One step synthesis of ethyl levulinate biofuel by ethanolysis of renewable furfuryl alcohol over hierarchical zeolite catalyst

Article abstract of DOI:10.1039/c5ra13520f

Ethanolysis of renewable furfuryl alcohol (FAL) to ethyl levulinate (EL) biofuel over various zeolites viz. H-ZSM-5 (microporous, medium pore), Hierarchical-HZ-5 (combination of micro- and meso pore), H-Beta (microporous, large pore) and Ultra Stable Y (USY, microporous, large pore) was studied in detail. To the best of our knowledge, probably for the first time, Hierarchical-HZ-5 synthesized by desilication post-treatment has been employed as a heterogeneous catalyst for ethanolysis of FAL. The synthesized catalysts were characterized by powder X-ray diffraction (PXRD), temperature programmed NH₃ desorption (TPAD), Energy dispersive X-ray analysis (EDAX), etc. Response surface methodology (RSM) with Box-Behnken experimental design (BBD) was used to investigate the influence of three crucial process variables of ethanolysis such as ethanol to FAL molar ratio, percent catalyst loading and reaction temperature on EL yield. The optimization tool of design expert software was employed to obtain the optimum reaction parameters for FAL ethanolysis over Hierarchical-HZ-5 catalyst. Three intermediates of FAL ethanolysis reaction such as, ethoxymethylfuran (EMF), 4,5,5-triethoxypentan-2-one and diethyl ether (DEE) have been identified and quantified from the product mixture with the aid of Gas Chromatography-Mass Spectroscopy (GC-MS). Hierarchical-HZ-5 was found to be a potential catalyst for ethanolysis of FAL with 73% EL yield and 26% EMF yield at optimized process parameters.

Full text of DOI:10.1039/c5ra13520f

RSC Advances
View Article Online
View Journal | View Issue
PAPER
One step synthesis of ethyl levulinate biofuel by
ethanolysis of renewable furfuryl alcohol over
hierarchical zeolite catalyst†
Cite this: RSC Adv., 2015, 5, 79224
*
Kakasaheb Y. Nandiwale, Ashwini M. Pande and Vijay V. Bokade
Ethanolysis of renewable furfuryl alcohol (FAL) to ethyl levulinate (EL) biofuel over various zeolites viz.
H-ZSM-5 (microporous, medium pore), Hierarchical-HZ-5 (combination of micro- and meso pore),
H-Beta (microporous, large pore) and Ultra Stable Y (USY, microporous, large pore) was studied in detail.
To the best of our knowledge, probably for the first time, Hierarchical-HZ-5 synthesized by desilication
post-treatment has been employed as a heterogeneous catalyst for ethanolysis of FAL. The synthesized
catalysts were characterized by powder X-ray diffraction (PXRD), temperature programmed NH3
desorption (TPAD), Energy dispersive X-ray analysis (EDAX), etc. Response surface methodology (RSM)
with Box–Behnken experimental design (BBD) was used to investigate the influence of three crucial
process variables of ethanolysis such as ethanol to FAL molar ratio, percent catalyst loading and reaction
temperature on EL yield. The optimization tool of design expert software was employed to obtain the
optimum reaction parameters for FAL ethanolysis over Hierarchical-HZ-5 catalyst. Three intermediates
of FAL ethanolysis reaction such as, ethoxymethylfuran (EMF), 4,5,5-triethoxypentan-2-one and diethyl
ether (DEE) have been identified and quantified from the product mixture with the aid of Gas
Chromatography-Mass Spectroscopy (GC-MS). Hierarchical-HZ-5 was found to be a potential catalyst
for ethanolysis of FAL with 73% EL yield and 26% EMF yield at optimized process parameters.
