Renewable Energy 99 (2016) 622e630

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Biodiesel production over lime. Catalytic contributions of bulk phases

and surface Ca species formed during reaction
Ana Paula Soares Dias a, *, Jaime Puna a, b, Joa
~o Gomes a, b, Maria Joana Neiva Correia a,
~o Bordado a
Joa
a
LAETA, IDMEC, CERENA e Instituto Superior T ecnico, Universidade de Lisboa, Av. Rovisco Pais, s/n, 1049-001 Lisboa, Portugal
b
ISEL e Instituto Superior de Engenharia de Lisboa, Chemical Engineering Department/CEEQ, R. Conselheiro Emdio Navarro 1, 1959-007 Lisboa, Portugal

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

Article history: Lime is pointed out as an effective catalyst for biodiesel production by oil methanolysis. Several Ca phases
Received 18 November 2015 are formed during reaction. Each Ca phase has different contribution to the catalyzed process.
Received in revised form Using CaO as a catalyst, S shape kinetics curve was observed and the induction period can be ascribed
1 June 2016
to the Ca(OH)2 formation. When Ca(OH)2, prepared by contacting CaO with H2O, is used as catalyst the
Accepted 15 July 2016
initial period with a slow rate of transesterication has almost vanished. Besides, if the catalyst surface is
totally converted into methoxide species the induction period is longer than the analogous obtained with
CaO. This is an indication that the methoxide species strongly bonded to Ca are less reactive.
Keywords:
Biodiesel
The calcium diglyceroxide material (CaO_diglyc), prepared by contacting CaO with a mixture of
Lime methanol and glycerol, displays a totally different kinetics curve with no induction period. The faster
Calcium hydroxide kinetics and the Ca species detected in the glycerin phase seem to underline a non-negligible homo-
Calcium methoxide geneous process contribution.
Calcium diglyceroxide The characterization of the post-reaction catalysts underlines the relevance of the surface and bulk
Active phase catalyst modications. Calcium hydroxide can be pointed out as the active phase whereas calcium
diglyceroxide is responsible for the catalyst deactivation due to calcium leaching.
2016 Elsevier Ltd. All rights reserved.

1. Introduction Therefore, the substitution of homogeneous catalysts by hetero-

Fuels production from renewable materials has been attracting
increasing interest over the last few decades as they are considered
to be part of the solution for some of the most important challenges
of modern life, which include concerns over energy security and
the need to mitigate climate changes [1]. Despite the problems
associated with rst generation biofuels like the fuel versus food
debate and land use, rst generation biodiesel, a mixture of long
chain fatty acids methyl or ethyl esters, is the presently available
and feasible alternative to substitute fossil diesel in internal com-
bustion engines [2]. There are a large number of commercial plants
producing biodiesel by transesterication of vegetable oils and fats
using homogenous basic catalysts [3]. However, homogeneous
processes have several drawbacks related to phases separation and
washing steps and to the impossibility of catalyst reutilization.
* Corresponding author.
E-mail address: apsoares@tecnico.ulisboa.pt (A.P. Soares Dias).
geneous ones may allow to achieve more efcient and economical
biodiesel production processes [4,5]. Several materials are referred
in the literature as potential catalysts for transesterication re-
actions, including biodiesel production from vegetable oils or fats
[6,7]. Among them, lime, a low cost material, is indicated as pre-
senting interesting catalytic performances for the methanolysis of
vegetable oils [8]. When lime is used as catalyst, several Ca phases
might be formed in the reaction medium (calcium hydroxide, cal-
cium diglyceroxide and calcium methoxide) and these phases may
have different contributions to the catalytic process [9]. Raafat [10]
stated that the reactivity order Ca(OH)2<CaO < Ca(OCH3)2 agrees
with the Lewis basic theory: the methoxides of alkaline earths
metals are more basic than their oxides which are more basic than
their hydroxides.
The high reactivity of methanol with calcium oxide, even at
room temperature, compelled Kouzu et al. [11] to argue that cal-
cium methoxide is the active phase, instead of CaO, during oil
methanolysis. Other researchers also studied the oil methanolysis
over calcium methoxide and underlined its excellent catalytic

