Bioresource Technology 100 (2009) 683689

Contents lists available at ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Production of biodiesel using a continuous gasliquid reactor

Sam Behzadi, Mohammed M. Farid *
Department of Chemical and Materials Engineering, University of Auckland, Private Bag, 92019 Auckland, New Zealand

a r t i c l e

i n f o

Article history:
Received 4 April 2008
Received in revised form 17 June 2008
Accepted 17 June 2008
Available online 30 July 2008
Keywords:
Biodiesel
Methanol
Sodium methoxide
Sodium hydroxide
Spray reactor

a b s t r a c t
A novel continuous reactor process has been developed for the production of biodiesel from fats and oils.
The key feature of the process is its ability to operate continuously with a high reaction rate, potentially
requiring less post reaction cleaning and product/reactant separation than currently established processes. This was achieved by atomising the heated oil/fat and then spraying it into a reaction chamber
lled with methanol vapor in a counter current ow arrangement. This allows the continuous separation
of product and the excess methanol stream in the reactor. The overall conversion based on a single cycle
of this process has been between 50% and 96% of the feed stock materials. Conversions of 9496% were
achieved while operating with 57 g of sodium methoxide/L of methanol at methanol ow rate of 17.2 L/
h and oil ow rate of 10 L/h. Additional variations in the reactant stoichiometry (i.e. reactant ow rates),
catalyst type/concentration, and reaction temperature on the overall product conversion were
investigated.
2008 Elsevier Ltd. All rights reserved.

1. Introduction
Biodiesel is a fuel derived from the transesterication of fats
and oils (Ma and Hanna, 1999; Srivastava and Prasad, 2000;
Knothe et al., 2005; Pahl, 2005; Van Gerpen, 2005; Mittelbach
and Remschmidt, 2006). This fuel has similar properties to that
of diesel produced from crude oil and can be used directly to
run existing diesel engines or as a mixture with crude oil diesel.
The main advantages of using biodiesel is that it is biodegradable, can be used without modifying existing engines, and produces less harmful gas emissions such as sulfur oxide (Knothe
et al., 2005; Pahl, 2005; Van Gerpen, 2005). However, biodiesel
is still more expensive than conventional petroleum derived diesel which is attributed to the higher feedstock and processing
costs. By creating a continuous process it may be possible to reduce the production cost and hence reduce the over all cost of
biodiesel, making the price of biodiesel competitive with respect
to fossil fuels (Ma and Hanna, 1999; Srivastava and Prasad, 2000;
Knothe et al., 2005; Pahl, 2005; Van Gerpen, 2005). At present
several countries such Brazil, United States, Germany, Australia,
Italy and Austria are already using biofuels such as bioethanol
and biodiesel. It is expected that this trend will grow and more
countries will use biofuels (Bender, 1999; Korbitz, 1999; International Energy Agency, 2004; Dorado et al., 2006; Haas et al.,
2006). However, for this to take place widely, the economic viability of these fuels need to be improved. In most countries production costs have been reduced by government support through
* Corresponding author. Tel.: +64 937377599x8; fax: +64 9373757463.
E-mail address: m.farid@auckland.ac.nz (M.M. Farid).
0960-8524/$ - see front matter 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2008.06.037

tax rebates. It is expected that, the development of a new continuous process will lower fuel production costs and contribute to
our growing energy needs. Biodiesel is generally manufactured
using batch reactors. However, this trend is changing and more
continuous processes have been examined and used either in a
laboratory or industrial scale due to the increase in biodiesel
demand.
1.1. Transesterication and reaction variables
The transesterication reaction is normally a sequence of three
consecutive reversible reactions (Marchetti et al., 2007). In this
process triglyceride is converted stepwise into diglyceride, monoglyceride, and, nally, glycerol in which 1 mol of alkyl esters is
formed in each step. The overall reaction is also characterized by
three control stages: mass transfer, kinetic and equilibrium controlled. The mass transfer stage is the slowest of the three stages
due to the poor immiscibility of the two reactants (i.e. methanol
and triglycerides). At the completion of the mass transfer stage
the process is controlled by the kinetic stage. The forward and reverse reactions for this stage can either follow a second order
mechanism based on a alcohol to oil mole ratio of 6:1 or a pseudo-rst order for higher alcohol to oil mole ratio such as 30:1
(Freedman et al., 1986; Vicente et al., 2005). A higher reaction rate
can be obtained when using higher alcohol to oil ratio. However,
the alcohol separation and purication at the end of the esterication process is complicated and costly. In addition, at higher alcoholoil ratios the separation of glycerol from ester becomes
difcult (Meher et al., 2006). Due to these limitations majority of
batch processes use a 6:1 alcohol to oil mole ratio.

