You are on page 1of 8

Journal of Environmental Chemical Engineering 10 (2022) 108191

Contents lists available at ScienceDirect

Journal of Environmental Chemical Engineering


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

Zinc oxide-co-sodium zirconate: A fast heterogeneous catalyst for biodiesel


production from soybean oil
Ricardo Rodríguez-Ramírez a, Fabiola S. Sosa-Rodríguez b, Jorge Vazquez-Arenas a, c, *
a
Centro Mexicano para la Producción más Limpia, Instituto Politécnico Nacional, Av. Acueducto s/n, Col. La Laguna Ticomán, Ciudad de México 07340, Mexico
b
Research Area of Growth and Environment, Metropolitan Autonomous University, Azcapotzalco (UAM-A), Av. San Pablo 180, Mexico City 02200, Mexico
c
Laboratorio Nacional de Desarrollo y Aseguramiento de la Calidad de Biocombustibles (LaNDACBio), México

A R T I C L E I N F O A B S T R A C T

Editor: Apostolos Giannis In this study, Zinc Oxide-co-Sodium Zirconate (ZnO⋅Na2ZrO3) is synthesized utilizing a solid-state method with
the aim of performing a fast transesterification of triglycerides contained in soybean oil, to be competitive against
Keywords: the homogeneous process employing caustic soda. This procedure enabled to synthesize a composite catalyst
Zinc Oxide with enhanced properties for this application. The material is characterized using Brunauer–Emmett–Teller
Sodium Zirconate
(BET) surface area analysis isotherms, X-Ray Diffraction (XRD), and Scanning Electron Microscopy (SEM)
Heterogeneous catalysis
coupled to Energy Dispersive X-Ray Spectroscopy (EDXS), finding a denser structure with sodium zirconate la­
Transesterification of triglycerides, Biodiesel
Box-Behnken design mellas collapsed each other, with a surface area of 0.7 m2 g− 1, and a weight composition close to the theoretical
value, without the presence of impurities. Using this composite catalyst, 97% of biodiesel conversion is produced
in 15 min at 65 ◦ C (i.e. methanol boiling point), with 4.5 wt% catalyst, methanol: oil molar ratio of 14:1, and
stirring speed of 500 rpm, as optimal conditions obtained from a Box-Behnken Experimental Design (BBD) along
with a Response Surface Methodology (RSM). The chemical composition of biodiesel is evaluated through FT-IR
spectroscopy (FT-IR) and Proton nuclear magnetic resonance (1H NMR) measurements.

1. Introduction to obtain good quality biodiesel), thus, heterogeneous catalysts are a


better alternative to achieve cleaner reactions with a good quality of
Climate change and imminent fossil oil shortage have become a biodiesel. Nevertheless, to compete against homogeneous catalysis,
global concern, whereby renewable bioresources have been envisaged heterogeneous materials must decrease the residence time and corro­
as the main alternative to replace these non-renewable sources of en­ sion, while increasing overall reusability. Additionally, these materials
ergy, since they present a closed carbon cycle and analog properties to need to display surface basic sites with the ability to remove the proton
petroleum derivatives. In this direction, biodiesel is a good alternative from the methanol, forming a very reactive CH3O- species, enhancing
which has efficiently started this replacement some decades ago, since the nucleophilic attack to the triglyceride (i.e. fatty acids) to produce a
35% of the consumed fossil fuels is petro-diesel. This compound displays fatty acid methyl ester[6]. Thus, alkaline regions adjacent to an acid site
similar physicochemical properties as its non-renewable counterpart (i. is a prerequisite for the catalyst [7]; whence lamellar structures are
e. carbon content, vapor pressure, flammability, pressure ignition, vis­ required with satisfactory particle size to sediment after the biodiesel
cosity, flashpoint, calorific value) [1–4]. Also, burning this biofuel does production. Different catalytic surface have been envisaged for thes
not release sulfur into the environment, presenting high biodegrad­ purposes including zeolites, Montmorillonite, CaO, MgO, among others
ability and low to null toxicity. [7–9]. It is important to make a distinction between different types of
In order to produce biodiesel, a vegetable oil must be typically mixed catalyst: a binary catalyst is a material with two different atoms
and heated with short-chain alcohols (e.g. methanol, ethanol) in the comprising its structure (i.e. ZnO); a ternary catalyst is comprised by
presence of a catalyst (usually dissolved NaOH at the industrial level) three different atoms (i.e. Na2ZrO3) or more; a composite material is
[5]. NaOH presents several problems regarding the transesterification comprised by two or more separate compounds in enough quantities not
reaction, (i.e. high corrosivity, carboxylate side reaction, washing steps to be considered a dopant, chemisorbed or phisisorbed to each other,

* Corresponding author at: Centro Mexicano para la Producción más Limpia, Instituto Politécnico Nacional, Av. Acueducto s/n, Col. La Laguna Ticomán, Ciudad de
México 07340, Mexico.
E-mail addresses: jgvazquez@ipn.mx, jorge_gva@hotmail.com (J. Vazquez-Arenas).