http://dx.doi.org/10.1016/j.renene.2016.07.033
0960-1481/ 2016 Elsevier Ltd. All rights reserved.
 A.P. Soares Dias et al. / Renewable Energy 99 (2016) 622e630
FAME yield
Reacon me
Fig. 1. Typical kinetics curve of oil methanolysis over lime catalyst (adapted from
Ref. [6]).
behavior [12]. Calcium hydroxide was also referred to be active for
biodiesel production through a mechanism involving the H
abstraction of CH3OH by surface eOH groups [10]. Some authors
[13] reported a positive effect for small water contents which can be
endorsed by an improvement of surface eOH group availability.
Higher water contents promote undesirable soap formation.
A new Ca phase, calcium glycerolate, with excellent catalytic
performances was recently reported by Rayero et al. [14].
The existence of several Ca species in the reaction medium, with
dissimilar catalytic behavior, potentially contributes to the S-shape
kinetics curves (Fig. 1) observed when oil methanolysis is carried
out over lime catalysts [15]. Most of the researchers ascribed the
sigmoid kinetics curves to the existence of mass transfer limitations
for the initial reaction period followed by a kinetics control for the
later period. Recently Csernica and Hsu [16] reported, for homo-
geneous catalyzed oil methanolysis, that the S-shape kinetics curve,
usually ascribed to a transition from a mass transfer controlled
regime, is, in fact, ascribable to the transition from two-phases to a
single phase system. Analogous results, for homogenous catalysis,
were reported by Likosar and Levec [17] and Likosar et al. [18].
In order to clarify the sigmoid form of the kinetic curve, the
rapeseed oil methanolysis was carried out over CaO, Ca(OH)2 and
calcium diglyceroxide. The catalysts, fresh and post-reaction, were
characterized by XRD and HATR-FTIR in order to identify the bulk
and surface Ca species formed. The role of each Ca phase in the
catalytic behavior during biodiesel production and in the quality of
the glycerin phase is discussed in the following sections.
2. Experimental
Methanol (99%) and the rapeseed oil (average composition
Table 1) were supplied by Iberol S.A., a Portuguese biodiesel pro-
ducer. The oil is used as raw-material in the industrial process and it
Table 1
Characteristics of the rapeseed oil provided by Iberol (local biodiesel industry).
FFA (%)
C14:0 C16:0 C16:1 C18:0 C18:1 C18:2
0.1 4.2 0.3 2.1 62.8 19.2
Iodine value: 112
Acidity: 0.55 mg KOH/g oil
Water: 334 ppm
623
is therefore semi-rened with an acidity of 0.55 mg KOH/g oil and a
water content of 334 mg/kg.
2.1. Preparation of the catalysts
Commercial CaO, from a local producer (Lusical), was used as
raw material for catalysts preparation. The as received material, a
chalky-white material with a wide granulometric distribution, was
grounded in a pyrex mortar and dried at 120  C overnight. The
calcium diglyceroxide was prepared by contacting (5 h) the CaO
with a mixture containing equal amounts of glycerin and methanol
(methanol reux temperature). A calcium methoxide sample was
prepared contacting CaO with methanol in analogous conditions. A
similar procedure was used to obtain calcium hydroxide samples,
where a water suspension of the grounded CaO was stirred during
5 h at 62  C. For each catalyst batch 20 g of raw CaO (dried overnight
at 120  C) and 150 mL of liquid (water, methanol or methanol/
glycerin) were used.
All the prepared samples were dried overnight at 120  C before
the catalytic tests. The post reaction catalysts (named in the next
sections as PR or PR#batch number) were washed with methanol
(around 20 mL during ltration) and dried overnight at 120  C.
2.2. Characterization of the catalysts
The surface area (BJH isotherm) of the raw CaO was assessed by
N2 physisorption, at 77 K, using a Micromeritics ASAP 2010 appa-
ratus. Particles size distributions of the fresh and post reaction
catalysts were assessed by laser diffraction (blue radiation, 455 nm)
based on Lorenz Mie law using a Malvern Mastersizer 2000
equipment. Water dispersions (with 10e20% of obscurity) of the
materials were prepared using ultrasound. The granulometric dis-
tributions were computed taking into account the refractive index
of CaO (1.838). The external surface areas of the catalysts were
computed considering spherical particles.
The surface basicity was assessed using basic Hammett in-
dicators in methanolic solutions. Experimental details are given
elsewhere [19].
The XRD patterns of fresh and post reaction (PR#batch number)
catalysts were recorded with a Rigaku Geigerex diffractometer
with Cu Ka radiation at 40 kV and 40 mA (2 /min).
In order to identify the reaction species (hydroxyl, methoxide
and others) on the catalysts surface infrared spectra were collected,
for fresh and post reaction catalysts, with a resolution of 16 cm 1,
using a FT-MIR equipment from BOMEN (FTLA2000-100, ABB) with
a DTGS detector. A horizontal total attenuated reection accessory
(HATR), from PIKE Technologies, with a ZnSe crystal was used.
Sixty-four scans were accumulated for each spectrum to obtain an
acceptable signal-to-noise ratio. Furthermore, the spectra were
recorded in duplicate for each sample and the average of the two
measurements was used for analysis.
2.3. Methanolysis tests
The methanolysis of rapeseed oil tests were carried out in a
500 mL three-neck round bottom ask equipped with a
C18:3 C20:0 C20:1 C22:0 C24:0 C24:1
9.0 0.7 1.2 0.3 0.2 0.1
624 A.P. Soares Dias et al. / Renewable Energy 99 (2016) 622e630

