J. & Radlein, D. (1985). Liquid products from the
fast pyrolysis of wood and cellulose. Industrial Eng. Chem. Process
Design Dev., 24, 581–588.
42. EU Contract NNE5-2001-00744, BIOTOX. Final publishable report.
43. Oasmaa, A., Leppämäki, E., Koponen, P., Levander, J. & Tapola, E.
(1997). Physical characterisation of biomass-based pyrolysis liquids, 46.
Technical Research Centre of Finland 1997.VTT Publication 306.
44. Diebold, J. P. & Czernik, S. (1997). Additives to lower and stabilize the
viscosity of pyrolysis oils during storage. Energy Fuels, 11: 662–665.
45. Diebold, J. P. (1999). A review of the chemical and physical mechanisms
of the storage stability of fast pyrolysis bio-oils, 1–43. Lakewood:
Thermalchemie.
46. Canadian Patent CA 2165858.
47. Henrich, E., Dinjus, E. & Weirich, F. (2002). A new concept for biomass
gasification at high pressure. Proceedings of the 12th European
Conference and Technology Exhibition on Biomass for Energy and
Industry 2002, Amsterdam, Netherlands, 17–22 June.
48. Wang, D., Czernik, S. & Chornet, E. (1998). Production of hydrogen
from biomass by catalytic steam reforming of fast pyrolysis oils. Energy
Fuels, 12(1), 19–24.
49. Herzog, H. J. (2001). What future for carbon capture and sequestration?
Env. Sci. Technol., 35(7), 148–153.

View publication stats