https://doi.org/10.1016/j.jece.2022.108191
Received 28 February 2022; Received in revised form 16 June 2022; Accepted 29 June 2022
Available online 1 July 2022
2213-3437/© 2022 Elsevier Ltd. All rights reserved.
R. Rodríguez-Ramírez et al. Journal of Environmental Chemical Engineering 10 (2022) 108191

making a difference of the phases constituting the catalyst (i.e. using a heating ramp of 5 ºC min− 1 in a ceramic crucible within a muffle
TiO2⋅ZnO). (Thermo Scientific). This temperature was maintained during 6 h.
Some of the most notable heterogeneous catalysts include the qua­
ternary Na2ZnSiO4 which obtained > 99% of biodiesel conversion in 45 2.2. Material Characterizations
min [10]; and Na2ZrO3 which produced 98% of biodiesel conversion
after 3 h, which after dopping with Cs enabled to decrease the reaction The crystal structure of the composite was evaluated using a X-ray
time to 15 min, but increasing the general cost by using non-abundant Bruker D-8 Advance diffractometer, with Co-Kα radiation (λ = 1.79 Å)
materials and losing reusability immediately after the reaction [11, under the Bragg-Brentano configuration at 40 kV and 30. SEM micro­
12]. A comparison between these materials is displayed in Table 1, graphs were analyzed to assess the catalyst texture in a Zeiss SUPRA 55-
including a comparison to NaOH, which is the most used homogeneous VP microscope (high-field emission) using 10.0 kV as acceleration
catalyst. voltage. Elemental analysis was also performed using EDXS capabilities.
ZnO is a widely accessible material which does not present high The surface area (SSA) of the composite catalyst was determined
activity towards the transesterification reaction until combined with estimating N2 adsorption-desorption isotherms in the range of relative
other elements (i.e. Na2ZnSiO4) [10,13], thus, a first attempt is con­ pressure from 0.05 to 0.3 bar at − 196 ◦ C. A Quantachrome Autosorb iQ
ducted in this work to increase the activity of Na2ZrO3 by adding ZnO as gas analyzer was used for these purposes, while the SSA was evaluated
co-catalyzer, expecting that the addition of more Lewis acid sites would based on the Brunawer-Emmett-Teller model.
increase the mobility of ions in the new composite material, directly
affecting the reaction rate. It is expected that this structural change
2.3. Transesterification Reaction
forming the co-catalyst remarkably expedites the transesterification of
triglicerydes to produce biodiesel, in order to be more sustainable
The transesterification reaction was carried out as detailed in the
against the consolidated homogeneous process of transesterification.
following procedure. 11.3 g of Nutrioli Soybean Oil were weighted (and
Thus, a solid-state method is used to synthesize the ZnO⋅Na2ZrO3 com­
kept constant for the sake of the statistical analysis), and added into a
posite, while its structure, texture and composition are characterized
three neck round bottom flask coupled to a condenser (to keep the re­
using: SEM-EDXS, BET surface area analysis, XRD, respectively. The
action under reflux), magnetic stirring, and a thermometer. The tem­
catalyst is tested towards the triglycerides transesterification using
perature was stabilized at 65ºC. A certain amount of catalyst and
soybean oil, wherein the chemical composition of biodiesel is evaluated
methanol determined by the experiments were added to the flask,
through FT-IR and 1H NMR measurements. Soybean oil is herein used
closing the reactor with a glass stopper afterwards. Once the trans­
since it is an abundant and available source of triglycerides compared to
esterification reaction concluded, the flask was removed from the heat,
other oils from non edible plants like Jatropha curcas L. oil, which
and introduced to an ice bath to quench the system. The content of the
additionally requires several unit operations involving seed peeling,
reactor was transferred to two 15 mL centrifuge tubes. The mixture was
extraction and purification of the oil, which are beyond the scope of the
centrifuged at 4500 rpm for 15 min to separate the three phases corre­
present manuscript. Although the global biodiesel production is high
sponding to the FAMEs, the glycerol-methanol mixture, and the catalyst.
with this oil due to high availability in many geographical areas, ined­
Subsequently, the FAMEs were carefully extracted using a pipette,
ible energy crop should be promoted to generate more sustainable bio­
transferred to a container, and dried in a stove at 65ºC overnight to
diesel at industrial levels. A factorial design (Box-Behnken) is utilized to
evaporate any methanol remaining.
carry out an optimization of the following parameters: methanol: cata­
lyst load (1–10%), stirring rate (200–700 RPM), and oil ratio (3:1–16:1).
Subsequently, the optimal conditions obtained from the BBD results, and 2.4. Reaction kinetics
analyzed with a RSM are evaluated at 5, 15, 20, 30, and 60 min. These
maximum conversion conditions are then used to account for the cata­ The optimal reaction time was evaluated by measuring the biodiesel
lyst reusability. conversion at 5, 15, 20, 30, and 60 min by using the experimental
conditions reported by Santiago-Torres et al. for Sodium Zirconate at
2. Materials and methods 65ºC, 16:1 Methanol: Oil Ratio, 3 wt% catalyst, with a stirring speed of
500 RPM [9].
2.1. Synthesis of the ZnO⋅Na2ZrO3 catalyst
2.5. Box-Behnken experimental design
A solid-state method was used to synthesize the ZnO⋅Na2ZrO3 com­
posite using Sodium Carbonate (Na2CO3), Zinc Oxide (ZnO) and Zirco­ The BBD (Table 2) was performed to determined the best operating
nium Oxide (ZrO2). All Merck reagents, (~99% purity) in weight conditions to evaluate the transesterification reaction using the
percentages of 50%, 30%, and 20%, respectively. These chemical re­ ZnO⋅Na2ZrO3 catalyst. This type of design involves midpoints of the
agents were well blended in an agate mortar, and calcinated at 700 ◦ C system edges and center,whereby drastically reduces the amount of
experiments to collect [14]. The reaction was conducted using dry

