Chemical Engineering Journal 210 (2012) 173–181

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier.com/locate/cej

Partial hydrogenation of polyunsaturated fatty acid methyl esters over

Pd/activated carbon: Effect of type of reactor
Natthida Numwong a, Apanee Luengnaruemitchai a,b,⇑, Nuwong Chollacoop c, Yuji Yoshimura d
a
The Petroleum and Petrochemical College, Chulalongkorn University, Soi Chula 12, Phayathai Rd., Pathumwan, Bangkok 10330, Thailand
b
Center of Excellence on Petrochemical and Materials Technology, Soi Chula 12, Phayathai Rd., Pathumwan, Bangkok 10330, Thailand
c
National Metal and Materials Technology Center, 114 Thailand Science Park, Paholyothin Rd., Klong 1, Klong Luang, Pathumthani 12120, Thailand
d
National Institute of Advanced Industrial Science and Technology, Tsukuba Central 5, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8565, Japan

h i g h l i g h t s

" Oxidative stability of biodiesel was improved by partial hydrogenation.

" Type of reactor had an important effect on the conversion and selectivity of C18:1.
" Flow reactor provided hydrogenation rates 4–5 times higher than batch reactor.
" At high conversion, batch reactor provided higher C18:1 selectivity.

a r t i c l e i n f o a b s t r a c t

Article history: Partial hydrogenation of polyunsaturated FAMEs has been investigated on a Pd/C catalyst. The effect of
Received 28 June 2012 type of reactor: batch and continuous flow reactors, on the FAME composition and properties of a biodie-
Received in revised form 8 August 2012 sel product were studied. In addition, many characterization techniques such as XRD, BET, FE-SEM, and
Accepted 14 August 2012
CO-chemisorption were used to examine the prepared catalysts. The result showed that the partial
Available online 7 September 2012
hydrogenation of polyunsaturated FAMEs in a batch-type reactor provides higher selectivity towards
C18:1 than that of a continuous-flow reactor. However, at the low conversion (78%); selectivity of
Keywords:
C18:1 obtained from both types of reactors were almost the same.
Partial hydrogenation
Pd/activated carbon
Ó 2012 Elsevier B.V. All rights reserved.
Biodiesel
Oxidative stability
Reactor

1. Introduction aldehydes, ketones, and acids that change biodiesel properties

and affect the combustion process [4]. Therefore, the saturation
Biodiesel or fatty acid methyl ester (FAME) is known to be one of polyunsaturated FAMEs by partial hydrogenation is a promising
of the most promising alternative fuels due to its biodegradability, way to improve its stability and enhance its utilization potential.
lower emissions of SO2, CO, un-burnt hydrocarbon and particulate The catalysts used in commercial hydrogenation of triglycerides
matters, and higher cetane number as compared to petroleum- are usually Ni catalysts supported on silica or alumina, and mostly
based diesel [1,2]. However, some biodiesel properties such as used in the slurry phases. These Ni catalysts are active in hydroge-
oxidative stability and cold flow properties depend on the natural nation and cheaper than noble metal catalysts, but more severe
characteristics of the starting oil [3]. Biodiesel produced from hydrogenation pressure is required [5,6]. In this respect, noble
material that contains higher unsaturated fatty acid composition metal catalysts such as Pd seem to be the most promising [7]. Pd
has a lower oxidative stability. In contrast, the higher the saturated supported on carbon materials has been extensively employed as
fatty acid composition, the worse the cold flow property becomes. a catalyst for hydrogenation reaction because of its advantages,
The oxidation of the unsaturated FAMEs produces peroxides, e.g., high activity, mild process condition [8], availability of carbon
support, and simplicity of recovery of Pd metal by just burning off
the carbon component [9]. Several recent publications have
⇑ Corresponding author at: The Petroleum and Petrochemical College, addressed the activity of Pd supported on various types of carbon
Chulalongkorn University, Soi Chula 12, Phayathai Rd., Pathumwan, Bangkok
materials such as activated carbon [9–13], nanocomposite carbon
10330, Thailand. Tel.: +66 2 2184148; fax: +66 2 2154459.
E-mail address: apanee.l@chula.ac.th (A. Luengnaruemitchai).
[14], carbon nanofibers [15–17], and carbon nanotube [18]. Of

1385-8947/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cej.2012.08.034
174 N. Numwong et al. / Chemical Engineering Journal 210 (2012) 173–181