condenser and a mechanical stirrer. The reaction temperature

was kept at 60  C by a water bath. Typically, as described else-
where, a mixture of 100 g of oil with 12:1 (molar ratio) of
methanol (p.a.) was used as reaction mixture. For each test 5%
(w/w, oil basis) of catalyst was used [19]. At the end of the re-
action period, the catalyst was removed by ltration and the
reaction mixture was transferred into a decantation funnel. The
esters phase (crude biodiesel) was washed with water, with a
0.1 M HCl solution and again with water to provide a puried
biodiesel. The washed methyl esters were then centrifuged
(Sigma 4K10, Osterode) Osterodeam Harz, Germany and dried at
80e90  C under vacuum (200 mbar) using a rotary evaporator
(RE111; Bchi, Flawil, Switzerland). The FAME (Fatty Acids Methyl
Esters) content of the biodiesel phase after purication was
evaluated by Near Infrared Spectroscopy (NIR), according to the
procedure described elsewhere [20]. The spectra of the biodiesel
samples were acquired using an ABB BOMEM MB160 (Zurich,
Switzerland) spectrometer equipped with an InGaAs detector
and a transectance probe from SOLVIAS (Basel, Switzerland).
The spectra were recorded in duplicate for each sample at room
temperature, with the aid of the Galactic Grams software package Fig. 2. Granulometric distribution (by laser diffraction) of raw and post reaction (4 h)
(Galactic Industries, Salem, NH, USA), in the wave number range CaO catalyst.
of 12000e4000 cm1, with a spectral resolution of 16 cm1. As
described, gas chromatography was the reference method used The catalysts characterization using Hammett indicators
to derive the NIRS calibration model. The spectral region be- showed, for the raw CaO and prepared catalysts, basicity strengths
tween 6102 and 5880 cm1 was used for determining the FAME in accordance with previously published data [10,22]. With H_ in
content of the samples. Within the calibration range (78.4e99.3% the range 7.2e17.6 the catalysts basicity was ordered as follows:
of FAME), NIR spectroscopy allows the determination of the
FAME content of biodiesel with an error of 1%, which is lower CaO_H2O < CaO < Ca_MeOH; Ca_diglyc
than the maximum error expected for the GC-reference method
error (1.5%) [21]. The XRD samples of fresh catalysts in Fig. 3 show lines ascribable
to lime, for the bare CaO material slightly contaminated with
3. Results and discussion