Table 1
Comparisons between Zinc and Zirconium based heterogeneous catalysts, and classical Sodium hydroxide used in homogeneous processes for biodiesel production.
Catalyst structure Time (h) Temperature (ºC) Methanol:Oil Ratio Catalyst load Conversion (%) Ref.
(wt%)

NaOH 1 60 6:1 0.5 97 [1]


ZnO 8 150 50:1 1 94.8 [2]
Na2ZrO3 3 65 16:1 3 98.3 [3]
Cs⋅Na2ZrO3 15 min 65 30:1 1 98.8 [4]
Na2ZnSiO4 45 min 65 14:1 5.3 > 99 [5]
SZr-Ti-Yb 1 150 6:1 2 97.98 [21]
Zr-SBA-15 6 209 50:1 12.45 92 [22]
Zn-CaO 4 40 20:1 5 96 [23]
Zn/Al/Co Complex 8 160 30:1 10 91 [24]
ZrSnO4 1 120 150:1 2 74 [25]

2
R. Rodríguez-Ramírez et al. Journal of Environmental Chemical Engineering 10 (2022) 108191

Table 2 Table 3
Box-Behnken design estimated to establish the experimental conditions of the Functional groups, vibrational modes and wavenumber described for soybean
transesterification reaction using ZnO⋅Na2ZrO3 as heterogeneous catalyst, dis­ oil and biodiesel produced from the transesterification reaction in the FT-IR
playing the screened factors and response variable (conversion). analysis (Fig. 7).
Conversion (%) Stirring rate (RPM) Catalyst load (%wt) MeOH:Oil ratio Functional group Wavenumber Vibrational Soybean Biodiesel
(cm− 1) mode oil
96.1 700 5.5 16
96.0 700 10 9.5 O-CH2-C 1098 Stretching Present Absent
95.7 450 5.5 9.5 (Asymmetric,
95.7 450 5.5 9.5 Axial)
95.7 450 10 16 O-CH3 1197 Stretching Absent Present
94.3 450 5.5 9.5 O-CH2 1370 Bending Present Absent
91.9 200 5.5 16 (Glycerol)
90.2 200 10 9.5 -CH3 1437 Bending Absent Present
73.6 450 1 16 (asymmetric)
55.8 700 1 9.5 -C–
–O 1743 Stretching Present Present
53.6 450 10 3 -CH2 (Symmetric 3009, 2925, Stretching Present Present
48.0 700 5.5 3 and 2855
32.9 200 5.5 3 Asymmetric),
14.6 200 1 9.5 CH3
8.5 450 1 3 (Asymmetric)

