Ultrasound–assisted biodiesel production using heterogeneous base catalyst and mixed non–edible oils

Accepted Manuscript Ultrasound–Assisted Biodiesel Production Using Heterogeneous Base Catalyst and Mixed Non–edible Oils Ritesh S. Malani, Vivek Shinde, Sumedh Ayachit, Arun Goyal, Vijayanand S. Moholkar PII: DOI: Reference:

S1350-4177(18)31132-5 https://doi.org/10.1016/j.ultsonch.2018.11.021 ULTSON 4390

To appear in:

Ultrasonics Sonochemistry

Received Date: Revised Date: Accepted Date:

25 July 2018 14 November 2018 24 November 2018

Please cite this article as: R.S. Malani, V. Shinde, S. Ayachit, A. Goyal, V.S. Moholkar, Ultrasound–Assisted Biodiesel Production Using Heterogeneous Base Catalyst and Mixed Non–edible Oils, Ultrasonics Sonochemistry (2018), doi: https://doi.org/10.1016/j.ultsonch.2018.11.021

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Ultrasound–Assisted Biodiesel Production Using Heterogeneous Base Catalyst and Mixed Non–edible Oils

1 2 3

Ritesh S. Malani,1 Vivek Shinde,2,* Sumedh Ayachit,2,* Arun Goyal,1,3 Vijayanand S. Moholkar 1,4,#

4 5 6 7

1

Center for Energy, 3 Department of Biosciences and Bioengineering, 4 Department of Chemical Engineering, Indian Institute of Technology Guwahati, Assam – 781 039, India

8 9

2

10

*

11 12

#

University Department of Chemical Technology, Sant Gadge Baba Amravati University, Amravati, Maharashtra – 444 602, India Authors with equal contribution

Author for correspondence E–mail: [email protected], Phone: +91 361 258 2300, Fax: +91 361 258 2291.

13

1

1

Abstract

2

In the present study, the ultrasound–assisted biodiesel production from mixed

3

feedstock of non–edible oils in presence of KI impregnated ZnO as a catalyst in batch reactor

4

was investigated. The production was optimized by using two approaches (1) feedstock

5

optimization and (2) process parameters optimization. Various non–edible oils at optimum

6

volumetric ratio were blended and used as feedstock for transesterification reaction. Biodiesel

7

yield was optimized by Box–Behnken statistical design. The maximum triglyceride

8

conversion of 92.35 ± 1.08% was achieved at optimized conditions of catalyst loading = 7 %

9

(w/w); alcohol/oil molar ratio = 11.68:1 and reaction temperature = 59oC. Transesterification

10

process with mechanical agitation was used as base case for identification of role of

11

sonication in the process. The transesterification process was analysed for kinetic behaviour

12

using pseudo first order kinetics and Eley–Rideal mechanism based model. Overall activation

13

energy of transesterification process for mechanically agitated and ultrasound–assisted

14

systems was calculated as 135.4 and 123.65 kJ/mol, respectively. However, the sum of

15

activation energies of three reaction steps of Eley–Rideal mechanism (64.69 kJ/mol and

16

46.63 kJ/mol, for mechanically agitated and ultrasound–assisted system, respectively) was

17

much lower. This discrepancy is attributed to mass transfer limitations in the system, even in

18

presence of sonication.

19

Keywords: Heterogeneous catalyst, Mixed–oil feedstock, Ultrasound, Biodiesel, Kinetic

20

modelling

2

1

1.

Introduction

2

Depleting petroleum reserves and increasing environmental pollution are the driving

3

forces of quest for alternative fuels derived from renewable sources to meet energy demand

4

for sustainable development. Liquid biofuels such as bioethanol, biobutanol, biodiesel

5

derived from agricultural crops and residues have been studied extensively in recent years by

6

researchers [1]. Among all the biofuels, biodiesel emerged as potential alternative for

7

transportation fuel. Biodiesel is a fatty acid methyl ester, which can be obtained from

8

esterification or transesterification of fatty acid or edible/non–edible oil in presence of

9

acid/base catalyst and alcohol such as methanol or ethanol. Biodiesel has several advantages

10

over petroleum diesel, such as no net carbon contribution to environment, high flash point,

11

complete combustion with low sulfur exhaust and high cetane number. On the other hand,

12

biodiesel has higher viscosity and lesser calorific value as compared with the petroleum

13

diesel [2,3]. Biodiesel blends (5–30%) with conventional diesel can be used directly in diesel

14

engines without significant modifications [4,5]. Conventionally, biodiesel is synthesized by

15

transesterification reaction of vegetable oil using homogeneous base (NaOH/KOH) catalyst

16

[6,7]. However, use of homogenous catalyst results in corrosion of reaction vessel, and also

17

excess water requirement in downstream processing for removal of spent catalyst from

18

biodiesel. Moreover, fraction of catalyst retained in by–product glycerol act as contaminants

19

and lower its merchantability [2,8,9]. Alternate solution to these issues is the use of

20

heterogeneous catalysts for transesterification [10,11]. The heterogeneous catalysts – either

21

basic or acidic in nature – can be easily separated from the reaction mixture for reuse, without

22

contamination of the by–product glycerol. which makes it merchantable by-product which

23

could improve the economics of biodiesel production. Thus, application of heterogeneous

24

catalyst is attractive and viable alternative for large scale biodiesel production in coming

25

years. With proper design of process, the recyclability of catalyst can be further increased.