684

S. Behzadi, M.M. Farid / Bioresource Technology 100 (2009) 683689

Both kinetic and mass transfer stages can be improved through

the use of higher reaction temperatures and vigorous mixing
(Noureddini and Zhu, 1997; Darnoko and Cheryan, 2000b; Vicente
et al., 2005). As in any reaction, the increase in temperature increases reaction rate exponentially. An increase in temperature
will also allow the reactants to be more miscible thus allowing a
higher rate of reaction to take place. The solubility of methanol
and ethanol in beef tallow at different temperatures was thoroughly studied by Ma et al. (1998). They found that the solubility
of the alcohols increased in the triglyceride phase as the temperature increased. However, the operating temperature is limited by
the boiling point of the alcohol used as a reactant (e.g. 67 C at
atmospheric pressure if methanol is used).
Darnoko and Cheryan (2000b) studied the kinetics of palm oil
transesterication in a batch reactor. Their study illustrated that
while the overall conversion of the process did not change with
temperature, the rate of the transesterication process was increased with temperature. The overall reaction kinetics is dependent on the individual rate constants for the conversion of
triglyceride to diglyceride, monoglyceride and alcohol ester. Based
on the rate constants obtained, the conversion of triglyceride to
diglyceride was the slowest reaction in transesterication. The
time needed for the mass transfer to occur is shortened as temperature is increased, leading to a higher rate of transesterication at
higher temperatures (Noureddini and Zhu, 1997; Darnoko and
Cheryan, 2000b).
Alternatively, vigorous mixing is utilized to increase the rate of
collision between the reactants and to homogenize the reaction
mixture (Noureddini and Zhu, 1997; Vicente et al., 2005). The alcohol (e.g. methanol) and the triglyceride source (typically vegetable
oil/animal fat) are immiscible and tend to form two layers. Vigorous mixing increases the mass transfer rate by dispersing the alcohol as ne droplets in the triglyceride phase increasing the contact
surface area between the two immiscible reactants (Stamenkovic
et al., 2007). A study conducted by Vicente et al. (2005) on the effect of impeller speed (300700 rpm) and reaction temperature (25
and 65 C) on the transesterication of sunower oil over a 1-min
period, illustrated that the triglyceride conversion is a function of
the impeller speed and reaction temperature. They found that
the delay in the formation of methyl ester became less with the increase of impeller speed from 300 to 600 rpm, indicating that the
mass transfer resistance became less important with increase in
stirring rate. The triglyceride conversion reached its maximum value at 600 rpm. The rate of methyl ester formation was also increased as the reaction temperature increased signicantly from
25 to 65 C (Vicente et al., 2005).
A secondary solvent (i.e. co-solvent) which is soluble in both
reactants (i.e. Alcohol and Triglyceride) has also been used as an
alternative means of creating a homogeneous phase. A common
co-solvent for this application has been THF (tetrahydrofuran),
with a boiling point of 67 C (Boocock et al., 1996, 1998; Ma et
al., 1998; Boocock, 2001; Vicente et al., 2005). A project by Boocock
et al. (1996) studied the effects of using a secondary solvent on the
transesterication of soybean oil to biodiesel. The co-solvent (THF)
was found to increase the reaction yield due to the reduction of
mass transfer resistance. In summary, co-solvents improve the solubility of the alcohol in the triglyceride phase allowing better mixing of the two phases and hence more reactions to take place.
However, it should be noted that when co-solvent is used the processing cost is higher due to extra processing equipment required
for the separation of the cosolvent.
1.2. Description of existing biodiesel process
Reactors are usually characterized as batch or continuous which
can be either a continuous stirred tank reactors (CSTRs) or plug