methanol (J.T. Baker), and NOM051-SCFI-SSA12010 soybean oil 3. Results and discussion
(Nutrioli brand) in a 25 mL double-neck bottom flask coupled to a
condenser, a water chiller, and a thermometer performing from − 3.1. Synthesis and characterization of the co-catalyst
20–150 ºC. The methanol reflux was promoted by continuous recircu­
lation of cool water within the condenser. The reaction temperature was The synthesis procedure was conducted in a porcelain crucible, ac­
kept at 65 ºC using an oil bath, which was chosen due to the methanol cording to the following expected reaction:
boiling point. Magnetic stirring was applied in the system using a plate.
Na2 CO3 + ZnO + ZrO2 →ZnO⋅Na2 ZrO3 + CO2 (1)
In the experimental design, the catalyst load was modified from 1 to
10 wt% relative to the oil weight, 3:1–16:1 in the methanol: oil ratio, where the decarbonization of sodium carbonate was the first reaction
while the stirring rate was varied from 200 to 700 rpm. These conditions producing sodium oxide (Na2O), as described below:
where stablished according to the following criteria: 200 rpm was the
minimum instrumental speed, 700 rpm was the speed before a chaotic Na2 CO3 ⇒Δ Na2 O + CO2 (2)
movement of the stirring bar started to occur, 1% wt. was the minimum
Subsequently, the eutectic reaction of the sodium oxide with the
weight measurable with accuracy and 10% wt. was the maximum
zirconium oxide produces sodium zirconate (reaction 3), fixing over the
amount of catalyst it could be inserted to keep all the reactions under a
surface of the unreacted zinc oxide, forming the final composite (Re­
single catalyst synthesis batch, 3:1 molar ratio of MeOH:Oil is the stoi­
action 4). The final catalyst is a hard and white powder.
chometric quantity for the transesterification reaction and 16:1 is
considered as a safe margin because molar ratios over 18:1 could reduce Na2 O + ZrO2 ⇒Δ Na2 ZrO3,eut. (3)
FAME conversion due to a formation of glycerol-methanol emulsion
[15]. All experiments were individually performed to circumvent ZnO + Na2 ZrO3,eut. →ZnO⋅Na2 ZrO3 (4)
influencing the reactor volume. FT-IR analysis was conducted to calcu­
Fig. 1a describes the diffractogram obtained from the XRD analysis of
late the triglyceride conversion as described in Section 2.5. Optimal
the as-synthesized powder. Two different crystalline phases, ZnO and
conditions were carried out in a final batch according to the surface
Na2ZrO3 are detected from this evaluation, according to the ICOD
response methodology using the conversion values reported in Table 2.
01–075–1526 and ICOD 00–008–0242 crystalographic cards, respec­
Additional re-usage experiments were carried out keeping the catalyst in
tively. The ZnO phase presents a prominent plane of (1 0 1) with a D-
the reactor, only incorporating a new batch of methanol and oil during 5
spacing of 2.4546 Å, which is commonly observed in pristine ZnO
consecutive cycles. The final products were likewise evaluated using the
samples. Regarding the Na2ZrO3, previous research describes an
analytical techniques aforementioned.
important plane of (0 0 14) with D-spacing of 5.3761 Å, followed by the
(− 2 1 2) plane with a D-spacing of 2.2983 Å, where the former one
2.6. Biodiesel characterization
adopted a higher predominance.
On the other hand, SEM micrographs were obtained from the syn­
The fatty acid methyl esters (FAME) and soybean oil were analyzed
thesized composite. Fig. 2a describes that the material is homogeneously
using a Bruker FTIR ALPHA spectrometer with an attenuated total
distributed in particles of an average of 50 µm, while the lamellar
reflection (ATR) accessory. The vibration of the methyl in the ester (CH3-
structure of the sodium zirconate starts to distinguish itself from the
O-) at 1437 cm− 1 was used to construct the calibration curve, and
particle, showing a particular lateral growth from the ZnO crystal.
calculating the conversion of triglycerides (Labeled FT-IR conversion).
Likewise, it is evident that the general structure of the material is non-
Each FT-IR vibration was compared against Table 3 [16,17]. Proton
porous and dense, with the lamellas occupying the majority of the
Nuclear Magnetic Resonance (1H NMR) was conducted to determine the
particle surface. The EDXS analysis (Table 4) exposes that these aggre­
fine structure of biodiesel, and then estimating the selectivity of the
gates are constituted by 34 wt% oxygen, 23.41 wt% Sodium, 23.37 wt%
transesterification reaction. This analysis was performed in a Bruker
Zinc, and 19.22 wt% Zirconium, showing the total elemental composi­
Ascend 500 equipment by dissolving 30 mg biodiesel in 1 mL CDCl3. 1H
tion of the material. No contaminant elements were observed in this
NMR spectra were obtained for cycles 1,2, 3 and 5, and the selectivity
evaluation, thus, confirming a successful synthesis of the composite
was estimated using area under the curve of signals at 4.25 and 3.65
material.
ppm [18].
The isotherm described in Fig. 3 fits to a Type-2 according to the
IUPAC data, which is particular of dense nonporous materials. Little to
no hysteresis is observed, indicating that the capillary absorption of N2 is

3
R. Rodríguez-Ramírez et al. Journal of Environmental Chemical Engineering 10 (2022) 108191

Table 4
Elemental analysis calculated for the as-synthesized and exhausted catalysts
using SEM-EDXS.
Element As-synthesized catalyst Exhausted catalyst

wt% at% wt% at%

OK 34.00 57.26 27.36 56.67


NaK 23.41 27.43 8.82 12.71
ZnK 23.37 09.63 51.81 26.26
ZrK 19.22 05.68 12.01 4.36

Fig. 1. XRD patterns determined from the composite catalyst: a) as-


synthesized, b) exhausted after 5 cycles.

minimal [19]. A specific surface area about 0.7 m2 g− 1 was obtained for
the ZnO⋅Na2ZrO3 composite catalyst, using the
Brunnauer-Emmett-Teller model in the range from 0.05 to 0.35 of Fig. 3. Type II isotherm of as-synthesized catalyst obtained by Nitrogen
relative pressure (adsorption values). Volumetric Adsorption.

3.2. Reaction kinetics group of the triglyceride. Accordingly, this system is optimized assuming
this reaction, as the limiting step, and the mass transference of CH3O- in
The evaluation of the reaction time (Fig. 4, Table 5) shows that the solution; whereby the catalyst load, the available methanol and stirring
synthesized composite (ZnO⋅Na2ZrO3) achieves the transesterification speed of the reactor solution are herein assumed as factors in the BBD.
reaction with a conversion of 78% within the first 5 min. Once the re­ This BBD was computed in the Minitab software considering the afore­
action proceeded for 15 min, the reaction conversion accomplished mentioned variables as screening factors (Table 2). Subsequently,
91%, showing a minimal increase in the following minutes (until optimal conditions were evaluated using the output variables (conver­
reaching 96% after 60 min), establishing the ideal reaction time at sion of triglycerides) of this design which were measured with FT-IR
15 min. (refer to Section 2.3 for further details).
The pre-statistical evaluation of the BBD design exposes that the
molar ratio between methanol and soybean oil, and the catalyst load in
3.3. Box-Behnken experimental design the reaction are the two most important variables, since the reaction
reached a conversion higher than 90% when these two variables
As previously described with the Na2ZnSiO4 material, the limiting remained over the middle point (9.5:1 and 5.5 wt%, respectively). If any
step of the transesterification reaction is the methanol deprotonation to of these two variables are maintained in the low value, the conversion
form the methoxy ion (CH3O-) on the catalyst surface, since it can re- drastically drops below 75%, meaning that both variables have a similar
protonate and completely halt the entire mechanism [10]. Subse­ relevance in the reaction. The stirring rate does not show a relevant role
quently, the CH3O- performs a nucleophilic attack over the carbonyl

Fig. 2. SEM micrographs determined from the composite catalyst using a secondary mode at magnifications of 5000X: a) as-synthesized, b) exhausted after 5 cycles.