3

1

Several studies have reported the use of heterogeneous base catalysts synthesized with

2

different methods using silica, zinc oxide, zirconia, zeolites, alumina, aluminosilicates, clays,

3

activated carbon, etc. as supports, and alkali and alkaline earth oxides or their salts (like

4

KOH, KF, KI, KNO3, K2CO3, NaOH, CaO, Ba(OH)2) as functional part [12-24]. ZnO is one

5

of the supports which is stable, non–corrosive, economical and available easily. It is

6

amphoteric in nature, which means it can be transformed into a basic as well as an acidic

7

catalyst by reacting with different promoters such as alkaline earth metals/alkali or sulphates

8

[25]. Reaction systems with heterogeneous catalyst suffer from slower kinetics due to strong

9

mass transfer limitation in 3–phase (solid–liquid–liquid) reaction mixture. The mass transfer

10

barrier could be overcome using intense mixing, increasing temperature or pressure,

11

application of co–solvents etc. [26]. In the present study, the ultrasound irradiation was

12

applied for boosting the reaction kinetics. Ultrasound and cavitation (which is nucleation,

13

growth and implosive collapse of gas/vapour bubble driven by pressure variation induced by

14

ultrasound) have distinct physical and chemical effects on reaction system [27,28]. The

15

physical effect of sonication involves generation of strong micro-convection in the system,

16

while the chemical effect involves generation of highly reactive radicals during transient

17

collapse of the gas/vapour bubbles [28].

18

A major problem for large–scale production of biodiesel is availability of feedstock.

19

Countries like India, edible oils with high price cannot used as feedstock for economic

20

biodiesel production. However, non–edible oils such as Rubber seeds, Mahua, Neem,

21

Karanja, Jatropha, Cassava, Castor etc. can be utilized as alternate low–cost feedstock [29].

22

Availability of a single non–edible feedstock throughout the year is another challenge for

23

viable large scale production of biodiesel. As a possible solution to this issue, in present

24

study, a blend of different non–edible oils was used as feedstock for transesterification

25

reaction using an in–house synthesized heterogeneous base catalyst [30]. This study has dual

4

1

approach, viz. (1) optimization of transesterification process using statistical methods, and (2)

2

mechanistic investigation of the process with kinetic and Arrhenius analysis through

3

mathematical model based on Eley–Rideal mechanism.

4 5

2.

6

2.1

Material and methods Materials and chemicals

7

Zinc oxide (> 99%, AR Grade) used as support material and potassium iodide (> 99%,

8

AR Grade) used as a source of potassium were procured from Himedia Ltd., India. Methanol

9

(99%, ACS grade) and sulfuric acid (conc. 98%, AR grade) were procured from Merck,

10

India. Methanol was distilled to get anhydrous methanol.

11

Crude Jatropha curcus oil and crude castor oil were procured from local farmers. Crude

12

palm oil was procured from local market of Guwahati. Rubber seeds were also procured from

13

local farmers and the oil from the seeds was extracted using Soxhlet extraction method using

14

n–hexane as solvent. Approx. 100 g of dried fine crushed powder (particle size 1–2 mm) of

15

rubber seeds was placed in a thimble covered with blotting paper, and 1.25 L n–hexane was

16

used for oil–extraction. The heating rate was manually controlled so as to complete one

17

extraction cycle in 60–75 min. Each batch of rubber seed powder was subjected to 9 cycles of

18

extraction. Extracted oil was recovered using a rotary vacuum evaporator (Make: Buchi;

19

Model: Rotavapor R–300) for removal (and subsequent recycle) of solvent. Net average oil

20

yield per batch of extraction was 40.7 g (44.24 mL). This yield is comparable with the results

21

reported by Reshad et al. [29].

22

Waste cooking oil was collected from a restaurant on the Institute campus of IIT

23

Guwahati. The waste cooking oil was pre–treated prior to use for biodiesel production by

24

heating up to 180oC and cooling to room temperature. This procedure was carried out twice

25

followed by filtration. The pre–treatment was aimed at removal of moisture and other

5

1

suspended impurities in waste cooking oil. The waste cooking oil mainly consisted of 60.1%

2

of unsaturated fatty acids (26.16% Oleic acid; 25.95% Linoleic acid and 7.99% Linolenic

3

acid) and 39.9% of saturated fatty acid (5.78% Palmitic acid and 34.12% Myristic acid)

4

determined using GC–MS analysis (Make: PerkinElmer, USA; Model: Clarus 680 GC &

5

Clarus 600C MS). The viscosity of oil samples was measured by using rheometer (Make: M/s

6

Thermo Electron, Germany; Model: Rheostress RS 1) at 40oC. Acid value (AV) and

7

saponification value (SV) of all the oils were analysed using titration method and average

8

molecular weight was determined by using the following expression [31]:

9

Average molecular weight = 56.1  3  1000 /( SV  AV ) .

10

The physical properties of individual oils are listed in supplementary material (as Table S.1).