the many types of carbon materials, activated carbon is still the The specific surface area, pore volume, and pore size distribu-
most commonly used as a catalyst support, because of its large sur- tion of the support and catalysts were determined by N2 physisorp-
face area (500–1200 m2/g) and low cost [19]. tion using a Quantachrome Autosorb-1 MP surface area analyzer.
In addition, it is generally known that the type of reactor affects Before analyzing, the sample was heated in a vacuum atmosphere
the conversion of the reactant and selectivity of the product after at 250 °C overnight to eliminate volatile species that adsorbed on
hydrogenation. Many researchers have investigated liquid-phase the surface.
hydrogenation in a batch-type reactor [8,11,14,20,21] for the par- Pd dispersion of the prepared catalysts was determined by
tial hydrogenation purpose, e.g., maximizing the intermediate pulse chemisorption of 10% CO/He at room temperature, using
hydrogenated products and minimizing the deeply hydrogenated Temperature-Programmed Desorption/Oxidation/Reduction (TPD/
products. In a batch-type reactor where hydrogenation catalysts R/O) Ohkura R6015. Prior to measurements, each sample was pre-
are used in the slurry phases, a stirrer could enhance the rate of treated with hydrogen at 120 °C for 1 h and followed by purging
reaction, but decrease the lifetime of the catalyst due to mechani- with helium at the same temperature for 10 min. The Pd dispersion
cal damages [22,23]. This attrition of catalysts will also cause some was calculated by assuming a stoichiometry of CO:Pd = 1:1.
problems in the filterability of hydrogenated products. However, li- Furthermore, a Field Emission-Scanning Electron Microscope
quid-phase hydrogenation in a fixed-bed reactor has not been (FE-SEM), Hitachi Model S4800, was utilized to identify the micro-
widely studied. The advantages of fixed-bed reactors are easy sep- structure and capture a micrograph of the catalyst morphology.
aration of the catalyst and products, and lowering mechanical The distribution of the Pd particle size was determined by using
damage of the catalyst particles. SemaphoreÒ.
On the basis of these facts, this study focused on the improve-
ment of oxidative stability of biodiesel by partial hydrogenation
2.4. Partial hydrogenation of polyunsaturated FAMEs
of polyunsaturated FAMEs over Pd/activated carbon. In order to
investigate the influence of the reactor type on the FAME composi-
2.4.1. Batch-type reactor
tion and properties of the hydrogenated biodiesel product, two
Partial hydrogenation of polyunsaturated FAMEs was carried
types of reactors, batch and continuous flow reactors, were stud-
out in a 300 ml stainless steel semi-batch reactor at a temperature
ied. Moreover, details of the characterization of the Pd/C catalyst
and hydrogen partial pressure of 120 °C and 0.4 MPa, respectively.
by XRD, BET, FE-SEM, and CO-chemisorption techniques were also
Stirring rate was maintained at 500 rpm and the flow rate of
reported.
hydrogen gas was 50 ml/min. Firstly, 1.5 g of Pd/C catalyst was
placed into the reactor. Then, the system was purged with nitrogen
2. Materials and methods to remove the remaining air. After that, 100 g of feed biodiesel was
fed into the reactor by a peristaltic pump. The reaction was started
2.1. Materials by increasing the temperature and pressure to the desired point.
Finally, the liquid products were collected every 30 min. The sche-
A commercial activated carbon purchased from Fluka Company matic of the partial hydrogenation experiment in batch-type reac-
with a BET surface area of 836.6 m2/g and particle size 640 lm was tor is shown in Fig. 1a.
used as a support. Palladium (II) nitrate dihydrate [Pd(NO3)22H2O]
40% purum was purchased from Sigma Aldrich and used as a Pd
2.4.2. Continuous flow-type reactor
precursor.
The experiments were performed using a stainless steel up-flow
Gases used in the partial hydrogenation experiment were high
reactor (Fig. 1b) at a temperature range of 80–120 °C and hydrogen
purity (99.99%) hydrogen, and high purity (99.99%) nitrogen sup-
partial pressure of 0.1–0.4 MPa. The flow rate of hydrogen gas was
plied from Thai Industrial Gases Public Co., Ltd. The biodiesel used
fixed at 200 ml/min and was controlled by a mass flow controller.
in this work was obtained from Veerasuwan Co., Ltd., Thailand. In
First, the Pd/C catalyst (0.2 g) was mixed with methyl cellulose and
addition, tetrahydrofuran (stabilizer free, Wako) was used as a sol-
silica diluents. After that, the mixture was pressed and sieved to
vent for GC.
600 lm–150 mm, and packed inside the reactor. Blank tests using
methyl cellulose and silica diluents were run. It was confirmed that
2.2. Catalyst preparation almost no hydrogenation reactions occurred within the experi-
mental conditions. Second, the reactor was assembled to the sys-
The 2 wt.% Pd on the activated carbon catalyst was prepared by tem and pressurized to the desired pressure. After that, the
incipient wetness impregnation (IWI) using a Pd(NO3)22H2O pre- temperature was increased to the desired point and was controlled
cursor. Firstly, activated carbon was dried in an oven at 110 °C by a temperature controller. Then, the feed biodiesel was passed
for 24 h to remove the absorbed water. Secondly, the activated car- through the catalyst bed by a high pressure pump with a flow rate
bon was impregnated with a solution of Pd precursor for 24 h at in the range of 20–180 g/h. Finally, the liquid products were col-
room temperature to ensure that Pd diffused and dispersed thor- lected every hour for 5 h of reaction.
oughly on the surface of the activated carbon. After that, it was FAME composition in the biodiesel before and after the partial
dried in an oven at 110 °C for 24 h. Finally, the dried catalyst was hydrogenation reaction was determined by using a Hewlett Pack-
calcined at 500 °C for 2 h under a nitrogen flow. ard gas chromatograph 6890N equipped with a flame ionization
detector (GC-FID). A HP-88 (100 m  250 lm  0.2 lm) capillary
2.3. Catalyst characterization column was used. Samples of 1 ll were injected under the follow-
ing conditions: the carrier gas was helium with a flow rate of
In order to identify the internal structure and crystallinity of the 2.4 ml/min, the injector temperature was 200 °C with a split ratio
support and catalysts, a Rigaku DMAX2200 X-ray diffractometer of 75:1 and the detector temperature was 230 °C. The sample
was used. Cu Ka radiation (1.5405 Å) was used as an X-ray source was injected at an oven temperature of 155 °C. After an isothermal
at running conditions, for the X-ray tube, of 40 kV and 30 mA. The period of 20 min, the GC oven was heated to 230 °C at a rate of
detector scanned the intensity of diffracted radiation from the 2 °C/min and held for 2.5 min with the total running time of
sample as a function of 2-Theta in the range of 5–90° with a scan 60 min. FAME composition was identified by reference to the
speed of 5°/s. retention time.
 N. Numwong et al. / Chemical Engineering Journal 210 (2012) 173–181 175

(a)

Peristaltic pump

Feed biodiesel

BP

Back pressure
Hydrogen regulator
FC

Mass flow
controller

Product
sampling

Nitrogen

(b)

FC

Mass flow BP

controller
Back pressure
regulator

Hydrogen
High pressure
pump
Sample Non-sample
Feed biodiesel collector collector

Fig. 1. Schematic of partial hydrogenation in batch-type, (a) and continuous flow-type reactor, (b).