4
R. Rodríguez-Ramírez et al. Journal of Environmental Chemical Engineering 10 (2022) 108191

respectively. In this case, the results of this analysis show that both, the
catalyst content and the MeOH: Oil ratio directly control the reaction
process in a similar way, as previously discussed. We propose that the
desorption of the methoxy ion (MetO-) from the surface proceeds at a
relatively fast pace, releasing the active sites more efficiently than in the
Na2ZrO3 case, close to the addition of Zn2+ lewis acid sites to Na2SiO3
presented elsewhere [10], with the fundamental difference of being in a
separate phase. This might facilitate the anion release thanks to an ion
mobility, while reducing the probability of re-protonation, therefore, the
more methanol in the reaction vessel, the highest concentration of
methoxy ions. The stirring rate is not important to maximize the con­
version, as formerly discussed in the conversion shown in Table 2.
Conversion contour plots accounting for methanol:oil molar pro­
portion as a function of the catalyst content (Fig. 5) were calculated at
constant stirring values of 200, 450, 500, and 700 RPM, to demonstrate
the system optimization process under determined ranges of the modi­
fied parameters. These graphs indicate that the extension of higher
conversion region can not be attained whether the stirring rate is kept at
a minimum, but it is achievable once we get to the middle point
(450 rpm), and starts reducing its size with the increase of the stirring
Fig. 4. Triglyceride conversions into FAMEs as a function of the trans­
speed. Nevertheless, it is required more than 9.5:1 methanol: oil ratio,
esterification reaction time (5, 15, 20, 30, 60 min).
and more than 4.2 wt% of catalyst to maintain an optimal conversion.
For this purpose, the minimum amount of catalyst possible was selected
Table 5 and balanced the response variable with the MetOH: Oil ratio to obtain
Conversion percentages obtained at different times an optimal conversion, giving the optimal conditions at 500 rpm, 4.4 wt
during the transesterification reaction using the com­ % catalyst and 14:1 methanol:oil proportion. These are tractable con­
posite catalyst. ditions compared to many heterogeneous catalysts (refer to Table 1)
Time (min) Conversion (%) including CaO use, described at 18:1 methanol:oil molar proportion,
while using 10 wt% of catalyst content [4].
5 78.3
15 90.5 The optimal conditions obtained from the BBD and RSM were tested
20 92.8 along with the residence time in a transesterification reaction to ratify
30 95.5 the estimates of the statistic evaluation. The reaction products were
60 96.1
analyzed by FT-IR analysis and 1H NMR to determine conversion and
selectivity from the transesterification mechanism (spectra described as
in the response variable. Most likely, this result is generated since con­ cycle 1 in Figs. 6 and 7, respectively). A 96.5% conversion was generated
vection inputs are more dominant than diffusion contributions (i.e. mass in the calibration curve using the FT-IR analysis [16,17] during the first
transfer), while the kinetic barriers have been satisfied at lower stirring cycle. The physicochemical properties of soybean oil and methyl esters
rates. Once the optimal conversion was obtained using the Minitab produced from the transesterification reaction of soybean oil are within
program (i.e. using the algorithm for RSM), the following polynomial the permissible limits of the quality standard ASTM D6751 [20].
equation (i.e. regression model) was determined as a function of coded The FT-IR spectra reported in Fig. 6 show the lack of -OH vibrations
variables: (typicaly displayed at 3500 cm− 1). Similarly, the absence of the O-CH2-C
group at 1098 cm− 1 in cycle one indicates the triglycerides depletion,
Conversion= − 125⋅8+0⋅2489 Stirring+21⋅49% Cat+13⋅67 Met:Oil-0⋅000174 and clear vibrations for CH3-O- and H-C-O- vibrations at 1437 and
Stirring*Stirring-1⋅000% Cat% Cat-0⋅406 Met:Oil*Met:Oil-0⋅00788 Stirring* 1197 cm− 1, respectively, considered as the most representative signals
% Cat-0⋅00168 Stirring*Met:Oil-0⋅196% Cat*Met:Oil (5) for biodiesel production. No signals were detected at 1570 cm− 1 corre­
sponding to carboxylates (COO-), confirming the absence of soaps. The
Table 6 displays the results of the analysis of variance (ANOVA) of
signals related to the C-H can be appreciated at 3010 and 2923 cm− 1;
the regression model (e.g. surface quadratic response) calculated based
while the C– – C bond of the unsaturated carbons can be found at
on the triglicerydes conversion, using ZnO⋅Na2ZrO3 as catalyst. Based on
2854 cm− 1, and the carboxyl group (C– – O) is sharply present at
this estimation, the methanol:oil molar proportion, catalyst load, and
1741 cm− 1.
stirring speed, exhibit inputs in the response equal 54%, 46%, and 17%,
Fig. 8 describes the 1H NMR analysis, where the methyl linoleate was
revealed as the main compound, followed by the coupling of CH=CH at
2.77 ppm, and the CH3-O singlet (3.64 ppm), consistent with the for­
Table 6
mation of omega-9 fatty acid. It is important to mention that the
Analysis of Variance (ANOVA) calculated with Minitab 17 for the response
surface quadratic model estimated for the conversion (response variable). coupling of CH-CH2 associated with the glyceride formation was not
found, compared against the NMR spectrum of the soybean oil. These
Source Analysis of variance
signals typically appear at 4.14 and 4.29 ppm, thus, indicating a FAMEs
df Sum of Mean F- P- Contribution conversion over 99%, while maintaining glycerides under the detection
Squares Square Value Value (%)
limits of the equipment.
Model 9 13617.3 1513.04 19.19 0.002 95.25
Stirring 1 551.9 551.94 7 0.046 16.61
3.4. Co-catalyst reusability
% Cat 1 4186 4185.96 53.1 0.001 45.75
Met:Oil 1 5737.7 5737.74 72.79 0 53.56
Lack-of- 3 392.8 130.95 202.1 0.005 The catalyst cyclability was estimated utilizing the optimal condi­
Fit tions of the system determined from previous sections for the trans­
Pure 2 1.3 0.65 esterification reaction. Thus, the triglicerydes conversion was evaluated
Error
during five consecutive cycles using the same co-catalyst only washed