Received 10th July 2015
Accepted 11th September 2015
DOI: 10.1039/c5ra13520f
www.rsc.org/advances
catalyzed hydrolysis of FAL to levulinic acid (LA). Followed by,
esterication LA with ethanol over acid catalyst. It has been
1 Introduction
reported that, the hydrolysis of FAL encounters polymerization
of FAL, leading to less production of LA.9,11 Also, the carboxylate
functional group in aqueous medium poisons the heteroge-
neous catalysts leading to loss in activity for subsequent reac-
tions.9–11 Route 2 is one step acid catalyzed ethanolysis of FAL to
EL.1,9–16 Ethanolysis would be an atom-economic and more
benecial compared to Route 1 (hydrolysis and esterication) as
The practical transformation of inexpensive and renewable
biomass into industrially important chemicals is one of the
important technological challenges for today's researchers.1 In
this context, furfural is an important platform chemical
obtained by hydrolysis and dehydration of hemicellulose has
been commercialized for decades.2 Hydrogenation of this
biomass derived furfural leads to furfuryl alcohol (FAL). The
threat of fossil fuel shortage and environmental concern is
stimulating the search of alternative fuels, hence efficient
conversion of FAL to ethyl levulinate (EL), an important
renewable oxygenate fuel additive would be sustainable green
process.3–5 Alkyl levulinates like EL have been already recog-
nized as one of the top 10 biorenery candidates in 2004 by
United States Department of Energy.6,7 Moreover, EL can also be
used as a precursor to produce g-valerolactone, which can be
converted to liquid alkanes and transportation fuels.8–11
There are two pathways which can be employed to transform
FAL to EL (Scheme 1): Route 1 involves two steps, rst acid
Catalysis and Inorganic Chemistry Division, CSIR-National Chemical Laboratory,
Pune-411008, India. E-mail: vv.bokade@ncl.res.in; Fax: +91-20-25902634; Tel: +91-
20-25902458
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c5ra13520f
Scheme 1 Acid-catalyzed transformation of furfuryl alcohol (FAL) to
ethyl levulinate (EL) biofuel.
79224 | RSC Adv., 2015, 5, 79224–79231
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
it reduces the polymerization of FAL, which ultimately gives
high EL yields.5,10,11 Also, one step process of ethanolysis of FAL
to EL would reduce process cost, hence it would be more
economical as compared to Route 1.
The ethanolysis of FAL has been reported by using strong
homogeneous acids.17,18 Although, strong homogeneous acids
such as HCl and H2SO4 nd effective in the ethanolysis reaction,
however they present the environmental concerns due to
extremely corrosive nature and being difficult to separate from
product mixture for recycle. Hence, it is technological demand
to establish environmentally sustainable process for ethanolysis
of FAL by designing highly active, efficient, reusable and
industrially benign heterogeneous catalyst. In this context,
performances of various catalysts such as propylsulfonic acid-
functionalized mesoporous silica,1 functionalized silica hollow
spheres,10,15 acid resin,11,12 zeolites,12 ionic liquids,13 and porous
aluminosilicates14 have been extensively investigated for etha-
nolysis. To the best of our knowledge, hierarchical H-ZSM-5 has
not been reported for ethanolysis of FAL. Hence, in view of
advantages of heterogeneous catalysts and to be aware of the
signicance and applications of FAL ethanolysis reaction from
industrial as well as academic point of view, it is thought of
research interest to evaluate the catalytic performance of
various zeolites such as H-Beta, USY, H-ZSM-5 and Hierarchical
H-ZSM-5 (Hierarchical-HZ-5) catalyst for FAL ethanolysis. These
zeolites were used due to their peculiar properties of inherent
acidity, porosity, surface area and temperature stability in
comparison with other zeolites.
Along with catalyst selection for the FAL ethanolysis reac-
tion, the optimization of process parameters is also utmost
important to recognize industrially benign catalyst. Response
surface methodology (RSM) is extensively used as an optimiza-
tion soware for various biomass conversion reactions such as
esterication and transesterication reactions.18–22 So far, RSM
has not been reported for ethanolysis reaction.
The present research involves insight study on inuence of
process parameters for ethanolysis of FAL with the help of RSM
design expert soware with Box–Behnken experimental design
(BBD) over Hierarchical-HZ-5, catalyst under our knowledge for
the rst time. The optimization tool of design expert soware is
used to obtain optimum process parameters for ethanolysis
reaction with aim to maximize EL yield. The inuences of three
critical process parameters such as molar ratio (ethanol to FAL),
percent catalyst loading and reaction temperature on EL yield
were investigated by BBD of RSM. The correlation between the
process variables is established using experimental and math-
ematical equations. The optimum reaction parameters
endorsed by RSM were validated by experiments. The optimized
reaction parameters were used to evaluate the reusability of
potential catalyst.
Fig. 1 Powder X-ray diffraction patterns of USY, H-Beta, H-ZSM-5,
Hierarchical-HZ-5 and used Hierarchical-HZ-5 catalyst.
diffraction pattern of H-Beta shows BEA phase, USY shows
typical FAU and H-ZSM-5 and Hierarchical-HZ-5 show MFI
phase (Fig. 1).
Fig. 2 depicts the N2 physisorption isotherms of H-ZSM-5
and Hierarchical-HZ-5 catalysts. The isotherm of the parent
H-ZSM-5 shows a plateau starting at a very low relative pressure
(type I), the characteristic of microporous zeolite structures. On
the other hand the N2 isotherms of Hierarchical-HZ-5 repre-
sents hysteresis loop at higher P/P0 value attributed to type I and
type IV isotherms which suggests the presence of both micro
and mesoporosity in Hierarchical-HZ-5 catalyst. Physico-
chemical properties of all catalysts is represented in Table 1.