ow reactors (PFRs) (Noureddini et al., 1998; Peterson et al.,

2002; Harvey et al., 2003; Van Gerpen et al., 2004; Van Gerpen,
2005; Knothe et al., 2005).
The most important consideration in reactor design is the extent of product conversion. Key reactor variables that dictate conversion and selectivity are temperature, pressure, reaction time
(residence time in continuous reactors), and degree of mixing. In
transesterication reaction, the selectivity of the reaction is not
negatively impacted by increasing temperature. However, an increase in the reaction temperature does impact the operating pressure. Since this reaction is usually conducted in liquid phase, the
pressure in the reactor must be maintained at a sufcient level that
keeps the methanol in the liquid phase. Depending on the reactor
type, this could become costly, which would increase the associated production cost of biodiesel. The conversion rate can be improved by increasing the reaction time. However, increasing the
reaction time will decrease the chemical throughput or increase
the reactor size. Another important parameter in the reactor design
is the degree of mixing. For batch reactors and CSTRs the degree of
mixing is directly related to the amount of energy introduced
through the impeller. However, a threshold exists where additional
mixing will not provide any performance enhancement. For PFRs,
the degree of mixing is dictated by the design of the reactor (i.e.
length and the use of static mixers) and the ow pattern created
within the reactor (i.e. turbulence or laminar) (Noureddini et al.,
1998; Harvey et al., 2003; Van Gerpen et al., 2004; Van Gerpen,
2005; He et al., 2005). The common continuous reactors investigated for the production of biodiesel are: plug ow reactor (Peterson et al., 2002), oscillatory ow reactor (Harvey et al., 2003),
reactive distillation column (He et al., 2006), continuous high temperature gasliquid reactor [Patent WO 01/88072 A1] (Nimcevic
and Gapes, 2004), and combined plug ow/CSTR arrangement
(Noureddini et al., 1998). The overall objectives of these continuous reactors were to reduce post reaction cleaning and total
processing time.
With the exception of the above mentioned patent, all biodiesel
reactors have an underlying similarity in that they operate in a liquid/liquid phase (Noureddini et al., 1998; Darnoko and Cheryan,
2000a; Peterson et al., 2002; Harvey et al., 2003; Nimcevic and
Gapes, 2004). The following summarizes the major limitations of
the CSTR or tubular reactors:
 The reaction temperature will be limited to alcohol boiling point
i.e. 65 C for methanol if the reactor is to be operated at atmospheric pressure.
 Signicant mass transfer resistance is expected in the industrial
size reactor even when using higher shear mixing.
Therefore, the aim of our research was to develop a continuous
process that does not have the previously mentioned limitations.
In this project the use of a gasliquid reactor to produce biodiesel
was adopted. The hot oil was sprayed as ne droplets into an atmosphere of vapor methanol in a counter current ow arrangement.
The effect of operating temperature, catalyst and the methanol
concentrations were studied.

2. Methods
2.1. Material
Soya bean oil, beef tallow and methanol were used in this research as the feedstock while sodium methoxide or sodium
hydroxide were used as catalysts. Beef tallow was supplied by
PVL Proteins Ltd., and Bakels Edible Oils (NZ) Ltd. The tallow contained less than 0.5% FFA and less than 1% moisture. Cooking grade

S. Behzadi, M.M. Farid / Bioresource Technology 100 (2009) 683689

soya bean oil was purchased from local food outlets. Technical
grade methanol was used as the alcohol source in all experiments.
It was purchased from Ajax Fine Chemicals Ltd. The methanol was
premixed with sodium hydroxide (NaOH) in the range of 010 g
NaOH/L methanol before being pumped into the reactor. Pre-mixing the sodium hydroxide with methanol allows the formation of
sodium methoxide needed to catalyze the reaction. In some of
the experiments, pure sodium methoxide was used as catalyst. Sodium hydroxide pellets, 99.98% and Sodium Methoxide 95% reagent grade were purchased from SigmaAldrich Co.
2.2. Equipment
It was identied that diesel injection pump would be a suitable
unit for generating small micro sized droplets of high surface area.
The surface area generated in this process is many order of magnitude higher than the contact area in any standard mixing processes
currently employed in biodiesel processes based on liquidliquid
reaction. This simplied the geometry of the unit and allowed
the reaction to be carried out in a spray like reactor unit. The potential for methanol circulation as illustrated in Fig. 1 was also a
key design parameter.
The reaction is carried out in a 2.3 m high  0.38 m diameter
stainless steel tank. The oil atomization is carried out using a high
pressure diesel injection pump while methanol is pumped using a
peristaltic pump.
The process operates by injecting oil or fat through a high pressure nozzle into a reactor lled with methanol vapor carrying the
catalyst and moving in counter current ow direction. This allows
direct contact between the two phases with large contact area
thereby promoting more heat and mass transfer. The high pressure
nozzles in the system atomize the oil into small droplets in the
range of 100200 lm diameter, thus signicantly increasing the
contact surface area of the methanol/oil. The desired reaction temperature is achieved by heating the reactor using a steam jacket
and preheating the reactants before being fed to the reactor. Since