In addition, a Thermo Nicolet Fourier Transform-Infrared spec- 3. Results and discussion

trometer (FT-IR) Nexus 670 was used to identify the chemical func-
tional groups of biodiesel before and after partial hydrogenation, in 3.1. Catalyst characterization
order to confirm the results, obtained from GC that the character-
istic peak, which represents C@C disappeared after the biodiesel The characteristics of the prepared catalysts were obtained
was partially hydrogenated. The liquid sample was placed on a using various analytical techniques: XRD, BET, FE-SEM, and CO-
Zn–Se sample holder. The IR spectra at 2 cm1 resolution with 32 chemisorption. A suitable calcination condition was already
scans were collected in the range of 3600–650 cm1. studied by our group [25]. The effect of catalyst calcination
Oxidative stability is one of the major issues indicating the environment (air and N2) and calcination temperature were inves-
content of polyunsaturated FAMEs. The oxidative stability of tigated. It was found that different calcination conditions exhibit
feed and hydrogenated biodiesel was tested by a Metrohm different Pd particle size, which plays an important role in the
743 Rancimat. Sample was aged at 110 °C under a constant air activity of the catalyst. The catalyst calcined under N2 at 500 °C
stream. The oxidative stability corresponds to the period of time for 2 h with 17 nm Pd particle size, provided a good catalytic
before FAMEs are degraded to such an extent that the formation activity for partial hydrogenation of polyunsaturated FAMEs.
of volatile acids can be recorded through an increase in conduc- Therefore, the calcination procedure using N2 at 500 °C for 2 h
tivity. This procedure was developed according to EN 14112 was used. The presence of crystalline Pd after calcination of Pd/C
[24]. catalyst was confirmed by XRD, as shown in Fig. 2b. The four main
Cold flow properties of biodiesel before and after partial characteristic peaks of crystalline Pd including plane (1 1 1), (2 0 0),
hydrogenation including cloud point and pour point were (2 2 0), and (3 1 1) at 2h of 40.20, 46.80, 68.25, and 82.10, respec-
investigated by using a Tanaka mini pour/cloud point tester Series tively, were observed. According to Wu et al. [26]; Pd (1 1 1), Pd
MCP-102, which was developed according to ASTM D6749. (2 0 0), Pd (2 2 0), and Pd (3 1 1) were found after Pd/activated
176 N. Numwong et al. / Chemical Engineering Journal 210 (2012) 173–181

Surface area, total pore volume, and average pore diameter of

the activated carbon support were analyzed by using an Auto-
sorb-1 MP surface area analyzer. The surface area of activated C
was 836.6 m2/g, with the total pore volume of 0.5871 ml/g, and
an average pore diameter of 2.807 nm, which can be indicated as
a mesoporous material (pore size: 2–50 nm). It could be suggested
that almost all of the Pd particles could migrate from the inside
pores of carbon to the outer surface of carbon during the calcina-
tion of Pd precursors.