2 Results and discussion
2.1 Catalyst characterizations and catalytic performance
The synthesized catalysts were characterized by XRD, BET and
TPAD. Fig. 1 shows powder X-ray diffraction patterns of H-Beta,
Fig. 2 N2 adsorption–desorption isotherms of H-ZSM-5 and Hierar-
USY, H-ZSM-5 and Hierarchical-HZ-5 catalysts. The X-ray chical-HZ-5 catalysts.
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 79224–79231 | 79225

View Article Online
RSC Advances
Paper
Table 1 Physico-chemical properties of catalysts
Si/Al ratioa
SBETb (m2 gꢀ1
)
VP (cm3 gꢀ1
c
)
DP (A)
Total aciditye (mmol gꢀ1
)
d
˚
Catalyst
USY
H-Beta
H-ZSM-5
Hierarchical-HZ-5
15
8.8
37
773
560
300.8
427.6
0.42
0.38
0.17
0.31
7.2
6.4
5.5
0.38
0.54
0.51
0.73
30.15
29.78
a
Si/Al ratio was estimated by EDAX. b Surface area (SBET) was calculated by Brunauer–Emmett–Teller equation. c Pore volume (VP) was determined
from single point desorption isotherm at P/P0 ¼ 0.9. d Pore diameter (DP) was calculated using Barrett–Joyner–Halenda desorption branch of the
isotherm. e Total acidity was determined with ammonia TPD.
Hierarchical-HZ-5 showed higher BET surface area, pore removal of silicon by desilication in Hierarchical-HZ-5 increases
volume, pore diameter and acidity as compared to parent the total acidity, this led to increase in EL yield, as this reaction
H-ZSM-5, this is attributed to removal of silicon from the is acid driven. Hierarchical-HZ-5 was also observed to be higher
framework without complete destruction of the lattice.23 Fig. S1 in surface area, pore volume than parent HZSM-5, which
(ESI†) represents the surface morphology of H-ZSM-5 and increases the accessibility of active sites in the Hierarchical-HZ-
Hierarchical-HZ-5 observed under FE-SEM. FE-SEM demon- 5. Amongst the evaluated catalysts, Hierarchical-HZ-5 was
strated that, the overall size and morphology of alkali treated found to be potential catalyst for ethanolysis of FAL. This
Hierarchical-HZ-5 zeolite was not signicantly affected.
comparison amongst micro-medium, micro-large and micro-
Catalytic performances of USY, H-Beta, H-ZSM-5 and meso porosity conrmed that the ethanolysis of FAL required
Hierarchical-HZ-5 catalysts were assessed for ethanolysis of FAL both micro-meso porosity, which helps to convert intermediate
to EL at identical set of reaction parameters: molar ratio into desired products. Hence, in view to maximize EL yield, RSM
(ethanol to FAL) of 8 : 1 and catalyst loading of 10 wt% of FAL, design with BBD was implied to investigate the inuence of
reaction temperature of 373 K and reaction time of 2 h (Fig. 3). various process parameters for an ethanolysis FAL to EL over
Three intermediates such as ethoxymethylfuran (EMF), 4,5,5- Hierarchical-HZ-5. Also, the optimized process parameters for
triethoxypentan-2-one and diethyl ether (DEE) have been iden- maximum EL yield and reusability of Hierarchical-HZ-5 catalyst
tied and quantied from product mixture with aid of GC-MS is presented later.
and GC. Hierarchical-HZ-5 was found to be most active
The ethanolysis of FAL was studied over Hierarchical-HZ-5
amongst the studied catalysts. The overall trend of EL yield with varying reaction time. Fig. 4 shows the progress in yield
obtained was USY (5%) < H-Beta (8%) < H-ZSM-5 (13%) < of EL, EMF, 4,5,5-triethoxypentan-2-one and DEE with respect to
Hierarchical-HZ-5 (19%). Hierarchical-HZ-5 (micro- and meso reaction time. It has been observed that at initial reaction time
pore) was observed to be more active as compared to micropo- up to 2 h, the formation of 4,5,5-triethoxypentan-2-one was
rous medium pore (H-ZSM-5) and microporous large pore more as compared to EL, however aer reaction time of 3.5 h,
(H-Beta and USY) zeolite. The activity trend obtained was the the DEE and 4,5,5-triethoxypentan-2-one were no longer
cumulative effect of physico-chemical properties such as observed in the reaction product. This is attributed to
surface area, pore volume and acidity of catalyst (Table 1). The
Fig. 3 Catalytic performance of various zeolites for ethanolysis of
furfuryl alcohol (FAL) at reaction conditions: molar ratio (ethanol to Fig. 4 Influence of reaction time on ethanolysis of FAL over Hierar-
FAL) of 8 : 1 and catalyst loading of 10 wt% of FAL, reaction temper- chical-HZ-5 at reaction conditions: molar ratio (ethanol to FAL) of 8 : 1
ature of 373 K and reaction time of 2 h.
and catalyst loading of 10 wt% of FAL and reaction temperature of 373 K.