685

the process operates at temperatures well above the alcohol boiling point with almost no mass transfer limitations, a higher rate
of reaction was achieved. This process can operate at any temperature below the degradation temperature of fat and oil (i.e.
6200 C). (Behzadi and Farid, 2007).
To increase the reaction rate, a high methanol to oil ratio was
used as in any other transesterifcation process. In this laboratory
reactor, the methanol is condensed, mixed with fresh liquid methanol and recycled back into the reactor. However, based on this
new technology, methanol vapor could be recycled in the industrial reactor without going through the energy intensive condensation/evaporation process. To improve reaction rate the reactants in
the laboratory reactor were pre-heated using a combination of
steam and electrical heating. The oil was injected into the reactor
at 100120 C while methanol was injected at 8595 C. This laboratory reactor was not designed as a pressure vessel and hence was
heated using saturated steam of 1 bar. This limited the operating
temperature to a maximum of 100 C. Based on this technology
and with suitable method of heating, industrial reactors could
operate at any temperature below the degradation temperature
of oil.
2.3. Procedure
Each experimental run was carried out for a period of 30
40 min to insure steady state operation. To determine reaction
mechanism, samples were collected from the reactor at different
axial and radial positions using a sampling device specially designed for this purpose (Fig. 2). Note that samples were collected
only after steady state operation had been reached.
In all experiments methanol ow rate of 100300 ml/min, oil/
fat ow rate of 100200 ml/min and catalyst concentration of 0
10 g/L of methanol were used. This methanol to oil ratio (1:20
fat/oil: methanol) was signicantly higher than what is used in
batch reactors. This high ratio does not reect a poor feature of
the process developed since the methanol is in the vapor phase

Fig. 1. Process ow diagram (Behzadi and Farid, 2007).

686

S. Behzadi, M.M. Farid / Bioresource Technology 100 (2009) 683689

Fig. 2. Sampling methodology.

and can be recycled at any ratio with minimum energy

consumption.
2.4. Product analysis
A gas chromatograph (GC) was the primary source of product
analysis accompanied by density and viscosity measurements. Viscosity was measured using a Paar-Physica viscometer and density
was measured by a 50 ml density bottle. These tests were carried
for both washed and un-washed methyl ester. The washing process
consisted of acid neutralization followed by water washing. This
was to remove sodium hydroxide and other impurities such as excess methanol, triglyceride, diglyceride and monoglyceride. A gas
chromatograph (Shimadzu GC17AAFW GC FID) equipped with a
ame ionization detector (FID) was used for product analysis (i.e.
methyl ester production and triglyceride conversion). A Rtx-35
(Restek, 30 m  0.25 mm  0.1 lm) GC column was used. Samples
were prepared by dissolving 1000 ll of product in 25 ml n-Hexane
with the addition of 200 ll internal standards. About 1 ll of the
sample was then injected into the column. The oven temperature
program was set at a starting temperature of 160 C (1 min) followed by an initial ramp of 25 C/min to 195 C and then at
10 C/min to 225 C. Then it was ramped at 25 C/min to 360 C
(12 min). The FID and injection temperatures were set at 370 C
and 355 C. GC was calibrated using stock solution of methyl olein,
mono-olein, di-olein, and tri-olein. Tricaprin was used as the internal standard.
3. Results and discussion
To understand the reaction mechanism the effect of catalyst
concentration and oil/methanol ow rates were examined. The
washed and unwashed products from this process had similar density and viscosity (i.e. dynamic and kinematic) to those obtained
from the batch process and were independent of the production

ow-rates examined. Based on these results, the current process

was found capable of producing biodiesel continuously at a ow
rate of 10 L/h. The data collected illustrated that the ow rates
tested had very little effect on product quality. This is a strong indication that the designed reactor can operate at ow rates higher
than the designed value of 10 L/h. To date, the process has been
operated using both vegetable oil and animal fat.
Table 1, shows the average conversion of biodiesel produced
and it associated physical parameters such as density and viscosity
for the current process for un-washed products. The results indicate that the products are close to the New Zealand and International biodiesel standard based on their physical parameters. As
illustrated in Table 1, the methyl ester content (%) for un-washed
product (i.e. raw products obtained from the reactor) is close to
the recommended value provided by NZS7500:2005 biodiesel
standard. It should be noted that the current process requires no
additional cleaning equipment for the removal and separation of
excess methanol from the product stream.
3.1. Catalyst concentration and methanol ow rate
The effects of catalyst concentration and methanol ow rate are
presented in Table 2 for three different oil ow rates. The conversion of triglyceride is dependent on both the catalyst concentration
and methanol ow rate. At a high methanol ow rate a much higher conversion was achieved. This is clearly demonstrated in Table
2, were the highest conversion was achieved at an oil ow rate
of 6 L/h and methanol ow rate of 8.6 L/h. These ow rates equate
to a mol ratio of 1:35 (oil:methanol). The reduction in methanol
ow rate reduced the overall conversion by 10%.
Similar to conventional processes, sodium methoxide and
sodium hydroxide were used as the catalyst. Prior to each experiment anhydrous sodium methoxide or sodium hydroxide powders
were dissolved in methanol. However, the reaction of sodium
hydroxide with methanol forms water, which is undesirable and

Table 1
Average methyl ester conversion, density and viscosity of the process
Property

Biodiesel Standard

Average

Unit

Ester content (%)