3.2. Partial hydrogenation of polyunsaturated FAMEs in a batch-type

reactor

In order to evaluate the performance of the catalysts, the partial

hydrogenation of biodiesel was operated and controlled at a condi-
tion of 120 °C, 0.4 MPa, 50 ml/min hydrogen flow rate, 500 rpm
stirring rate, and 1.5 wt.% of catalyst compared to the starting oil.
Fig. 2. XRD patterns of activated carbon support, (a) and Pd/C catalyst calcined
The percentages of C18:0, C18:1, C18:2, C18:3, and other FAMEs
under N2 at 500 °C, (b). (C12:0, C14:0, C16:0, C16:1, C17:0, C 20:0, C20:1, C22:0 and C24:0)
before and after partial hydrogenation reaction are shown in
Fig. 4a. It evidently showed that the higher the reaction time, the
carbon fibers were calcined under N2 atmosphere. In addition, the higher the conversion of unsaturated FAMEs (C18:3, C18:2 and
broad signal at about 25° which is assigned to the amorphous car- C18:1) while the other FAMEs composition was quite stable. As
bon, was also observed in the activated carbon support (Fig. 2a) shown in Fig. 4a, C18:3 and C18:2 were gradually hydrogenated
and the Pd/C catalyst (Fig. 2b) [27]. In order to evaluate hydrogena- as they decreased from 0.08% to 0.02% and 6.12% to 1.33%, respec-
tion activity of the prepared catalyst, the mean particle size of Pd tively; whereas C18:1 and C18:0 increased from 29.93% to 32.44%
was calculated from the peak width at half height of the Pd and 4.90% to 7.01%, respectively, after 1 h of reaction. After that, at
(1 1 1) diffraction peak by applying Scherrer’s equation [9], the 1.5 h of reaction; C18:3 and C18:2 were completely hydrogenated
mean Pd particle size was 17.75 nm. After that, FE-SEM was subse- as there were no significant compositions of them in the biodiesel
quently employed in order to investigate the microstructure and product; while a decreasing in composition of C18:1 due to its
morphology of the catalyst. FE-SEM micrograph of the Pd/C cata- hydrogenation to C18:0 could be observed. This suggests that this
lyst presented many small particles of Pd with a spherical shape; Pd/C catalyst seems to meet the purpose in the partial hydrogena-
highly disperse on the surface of the activated carbon support tion point of view, which showed the behavior of consecutive reac-
(Fig. 3). The average Pd particle size was 14.50 nm, which gave a tions from C18:3 to C18:2, C18:2 to C18:1, and then C18:1 to
result in agreement with the crystallite size calculated from XRD. C18:0, consequently. Many researchers [3,29,30] suggested that
Yeung and Wolf [28] also studied the crystallite size of Pt sup- the higher the amount of saturated FAME, the higher oxidative sta-
ported on highly oriented pyrolytic graphite by scanning tunneling bility was, and a worse cold flow property resulted. This suggests
microscope and found that the crystallite size of Pt is between 10 that the biodiesel product after 1.5 h of reaction would exhibit
and 20 nm. In addition, Pd dispersion and Pd particle size were fur- good fuel properties, especially oxidative stability and cold flow
ther identified by the CO-chemisorption technique, which were properties; where polyunsaturated FAMEs (C18:3 and C18:2) were
2.81% and 17.30 nm, respectively. It was shown that Pd particle fully converted to monounsaturated FAME (C18:1) with a minimal
sizes obtained from those three techniques were in agreement. increase of saturated FAME (C18:0). The FAME composition and
These comparisons also indicated that there would be a small some fuel properties of biodiesel feed and biodiesel product after
number of very small Pd particles over the Pd/C catalyst, which 1.5 h of batch reaction are reported in Table 1. It demonstrates that
would be detected by CO-chemisorption but not detected by FE- partial hydrogenation can improve the oxidative stability of bio-
SEM or XRD. diesel from 1.49 h to 32.50 h, which meets the Thai Standard,
which requires P10 h of oxidative stability even without addition
of any antioxidants. However, the hydrogenated biodiesel exhib-
ited an unfavorable pour point and cloud point, which shows a lim-
itation in cold flow properties of hydrogenated biodiesel. However,
this problem can be minimized by blending with petroleum diesel
in a proper composition [31].
Another interesting observation is an increase of trans-C18:1
composition with the reaction time as presented in Fig. 4b. This
indicates that some of the unreacted cis-C18:1 are transformed
into trans-C18:1 under this pressure and temperature [11]. This re-
sult is consistent with the work done by Agustin and co-workers
[32]. They found that hydrogenation and cis–trans isomerization
of the C@C bonds simultaneously take place during the hydrogena-
tion of vegetable oil or FAMEs.
Fig. 5a shows an FT-IR spectrum of biodiesel feed. The peak at
3006 cm1 corresponding to @CAH stretching in cis-configuration
indicates the presence of unsaturated FAMEs, where nearly all of
the double bonds in the biodiesel feed are in the cis-form [33].
The peaks at 1742 and 1169 cm1, which corresponds to C@O
Fig. 3. FE-SEM micrograph of Pd/C catalyst. stretching and (C@O)AO stretching of FAME were also detected,
 N. Numwong et al. / Chemical Engineering Journal 210 (2012) 173–181 177

Fig. 4. Total FAMEs, (a) and C18 FAMEs composition, (b) of biodiesel after partial hydrogenation in a batch-type reactor as a function of reaction time (reaction conditions:
120 °C, 0.4 MPa, 50 ml/min H2 flow rate, 500 rpm stirring rate, and 1.5 wt.% catalyst compared with starting oil).

creased with hydrogenation time; that is consistent with the re-

Table 1 sults obtained from GC.
FAMEs composition and some fuel properties of biodiesel feed and biodiesel product Moreover, it was found that the correlation between the loga-
after 1.5 h of hydrogenation in a batch-type reactor.
rithm of the ratio of cis/trans C18:1 with the ratio of C18:2 &
Biodiesel feed Biodiesel product after 1.5 h C18:3/C18:1 is linear (not shown here). This suggests that as C18:3
FAMEs composition (%) and C18:2 were hydrogenated to C18:1, a higher isomerization of
Saturated FAMEs 61.92 66.01 cis-C18:1 to trans-C18:1 took place. This result is in accordance with
Monounsaturated FAMEs 30.12 31.63 the work done by Nikolaou and co-workers [4]; they found that
trans-Monounsaturated FAMEs 0.11 17.36 C18:1 cis/trans ratio presented a high exponential correlation with
cis-Monounsaturated FAMEs 30.01 14.27
Diunsaturated FAMEs 6.12 0.31
the conversion during hydrogenation of sunflower oil using Rh sul-
Triunsaturated FAMEs 0.08 0.00 fonated triphenylphosphite complexes. The trend of activity, selec-
Fuel properties
tivity, and cis–trans isomerization of precious metal catalysts has
Oxidative stability (h) 1.49 32.50 been reported by Dijkstra [36]. Pd, which is the most active metal,
Pour point (°C) 16.0 22.0 also causes the most cis–trans isomerization and migration of double
Cloud point (°C) 16.0 23.0 bonds. The presence of trans-isomers leads to a much higher melting
point than those of the corresponding cis-isomers [37]. The melting
respectively [34]. Fig. 5b shows the disappearance of a peak of point of trans-C18:1 is 9 °C, whereas it is only 20 °C for cis-C18:1
@CAH stretching at 3006 cm1 in the biodiesel product with an in- [38]. Therefore, for our future work, we will modify the catalysts to
crease in hydrogenation time: 0, 0.5, 1.5, and 3 h in a batch reactor. minimize the trans-isomers formation for improving the cold flow
The disappearance of this peak monitors the saturation of double properties of the partially hydrogenated products.
bond during hydrogenation reaction and the isomerization of
remaining double bonds in the original cis-form. This explanation 3.3. Partial hydrogenation of polyunsaturated FAMEs in a continuous
can be confirmed by the appearance of the peak at 966 cm1 after flow reactor
partial hydrogenation, as shown in Fig. 5c, which can be attributed
to CAH out of plane in trans-configuration [35]. It is clearly seen The partial hydrogenation of polyunsaturated FAMEs in a
that the peak intensity of trans-configuration at 966 cm1 in- continuous flow reactor was conducted at various parameters:
178 N. Numwong et al. / Chemical Engineering Journal 210 (2012) 173–181

Fig. 5. FT-IR spectra of (a) biodiesel feed, (b)@CAH stretching, and (c) CAH out of plane in trans-configuration in biodiesel product after partial hydrogenation in a batch
reactor at 0, 0.5, 1.5, and 3 h.

reaction temperature, hydrogen partial pressure, and biodiesel

feed flow rate.