79226 | RSC Adv., 2015, 5, 79224–79231
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
Table 2 Selected variables and coded levels used in the Box–Behnken
design
utilization of 4,5,5-triethoxypentan-2-one and DEE in formation
of EL. The EMF was observed in reaction product even aer 4 h
with 49% yield. The results obtained about the product distri-
bution are in consistent with the reported literature.11,13 EL yield
reached at its maximum value of 41% at reaction time of 4 h and
thereaer it remained unchanged. Hence, reaction time of 4 h
with 41% EL yield and 49% EMF yield was considered as
optimum reaction time and used in all further experiments.
Coded levels
Variables
Symbol
ꢀ1
0
+1
12
30
413
Molar ratio (ethanol to FAL)
Catalyst loading (%)
Reaction temperature (K)
X1
X2
X3
4
10
373
8
20
393
2.2 Statistical analysis of RSM and inuence of process
signal to noise ratio) >4 is desirable. In present study, the
adequate precision ratio of 58.35 meant an adequate signal and
hence model could be employed to navigate the design space. In
addition, a low value of the coefficient of variation (CV ¼ 1.88%)
demonstrated high degree of precision and great deal of
experimental reliability. Furthermore, in present model, a
minimum of three Lack of Fit degrees of freedom (Df) and four
Df for ‘Pure Error’ conrmed a validity of ‘Lack of Fit‘ test
(Table S2†). The above statistical analysis implied that the
model is adequate to predict the EL yield (Y) within the scope of
the variable investigated.
parameters
2.2.1 The model tting and statistical analysis. The etha-
nolysis of FAL to EL over Hierarchical-HZ-5 was optimized
through RSM approach. The seventeen designed experiments in
order to optimize three process parameters in the BBD were
represented in Table S1 (ESI†). Table S1† and Fig. 5 implied that
there was no noticeable variation among the actual and pre-
dicted response values.
By applying the multiple regression analysis on the experi-
mental outcome, the responsive variable (Y) and the three test
variables were related in terms of coded factors by second-order
polynomial eqn (1):
The p-values less than 0.05 (p < 0.05) indicate the model term
is signicant.19–21 In present study, linear term of catalyst loading
(X2, F-value ¼ 1152) was found to be more important than molar
ratio (X1, F-value ¼ 512) and reaction temperature (X3, F-value ¼
84.5). The interaction between molar ratio and catalyst loading
(X1X2, F-value ¼ 72.25) has more inuence on response (Y) than
the catalyst loading and reaction temperature (X2X3, F-value ¼
30.25). However, the interaction between molar ratio and reac-
tion temperature (X1X3) is not signicant, owing to low F-value of
0.25 and high p-value of 0.63. Moreover, the quadratic term of
Y ¼ +54 ꢀ 8X1 + 12X2 + 3.25X3 + 4.25X1X2 + 0.25X1X3
2
2
2
+ 2.75X2X3 ꢀ 2.12X1 – 0.13X2 + 0.38X3
(1)
where X1, X2 and X3 are the coded process variables for etha-
nolysis and Y indicates the response (EL yield) (Table 2).
Statistical testing of regression equation was performed by
analysis of variance (ANOVA) and Fisher F-test (Table S2†). At
95% of the condence level, the model F-value of 207.86 with
very low probability value (p < 0.001) indicated that the model
was reliable to predict the yield of EL. The predicted R2
(R2-predicted ¼ 0.9404) was in reasonable agreement with the
adjusted R2 (R2-adjusted ¼ 0.9915). Adequate precision (the
2
molar ratio (X1 ) with p-value of 0.003 is signicant term, while
X22 and X32 are not signicant terms (p > 0.05).
2.2.2 Analysis of the response surface plots. For better
visualization of the statistically signicant process parameters
Fig. 5 Plot of actual versus predicted values of EL yield over Hierarchical-HZ-5 catalyst.