Density at 15 C
Kinematic viscosity at 40 C
Dynamic viscosity at 40 C

NZS 7500:2005
NZS 7500:2005
NZS 7500:2005

8596
870.6
4.83
4.2

% (m/m)
kg/m3
mm2/s
mPa S

Limits
Minimum

Maximum

96.5
860
2.00

900
6.00

687

S. Behzadi, M.M. Farid / Bioresource Technology 100 (2009) 683689

Table 2
Effect of catalyst concentration, oil ow (Beef Tallow), and methanol ow rate on
biodiesel conversion
Methanol ow
rate (L/h)

Catalyst concentration
(g NaOH/L of methanol)

Conversion (%) at different oil ow

rates (L/h)
6

10

6.5

1
2
3
4
5
6
7
8
9

58.2
69.5
76.3
82.6
92.85
93.10
89.52
88.41
86.86

57.8
64.6
70.1
77.2
88.64
86.25
84.81
84.38
82.50

48.6
56.7
66.5
75.6
86.6
84.0
82.7
82.4
80.1

1
2
3
4
5
6
7
8
9

66.8
75.3
79.8
89.7
96.11
94.74
92.45
91.76
90.59

63.3
72.6
76.5
87.6
95.33
92.14
91.41
89.01
86.88

59.7
70.2
79.2
85.5
92.0
90.0
88.3
86.8
84.0

8.6

Table 4
Effect methanol ow rate and catalyst concentration, on the biodiesel production
Methanol ow
rate (L/h)

Catalyst type

Catalyst
concentration
(g /L of methanol)

Oil ow rate (10 L/h)

Conversion (%)

8.6

Sodium Methoxide

5
6
7

93.4
92.0
90.6

17.2

Sodium Methoxide

5
6
7

96.15
94.85
94.81

As illustrated in Table 2, a change in methanol concentration

had a signicant effect on biodiesel conversion. Therefore, additional experiments were conducted at oil ow rate of 10 L/h, methanol ow rate of 17.2 L/h, and catalyst concentration of 57 g
sodium methoxide/L of methanol. Table 4 shows a signicant increase in the conversion to methyl ester when methanol ow rate
was increased. As discussed earlier the increase in methanol concentration allows the reaction equilibrium to be pushed forward.
3.2. Product distribution

enhances the saponication reaction. The catalyst concentration

range was determined from pilot studies conducted in this project,
which ascertained the minimum amount of catalyst required to
initiate the reaction. As illustrated in Table 2 a catalyst concentration of 59 g NaOH/L of methanol had the highest overall
conversion. The overall conversion decreased as the catalyst concentration increased above 9 g catalyst/L of methanol. This is assumed to have been caused by the formation of sodium soaps.
Vicente et al. (2004) have previously reported that with sodium
hydroxide as a catalyst, high yield losses are experienced due to
triglyceride saponication and methyl ester dissolution in the glycerol phase. Our results indicate that the formation of sodium soaps
increases with catalyst concentration. This trend was observed in
all experiments. The results suggest that a catalyst concentration
of 57 g sodium hydroxide/L of methanol is optimal for high yields
of biodiesel. Based on these ndings, all subsequent experiments
were then performed using 57 g sodium hydroxide or sodium
methoxide/L of methanol, at various oil and methanol ow rates.
To minimize the undesirable saponication reaction, pure sodium methoxide and soya bean oil were used in subsequent experiments. Experiments were conducted using oil ow rate of 10 L/h
and a methanol ow rate of 8.6 L/h, at a catalyst concentration of
57 g sodium methoxide/L of methanol. It should be noted that
this change had no effect on the overall reaction mechanism. As expected the conversion of biodiesel had increased with the introduction of sodium methoxide (refer to Table 3). Based on these
ndings, subsequent experiments were then performed using 5
7 g sodium methoxide /L of methanol, at various methanol ow
rates.

Samples were collected from various radial and axial points in

the reactor to determine the reaction mechanism. The conversion
was higher at central part of the reactor compared to locations near
the wall (refer to Table 5). This is probably caused by the coalescence of the droplets into a liquid lm on the reactor wall. Formation of this lm decreases the contact surface area between the two
reactants. It also reduces the heat and mass transfer that is vital for
a rapid reaction. This effect could be reduced through the use of a
larger diameter reactor or a smaller injection angle. The conversion
of methyl ester was found to increase as the droplets moved further away from the injection point.
3.3. Effect of injection temperature
To increase the reaction rate and the overall conversion of
methyl ester the effect of higher injection temperature was stud-

Table 5
Product distribution inside the reactor (5 g sodium methoxide/L of methanol, 17.2 L/h
methanol and 10 L/h oil)
Sampling port