3.3.1. Effect of reaction temperature

To study the effect of temperature; hydrogen partial pressure,
hydrogen flow rate, and biodiesel feed flow rate, respectively, were
fixed at 0.4 MPa, 200 ml/min, and 90 g/h, while the reaction tem-
perature was varied from 80 to 120 °C. It was found that the con-
version of C18:2 and C18:3 increased with the higher reaction
temperature. The conversions of C18:2 and C18:3 at 80 °C,
100 °C, and 120 °C were 53%, 66%, and 73%, respectively. Fig. 6
shows that the increase in temperature caused an increase of
trans-C18:1 and C18:0. This was due to the transformation of cis-
C18:1 to trans-C18:1 as the result of the higher temperature and
higher reaction rate at the higher temperature according to kinet-
ics principle.
Fig. 6. Effect of reaction temperature on composition of C18 FAMEs (reaction
conditions: 0.4 MPa, 200 ml/min H2 flow rate, and 90 g/h biodiesel feed flow rate).
3.3.2. Effect of hydrogen partial pressure
The effect of hydrogen partial pressure was investigated by
varying hydrogen partial pressure from 0.1 to 0.4 MPa; whereas partial pressure also had a significant effect on the conversion of
temperature, hydrogen flow rate, and biodiesel feed flow rate were polyunsaturated FAMEs. With the hydrogen partial pressure being
fixed at 120 °C, 200 ml/min, and 90 g/h, sequentially. Hydrogen increased from 0.1 to 0.4 MPa, the conversion of C18:2 and C18:3
 N. Numwong et al. / Chemical Engineering Journal 210 (2012) 173–181 179

Fig. 8. Effect of contact time (1/WHSV) on composition of C18 FAMEs (reaction

conditions: 120 °C, 0.4 MPa, 200 ml/min H2 flow rate).

Fig. 7. Effect of hydrogen partial pressure on composition of C18 FAMEs (reaction

conditions: 120 °C, 200 ml/min H2 flow rate, and 90 g/h biodiesel feed flow rate).

increased. It was 55%, 60%, and 71% C18:2 and C18:3 conversion for
0.1 MPa, 0.2 MPa, and 0.4 MPa, respectively. This can be explained
by the increase of hydrogen partial pressure that resulted in more
presence of hydrogen, which caused the hydrogenation reaction to
accelerate faster. Moreover, it showed that at a higher hydrogen
partial pressure, the higher amount of trans-C18:1 and C18:0 was
produced, as shown in Fig. 7. This is due to the transformation of
cis-C18:1 to trans-C18:1 under the higher pressure, which was in
accordance with the result obtained at the higher reaction
temperature.

3.3.3. Effect of biodiesel feed flow rate

The biodiesel feed flow rate was varied from 20 to 180 g/h;
where the other parameters were fixed at 120 °C, 0.4 MPa of
hydrogen partial pressure, and a 200 ml/min hydrogen flow rate.
As shown in Table 2, the conversions in a continuous flow reactor
were equivalent to that in the batch reactor, even at the short con-
tact time, i.e., average contact time of 0.01 h for the biodiesel flow
rate of 20 g/h, using 0.2 g of Pd/C catalyst. On the contrary, 100 g of
biodiesel was hydrogenated over 1.5 g of Pd/C catalyst within the
reaction time of 2 h in a batch reactor. These data suggest that con-
tinuous flow reactor was more preferable than the batch-type
reactor in terms of the hydrogenation-processing rate per unit
amount of Pd/C catalyst. Fig. 8 shows a comparison of C18 FAMEs
as a function of contact time. With an increase in contact time, the
amount of C18:3 and C18:2 gradually decreased, whereas the
amount of trans-C18:1 and C18:0 increased. Deeper hydrogenation
of trans-C18:1 into C18:0 was significant under the longer contact
time between the biodiesel and catalysts.
The oxidative stability of biodiesel products obtained from
different biodiesel feed flow rates were investigated. Fig. 9a shows

Table 2
C18:2 and C18:3 conversions of batch and continuous-flow reactor at different
reaction and contact times. Fig. 9. Oxidative stability (a) and cold flow properties (b) of biodiesel feed and
biodiesel product operated at different biodiesel feed flow rates.
Batch Continuous-flow
Reaction Conversion Biodiesel feed flow Contact Conversion
time (h) (%) rate (g/h) time (h) (%) the oxidative stability as a function of composition of polyunsatu-
0.5 58.6 180 0.0011 44.2 rated FAMEs (C18:2 and C18:3). After hydrogenation; the compo-
1 78.3 120 0.0017 57.2 sition of C18:2 and C18:3 decreased with an increase of oxidative
1.5 94.5 90 0.0022 69.7
stability. At the higher degree of hydrogenation (the lower biodie-
2 97.1 60 0.0033 77.5
3 97.9 50 0.0040 88.5 sel feed flow rate), the lower composition of C18:2 and C18:3 and
4 98.3 40 0.0050 90.9 resulted in higher oxidative stability. It could be suggested that to
20 0.0100 96.2 meet the Thai standard, which requires 10 h of oxidative stability;
the composition of C18:2 and C18:3 should be less than 0.65%.
180 N. Numwong et al. / Chemical Engineering Journal 210 (2012) 173–181