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 79224–79231 | 79227

View Article Online
RSC Advances
Paper
depicted from ANOVA, the three-dimensional (3D) response The interaction effect of catalyst loading and molar ratio on EL
surface plots and two dimensional (2D) contour curves were yield was found to be signicant with shape of 2D contour curve
represented as Fig. 6–8. These types of plot represent the effect ellipse mound and the low p-value (<0.0001) of the interaction
of two process parameters on response (EL yield) at one time term (X1X2) (Table S2†).
while other one variable is maintained at its zero level (central
The inuence of molar ratio and reaction temperature on EL
value). All these experiments were performed at constant reac- yield at a constant catalyst loading of 20% and reaction time of
tion time of 5 h. The inuence of correlation among catalyst 5 h is depicted as Fig. 7. With increase in reaction temperature
loading and molar ratio (Fig. 6), molar ratio and reaction from 373 to 413 K at constant molar ratio of 4 : 1, the EL yield
temperature (Fig. 7) and catalyst loading and reaction temper- increased from 57 to 62%. This is attributed to increase in rate
ature (Fig. 8) at constant reaction time of 4 h are indicated by 3D of polymerization of FAL to EL at elevated temperature.
response surface plots and 2D contour plots.
However, the interaction between reaction temperature and
Fig. 6 represents the effects of molar ratio (ethanol to FAL) molar ratio has no inuence on EL yield which is also indicated
and catalyst loading (wt% of FAL) on EL yield at reaction by high p-value (0.6324) of interaction term (X1X3) (Table S2†).
temperature of 393 K and reaction time of 5 h. It can be seen
Fig. 8 shows effect of different catalyst loading (%) and reaction
from Fig. 6 that at constant molar ratio of 4 with increase in temperature on the EL yield in 3D surface response and 2D
loading of Hierarchical-HZ-5 catalyst from 10 to 30%, the EL interaction plots at identical molar ratio of 8 : 1 and reaction time
yield increased from 53 to 68%. In contrary, with same rise in of 5 h. It is obvious from the gure that at any designed quantity of
catalyst loading (10–30%), the EMF yield decreased from 45 to catalyst loading from 10 to 30%, the EL yield was found to be
26%. The lower catalyst loading (10%) showed less activity (53% increased proportionally with reaction temperature. At xed
EL yield and 45% EMF yield), indicating the need of catalytically reaction temperature of 413 K, with increase in catalyst loading
active acid sites for EL formation. The catalyst loading of 30% from 10 to 30%, the EL yield increased from 43 to 73%. The
provides more catalytically active acid sites which help in con- individual terms of catalyst loading (X2) and reaction temperature
verting intermediates such as EMF to EL (Table 1). The EL yield (X3) with p-value of <0.0001 have more signicant effect on EL yield
is found to be proportional to catalyst amount used which than the interaction term X2X3 with p-value of 0.0009 (Table S2†).
indicates that the reaction follows pure heterogeneous mecha-
2.2.3 Multiple response optimization of FAL ethanolysis
nism. Moreover, as indicated by low p-value (<0.0001) (Table and model validation. The optimum process parameters for
S2†), the catalyst loading is signicantly inuencing process ethanolysis of FAL over Hierarchical-HZ-5 catalyst were
parameter in ethanolysis of FAL. However, with increase in obtained by using numerical algorithm built in the Design-
molar ratio of ethanol to FAL from 4 : 1 to 12 : 1 at constant Expert® Version 8.0.7.1 soware. The three independent reac-
catalyst loading of 30%, the EL yield decreased from 68 to 59% tion parameters depicted in Table 2 were set in the range among
and EMF yield increased from 26 to 36%. The higher dilution of low (ꢀ1) and high (+1) while the EL yield (response) was set to
FAL at identical catalyst acid sites leads to more formation of the be at maximum value.18,19 Table 3 shows the predicted and
intermediate EMF due to partial polymerization of FAL. This experimental EL yield at optimized process parameters. EL yield
implies that the higher yield of EL can be obtained at lower of 73% is in ne agreement with the predicted value (73.4%),
molar ratio, which would save operating, equipment and energy with the experimental error less than ꢁ5%. This implied that
cost associated with the separation of unreacted ethanol from the proposed RSM statistical model valid and accurate to
nal product mixture making the process industrially benign. predict EL yield.
Fig. 6 Response surface (3D) and contour (2D) plot of EL yield as a function of molar ratio (ethanol to FAL) and catalyst loading at reaction
temperature of 393 K and reaction time of 4 h.
79228 | RSC Adv., 2015, 5, 79224–79231
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
Fig. 7 Response surface (3D) and contour (2D) plot of EL yield as a function of molar ratio (ethanol to FAL) and reaction temperature at catalyst
loading of 20% and reaction time of 4 h.
elemental analysis of used Hierarchical-HZ-5 by EDAX showed
Si/Al ratio of 30, indicating no further desilication of catalyst
during the reaction. This indicates that the Hierarchical-HZ-5
catalyst is stable, highly active and has a potential of its
application.