2
3
4
5

Distance from
oil injection (m)

Conversion (%) at different sampling location

1 (centre)

4 (reactor wall)

0.816
1.149
1.482
1.815

19.95
60.73
94.99
93.74

17.53
62.22
93.72
94.37

19.61
62.62
93.66
96.66

17.11
47.58
92.75
95.06

Table 6
Effect of injection temperature and catalyst concentration on methyl ester conversion

Table 3
Effect of catalyst type and concentration, on the biodiesel production
Catalyst type

Catalyst concentration
(g /L of methanol)

Oil ow rate (10 L/h)

Conversion (%)

Sodium hydroxide

5
6
7

92.0
90.0
88.3

Sodium methoxide

5
6
7

93.4
92.0
90.6

Catalyst
concentration
(g sodium
methoxide/ L
of methanol)

Injection temperature (C)

ME

MG

DG

TG

ME

MG

DG

TG

0
1
2
3
4
5

0
62.29
74.50
79.20
85.98
94.25

0
7.59
2.35
2.42
1.25
0.70

0
1.80
2.96
1.94
2.02
0.21

100
28.32
20.19
16.44
10.75
4.84

0
67.77
78.52
84.36
89.35
96.15

0
8.43
2.35
1.41
1.84
0.47

0
4.54
2.07
1.48
0.25
0.14

100
19.25
17.06
12.75
8.56
3.25

90

120

ME, methyl ester; MG, mono-glyceride; DG, di-glyceride, and TG, tri-glyceride.

688

S. Behzadi, M.M. Farid / Bioresource Technology 100 (2009) 683689

ied. Experiments were conducted using oil ow rate of 10 L/h and

methanol ow rate of 17.2 L/h, at a catalyst concentration of 05 g
sodium methoxide/L of methanol and injection temperatures of 90
and 120 C. As illustrated in Table 6, as the oil injection temperature increased from 90 to 120 C the overall conversion also increased. The highest conversion was achieved at a catalyst
concentration of 5 g/L methanol and an injection temperature of
120 C. This increase in conversion could be due to the signicant
decrease in the droplet diameter with increase in injection temperature, in addition to the usual effect of temperature on reaction
kinetics. At 90 C the average oil droplet size was 170 lm while
at 120 C the average oil droplet size was 120 lm. This is signicant increase in contact surface area.
Although the increase in injection temperature requires additional energy the current process requires less energy to operate
than that of a batch process. The current gasliquid process requires about 296 kJ/kg of biodiesel while a batch process requires
about 348 kJ/kg of biodiesel. These were calculated based on the
heat of reaction and the total energy required to heat and separate
feed and product streams. The energy required for mixing reactants, atomising the feed oil and initially heating the two reaction
vessels are negligible and thus ignored. These calculations also assume minimal heat loss from the two processes. For reference the
energy content of biodiesel from beef tallow: 39,949 kJ/kg and soya
bean oil: 40,080 kJ/kg, while diesel fuel has an energy content of
45,509 kJ/kg (Ali et al., 1995).
3.4. External product validation
In order to get independent verication, samples were analyzed
by Biofuel Testing New Zealand, Independent Petroleum Laboratory Ltd. (Table 7). In accordance with EN 14103 and ASTM D
6584 standards the total ester and glycerol content for washed
and unwashed beef tallow and soya bean oil based biodiesel were
measured. Operating condition used for these samples were; oil
ow rate of 10 L/h and methanol ow rate of 17.2 L/h, at a catalyst
concentration of 5 g sodium methoxide/L of methanol and injection temperature of 120 C. The results obtained validated previous
in-house tests conducted in this project. As illustrated in Table 7,
the washed and un-washed samples from both feedstocks contained high concentration of ester. However, the biodiesel from
soya bean had slightly higher ester content and lower total glycerol
content (i.e. free glycerol, mono, di and triglyceride) than biodiesel
from beef tallow. As illustrated in Table 7, sample C (washed biodiesel from soya bean oil) was in accordance with the standards.
This difference could be attributed to the physical property of the
two feedstocks. Soya bean oil has a lower viscosity, thus a better