Moreover, cold flow properties of biodiesel as a function of sat-

urated FAMEs (C18:0) composition in the biodiesel product, oper-
ated at different biodiesel feed flow rates, were also investigated as
presented in Fig. 9b. With a higher degree of hydrogenation (utiliz-
ing a lower biodiesel feed flow rate), the higher composition of
C18:0 and resulted in an increasing of cloud point and pour point.
According to the above results, it could be concluded that reac-
tion conditions of contact time of 0.004 h where biodiesel feed flow
rate was 50 g/h and the amount of Pd/C catalyst was 0.2 g, would
be optimal for partial hydrogenation in a continuous flow reactor,
where we could obtain biodiesel products with oxidative stability
that meet the Thai standard without the addition of any antioxi-
dants. At this contact time, the biodiesel product composed of
11.5% of C18:0 with 20 °C and 21 °C pour point and cloud point,
respectively. Therefore, the critical composition of C18 FAMEs to
Fig. 10. Comparison of C18 FAMEs composition of biodiesel feed and product
get biodiesel products that pass the Standard are as follows:
obtained from batch (B) and continuous-flow (C) reactors at different conversions of
C18:2 and C18:3 6 0.65%, C18:1 P 25%, and C18:0 6 12%. C18:2 and C18:3.

3.4. Difference in catalytic performances between a batch-type reactor

and continuous-flow reactor at the high conversion (>80%); the batch reactor provides higher
selectivity towards cis-C18:1. Nevertheless, at the lower conver-
The catalytic performances of 2 wt.% Pd/C catalyst in a batch- sion; both types of reactor provide product with almost the same
type reactor and continuous-flow reactor were compared by fixing cis-C18:1 selectivity. This was consistent with the result of selec-
the reaction temperature and hydrogen partial pressure at 120 °C tivity towards C18:1 obtained from both types of reactor, which
and 0.4 MPa, respectively. could be explained by the different in contact probability between
In order to compare the partial hydrogenation of polyunsaturated catalyst and oil in both types of reactor.
FAMEs in batch and continuous flow reactors, the FAME composition Thus, it can be suggested that, to obtain the high conversion and
of biodiesel product at the same conversion were compared. Fig. 10 high selectivity towards C18:1 and cis-C18:1, the partial hydrogena-
displays the comparison of C18 FAMEs composition in the biodiesel tion of polyunsaturated FAMEs in a batch-type reactor is better.
feed and a biodiesel product obtained from batch and continuous- However, to obtain the same selectivity at a lower conversion, oper-
flow reactors at different conversions. At the lowest conversion of ating in a continuous flow reactor is better because it requires much
C18:2 and C18:3 (58%), selectivity of C18:1 obtained from the lower residence time when compared to a batch-type reactor.
batch-type reactor and continuous flow reactor was almost the
same, although the hydrogenation processing rate was WBDF/Wcat/
4. Conclusions
4t = 100 g/1.5 g/0.5 h = 133 h1 and WHSV = 120 g/0.2 g/1 h = 600
h1, respectively. At the conversion of 78%, selectivity of C18:1 ob-
Partial hydrogenation of polyunsaturated FAMEs has been
tained from both types of reactor was still the same, even though
investigated on a Pd/C catalyst. It was found that this catalyst pro-
the hydrogenation processing rate of the batch reactor (WBDF/Wcat/
vides good partial hydrogenation activity, which showed the
4t = 100 g/1.5 g/1 h = 67 h1) was lower than that of the continuous
behavior of consecutive reactions from C18:3 to C18:2, C18:1,
flow reactor (WHSV = 60 g/0.2 g/1 h = 300 h1). On the other hand,
and C18:0, resulting in an improvement of the oxidative stability.
at the conversion of 94%, a biodiesel product obtained from a
Moreover, the results showed that the type of reactor used had
batch-type reactor provided higher selectivity towards C18:1,
an important effect on the conversion and selectivity of C18:1. It
whereas the hydrogenation processing rate of the batch and contin-
was found that the partial hydrogenation in a continuous flow
uous flow reactors was WBDF/Wcat/4t = 100 g/1.5 g/1.5 h = 44 h1
reactor provided hydrogenation rates 4–5 times higher than in a
and WHSV = 40 g/0.2 g/1 h = 200 h1, respectively.
batch type reactor. However, the partial hydrogenation in a
This suggests that a continuous flow reactor could hydrogenate
batch-type reactor provided higher selectivity towards C18:1 at a
biodiesel about 4–5 times faster than a batch-type reactor. These
high conversion. This can be explained by the possibility of oil con-
differences in reactivity between a batch-type reactor and contin-
tact with the catalyst surface in a batch reactor being lower than
uous flow reactor could be explained by the possibility of oil con-
that of a continuous flow reactor. This lesser contact between oil
tact with the catalyst surface where molecular hydrogen is
and catalyst in a batch type reactor would suppress the deeper
activated, i.e., more frequent contact in a continuous flow reactor
hydrogenation of intermediate products (C18:1) into saturated
and less contact in a batch reactor. This lesser contact could sup-
products (C18:0). In contrast, at the lower conversion; selectivity
press the deeper hydrogenation of the hydrogenated intermediate
of C18:1 obtained from both types of reactors were almost the
products (C18:1) into saturated products (C18:0). This explanation
same. In conclusion, it could be suggested that to obtain the high
can be confirmed from the higher selectivity towards C18:1 of the
conversion and high selectivity towards C18:1, partial hydrogena-
batch-type reactor at the conversion of 94%. Gómez-Quero et al.
tion of polyunsaturated FAMEs in a batch type reactor is better.
[39] also found that switching from a batch to continuous reactor
However, to obtain the same selectivity at a lower conversion;
resulted in an increase in hydrodechlorination rate of 2,4-dichloro-
operating in continuous flow reactor is better because it requires
phenol over Pd/Al2O3.
much lower residence time when compared to the batch-type
In addition, the selectivity of cis-C18:1 in biodiesel products ob-
reactor.
tained from batch and continuous flow reactors as a function of
conversion of C18:2 and C18:3 were also considered. It obviously
presented that at higher C18:2 and C18:3 conversion, the lower Acknowledgements
cis-C18:1 selectivity was obtained. This was due to the further
hydrogenation of cis- and trans-C18:1 to C18:0 at higher C18:2 The authors gratefully acknowledge the financial support from
and C18:3 conversion. Furthermore, the results also showed that the Chulalongkorn University Dutsadi Phipat Endowment Fund,
 N. Numwong et al. / Chemical Engineering Journal 210 (2012) 173–181 181