5
R. Rodríguez-Ramírez et al. Journal of Environmental Chemical Engineering 10 (2022) 108191

Fig. 5. Contour graphs of triglyceride conversion accounting for the methanol:oil molar proportion in terms of the catalyst content at constant stirring rates of 200,
450, 500 and 700 RPM.

with methanol (Fig. 8). A new batch of methanol and triglycerides was
only incorporated in the reactor. Fig. 6 shows the reaction products
determined by FT-IR analysis, where the main signals related to the
methyl esters (i.e. H-CH2-O group at 1197 cm− 1, and CH3-O- vibration at
1437 cm− 1) are strongly observed in the first two cycles, starting to drop
from this point ahead.
Fig. 7 shows the 1H NMR spectra of the soybean oil and the reaction
products for each cycle. As observed, the fine signals are mainly asso­
ciated with the methyl linoleate at 3.65 ppm (e.g. fatty acid methyl ester
which belongs of the oil), corresponding to the CH3-O- group. The
glyceride protons signals which reveal an incomplete transesterification
reaction vibrate around 4.25 ppm. Thus, the selectivity can be estimated
using the area under the curves of 3.65 and 4.25 signals [18]. The
selectivity (1H NMR) was compared against the conversion calculated by
FTIR, as displayed in Table 7. Indeed, the selectivity values calculated by
1
H NMR are higher compared to the conversion percentages obtained
from the FTIR method (e.g. the glyceride vibrations are not displayed in
cycles 1 and 2). These variations are mostly produced due to a greater
sensitivity of 1H NMR to minor fluctuations in concentrations within the
reaction system, whereby their results are more accurate.
A comparison between the diffractograms collected from the as-
synthesized (Fig. 1a) and exhausted catalyst (Fig. 1b) shows that the
presence of the deprotonating phase (sodium zirconate) has almost
disappeared, possibly due to the affinity of the material with methanol
or the dissolution of the zirconium phase caused by the consumption of
sodium from the glycerolates (instead of protons). Likewise, the com­
parison of SEM micrographs of the new catalyst (Fig. 2a) compared to
the exhausted catalyst (Fig. 2b) reveals that the surface texture and
morphology of the sodium zirconate has significantly changed, reducing
its overall crystal size and becoming amorphous. According to the EDXS
analysis presented for these samples in Table 4, the overall concentra­
tion of sodium zirconate has fallen under 5 at%, supporting the results
obtained in the XRD analysis. Thus, Sodium Zirconate and ZnO must be
Fig. 6. FT-IR spectra collected for the soybean oil, and products of the trans­
esterification reaction during five consecutive cycles. strongly bonded to improve its reusability.

6
R. Rodríguez-Ramírez et al. Journal of Environmental Chemical Engineering 10 (2022) 108191

Table 7
Comparison of the biodiesel conversion and selectivity obtained from the
analysis of the FT-IR and 1 H NMR spectra, respectively, using optimized con­
ditions after each cycle.
Cycle Conversion (FT-IR) Selectiviy (NMR)

1 96.5 100.0
2 93.0 99.5
3 68.8 80.7
4 73.1 –
5 43.2 62.2

3.5. Reaction mechanism proposal

The reaction mechanism for this material is considered to start in the


same way as the traditional heterogeneous transesterification reaction
[10]. First, methanol approaches the most basic material as shown in
Fig. 9 (reaction 1), deprotonating and forming a metastable intermedi­
ate with the catalyst, with the proton being held in the alkaline oxygen
and the methoxy ion in the lewis acid site. In this case, it seems that the
fast kinetics is facilitated by the migration of the proton of the methanol
from the zirconate oxygens to the ones situated in the zinc oxide (Fig. 9,
reaction 2), which are not suitable for the methanol deprotonation,
increasing the distance between them, decreasing the influence of each
other, and holding the proton until another anion reacts with it. This is
reflected in a lower recombination and more methoxy ions in solution,
arriving to attack the glycerol carboxyl group. The next stages proceed
as in the conventional heterogeneous transesterification (Fig. 9, reaction
3).
As previously reported, poisoning of the surface of the material with
glycerol may be the main cause of the decrease in the FAMEs conversion
rate. Nevertheles, if the hypothesis for proton migration is true, another
limiting step for the reaction could be the sequestration of sodium by the
methoxy ion, forming sodium alkoxide, which could suppress the ac­
tivity of the alkaline material.