11

2.2

12

Catalyst preparation: Wet impregnation method was used to synthesis the catalyst as

13

described in our earlier work [32]. Briefly, 25 g of ZnO was mixed with 75 mL of 35 %

14

(w/w) KI solution in beaker under constant mixing (Tarson–spinot, Model MC–02) at 150

15

rpm and heating at 100oC for 3 h. The semi–solid impregnate was dried overnight in oven at

16

110oC, followed by calcination in muffle furnace at 500oC for 3 h with temperature rise of

17

10oC/min. The calcined catalyst was cooled and stored in vacuum desiccator for further

18

characterization.

19

Characterization of catalyst: Pristine ZnO and synthesized KI/ZnO catalyst were

20

characterized by X–ray diffraction (XRD) analysis by using a Philips equipment (Model:

21

X’Pert Pro MPD) 2θ range = 10°–80°, step size = 0.01o, scanning speed = 5o min–1 and

22

compared with standard JCPDS data files to identify different phases. The surface elemental

23

compositions of the pristine ZnO and synthesized catalyst were measured by X–ray

24

photoelectron spectroscopy (XPS, Auger Electron Spectroscopy (AES) Module, Model: PHI

25

5000 Versa Probe II, FEI Inc.). Surface morphologies of ZnO and KI/ZnO were analysed by

KI impregnated ZnO catalyst synthesis and characterization

6

1

field emission scanning electron microscopy (FE–SEM, Zeiss, Model: Sigma). The surface

2

area of support ZnO and KI impregnated ZnO catalyst were determined by using N2

3

adsorption–desorption isotherms, using surface area and pore analyzer (Make: Quanta

4

chrome, Model: Autosorb–IQ MP). Prior to measurements, the samples were degassed at

5

150°C for 3 h in vacuum. The specific surface area of the samples was calculated from the

6

adsorption data using Brunauer–Emmett–Teller (BET) method. The Barrett–Joyner–Halenda

7

(BJH) model was used to determine the pore size distribution from adsorption isotherms. The

8

presence of basic sites (K2O) on the catalyst surface was determined using Flame photometric

9

study using Flame Photometer (Make: Systronics; Model: 128). 0.25 g catalyst was mixed

10

with 40 mL of 0.05 M HCl and digested at 80oC for 1 h using hot plate (Remi, 2 MLH). The

11

solution was filtered using Whatman 40 filter paper followed by dilution using deionised

12

water to total volume 200 mL. Standard (10 ppm potassium) sample was prepared using KCl

13

(AR Grade, Himedia, India) for calibration of Flame Photometer [33,34].

14

2.3

Preliminary esterification process

15

All non-edible oils used in present study have high acid values (i.e. AV > 2 mg KOH/g)

16

that restricts direct transesterification with base catalyst. Therefore, the oils were initially

17

subjected to esterification process to reduce the free fatty acid content. The esterification

18

experiments were carried out with conc. H2SO4 [6]. The esterification reaction was carried

19

out for 1 h, at alcohol/oil molar ratio of 15:1 with catalyst concentration of 5% (w/w) at 65oC

20

in a 2–neck round bottom flask (250 mL) fixed with a reflux condenser using mechanical

21

agitation at 400 rpm (Tarson–Spinot, Model MC–02). After completion of reaction, the

22

mixture was separated in two phases (organic and aqueous) using a separating funnel. The

23

traces of acids and methanol were removed from the organic layer with hot water wash. The

24

traces of moisture in the organic layer were removed by passing it over activated silica (230–

25

400 mesh, Merck, India) and Whatmann 40 filter paper. The filtrate was stored in clean and

7

1

air–tight containers. The acid values of esterified oils were determined using titration method.

2 3

2.4

Transesterification process

4

Transesterification experiments were performed in two stages aimed at: (1)

5

optimization of the composition of feedstock, and (2) optimization of process parameters. In

6

the first part, optimization of feedstock composition was done using mixture pseudo–

7

component design with Minitab 16 software (trail version). The limits (volume fraction)

8

selected for different oils were as follows Jatropha oil (0.05; 0.5), Castor oil (0.05; 0.5),

9

Rubber seed oil (0.05; 0.5), Waste cooking oil (0.2; 0.65) and Palm oil (0.2; 0.65). A total 36

10

experimental sets were obtained for different feedstock compositions. Transesterification

11

reaction of these feedstocks was carried out using following parameters: catalyst loading =

12

3% w/w, molar ratio = 9:1, and reaction temperature = 60oC. The feedstock compositions and

13

the corresponding transesterification yields in 36 experimental sets are listed in Table 1.

14

In second part, process optimization was done for the optimum feedstock composition

15

using Box–Behnken statistical design coupled with response surface methodology (RSM).

16

Three process parameters, viz. catalyst loading, molar ratio and reaction temperature were

17

selected for optimization of transesterification yield. The Box–Behnken statistical design

18

consisted of 15 experimental sets with different combinations of individual parameters

19

(further details of levels and combination of parameters in each experimental set are given in

20

Table 2A and B).

21

All transesterification experiments (in both stages of feedstock optimization and

22

process optimization) were carried out for 1 h in batch mode using ultrasound bath (Make:

23

Elma Transsonic, Germany, Model: T–460, total capacity: 2 L, power: 35 W, frequency: 35

24

kHz). The ultrasound bath was filled by distilled water up to 2/3rd of total volume, which

25

worked as the medium for transmission of ultrasound. Transesterification experiments were

8

1

performed in 25 mL two neck round bottom borosilicate flask with a batch size of 15 mL. A

2

coiled condenser was fitted to round bottom flask for refluxing of methanol. Circulation

3

water bath (Jeio Tech, Lab Companion RW 0525G) was used to control the reaction

4

temperature. The power (or pressure amplitude) of ultrasound varies point to point in bath,

5

hence the position of reaction flask was cautiously maintained same in all experiments [35].