Chulalongkorn University, Thailand, and Center of Excellence on [17] D.B. Thakur, R.M. Tiggelaar, J.G.E. Gardeniers, L. Lefferts, K. Seshan, Carbon
nanofiber based catalyst supports to be used in microreactors: synthesis and
Petrochemical and Materials Technology, Thailand.
characterization, Chem. Eng. J. 160 (2010) 899–908.
[18] C. Ge, Y. Li, J. Zhao, R. Zhou, Carbon nanotubes supported palladium catalysts
for selective hydrogenation of cinnamaldehyde under atmospheric pressure,
Appendix A. Supplementary material Indian J. Chem. 49A (2010) 281–287.
[19] M.L. Toebes, J.A.V. Dillen, K.P. de Jong, Synthesis of supported palladium
catalysts, J. Mol. Catal. A: Chem. 173 (2001) 75–98.
Supplementary data associated with this article can be found, in [20] I.V. Deliy, N.V. Maksimchuk, R. Psaro, N. Ravasio, V.D. Santo, S. Recchia, E.A.
the online version, at http://dx.doi.org/10.1016/j.cej.2012.08.034. Paukshtis, A.V. Golovin, V.A. Semikolenov, Kinetic peculiarities of cis/trans
methyl oleate formation during hydrogenation of methyl linoleate over Pd/
MgO, Appl. Catal. A 279 (2005) 99–107.
[21] A.F. Perez-Cadenas, F. Kapteijn, M.P. Zieverink M, J.A. Moulijn, Selective
References hydrogenation of fatty acid methyl esters over palladium on carbon-based
monoliths structural control of activity and selectivity, Catal. Today 128 (2007)
[1] N. Ellis, F. Guan, T. Chen, C. Poon, Monitoring biodiesel production 13–17.
(transesterification) using in situ viscometer, Chem. Eng. J. 138 (2008) 200– [22] J. Ni, F.C. Meunier, Esterification of free fatty acids in sunflower oil over solid
206. acid catalysts using batch and fixed bed-reactors, Appl. Catal. A 333 (2007)
[2] Y.H. Taufiq-Yap, H.V. Lee, R. Yunus, J.C. Juan, Transesterification of non-edible 122–130.
Jatropha curcas oil to biodiesel using binary Ca–Mg mixed oxide catalyst: [23] N. Shibasaki-Kitakawa, H. Honda, H. Kuribayashi, T. Toda, T. Fukumura, T.
effect of stoichiometric composition, Chem. Eng. J. 178 (2011) 342–347. Yonemoto, Biodiesel production using anionic ion-exchange resin as
[3] T. Sonthisawate, A. Suemanotham, Y. Yoshimura, T. Makoto, Y. Abe, Upgrading heterogeneous catalyst, Bioresour. Technol. 98 (2007) 416–421.
of biodiesel fuel quality by partial hydrogenation Process, Global Warming [24] M.J. Ramos, C.M. Fernandez, A. Casas, L. Rodriguez, A. Perez, Influence of fatty
Conference, Biodiversity and their Sustainable Use, 2009, pp. 90–97. acid composition of raw materials on biodiesel properties, Bioresour. Technol.
[4] N. Nikolaou, C.E. Papadopoulos, A. Lazaridou, A. Koutsoumba, A. Bouriazos, G. 100 (2009) 261–268.
Papadogianakis, Partial hydrogenation of methyl esters of sunflower oil [25] N. Numwong, A. Luengnaruemitchai, Y. Yoshimura, Effects of palladium
catalyzed by highly active rhodium sulfonated triphenylphosphite precursor and catalyst calcination condition on the catalytic activity of
complexes, Catal. Commun. 10 (2009) 451–455. activated carbon-supported palladium in partial hydrogenation of
[5] H.W.B. Patterson, in: G.R. List, J.W. King (Eds.), Hydrogenation of Fats and Oils: polyunsaturated fatty acid methyl esters, Renew Energy Conf (2010).
Theory and Practice, second ed., American Oil Chemist’s Society Press, 2011, [26] X. Wu, N.C. Gallego, C.I. Contescu, H. Tekinalp, V.V. Bhat, F.S. Baker, M.C. Thies,
pp. 169–187. The effect of processing conditions on microstructure of Pd-containing
[6] R.R. Allen, M.W. Formo, R.G. Krishnamurthy, G.N. McDermott, F.A. Norris, activated carbon fibers, Carbon 46 (2008) 54–61.
N.O.V. Sonntag, Bailey’s Industrial Oil and Fat Products, in: D. Swern (Ed.), [27] H. Wu, L. Zhuo, Q. He, X. Liao, B. Shi, Heterogeneous hydrogenation of
fourth ed., A Wiley-Interscience Publication, John Wiley & Sons, 1982, pp. 1– nitrobenzenes over recyclable Pd(0) nanoparticles catalysts stabilized by
95. polyphenol-grafted collagen fibers, Appl. Catal. A 366 (2009) 44–56.
[7] M. Zajcew, The hydrogenation of fatty oils with palladium catalyst IV. Pilot- [28] K.L. Yeung, E.E. Wolf, Scanning tunneling microscopy studies of size and