3.1. Characterization of the catalysts

The raw CaO catalyst morphology was previously examined by

scanning electron microscopy [22]. The micrographies showed
non-uniform crystals in shape and size. Additionally the Ca-
diglyceroxide was easy to identify by SEM due to the presented
cubic morphology.
The N2 adsorption/desorption isotherms of CaO showed a
pattern belonging to a macroporous material. Taking the BJH
isotherm the computed surface area was 15.03 m2/g with a porous
volume of 0.06 cm3/g and average porous size of 16.85 nm. The
granulometric distribution (Fig. 2), assessed by laser diffraction,
showed a multimodal distribution with particles smaller than
100 mm and a surface weighted mean diameter of 11.8 mm.
The external surface area computed for spherical particles was
0.51 m2/g. Taking into account the size of the oil molecules only the
external surface area will be effective for reaction, as the oil mol-
ecules cannot diffuse into the catalyst poros [20]. The post reaction
CaO catalyst (4 h of reaction) showed a mean granulometric dis-
tribution presented in Fig. 2 that corresponds to a surface weighted
mean diameter of 8.2 mm . Fig. 2 also shows that there was a
decrease of the percentage of particles with a diameter smaller
than 1 mm, which have a larger surface area. The rise of the per-
centage of particles with a diameter between 1 mm and 10 mm was
probably due to the formation of different phases, like calcium
methoxide and diglyceroxide during the reaction [19].
The catalysts prepared by contacting the raw CaO with water
(CaO_H2O), methanol (CaO_MeOH) and glycerin/methanol
(CaO_diglyc) presented granulometric distribution curves shifted
towards larger diameters with external surface areas lower than Fig. 3. XRD patterns of catalysts (calcium methoxide JCPDS 00-20-1565;
1 m2/g. C-calcite).
 A.P. Soares Dias et al. / Renewable Energy 99 (2016) 622e630
Fig. 4. XRD patterns of post reaction CaO and CaO_MeOH catalysts (62  C; Wcat/Woil 5%; MeOH/Oil 12 molar ratio; 4 h).
calcium carbonate (calcite). Regarding the samples prepared by
contacting the CaO with water, methanol and glycerin/methanol
the expected diffraction lines were present, except for the CaO_-
methanol sample, with the main XRD lines belonging to Ca(OH)2.
The Ca_Diglyc sample presents a XRD pattern which can be due to
the overlay of lines ascribable to calcium diglyceroxide and calcium
methoxide. The catalyst prepared by contacting the bare CaO with
methanol only displays the lines ascribable to portlandite
(Ca(OH)2). Surprising well crystallized Ca methoxide product, using
similar preparation procedure, is reported by several authors
[12,23]. Possibly, the absence of bulk calcium methoxide is due to
the water content of methanol. Lines belonging to calcium meth-
oxide, in addition to calcium diglyceroxide lines, are well visible for
the catalyst prepared in the presence of methanol/glycerin mixture.
The post reaction CaO samples showed patterns mainly ascrib-
able to calcium hydroxide (Fig. 4). Similar results were observed for
catalysts prepared by hydration and by contacting the CaO with
methanol (Fig. 5). The catalysts prepared by contacting the com-
mercial CaO with a mixture of glycerol and methanol showed a XRD
pattern (Fig. 5), attributable to calcium methoxide and diglycer-
oxide unchanged during the reaction. From the crystallinity point
of view, the hydrated calcium seems to be highly stable in reaction
conditions. Recently Sa nchez-Cant et al. [11] reported a high
active hydrated lime catalyst, but observed crystallinity changes
after the rst 60 min of reaction. They noticed a lattice expansion,
with displacement of the XRD lines towards small Bragg angles,
which they ascribed to the leaching of catalyst components (Ca
species).
The surface of fresh and post reaction catalysts was
625
characterized by FTIR using HATR mode. The collected IR spectra
are displayed in Figs. 6 and 7. The HATR-FTIR spectrum of bare CaO
material, after drying overnight at 120  C, presented a low intense
band around 3660 cm1, characteristic of the stretching vibration of
the OH group, ascribable to CaeOH species thus showing partial
surface hydroxylation [23]. The intensity of CaeOH band increased
in the post reaction CaO samples, but this band becomes less
intense with the catalyst deactivation. This result seems to indicate
the CaeOH surface species has active sites for the CH3OH activation.
The methanol activation, over the catalyst surface, is referred to be
the rst step of the reaction mechanism [24e26]. Concerning the
other catalysts, only the CaO_diglyc presented a dissimilar spec-
trum ascribable to calcium diglyceroxide [27,28].
The post reaction CaO catalysts showed complex IR spectra
(Fig. 6), with reectance bands belonging to surface Ca species
formed during reaction and to adsorbed oil or/and FAME. They
ascribed the IR features to chelating carboxylated species. The
intense band at 1750 cm1, characteristic of the carbonyl C]O
stretching, belongs to oil or FAME species on the catalyst surface.
The intensity of this band decreases with increasing number of
catalyst reutilizations, following the same trend of the catalyst ac-
tivity: lower catalyst activity corresponds to a lower intensity of
this band. Additionally, post reaction CaO samples showed bands in
the range 3000e2750 cm1 ascribable to eCH3 vibrations from the
surface adsorbed FAME, since the relative intensity of this band has
the same trend of the band at 1750 cm1. Analogous IR reectance
spectra in the range 1800e800, with exception for the band at
1750 cm1, were observed by Le on-Reina et al. [28] for CaO cata-
lysts prepared by contacting lime with glyceroxides. The intensity
626 A.P. Soares Dias et al. / Renewable Energy 99 (2016) 622e630