6

The pressure amplitude and energy dissipation of the ultrasound wave in the bath was

7

characterized using calorimetry study as described in earlier papers [6,36,37]. The acoustic

8

pressure amplitude generated in the bath was calculated as 150 kPa. The results of

9

calorimetric study show that the actual power transferred to bath was 18.6 W with an energy

10

density of 9.3 W/L, from the theoretical supplied power of 35 W. All experiments were performed in duplicate to confirm the reproducibility.

11 12

Reusability of catalyst: The catalyst was also tested for reusability. The catalyst was

13

recovered after transesterification reaction, by centrifuging the reaction mixture at 6000g for

14

15 min at 25oC. The recovered catalyst was washed three times using 5 mL of n–hexane

15

solvent to remove any impurity from the catalyst surface viz., oil, methanol or glycerol.

16

Catalyst was then dried in oven at 110oC for 1 h followed by calcination at 500oC for 3 h. The

17

regenerated catalyst was used for transesterification process at optimized process conditions.

18

The transesterification experiments were repeated with recycled catalyst till the biodiesel

19

yield reduced to half of fresh catalyst. The regenerated recycled catalyst was analysed for

20

BET, FE–SEM and basicity determination using Flame photometric analysis after the 5th

21

(final) cycle.

22

2.5

Kinetic and Arrhenius analysis

23

To study the mechanistic aspects of ultrasound-induced intensification of

24

transesterification reaction, the experiments were performed in two categories, viz. (1) control

25

(using mechanical agitation), and (2) test (using sonication). Mechanical agitation in control

9

1

experiments was provided at 400 rpm using magnetic stirrer (Tarson–spinot, Model MC–02).

2

The time profiles of reactants and products were determined by withdrawing 500 μL

3

samples of reaction mixture at intervals of 15, 30, 45 and 60 min and analysing them for

4

triglyceride conversion by 1H NMR. To determine the Arrhenius parameters (such as

5

activation energy and frequency factor), kinetic constants of transesterification were obtained

6

by conducting transesterification reaction at temperatures of 54o and 49oC, in addition to the

7

optimum temperature predicted through Box-Behnken method.

8

The Arrhenius analysis was carried out with two approaches, viz. (1) the overall

9

transesterification process and (2) individual reaction steps of triglyceride conversion to

10

FAME and glycerol. In first approach, time profiles of overall triglyceride conversion were

11

fitted to pseudo first order kinetic model to determine overall kinetic constant. The pseudo

12

first order kinetic expression is: ln 1  X   kt , where k is the pseudo first order kinetic

13

constant and X is the conversion of limiting reactant.

14

In the second part, the time profiles of two reactants (triglyceride and methanol) and

15

two products (fatty acid methyl ester (FAME) and glycerol) were fitted to Eley–Rideal

16

kinetic model for transesterification process. This model has been explained in detail in our

17

earlier paper [38]. The time profiles of methanol, FAME and glycerol were obtained

18

stoichiometrically using the overall triglyceride conversion. All 4 profiles were fitted to

19

kinetic expressions of Eley–Rideal model using 4th order Runge–Kutta ordinary differential

20

equation solver coupled with Genetic Algorithm. Finally, Arrhenius parameters (using kinetic

21

expression: k  A exp   Ea RT  ) were obtained by plotting ln k versus 1/T.

22

Analytical method: The gross conversion of triglycerides to biodiesel was determined by 1H

23

NMR (Nuclear Magnetic Resonance) spectroscopy (Bruker Advance III HD Ascend 600

24

MHz) using TMS (tetramethylsilane) as internal standard and CDCl3 (Merck, India) as a

25

solvent [6]. The following equation was used for calculating gross molar conversion of

10

1

triglycerides to biodiesel [39,40]:

2

X   2  AME   100 3  A CH 2 , where, AME = integration value of methyl esters protons,

3

the strong singlet peak at 3.6 ppm, A CH 2 = integration value of methylene protons at 2.3

4

ppm.

5

3.





Eley–Rideal kinetic model

6

Transesterification reaction with heterogeneous catalyst involves adsorption of one or

7

more reactants on the surface/ active sites of the catalyst followed by reaction with other

8

reactant. Different mechanisms have been proposed to model heterogeneously catalysed

9

reactions, one of which is the Eley–Rideal mechanism. In this mechanism, one of the

10

reactants from gas/liquid phase will get absorbed on the surface/ active site of the catalyst and

11

other reactant will react directly from the bulk. The reaction occurs on the catalyst surface/

12

active sites resulting in product in adsorbed form. The product subsequently desorbs from the

13

surface/active site, making it free for continuation of the reaction. The Eley–Rideal

14

mechanism treats each of the above steps as independent and rate–determining step.