atomisation characteristics which allows the generation of smaller

droplets. This allows for better mass transfer and higher reaction
rates. In addition, less saponication occurs with soya bean oil than
beef tallow. In summary, the results obtained indicate that current
process is capable of producing biodiesel similar to that produced
by previously described batch and continuous processes.
3.5. Catalyst supply and reaction mechanism
Conventionally, catalyst is introduced into the transesterication reaction through the methanol phase. In the technology presented in this paper, sodium hydroxide or sodium methoxide is
also pre-dissolved in the methanol phase before being injected into
the reactor as a vapor.
The transesterication reaction is initiated by a nucleophilic attack of the alkoxide ion on the carbonyl carbon atom of the triglyceride molecule, resulting in a tetrahedral intermediate.
Subsequently this intermediate splits into a alkyl ester and the corresponding anion of the diglyceride. The latter deprotonates the
catalyst, thus regenerating the active species, e.g. the diglyceride
molecule and the catalyst. This facilitates the reaction of a second
alcohol molecule with the catalyst, thus starting another catalytic
cycle to convert diglycerides and monoglycerides to a mixture of
alkyl esters and glycerol (Schuchardta et al., 1998). The same principles apply to the current process.
In the current continuous process methanol is introduced as a
vapor while oil as liquid droplets. It is believed that the sodium
methoxide is introduced into the reactor as nally dispersed droplet carried out by the methanol vapor. Inside the reactor, the catalyst n mist activates the reaction between the oil droplet and
methanol gas. As stated earlier it is the alkoxide ion that initiates
the reaction and not the methanol.
4. Conclusions
The process developed in this work for esterication reaction
has reduced reaction time from many minutes to few seconds.
The high reaction rate maybe attributed to the following:
 The use of atomized oil/fat. The atomization process increases
the oil/methanol contact area by producing micro sized droplets
(100200 lm droplets) and therefore increasing the heat and
mass transfer that is vital for a rapid reaction.
 The use of higher reaction temperatures. This process can operate at any temperature below the degradation temperature of
fat and oil (i.e. 6200 C) and is not limited to the boiling point
of the methanol, which is 65 C.

Table 7
Total ester and glycerol content as measured by biofuel testing New Zealand, Independent Petroleum Laboratory Ltd.
Method

Test

Units (%)

Sample A
Beef tallow
washed

Sample B
Beef tallow
un-washed

Sample C
Soya bean
washed

Sample D
Soya bean
un-washed

NZS 7500:2005
B100
specications

EN 14103
EN 14103a
EN 14103
EN 14103a
ASTM D 6584
ASTM D 6584
ASTM D 6584
ASTM D 6584
ASTM D 6584
ASTM D 6584

Ester content
Ester content
Linolenic acid methyl ester
Linolenic acid methyl ester
Free glycerin
Bound glycerin
Monoglycerides
Diglycerides
Triglycerides
Total glycerol

w/w
w/w
w/w
w/w
w/w
w/w
w/w
w/w
w/w
w/w

88.2
94.1
1.8
1.8
0.029
0.496
0.47
0.78
2.47
0.525

87.9
93.5
1.8
1.7
0.038
0.543
0.76
0.72
2.29
0.581

98.2
99.3
5.4
5.4
<0.005
0.234
0.58
0.27
0.40
0.236

96.5
97.9
5.4
5.4
0.053
0.376
1.10
0.32
0.42
0.429

96.5
96.5
Maximum
Maximum
Maximum

Maximum

Maximum

12
12
0.02
0.80

0.25

The presence in the sample of C17 methyl ester and molecular weight material below C14 is causing this method to under report the FAME content. The modied method
takes into account the C17 in the sample.

S. Behzadi, M.M. Farid / Bioresource Technology 100 (2009) 683689

 The process allows the use of very high excess methanol since
unlike the batch process methanol vapor can be recycled back
to the reactor without the need for expensive separation processes and intensive energy.
Acknowledgements
We would like to thank the University of Auckland, Energy Centre for its nancial support for the publication of this journal paper.
We would also like to thank FLO-DRY Engineering limited for their
project support.
References
Ali, Y., Hanna, M.A., Cuppett, S.L., 1995. Fuel properties of tallow and soybean oil
esters. Journal of the American Oil Chemists Society 72 (12), 15571564.
Behzadi, S., Farid, M.M., 2007. Method of biodiesel production, Auckland
Uniservices Ltd., WO2007049979.
Bender, M., 1999. Economic feasibility review for community-scale farmer
cooperatives for biodiesel. Bioresource Technology 70 (1), 8187.
Boocock, D.G.B., Konar, S.K., Mao, V., Lee, C., Buligan, S., 1998. Fast formation of highpurity methyl esters from vegetable oils. Journal of the American Oil Chemists
Society 75 (9), 11671172.
Boocock, D.G.B., Konar, S.K., Mao, V., Sidi, H., 1996. Fast one-phase oil-rich processes
for the preparation of vegetable oil methyl esters. Biomass and Bioenergy 11 (1),
4350.
Boocock, D.G.B., 2001. Single-phase process for production of fatty acid methyl
esters from mixtures of triglcerides and fatty acids, Boocock, David Gavin
Broole, WO0112581.
Darnoko, D., Cheryan, M., 2000a. Continuous production of palm methyl esters.
Journal of the American Oil Chemists Society 77 (12), 12691272.
Darnoko, D., Cheryan, M., 2000b. Kinetics of palm oil transesterication in a batch
reactor. Journal of the American Oil Chemists Society 77 (12), 12631267.
Dorado, M.P., Cruz, F., Palomar, J.M., Lopez, F.J., 2006. An approach to the economics
of two vegetable oil-based biofuels in Spain. Renewable Energy 31 (8), 1231
1237.
Freedman, B., Buttereld, R.O., Pryde, E.H., 1986. Transesterication kinetics of
soybean oil. Journal of the American Oil Chemists Society 63 (10), 13751380.
Haas, M.J., McAloon, A.J., Yee, W.C., Foglia, T.A., 2006. A process model to estimate
biodiesel production costs. Bioresource Technology 97 (4), 671678.
Harvey, A.P., Mackley, M.R., Seliger, T., 2003. Process intensication of biodiesel
production using a continuous oscillatory ow reactor. Journal of Chemical
Technology and Biotechnology 78 (23), 338341.