plant preparation of shortening stocks, J. Am. Oil Chem. Soc. 37 (1960) 130– morphology of Pt graphite catalysts, J. Catal. 135 (1992) 13–26.
132. [29] G. Knothe, Dependence of biodiesel fuel properties on the structure of fatty
[8] J. Panpranot, O. Tangjitwattakorn, P. Praserthdam, J.G. Goodwin Jr, Effects of Pd acid alkyl esters, Fuel Process. Technol. 86 (2005) 1059–1070.
precursors on the catalytic activity and deactivation of silica-supported Pd [30] M.J. Ramos, C.M. Fernández, A. Casas, L. Rodríguez, A. Pérez, Influence of fatty
catalysts in liquid phase hydrogenation, Appl. Catal. A 292 (2005) 322–327. acid composition of raw materials on biodiesel properties, Bioresour. Technol.
[9] T. Harada, S. Ikeda, M. Miyazaki, T. Sakata, H. Mori, M. Matsumura, A simple 100 (2009) 261–268.
method for preparing highly active palladium catalysts loaded on various [31] S. Goto, M. Oguma, N. Chollacoop, L. Dowling, D. Sheedy, W. Zhang et al., EAS-
carbon supports for liquid-phase oxidation and hydrogenation reactions, J. ERIA biodiesel fuel trade handbook: 2010, 1st printed, Economic Research
Mol. Catal. A: Chem. 268 (2007) 59–64. Institute for Asian and East Asia (ERIA), 2008.
[10] H. Tamai, U. Nobuaki, H. Yasuda, Preparation of Pd supported mesoporous [32] F. Agustin, C. Perez, M.P.Z. Martijn, K. Freek, A.M. Jacob, High performance
activated carbons and their catalytic activity, Mater. Chem. Phys. 114 (2009) monolithic catalysts for hydrogenation reactions, Catal. Today 105 (2005)
10–13. 623–628.
[11] K. Wadumesthrig, S.O. Salley, K.Y.S. Ng, Effects of partial hydrogenation, [33] S. GuO, K.Y. Liew, J. Li, Catalytic activity of ruthenium nanoparticles supported
epoxidation, and hydroxylation on the fuel properties of fatty acid methyl on carbon nanotubes for hydrogenation of soybean oil, J. Am. Oil Chem. Soc. 86
esters, Fuel Process. Technol. 90 (2009) 1292–1299. (2009) 1141–1147.
[12] L. Wang, X. Chen, J. Liang, Y. Chen, X. Pu, Z. Tong, Kinetics of the catalytic [34] N.P.G. Roeges, A guide to complete interpretation of infrared spectra of organic
isomerization and disproportionation of rosin over carbon-supported structures, John Wiley & Sons, Chichester, 1994.
palladium, Chem. Eng. J. 152 (2009) 242–250. [35] J. Fritsche, H. Steinhart, in: R.E. McDonald, M.M. Mossoba (Eds.), New
[13] P. Mäki-Arvela, G. Martin, I. Simakova, A. Tokarev, J. Wärnå, J. Hemming, B. techniques and applications in lipid analysis, American Oil Chemist’s Society
Holmbom, T. Salmi, D.Yu. Murzin, Kinetics, catalyst deactivation and modeling Press, Champaingn IL, 1997, pp. 234–255.
in the hydrogenation of ß-sitosterol to ß-sitostanol over microporous and [36] A.J. Dijkstra, Revisiting the formation of trans isomers during partial
mesoporous carbon supported Pd catalysts, Chem. Eng. J. 154 (2009) 45–51. hydrogenation of triacylglycerol oils, Eur. J. Lipid Sci. Tech. 108 (2006) 249–
[14] I.L. Simakova, O.A. Simakova, A.V. Romaneko, D.Y. Murzin, Hydrogenation of 264.
vegetable oils over Pd on nanocomposite carbon catalysts, Ind. Eng. Chem. Res. [37] G. Knothe, ‘‘Designer’’ biodiesel: optimizing fatty ester composition to improve
47 (2008) 7219–7225. fuel properties, Energy Fuels 22 (2008) 1358–1364.
[15] R.F. Bueres, E. Asedegbega-Nieto, E. Diaz, S. Ordonez, F.V. Diez, Preparation of [38] G. Knothe, R.O. Dunn, A comprehensive evaluation of the melting points of
carbon nanofibres supported palladium catalysts for hydrodechlorination fatty acids and esters determined by differential scanning calorimetry, J. Am.
reactions, Catal. Commun. 9 (2008) 2080–2084. Oil Chem. Soc. 86 (2009) 843–856.
[16] J. Zhu, J. Zhou, T. Zhao, X. Zhou, D. Chen, W. Yuan, Carbon nanofiber-supported [39] S. Gómez-Quero, F. Cárdenas-Lizana, M.A. Keane, Liquid phase catalytic
palladium nanoparticles as potential recyclable catalyst for the Heck reaction, hydrodechlorination of 2,4-dichlorophenol over Pd/Al2O3: batch vs.
Appl. Catal. A 352 (2009) 243–250. continuous operation, Chem. Eng. J. 166 (2011) 1044–1051.