15

Transesterification reactions are usually carried out with alcohol/oil molar ratio higher

16

than the stoichiometric, so as to force greater conversion of triglycerides. In present study, it

17

was assumed that methanol (excess reactant) preferentially adsorbs on the surface/ active

18

sites of the catalyst. The triglyceride molecule reacts from the bulk to form diglyceride

19

molecule (in absorbed form) and an ester (FAME or biodiesel) molecule. The diglyceride

20

molecule later desorbed from the catalyst surface – to adsorb the next methanol molecule.

21

Similarly, the diglyceide molecule reacts with the next adsorbed methanol molecule to form

22

monoglyceride molecule (in adsorbed form) and ester molecule. In same way the

23

monoglyeride molecule react with third adsorbed methanol molecule resulting in formation

24

of final (last) ester molecule and the by–product glycerol molecule. Thus, the overall

25

transesterification process has three sub–processes viz. adsorption process, chemical reaction 11

1

(three reaction steps) and desorption process. The overall transesterification reaction is

2

written as below:

3 4

 3F  G  T  3CH 3OH   where, T – triglyceride, F – fatty acid methyl ester (or biodiesel) and G – glycerol.

5

The individual steps of transesterification process are based on the Eley–Rideal

6

mechanism (as discussed above), and their corresponding forward reaction rate expression

7

are given in Table 3. Certain assumptions were made in deriving the final rate expressions

8

and to solve them. The assumptions were as follows: (1) Catalyst surface was assumed to be

9

homogeneous without any contamination or inert species and the active sites were distributed

10

uniformly over the surface of catalyst. (2) For adsorbed species, quasi–steady state

11

assumption is made [41].

12

The final rate expression of transesterification reaction, i.e. (rate of consumption of

13

triglyceride, diglyceride, monoglyceride and methanol and rate of formation of FAME or

14

biodiesel and glycerol) was derived and discussed in detail in our earlier paper [38]. In

15

present study, the final rate expressions, are listed in Table 4.

16

Fitting of experimental data to kinetic model: To solve the rate expression and to get the

17

kinetic constants, four ordinary differential equations (ODE), i.e. rate expressions of

18

consumption of triglyceride and methanol, and rate expression of formation FAME and

19

glycerol, were solved simultaneously in MATLAB R2016b using 4th order Runge–Kutta

20

method coupled to GA (Genetic Algorithm) solver, as initial value problem (IVP). The four

21

differential equations consisted of 7 unknown parameters or kinetic rate constants, viz. k1, k2,

22

k3, k4, k5, k6 and k7. The upper and lower limits for each kinetic rate constant were specified in

23

GA solver on the basis of the results of Malani et al. [38] and Kapil et al. [41] The GA solver

24

gives the values of model parameters or kinetic rate constants within the specified limits.

25

These values were used by ODE solver to generate numerical solution for time profiles of

12

1

triglyceride, methanol, glycerol and FAME. The simulated profile was compared with the

2

experimental profile, and the values of seven kinetic rate constants (k1 to k7) were adjusted till

3

objective function reached minimum. The objective function (Obj) used for optimization and

4

minimizing the error was defined as [38]: Obj  min

5

experimental data points for concentrations of four species. The error (er) function was

6

defined as:

7

2 2 2 2 eri  Ti exp  Ti model    CH 3OH iexp  CH 3OH imodel    Fi exp  Fi model    Giexp  Gimodel    

 er  , where n is the number of n

i 1

i

1/ 2

8 9

4.

Results and discussion

10

4.1

11

XRD analysis: Pristine ZnO and KI/ZnO catalyst were characterised by X–Ray Diffraction to

12

identify the different phases. The X–ray diffractograms of pristine ZnO and KI/ZnO catalyst

13

have been reported in a previous study [32] and also provided in supplementary material (as

14

Fig S.1). Diffractogram of ZnO shows the different peaks at 2θ = 31.8o 1 0 0; 34.4o 0 0 2;

15

36.2o 1 0 1; 47.5o 1 0 2; 56.6o 1 0 1; 62.8o 1 0 3; 66.3o 1 1 2; 67.8o 1 1 2 and 69.1o

16

1 1 2 (as per standard JCPDS file no. 36–1451). The diffractogram of KI impregnated ZnO

17

shows the additional peaks at 2θ = 24.3o, 29.4o, 41.8o, and 77o which confirmed the presence

18

of K2O phase (as per JCPDS file no. 77–2176). K2O phase on catalyst surface, acts as active

19

sites for the transesterification reaction and imparts basicity to the catalyst [12,30].

20

XPS analysis: The XPS spectrum of KI/ZnO is shown in Fig. 1, which confirmed the

21

presence of Zn, O and K on the catalyst surface. The two peaks corresponding to binding

22

energies 291.55 eV and 295.15 eV in K 2p spectrum correlates with the presence of

23

potassium in +1 state. The two high intense peaks at 1022.35 eV and 1048.25 eV in Zn 2p

24

spectrum attributed to the presence of elemental Zn in +2 state. The binding energy of O 1s

Characterization of catalyst

13

1

for KI/ZnO catalyst in present study was found to be 533.90 eV, which exactly matches with

2

the reported values by Yadav et al. [23]. However, Yadav et al. [23] have also reported

3

binding energy of O 1s in pristine ZnO as 532.94 eV. Higher binding energy of O 1s in

4

catalyst, as compared to O 1s in pure ZnO could be attributed to distortion of ZnO lattice due

5

to introduction of K ions.