The main difference between the CaO, CaO_H2O and CaO_MeOH

post reaction samples was the observed lower intensity of the
bands ascribable to surface glyceroxide species (1250-1000 cm1
[20]), for the CaO_H2O samples.
The IR spectra of the CaO_Diglyc post reaction samples do not
present the band at 3660 cm1, characteristic of surface hydroxyl
groups, and also the bands in the range 3000e2750 cm1 and
1750 cm1 are absent. The absence of eOH surface species indicates
that methanol activation for CaO_Diglyc presents a different
mechanism. According to Kouzu et al. [29] the methanol activation
by calcium diglyceroxide proceeds through the hydrolysis of the
catalysts thus leading to a homogeneous mechanism:

CaC3 H7 O3 2 2H2 OCa2 2OH 2C3 H8 O3

OH CH3 OHH2 O CH3 O
A different methanol activation mechanism over calcium
 n-Reina et al. [28]. According to
diglyceroxide was reported by Leo
them the abstraction of a proton from methanol can be accom-
plished by interaction of a methanol molecule with two oxygen
atoms belonging to adjacent tetramer of Ca glyceroxide, being the
mechanism heterogeneous.

3.2. Methanolysis tests

The catalytic activity of the prepared materials was evaluated in

Fig. 5. XRD patterns of post reaction CaO_H2O and CaO_Diglyc catalysts (62  C; Wcat/
Woil 5%; MeOH/Oil 12 molar ratio).
of the bands, in this range, increases for samples used in multiple
reaction batches.
standard conditions described in the experimental section. The
main catalytic data are displayed in Fig. 8 and Table 2.
It is worth noting that the transesterication experiments have
a good repeatability. In fact, for example, ve replicates of the ex-
periments carried out using raw CaO as catalyst showed a standard
deviation of 2% of the FAME concentration, which is a remarkable
result having in mind the error of z1% associated with FAME
analysis.
As seen in Fig. 8, the dried bare CaO material presented a S shape
kinetics curve. The large induction period can be related to the
formation of the surface methoxide species. Such kinetics curve
usually indicates a change in the reaction mechanism. In fact, ac-
cording to previously published papers, in the initial phase of the
process the immiscibility of methanol with the vegetable oil causes
a slow mass transfer controlled process. The oil conversion into
FAME increases the miscibility between methanol and the oily
phase and so the transesterication becomes a faster chemical
controlled process [30].
On the contrary to the CaO sample, for the CaO_H2O catalyst the
induction period was absent. This behavior can be ascribed to the
fast methanol activation by the surface hydroxyl groups formed
during the preparation step (IR band at 3660 cm1, Fig. 6). Analo-
gous results have been published by Sa nchez-Cant et al. [31].
These authors studied the methanolysis of vegetable oil over
commercial hydrated lime and ascribed the active phase to Ca(OH)2
which presented Bro nsted basic sites for methanol activation.
However, the behavior observed for the CaO_H2O catalyst differs
from that reported by Stamenkovi c et al. [24]. They reported an S-
shape curve for calcium hydroxide catalyzed methanolysis of sun-
ower oil. The authors ascribed the reported behavior to the exis-
tence of mass transfer limitations due to the methanol/oil
immiscibility.
The sample prepared by contacting the CaO with methanol also
presented a sigmoid curve with an extended induction period.
Actually, this sample presented the longest induction period which
can be related to the low amount of the reactive surface methoxide
species due to the fact that they are strongly bonded to the bulk
catalyst phase as calcium methoxide. These species must be
replaced on the catalyst surface by activated methanol species
 A.P. Soares Dias et al. / Renewable Energy 99 (2016) 622e630 627