Fig. 7. NMR spectra stack of different consecutive cycles of the trans­


esterification reaction. The region of the molecule structure associated with the
chemical shift is displayed in the upper part of the graph.

Fig. 8. Conversion of triglycerides as a function of cycle estimated using FT-IR


analysis, describing the catalyst reusability. Fig. 9. Reaction mechanism for the transesterification reaction arising on the
synergetic interaction in the ZnO⋅Na2ZrO3 composite.

7
R. Rodríguez-Ramírez et al. Journal of Environmental Chemical Engineering 10 (2022) 108191

4. Conclusions [6] A. Naeem, I. Wali Khan, M. Farooq, T. Mahmood, I. Ud Din, Z. Ali Ghazi, et al.,
Kinetic and optimization study of sustainable biodiesel production from waste
cooking oil using novel heterogeneous solid base catalyst, Bioresour. Technol. 328
Heterogeneous biodiesel production presents important challenges (2021), 124831, https://doi.org/10.1016/j.biortech.2021.124831.
as overcoming the price and efficiency of the homogeneous process, to [7] P.D. Patil, S. Deng, Transesterification of Camelina sativa oil using heterogeneous
become an attractive and a green alternative for the industry. In order to metal oxide catalysts, Energy Fuels 23 (2009) 4619–4624, https://doi.org/
10.1021/ef900362y.
contribute to this goal, the synthesis and characterization of a [8] A. Navajas, I. Campo, G. Arzamendi, W.Y. Hernández, L.F. Bobadilla, M.
ZnO⋅Na2ZrO3 phase was conducted by a solid-state process. The XRD A. Centeno, et al., Synthesis of biodiesel from the methanolysis of sunflower oil
study revealed that the composite was successfully synthesized pre­ using PURAL® Mg–Al hydrotalcites as catalyst precursors, Appl. Catal. B: Environ.
100 (2010) 299–309, https://doi.org/10.1016/j.apcatb.2010.08.006.
senting a salient plane of (101) and (− 331) for ZnO and Na2ZrO3, [9] I.W. Khan, A. Naeem, M. Farooq, I.U. din, Z.A. Ghazi, T. Saeed, Reusable Na-SiO2@
respectively. The co-catalyst displayed an specific area around 0.7 m2 CeO2catalyst for efficient biodiesel production from non-edible wild olive oil as a
g− 1, and did not present any impurities that may affect the reaction. The new and potential feedstock, Energy Convers. Manag. 231 (2021), 113854, https://
doi.org/10.1016/j.enconman.2021.113854.
composite material presented an important activity since the reaction [10] R. Rodríguez-Ramírez, I. Romero-Ibarra, J. Vazquez-Arenas, Synthesis of sodium
occurred in 15 min with a conversion into biodiesel higher than 99%, zincsilicate (Na2ZnSiO4) and heterogeneous catalysis towards biodiesel production
according to the conditions calculated by a statistical optimization in a via Box-Behnken design, Fuel 280 (2020), 118668, https://doi.org/10.1016/j.
fuel.2020.118668.
BBD along with a RSM: 500 RPM, 4.4 wt% of catalyst load and 14:1 [11] N. Santiago-Torres, I.C. Romero-Ibarra, H. Pfeiffer, Sodium zirconate (Na2ZrO3) as
Methanol:oil proportion, using triglycerides from soybean oil at 65 ºC. a catalyst in a soybean oil transesterification reaction for biodiesel production, Fuel
The factors sensitivity adopted the following order: Methanol: oil ratio Process. Technol. 120 (2014) 34–39, https://doi.org/10.1016/j.
fuproc.2013.11.018.
≥ Catalyst load > >stirring rate. The reusability test of the trans­
[12] D.A. Torres-Rodríguez, I.C. Romero-Ibarra, I.A. Ibarra, H. Pfeiffer, Biodiesel
esterification reaction using ZnO⋅Na2ZrO3 showed decay in the con­ production from soybean and Jatropha oils using cesium impregnated sodium
version rate to 43.2% and 62% after five consecutive cycles, relying on zirconate as a heterogeneous base catalyst, Renew. Energy 93 (2016) 323–331,
FT-IR analysis and 1H NMR, respectively. A reaction mechanism was https://doi.org/10.1016/j.renene.2016.02.061.
[13] C. Molina, ZnO nanorods as catalyst for biodiesel production from olive oil,
proposed to explain the synergy arising in the ZnO⋅Na2ZrO3 composite, Electron. Theses Diss. (2013), https://doi.org/10.18297/etd/1000.
establishing the hypothesis of a facile desorption of the methoxy ion due [14] R.H. Myers, D.C. Montgomery. Response Surface Methodology: Process and
to the diffusion of protons to the ZnO phase. Product in Optimization Using Designed Experiments, 1st ed..,, John Wiley & Sons,
Inc.,, New York, NY, USA, 1995.
[15] V. Singh, L. Belova, B. Singh, Y.C. Sharma, Biodiesel production using a novel