Song et al.,10 used aryl sulfonic acid functionalized hollow
mesoporous carbon spheres (ArSO3H-HMCSs) for ethanolysis
of FAL to EL. They reported EL yield of 81.3% at ethanol to FAL
molar ratio of 60 : 1 and reaction temperature of 393 K. Use of
excess amount of ethanol would be uneconomical in large
scale process.10 Wang et al.,13 reported 90% EL yield with 1,3-
bis(sulfopropyl)1H-imidazol-3-ium hydrogenosulfate ionic
liquid. However, ionic liquids have known limitations such as
not easy to separate and reuse. Neves et al.,14 used various
porous aluminosilicates for ethanolysis of FAL to EL. They
reported Al-TUD-1 zeolite with 80% EL yield at 60 equivalent of
ethanol, reaction temperature of 413 K and reaction time of 24
2.3 Reusability of catalyst
The optimized process parameters summarized in Table 3 were
used to evaluate the reusability of Hierarchical-HZ-5 catalyst
for ethanolysis of FAL. The catalyst was separated by ltration
from product mixture and used for next catalytic run.
Hierarchical-HZ-5 catalyst was found to be stable for three
reaction cycles (fresh and two reuses) with same 73% EL yield
and 26% EMF yield (Fig. 9). Thereaer, for the fourth cycle the
EL yield decreased marginally from 73% to 70%. Four times
used Hierarchical-HZ-5 catalyst was characterized by XRD
(Fig. 1) and EDAX. The XRD pattern of used Hierarchical-HZ-5
in Fig. 1 indicated slight decrease in intensity of peaks as
compared to fresh Hierarchical-HZ-5. This is attributed to the
deposition of reaction intermediates to catalyst surface.
However, four times used Hierarchical-HZ-5 was found to be
still crystalline with no amorphous contribution. The
Fig. 8 Response surface (3D) and contour (2D) plot of EL yield as a function of catalyst loading (%) and reaction temperature (K) at molar ratio
(ethanol to FAL) of 8 : 1 and reaction time of 4 h.
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 79224–79231 | 79229

View Article Online
RSC Advances
Paper
Table 3 Most favorable process parameters for ethanolysis of FAL over Hierarchical-HZ-5 zeolite for reaction time 4 h and validation model
adequacy
Molar ratio (ethanol
Catalyst loading
Reaction temperature,
EL yield,
Process parameters
to FAL), X1
(wt%), X2
X3 (K)
Y (%)
Predicted
Experimental
5.7
6
29.9
30
412.4
413
73.4
73
4 Experimental section
4.1 Materials
Ultra Stable Y (USY) zeolite (Si/Al ¼ 15) was procured from
Zeolyst, USA. Ethyl levulinate (99.8%) and furfuryl alcohol (98%)
were obtained from Sigma Aldrich (USA). Ethanol (99.9%) was
obtained from Cympran Gludt BV, Belgium. H-Beta, H-ZSM-5,
Hierarchical-HZ-5 catalysts were synthesized at Catalyst Pilot
Plant, CSIR-NCL, Pune (India).
4.2 Catalyst synthesis and characterization
Reported procedures were used to synthesize H-Beta (Si/Al ¼ 8),24
and H-ZSM-5 (Si/Al ¼ 37).25 Hierarchical-HZ-5 was obtained by
following method 10 g of H-ZSM-5 was mixed with 300 mL 0.2 M
aq. NaOH in a ask and kept at 338 K for 30 min. Then, this
zeolite sample was subjected to threefold ion exchange with aq.
0.1 M ammonium nitrate (in the proportion of 10 mL gꢀ1 of
product for 5 h at 338 K). Finally, the sample was transformed
into proton form by calcinations in air at 823 K for 5 h.
Fig. 9 Reusability of Hierarchical-HZ-5 catalyst for ethanolysis of FAL
at reaction conditions: molar ratio (ethanol to FAL) of 6 : 1 and catalyst
loading of 30 wt% of FAL, reaction temperature of 413 K and reaction
time of 4 h.
The phase purity of all catalyst samples was conrmed by
powder X-ray diffraction patterns, which were recorded on X-ray
diffractometer (P Analytical PXRD system, Model X-Pert PRO-
1712) using CuKa radiation at a scanning rate of 0.0671/s in
the 2q ranging from 5 to 50ꢂ (Fig. 1). Low temperature (77 K)
nitrogen isotherms (adsorption and desorption) of H-ZSM-5
and Hierarchical-HZ-5 catalysts were recorded with Beckman
Coulter SA 3100 analyzer (CA, USA) (Fig. 2). The calcined sample
was degassed at 573 K for 10 h prior to the measurements. The
specic surface area and pore diameter of synthesized catalysts
was calculated using Brunaer–Emmett–Teller (BET) and Bar-
rettꢀJoynerꢀHalenda (BJH) method respectively (Table 1).