Fig. 6. HATR-FTIR spectra of post reaction CaO and CaO_H2O catalysts.

which will react with the oil molecules. In the literature calcium
methoxide is reported to be a good catalyst for oil methanolysis [12]
showing a sigmoid kinetics curve [23] by the above stated reasons.
Surprisingly, the Ca_MeOH sample presented a catalytic behavior
similar to that of glycerolate reported by Reyero et al. [15].
Fig. 8 also shows that calcium diglyceroxide presents a behavior
similar to that reported for CaO_H2O catalyst with no induction
period. In fact, as discussed above, calcium diglyceroxide catalysis
involves an important contribution of the homogeneous process,
with the correspondent Ca leaching, and has therefore a faster
kinetics.
The inuence of the operating variables on the quality of the
glycerine phase is a very important issue that can affect the eco-
nomics of the global process and, as reported in the literature [32],
the glycerin obtained using heterogeneous catalysts presents a
higher purity than the one obtained in homogeneous catalyzed
processes. Thus, the glycerin phases obtained using the prepared
catalysts were evaluated in order to outwit the possible contribu-
tion of homogeneous catalyzed processes.
The image in Fig. 9 shows the glycerin phases obtained using the
628 A.P. Soares Dias et al. / Renewable Energy 99 (2016) 622e630

Fig. 7. HATR-FTIR spectra of fresh and post reaction CaO_MeOH and CaO_Diglyc catalysts.

fresh catalysts after 3 h of reaction. The dark color of glycerine
phase produced when CaO_Diglyc was used as catalyst conrms
the homogeneous mechanism above reported. Furthermore, the
thermogravimetry analysis of this glycerin conrmed its contami-
nation with inorganic materials since a solid residue was obtained
after combustion at 1100  C (Fig. 10). This white solid residue was
analyzed by XRD and as shown in Fig. 10, the difractogram shows
the lines ascribable to lime. The glycerin obtained using the bare
CaO was contaminated with MONG (non-glycerin organic material)
[33], which is responsible for its dark color, but with non-
measurable inorganic residue.
Finally, the stability of the different catalysts in the reaction
medium was assessed by its reutilization in consecutive reaction
batches without any intermediate reactivation procedure. Data in
Table 1 show that CaO_Diglyc catalyst suffers a fast deactivation
that led to the signicant decrease of the FAME yield in the 2nd
reaction batch. This activity decay can be attributed to the catalyst
dissolution in the reaction medium accompanied by the loss of
crystallinity of the insolubilized material. For the other catalysts,
the deactivation process is slower and can also be related to the
 A.P. Soares Dias et al. / Renewable Energy 99 (2016) 622e630 629

crystallinity loss and to the decrease of surface eOH species

(3660 cm1 IR absorption band) availability because these species
are responsible for the methanol activation.
Taking into account the above data on the catalytic behavior for
the different Ca phases, it can be hypothesized that the S-shape
kinetics curve, usually observed for the lime catalysts, can not only
result from the existence of mass transfer limitations, but also from
the contribution of the different phases formed in the reaction
medium. Furthermore, the results allow to conclude that calcium
hydroxide is the active phase since the post reaction catalysts, with
exception of CaO_Diglyc, were totally converted into Ca(OH)2.
Additionally, calcium diglyceroxide promotes a homogeneous
catalyzed mechanism that allows a high rate of the trans-
esterication reaction [34] but a fast deactivation because of the
catalyst leaching into the glycerin phase. The contamination of the
FAME phase with Ca species when CaO_Diglyc is used cannot be
discarded and seems to be a motivating issue for future research
work.
Fig. 8. FAME content (%) versus reaction time.