6

FE–SEM analysis: Fig. 2 shows the FE–SEM micrographs of support ZnO and KI/ZnO

7

catalyst particles. FE–SEM micrograph of pristine ZnO material shows the heterogeneous

8

cylindrical shaped particles in size range 100 to 500 nm. On the other hand, KI impregnated

9

ZnO catalyst micrograph shows larger particles with heterogeneity in shape and size as

10

compared to pristine ZnO. The size of KI/ZnO catalyst particles ranges from 300 to 700 nm

11

with agglomeration of particles. The agglomerated particles are in the range 1–2 µm.

12

Surface area and pore size analysis: Nitrogen adsorption–desorption isotherms for pure ZnO

13

and KI/ZnO catalyst are shown in Fig. 3. Pure ZnO and synthesized ZnO catalyst follow type

14

II isotherm, which predicts the non–porous structure of material. The average pore diameter

15

of pure ZnO and KI impregnated ZnO were 1.87 nm and 2.12 nm, respectively. The BET

16

surface areas of pure ZnO and KI impregnated ZnO were 6.327 m2/g and 3.853 m2/g,

17

respectively. The sharp reduction in surface area of synthesized catalyst was mainly due to

18

covering and agglomeration of KI over ZnO fine pores.

19

Flame photometric analysis: The basicity of synthesized catalyst was determined using

20

flame photometer. The results showed that the synthesized KI impregnated ZnO catalyst had

21

195.65 mg K/g catalyst.

22

4.2

Preliminary esterification experiments

23

Esterification of all non-edible oils with homogeneous acid catalyst, resulted in

24

average 80–90% reduction of acid value. These results are in line with our earlier studies [6].

25

The reduction in acid values and viscosities of different oils after esterification are listed in

14

1

Table S.2 provided in supplementary material. From Table S.2, it can be noted that the

2

highest reduction in acid values occurs in Jatropha and Rubber seed oil, but final values

3

achieved are still not within the limits and thus restricts the use of either (Jatropha or Rubber

4

seed) oil for biodiesel production with base catalyst in individual form. The least reduction in

5

acid value occurs in Castor oil, this may be due to its high viscosity, which restricts the

6

uniform mixing of oil, methanol and acid catalyst.

7

4.3

8

Feedstock optimization: The pseudo–component mixture design with 36 experimental sets

9

and corresponding triglyceride conversion with standard deviation is shown in Table 1. In

10

feedstock optimization the esterified oils are blended in different volumetric propositions.

11

From Table 1, it can be seen that, no single oil feedstock has significant dominance in blends.

12

The highest conversion of triglyceride was obtained with experimental set 11, followed by set

13

1 and set 10. Some major observations from these experiments are as follows:

Transesterification experiments

14

1. As the volume of Castor oil increases above 20% in the mixed feedstock, the

15

transesterification yield reduces significantly. The decrease in transesterification yield

16

attributed to high viscosity of castor oil, which increases the overall feedstock viscosity. This

17

is manifested in terms of mass transfer limitations between oil and methanol phase. This

18

result also suggests that in order to enhance the transesterification yield, volume fraction of

19

castor oil in mixed feedstock should be kept minimum (< 0.1 v/v).

20

2. Blending of Jatropha oil at moderate to higher volumetric ratio (> 0.275 v/v) has

21

adverse impact on transesterification yield. Reduction in transesterification yield is mainly

22

due to its high acid value even after esterification reaction. Thus low to moderate (< 0.275

23

v/v) fraction of Jatropha oil to should be blended to enhance the transesterification yield.

24

3. Waste cooking oil in mixed oil feedstock shows mixed response on transesterification

25

yield. In some experimental sets, moderate or higher volume fraction of waste cooking oil

15

1

shows good transesterification yield, while in other experimental sets it resulted in lower

2

transesterification yields. This mixed reaction depends on the volume fractions of other oils

3

especially volume fraction of Castor, Jatropha and Palm oil. Hence, the moderate volume

4

fraction (0.1 to 0.25 v/v) of waste cooking oil is favourable for higher yield of

5

transesterification reaction.

6

4. Blending of Rubber seed oil to mixed oil feedstock improved the yield of

7

transesterification reaction. This could be attributed to its relatively low viscosity, which

8

helps in mixing of other oils and methanol with each other. The highest yield of

9

transesterification was obtained at 50% volume of rubber seed oil in feedstock blend.

10

5. Addition of Palm oil to mixture feedstock showed effect as that of waste cooking oil,

11

i.e. for some experimental sets with higher volume fraction of palm oil, enhanced the

12

transesterification yield and for some experimental sets, it reduced the yield. Palm oil has

13

second highest viscosity after Castor oil among all oils used in feedstock. On the other hand,

14

the acid value of esterified palm oil is low (0.95 mg KOH/g), which helps promoting the

15

transesterification reaction. However, for volume fraction of palm oil > 0.2, overall viscosity

16

of feedstock increases leading to lower mass transfer between oil and methanol. Therefore,

17

moderate volume fraction (0.1 to 0.2 v/v) of Palm oil in blended feedstock results higher

18

transesterification yield.

19

On the basis of above observations and experimental results, the optimum volume

20

fractions of different oils in mixed oil feedstock were selected as experimental set 11. The

21

experimental set 1 and 10 are also good alternatives for large scale transesterification process

22

in case of non–availability of any oil feedstock.