Energy dispersive X-ray analysis (EDAX) was used to determine
the Si/Al ratio of the synthesized catalysts (Table 1). The overall
acidity of synthesized catalysts was measured by Temperature
Programmed Ammonia Desorption (TPAD) using a Micro-
meritics AutoChem (2910, USA) equipped with thermal
conductivity detector (Table 1). The surface morphology of
H-ZSM-5 and Hierarchical-HZ-5 was observed by Field Emission
Scanning Electron Microscope (FE-SEM, HITACHI, Model-
S4800 type II) (Fig. S1, ESI†).
h and 30 wt% of catalyst.14 Use of longer reaction time of 24 h
would lead to increase in the process cost.
The current process of catalytic synthesis of EL biofuel by
ethanolysis of renewable FAL over Hierarchical-HZ-5 catalyst
would be environmentally and industrially benign in perspec-
tive of high catalytic activity (73% EL yield) and high catalyst
stability (3 cycles) (Fig. 9), renewable synthetic route and devoid
of waste byproducts.
3 Conclusions
Hierarchical-HZ-5 zeolite was synthesized by alkali treatment to
H-ZSM-5, which improves the catalyst physicochemical prop-
erties such as surface area, acidity and create mesoporosity in
addition to micropores. This improved catalyst was used for the
one step ethanolysis of renewable FAL to EL biofuel. The reac-
tion parameters were optimized by using RSM design with BBD.
The optimized process parameters were validated with experi-
mental. The proposed RSM statistical model was validated well
with accuracy of 99.45%.
4.3 Catalytic evaluation and product analysis
Hierarchical-HZ-5 zeolite gave maximum catalytic activity of
73% EL yield and 26% EMF yield, with three times catalyst All the experiments were performed under reux in 50 mL two-
reusability. The present process of FAL ethanolysis over necked round bottom ask equipped with condenser, a
Hierarchical-HZ-5 follows green chemistry principles such as magnetic stirrer and thermostatic oil bath. The reaction was
renewable and sustainable routes and minimization of allowed to run for set time (1–5 h) at the set temperature
byproduct formation.
(373–413 K) and aer completion of reaction the ask was
79230 | RSC Adv., 2015, 5, 79224–79231
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
cooled with cold water to stop the reaction. The catalyst from by exploring the three dimensional (3D) response surfaces, two
liquid product mixture was removed by centrifugation. dimensional (2D) contour plots and computing the regression
The liquid reaction feed and product were analyzed by using equation.
GC, Varian-CP-3800, capillary column, SPB-5 (30 m length,
0.25 mm I.D. and 0.25 mm lm thickness) with nitrogen as a
carrier gas and Flame Ignition Detector (FID) in programmable
References
temperature range of 353 to 553 K. The products were quanti-
ed by an external standard method based on the average peak
area of each product under three parallel GC measurements of
each experiment. The concentrations of FAL, EMF and EL in
product mixture were calculated based on the standard curve
obtained using an authentic samples with an analytical error of
ꢁ2%. The reaction products were also conrmed by GC-MS
(Agilent-5977-AMSD). All the experiments were carried out in
duplicate and the average values with an error of ꢁ2% were
reported.
1 P. D. Cara, R. Ciriminna, N. R. Shiju, G. Rothenberg and
M. Pagliaro, ChemSusChem, 2014, 7, 835–840.
2 J. P. Lange, E. van der Heide, J. van Buijtenen and R. Price,
ChemSusChem, 2012, 5, 150–166.
3 C. Chang, G. Xu, W. Zhu, J. Bai and S. Fang, Fuel, 2015, 140,
365–370.
4 G. W. Huber, J. N. Chheda, C. J. Barrett and J. A. Dumesic,
Science, 2005, 308, 1446–1449.
5 Z. Zhang, K. Dong and Z. K. Zhao, ChemSusChem, 2011, 4,
112–118.
6 J. J. Bozell and G. R. Petersen, Green Chem., 2010, 12, 539–
554.
7 A. Demolis, N. Essayem and F. Rataboul, ACS Sustainable
Chem. Eng., 2014, 2, 1338–1352.