4. Conclusions
Table 2
FAME yields (%) for fresh and post reaction catalysts without intermediate As reported before lime catalyst presents good catalytic per-
regeneration. formances during the methanolysis of rapeseed oil allowing to
Catalyst PR#1 PR#2 PR#3 PR#4 PR#5 PR#6 obtain a FAME yield of 97% after 3 h of reaction at atmospheric
pressure and 60  C. The characterization of post reaction catalysts
CaO 97 97 96 96 97 62
CaO_H2O 98 98 98 e e e clearly shows that CaO is totally converted into calcium hydroxide
CaO_MeOH 98 96 96 95 e e during reaction being this phase the active phase. In fact when
CaO_Diglyc 95 85 e e e e calcium hydroxide is used as catalyst the initial induction period,
observed for CaO, vanished. Calcium methoxide and diglyceroxide
are also formed during reaction and behave differently from CaO.
Thus, the complex kinetics of oil methanolysis over lime catalyst is
due to the presence of several Ca phases. Each Ca phase presents a
different contribution to the global catalytic process as revealed by
the different kinetics curves obtained when the pure Ca phases
were used as catalyst.
The absence of S-shape kinetics curve, when calcium diglycer-
oxide was used as catalyst, appears to indicate an important
contribution of a homogeneous catalyzed process which was
conrmed by the presence of Ca in the glycerin phase. The glycerin
phase appearance is correlated with the reaction mechanism: dark
glycerin is obtained by homogeneous mechanism whereas het-
erogeneous one leads to a clear glycerin phase.
The results pointed out that the formation of calcium diglycer-
oxide is responsible for Ca leaching.
The characterization of the post reaction catalysts underlined
the relevance of the surface and bulk catalysts dynamics during
methanolysis.
Fig. 9. Glycerin phases obtained after 3 h of reaction using catalysts.

Fig. 10. Thermogram of the glycerin phase obtained using CaO_Diglyc catalysts (Commercial glycerin was used as standard). XRD diffractogram of the solid residue collected from
the ceramic pan after thermogravimetry.
630 A.P. Soares Dias et al. / Renewable Energy 99 (2016) 622e630