23

Process optimization: After initial screening of different oil mixtures for transesterification

24

reaction, an attempt was made to maximize the yield of transesterification reaction by varying

25

the operating parameters such as alcohol: oil molar ratio, operating temperature of reaction

16

1

and addition of catalyst to reaction mixture. The process optimization was done using Box–

2

Behnken statistical design coupled with response surface methodology. Statistical

3

optimization helps in understanding the interaction between two operating parameters.

4

The experimental transesterification yield and the model predicted yield for 15

5

experimental sets of Box–Behnken statistical design are tabulated in Table 2B. The

6

experimental yield consisted of the average of two runs and corresponding standard

7

deviation, as the experiment for each set was carried out in duplicate to validate

8

reproducibility of results. A quadratic equation was fitted to the experimental data using

9

coded values of process parameters as follows:

10 11

Y  86.34  8.49C  8.68M  14.46T  21.50C 2  16.66 M 2  22.54T 2  6.65C  M  9.08T  C  0.64 M  T

12

From the Table 2B, the experimental results showed a close match with model–predicted

13

values of triglyceride conversion, indicating that the model prediction matches well to the

14

experimental results. The ANOVA (analysis of variance) of the fitted model predicted

15

various coefficients of the quadratic model such as linear, square and interaction coefficients.

16

p– and t– values of these coefficients are provided in supplementary material (as Table S.3.).

17

The regression coefficients values, viz. R2 = 0.9998; R2 (predicted) = 0.9979; R2 (adjusted)

18

=0.9995 supports the validity of quadratic model selected for parameter optimization. The

19

ANOVA analysis predicted coefficients, large (absolute) t–stat value and p–value < 0.05

20

indicated the consequence of the coefficient and thus, corresponding process parameter. F–

21

values related to coefficients of linear, interaction and quadratic variables, indicated the

22

importance of the individual effect of corresponding optimization variable and the magnitude

23

of interaction between them. From ANOVA results the p–values of all coefficients are <

24

0.05, which indicates all the process parameters have significant impact on process

25

optimization. Based on p–values of interactions, the interaction between catalyst loading and

26

reaction temperature is most significant followed by the interaction between molar ratio and

17

1

catalyst, while the temperature and molar ratio showed the least interaction. The F–value and

2

p–value of Lack–of–Fit were 2.01 and 0.351 respectively, which denotes that Lack–of–Fit is

3

not significant with compared to pure error or in other way, the model is significant. The

4

response surface plots for the quadratic model are provided in supplementary material (as

5

Fig. S.2), which indicated the interaction between two process parameters on

6

transesterification yield, with retaining the third parameter at its centre point. The surface

7

plots are graphical presentations of quadratic equation. The top (dark) colour region displays

8

to the maximum transesterification yield. The surface plots indicated the strong interaction

9

between the parameters as same predicted by ANOVA analysis.

10

The optimum set of parameters predicted by the quadratic model was: catalyst loading

11

= 7% (w/w); molar ratio (alcohol/oil) = 11.68:1; temperature = 59oC. The validation

12

experiment was performed at optimum set of parameters. Triglyceride conversion in the

13

validation experiment was 92.35 ± 1.08%, which is very close to the triglyceride conversion

14

of 92.03% predicted by the model.

15

4.4

Kinetic modelling and Arrhenius analysis

16

Batch experiments were conducted at optimum conditions obtained from Box–

17

Behnken statistical optimization design to determine the kinetic parameters or rate constants.

18

Fig. 4 shows the representative 1H NMR spectrum of triglyceride conversion in control and

19

test experiments conducted at optimized conditions. The kinetic rate constants were

20

determined using pseudo first order kinetics (overall process) and Eley–Rideal model

21

(individual reaction steps) for both control and test conditions and are summarised in Table 5.

22

The experimental profiles of FAME or biodiesel yield were fitted to Eley–Rideal model as

23

shown in Fig. 5. As mentioned earlier, for Arrhenius analysis (activation energy

24

determination), the transesterification reaction (both test and control) were conducted at two

25

different temperatures (other than optimum temperature) 54o and 49oC with same molar ratio

18

1

(11.68:1) and catalyst loading (7% w/w). For calculating the activation energy of overall

2

transesterification process, the kinetic constants determined at three reaction temperatures

3

were used, whereas kinetic constants of three individual reaction steps were used for

4

determining the activation energy of these steps. All rate constants are tabulated in Tables 6A

5

and B. Activation energies calculated using Arrhenius plots (ln k vs 1/T, shown in Fig. 6) are

6

listed in Table 6A and B. The kinetic and Arrhenius analyses showed some interesting trends

7

and revealed facts of transesterification reaction and the role of ultrasound in

8

transesterification reaction. These findings are as follows:

9



The simulated FAME or biodiesel yield profile was compared with the experimental data.

10

A good match between the simulated results and experimental data could be seen from

11

Fig. 6, which validated the selected kinetic model.

12



It could be seen from Table 5, that the rate constants k2 – k7 showed 25–50%

13

enhancement, as mechanical agitation is replaced by ultrasound, while the rate constant k1

14

shows ~ 50% reduction when mechanical agitation is replaced by ultrasound.