CRediT authorship contribution statement heterogeneous catalyst, magnesium zirconate (Mg2Zr5O12): Process optimization
through response surface methodology (RSM), Energy Convers. Manag. 174 (2018)
198–207, https://doi.org/10.1016/j.enconman.2018.08.029.
Ricardo Rodríguez-Ramírez: Investigation, Experimental data [16] N.N. Mahamuni, Y.G. Adewuyi, Fourier Transform Infrared Spectroscopy (FTIR)
acquisition, Data treatment, Data curation. Fabiola S. Sosa-Rodríguez: Method To Monitor Soy Biodiesel and Soybean Oil in Transesterification Reactions,
Conceptualization, Collection and Discussion of XRD and SEM analyses. Petrodiesel− Biodiesel Blends, and Blend Adulteration with Soy Oil, Energy Fuels
23 (2009) 3773–3782, https://doi.org/10.1021/ef900130m.
Jorge Vazquez-Arenas: Methodology, Writing – original draft, Writing [17] K. Zhao, L. Shi, Z. Liu, J. Li, Quality Analysis of Reheated Oils by Fourier Transform
– review & editing. Infrared Spectroscopy, Atlantis Press,, 2015, pp. 245–249, https://doi.org/
10.2991/icectt-15.2015.47.
[18] A. Martínez, G.E. Mijangos, I.C. Romero-Ibarra, R. Hernández-Altamirano, V.
Declaration of Competing Interest
Y. Mena-Cervantes, In-situ transesterification of Jatropha curcas L. seeds using
homogeneous and heterogeneous basic catalysts, Fuel 235 (2019) 277–287,
The authors declare that they have no known competing financial https://doi.org/10.1016/j.fuel.2018.07.082.
interests or personal relationships that could have appeared to influence [19] S. Lowell. Powder surface area and porosity/, 3rd ed..,, Chapman and Hall,,
London, 1991.
the work reported in this paper. [20] Standard Specification for Biodiesel Fuel Blend Stock (B100) for Middle Distillate
Fuels n.d. 〈https://www.astm.org/d6751–20a.html〉 (accessed June 2, 2022).
Acknowledgments [21] Y.M. Sani, P.A. Alaba, A.O. Raji-Yahya, A.R. Abdul Aziz, W.M.A.W. Daud, Acidity
and catalytic performance of Yb-doped SO42− /Zr in comparison with SO42− /Zr
catalysts synthesized via different preparatory conditions for biodiesel production,
The authors thank the support from CONACyT “Ciencia Básica y/o J. Taiwan Inst. Chem. Eng. 59 (2016) 195–204, https://doi.org/10.1016/j.
Ciencia de Frontera. Modalidad: Paradigmas y Controversias de la jtice.2015.07.016.
[22] J. Iglesias, J.A. Melero, L.F. Bautista, G. Morales, R. Sánchez-Vázquez, Continuous
Ciencia 2022” grant no. 320252, and SIP-IPN 2194 (modules 20221207 production of biodiesel from low grade feedstock in presence of Zr-SBA-15:
and 20221694). Catalyst performance and resistance against deactivation, Catal. Today 234 (2014)
174–181, https://doi.org/10.1016/j.cattod.2014.01.004.
[23] M.J. Borah, A. Das, V. Das, N. Bhuyan, D. Deka, Transesterification of waste
References cooking oil for biodiesel production catalyzed by Zn substituted waste egg shell
derived CaO nanocatalyst, Fuel 242 (2019) 345–354, https://doi.org/10.1016/j.
[1] D02 Committee. Specification for Diesel Fuel Oils. ASTM International; n.d. 〈htt fuel.2019.01.060.
ps://doi.org/10.1520/D0975–19B〉. [24] C. Sun, Y. Hu, F. Sun, Y. Sun, G. Song, H. Chang, et al., Comparison of biodiesel
[2] D02 Committee. Specification for Diesel Fuel Oil, Biodiesel Blend (B6 to B20). production using a novel porous Zn/Al/Co complex oxide prepared from different
ASTM International; n.d. 〈https://doi.org/10.1520/D7467–19〉. methods: Physicochemical properties, reaction kinetic and thermodynamic studies,
[3] D02 Committee. Specification for Biodiesel Fuel Blend Stock (B100) for Middle Renew. Energy 181 (2022) 1419–1430, https://doi.org/10.1016/j.
Distillate Fuels. ASTM International; n.d. 〈https://doi.org/10.1520/D6751–19〉. renene.2021.09.122.
[4] G. Sahu, N.K. Gupta, A. Kotha, S. Saha, S. Datta, P. Chavan, et al., A Review on [25] S.M. Ibrahim, Preparation, characterization and application of novel surface-
Biodiesel Production through Heterogeneous Catalysis Route, ChemBioEng Rev. 5 modified ZrSnO4 as Sn-based TMOs catalysts for the stearic acid esterification with
(2018) 231–252, https://doi.org/10.1002/cben.201700014. methanol to biodiesel, Renew. Energy 173 (2021) 151–163, https://doi.org/
[5] V.K. Mishra, R. Goswami, A review of production, properties and advantages of 10.1016/j.renene.2021.03.134.
biodiesel, Biofuels 9 (2018) 273–289, https://doi.org/10.1080/
17597269.2017.1336350.

You might also like