689

He, B.B., Singh, A.P., Thompson, J.C., 2005. Experimental optimization of a continuousow reactive distillation reactor for biodiesel production. Transactions of the
American Society of Agricultural Engineers 48 (6), 22372243.
He, B.B., Singh, A.P., Thompson, J.C., 2006. A novel continuous-ow reactor using
reactive distillation for biodiesel production. Transactions of the ASABE 49 (1),
107112.
International Energy Agency, 2004. Biofuels for transport: an international
perspective. International Energy Agency, Paris, France.
Knothe, G., Van Gerpen, J.H., Krahl, J., 2005. The biodiesel handbook. AOCS Press,
Champaign, IL.
Korbitz, W., 1999. Biodiesel production in Europe and North America, an
encouraging prospect. Renewable Energy 16 (14), 10781083.
Ma, F., Clements, L.D., Hanna, M.A., 1998. Biodiesel fuel from animal fat. Ancillary
studies on transesterication of beef tallow. Industrial and Engineering
Chemistry Research 37 (9), 37683771.
Ma, F., Hanna, M.A., 1999. Biodiesel production: a review. Bioresource Technology
70 (1), 115.
Marchetti, J.M., Miguel, V.U., Errazu, A.F., 2007. Possible methods for biodiesel
production. Renewable and Sustainable Energy Reviews 11 (6), 13001311.
Meher, L.C., Vidya Sagar, D., Naik, S.N., 2006. Technical aspects of biodiesel
production by transesterication-a review. Renewable and Sustainable Energy
Reviews 10 (3), 248268.
Mittelbach, M., Remschmidt, C., 2006. Biodiesel: the comprehensive handbook. M.
Mittelbach, Austria.
Nimcevic, D., Gapes, R.J., 2004. Transesterication of fat. Nimcevic, Dragan & Gapes,
Richard. J, WO0188072.
Noureddini, H., Harkey, D., Medikonduru, V., 1998. Continuous process for the
conversion of vegetable oils into methyl esters of fatty acids. Journal of the
American Oil Chemists Society 75 (12), 17751783.
Noureddini, H., Zhu, D., 1997. Kinetics of transesterication of soybean oil. Journal
of the American Oil Chemists Society 74 (11), 14571463.
Pahl, G., 2005. Biodiesel: growing a new energy economy. Chelsea Green Publishers,
White River Junction, VT.
Peterson, C.L., Cook, J.L., Thompson, J.C., Taberski, J.S., 2002. Continuous ow
biodiesel production. Applied Engineering in Agriculture 18 (1), 511.
Schuchardta, U., Serchelia, R., Vargas, R.M., 1998. Transesterication of vegetable
oils: a review. Journal of the Brazilian Chemical Society 9 (1), 199210.
Srivastava, A., Prasad, R., 2000. Triglycerides-based diesel fuels. Renewable and
Sustainable Energy Reviews 4 (2), 111133.
Stamenkovic, O.S., Lazic, M.L., Todorovic, Z.B., Veljkovic, V.B., Skala, D.U., 2007. The
effect of agitation intensity on alkali-catalyzed methanolysis of sunower oil.
Bioresource Technology 98 (14), 26882699.
Van Gerpen, J., 2005. Biodiesel processing and production. Fuel Processing
Technology 86 (10), 10971107.
Van Gerpen, J., Shanks, B., Pruszko, R., Clements, D., Knothe, G., 2004. Biodiesel
production technology. National Renewable Energy Laboratory, CO, USA.
Vicente, G., Martinez, M., Aracil, J., Esteban, A., 2005. Kinetics of sunower oil methanolysis. Industrial and Engineering Chemistry Research 44 (15), 54475454.