4.4 Experimental design with Box–Behnken and
optimization by RSM
RSM with Design-Expert® Version 8.0.7.1 (Stat-Ease, Inc.,
Minneapolis, USA) in combination with Box–Behnken design
(BBD) was used to perform experimental design for ethanolysis
of FAL to EL biofuel over Hierarchical-HZ-5 catalyst. RSM was
developed to obtain optimum process parameters for ethanol-
ysis of FAL in view to maximize the EL yield. Three independent
process variables were employed to design the experiments,
namely molar ratio of ethanol to FAL (X1), percent catalyst
loading (X2) and reaction temperature (X3). The percent yield of
EL (Y) was the response/target parameter of this design.
The 33 Box–Behnken experimental design (BBD) containing
17 set of experimental runs comprising 12 factorial points and
5 center points were performed.19–21 The second-order quadratic
polynomial equation for the EL yield (Y) could be described by
eqn (2):21
8 J. Q. Bond, D. M. Alonso, D. Wang, R. M. West and
J. A. Dumesic, Science, 2010, 327, 1110–1114.
9 G. M. G. Maldonado, R. S. Assary, J. Dumesic and
L. A. Curtiss, Energy Environ. Sci., 2012, 5, 6981–6989.
10 D. Song, S. An, B. Lu, Y. Guo and J. Leng, Appl. Catal., B, 2015,
179, 445–457.
11 G. M. G. Maldonado, R. S. Assary, J. A. Dumesic and
L. A. Curtiss, Energy Environ. Sci., 2012, 5, 8990–8997.
12 J. P. Lange, W. D. van de Graaf and R. J. Haan,
ChemSusChem, 2009, 2, 437–441.
13 G. Wang, Z. Zhang and L. Song, Green Chem., 2014, 16, 1436–
1443.
14 P. Neves, S. Lima, M. Pillinger, S. M. Rocha, J. Rocha and
A. A. Valente, Catal. Today, 2013, 218–219, 76–84.
15 B. Lu, S. An, D. Song, F. Su, X. Yang and Y. Guo, Green Chem.,
2015, 17, 1767–1778.
k
k
kꢀ1
k
X
X
X X
Y ¼ b0 þ
biXi þ
biiXi2
þ
bijXiXj
(2)
i¼1
i¼1
i¼1 j¼2
16 B. Capai and G. Lartigau, US Patent 5,175,358, 1992.
where, Y is the value of predicted response (percentage EL 17 C. C. Hsu and D. W. Chasar, US Patent 4,236,021, 1980.
yield). The Xi and Xj are the uncoded independent variables. The 18 K. Y. Nandiwale, N. D. Galande and V. V. Bokade, RSC Adv.,
terms b0, bi, bii and bij are the intercept coefficient (offset), the
linear effect term, the squared effect term and interaction effect 19 K. Y. Nandiwale and V. V. Bokade, Ind. Eng. Chem. Res., 2014,
term, respectively. The k is total number of independent vari- 53, 18690–18698.
2015, 5, 17109–17116.
ables (in this case, k ¼ 3) used to study inuence on the EL yield. 20 P. Yin, W. Chen, W. Liu, H. Chen, R. Qu, X. Liu, Q. Tang and
Each process variable was coded into levels ꢀ1, 0 and +1 and
Q. Xu, Bioresour. Technol., 2013, 140, 146–151.
21 L. Ma, Y. Han, K. Sun, J. Lu and J. Ding, Energy Convers.
Manage., 2015, 98, 46–53.
shown in Table 2.
The adequacy of each factor of response was checked by
analysis of variance (ANOVA). The central composite rotatable 22 Y. Y. Gao, W. W. Chen, H. Lei, Y. Liu, X. Lin and R. Ruan,
design was employed to obtain second-order regression coeffi- Biomass Bioenergy, 2009, 33, 277–282.
cients (R2). Its signicance of coefficient of regression was 23 K. Y. Nandiwale, S. K. Sonar, P. S. Niphadkar, P. N. Joshi,
evaluated by the value of F-test. Fisher F-test in form of F-value
was employed to investigate tness of experimental model.
S. S. Deshpande, V. S. Patil and V. V. Bokade, Appl. Catal.,
A, 2013, 460–461, 90–98.
The P-value test also used to justify the signicance of 24 M. W. Kasture, P. S. Niphadkar, N. Sharanappa,
process parameters on the response (yield of EL biofuel). The
smallest P-value indicates the signicantly inuencing param-
S. P. Mirajkar, V. V. Bokade and P. N. Joshi, J. Catal., 2004,
227, 375–383.
eter on the response. The optimum process parameters for the 25 V. P. Shiralkar, P. N. Joshi, M. J. Eapen and B. S. Rao, Zeolites,
ethanolysis of FAL to EL over Hierarchical-HZ-5 were obtained
1991, 11, 511–516.
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 79224–79231 | 79231