References [19] J.F. Puna, J.F. Gomes, J.C. Bordado, M.J. Neiva Correia, A.P. Soares Dias, Appl.
Catal. A General 470 (2014) 451e457.
kos, M. Fari, Renew. Sust. Energy Rev. 32 [20] A.P. Soares Dias, J. Bernardo, P. Felizardo, M.J. Neiva Correia, Fuel Proc.
[1] J. Popp, A. Lakner, M. Harangi-Ra
Technol. 102 (2012) 146e155.
(2014) 559e578, http://dx.doi.org/10.1016/j.rser.2014.01.056.
[21] P. Felizardo, P. Baptista, M.S. Uva, J.C. Menezes, M.J. Neiva Correia, J. Near
[2] R. Luque, J.C. Lovett, B. Datta, J. Clancy, J.M. Campelo, Energy Environ. Sci. 3
Infrared Spectrosc. 15 (2007) 97e105.
(2010) 1706e1721.
[22] A.P. Soares Dias, J. Puna, M.J. Neiva Correia, I. Nogueira, J. Gomes, J. Bordado,
[3] S. Semwal, A.K. Arora, R.P. Badoni, D.K. Tuli, Bioresour. Technol. 102 (2011)
Fuel Proc. Technol. 116 (2013) 94e100.
2151e2216.
[23] V.G. Deshmane, Y.G. Adewuyi, Fuel 107 (2013) 474e482.
[4] Y.M. Sani, W.M.A.W. Daud, A.R. Abdul Aziz, Appl. Catal. A General 470 (2014)
[24] W.N.N.W. Wan, N.A.S. Amin, Fuel Proc. Technol. 92 (2011) 2397e2405.
140e161.  Kesi
[25] I. Lukic, Z. c, S. Maksimovi c, M. Zdujic, H. Liu, J. Krsti
c, D. Skala, Fuel 113
[5] J. Konwar, J. Boro, D. Deka, Renew. Sust. Energy Rev. 29 (2014) 546e564.
(2013) 367e378.
[6] A.F. Lee, J.A. Bennett, J.C. Manayil, K. Wilson, Chem. Soc. Rev. 43 (2014)
[26] O.S. Stamenkovi c, V.B. Veljkovi
c, Z.B. Todorovic, M.L. Lazi c, I.B. Bankovi
c-Ili
c,
7887e7916.
D.U. Skala, Bioresour. Technol. 101 (2010) 4423e4430.
[7] A.F. Lee, K. Wilson, Catal. Today 242 (2015) 3e18.
[27] C. Li, Z. Huang, Y. He, D. Zhou, C. Du, S. Zhang, J. Chen, Adv. Mater. Res. 666
[8] M. Kouzu, J.S. Hidaka, Fuel 93 (2015) 1e12.
 pez Granados, A.C. Alba-Rubio, F. Vila, D. Martn Alonso, R. Mariscal, (2013) 93e102.
[9] M. Lo
[28] L. Leo n-Reina, A. Cabeza, J. Rius, P. Maireles-Torres, A.C. Alba-Rubio, M. Lo  pez
J. Catal. 276 (2015) 229e236.
Granados, J. Catal. 300 (2013) 30e36.
[10] A.A. Raafat, Int. J. Environ. Sci. Technol. 8 (2011) 203e221.
[29] M. Kouzu, S.Y. Yamanaka, J.S. Hidaka, M. Tsunomori, Appl. Catal. A General
[11] M. Kouzu, M. Tsunomori, S. Yamanaka, J. Hidaka, Adv. Powder Technol. 21
355 (2009) 94e99.
(2010) 488e494.
[30] J.Y. Park, D.K. Kim, Z.M. Wang, J.S. Lee, Appl. Biochem. Biotech. 154 (2009)
[12] X. Liu, X. Piao, Y. Wang, S. Zhu, H. He, Fuel 87 (2008) 1076e1082.
nchez-Cant, L.M. Pe rez-Daz, I. Pala-Rosas, E. Cadena-Torres, L. Jua rez- 246e252.
[13] M. Sa
[31] M. S anchez-Cant, L.M. Pe rez-Daz, R. Rosales, E. Ramrez, A. Apreza-Sies,
Amador, E. Rubio-Rosas, M. Rodrguez-Acosta, J.S. Valente, Fuel 110 (2013)
I. Pala-Rosas, E. Rubio-Rosas, M. Aguilar-Franco, J.S. Valente, Energy Fuels 25
54e62.
(2011) 3275e3282.
[14] I. Reyero, G. Arzamendi, L.M. Ganda, Chem. Eng. Res. Des. 92 (2014)
[32] L. Bournay, D. Casanave, B. Delfort, G. Hillion, J.A. Chodorge, Catal. Today 106
1519e1530.
(2005) 190e192.
[15] V.B. Veljkovi c, O.S. Stamenkovi c, Z.B. Todorovi
c, M.L. Lazi
c, D.U. Skala, Fuel 88
[33] J.F. Puna, M.J. Neiva Correia, A.P. Soares Dias, J. Gomes, J. Bordado, Reac. Kinet.
(2009) 1554e1562.
Mech. Catal. 109 (2013) 405e415.
[16] S.N. Csernica, J.T. Hsu, Ind. Eng. Chem. Res. 51 (2012) 6340e6349.
[34] G.O. Ferrero, M.F. Almeida, M.C.M. Alvim-Ferraz, J.M. Dias, Fuel Proc. Technol.
[17] B. Likozar, J. Levec, Fuel Proc. Technol. 122 (2014) 30e41.
121 (2014) 114e118.
[18] B. Likozar, A. Pohar, J. Levec, Fuel Proc. Technol. 142 (2016) 326e336.