15



Among all the rate constants the value of k1 is ~ 10–100 times smaller than other rate

16

constants, which indicated that the adsorption of methanol on the catalyst surface is the

17

slowest or rate limiting step in transesterification process.

18



The sum of activation energies of three individual reaction steps (r2, r3, r4) was

19

significantly smaller than the activation energy of overall transesterification process in

20

both test and control conditions. For the test experiments, the total activation energy for

21

three reactions steps was 46.63 kJ/mol, with overall activation energy of

22

transesterification was 123.65 kJ/mol. While for the control experiments, the total

23

activation energy for three reactions steps was 64.69 kJ/mol, with overall activation

24

energy of transesterification was 135.40 kJ/mol.

25



The activation energies of three individual reaction steps showed the trend: r2 > r3 > r4,

19

1

for both the test and control conditions, which means that successive transformation of

2

tri–, di– and mono– glycerides to biodiesel requires less activation energy.

3



A marginal reduction (~ 9%) in overall activation energy of transesterification process

4

was observed with application of ultrasound, which is in concurrence with observation of

5

Choudhury et al [10]. But, the reduction in activation energies of three individual reaction

6

steps is remarkable (~ 30%), when mechanical agitation was replaced with ultrasound as

7

listed in Table 6.

8 9

Plausible explanations to these trends, which may help in identifying the mechanistic role of ultrasound in transesterification reaction, are as follows:

10

1. The lowest value of methanol adsorption rate constant (k1) results in less adsorption of

11

methanol on catalyst active sites/surface. The probable reason is high reaction

12

temperature, which is close to boiling point of methanol, and does not favour adsorption

13

process. The rate of adsorption of methanol in test experiments is further reduced, due to

14

the shock waves generated by transient cavitation that cause desorption of adsorbed

15

species [36,42,43]. These phenomena are essentially manifested in terms of reduction in

16

value of k1. Moreover, this speculation can also be proven from the increase in the values

17

of kinetic rate constants k5, k6, and k7, in test experiments when compared to the control

18

experiments. The desorption rate of intermediates and by–product glycerol enhanced

19

significantly in the test experiments in presence of sonication.

20

2. The enhancement in rate constants of individual reaction steps with application of

21

sonication could be attributed to intense mixing between the two phases (oil and

22

methanol). The micro-convection generated by sonication causes fine emulsification of

23

the phases and reduction in the interfacial tension and consequently, the activation

24

energies.

25

3. The reduction in activation energies of three individual reaction steps (i.e. r2 > r3 > r4) is

20

1

again a mass transfer effect. Successive conversion of triglyceride to diglyceride, results

2

in lower interfacial tension and enhanced mixing with methanol phase. Similarly, this

3

effect is more prominent with conversion of diglyceride to monoglcyeride. This effect

4

was also demonstrated by Bhoi et al. [44]. Thus, the activation energies of the individual

5

reaction steps decreased from r2 to r4.

6

4. The marginal reduction in activation energy of overall transesterification process in

7

presence of sonication demonstrates influence of mass transfer in transesterification

8

reaction with heterogeneous catalyst. This result is further corroborated by comparing the

9

difference between the sum of activation energies of three individual reaction steps and

10

the overall activation energy in both control and test experiments. For mechanically

11

agitated system, the overall activation energy was 135.40 kJ/mol, whereas for ultrasound-

12

assisted system the overall activation energy was 123.65 kJ/mol. However, the sum of

13

activation energies for three reactions steps in control and test experiments were 64.69

14

and 46.63 kJ/mol, respectively. The discrepancy between the overall activation energy

15

and sum of activation energies of three reaction steps is attributed to mass transfer

16

limitation of the process. Notably, this discrepancy is higher for ultrasound-assisted

17

experiments, which demonstrates greater control of mass transfer limitation on the

18

transesterification reaction.

19

4.5

Reusability of the catalyst

20

The KI/ZnO catalyst was also tested for reusability at the end of each cycle of

21

transesterification reaction by separation from the reaction mixture through centrifuging the

22

reaction mixture followed by filtration through Whatmann filter paper. This procedure

23

assured that there are no traces of solid residue of catalyst present in biodiesel. The activity of

24

the catalyst was analysed on the basis of triglyceride conversion in successive reaction cycles

25

as compared with fresh catalyst. The results of these experiments are given in Fig. 7. It could

21

1

be inferred from Fig. 7 that KI/ZnO catalyst retained ~ 50% of its initial activity after 5

2

successive reaction cycles. The reused/ recycled catalyst (after 5th cycle) was analysed for

3

BET and FE–SEM analysis to determine changes in surface morphology. The BET surface

4

area of reused catalyst was found to be 2.426 m2/g, lower as compared to fresh catalyst. The

5

FE–SEM micrograph of recycled catalyst (Fig. S.3 in supplementary material) shows that the

6

catalyst had morphology similar to pure KI/ZnO catalyst with higher agglomeration of

7

particles. The recycled catalyst was also tested for Flame photometric analysis. The results

8

indicated that recycled catalyst had 127.68 mg K/g catalyst after 5 cycles of reuse. The

9

reduction in potassium from catalyst was the major reason for loss of catalytic activity of

10

catalyst. A probable cause of loss of catalyst activity is presence of impurities (triglyceride or

11

glycerol) on catalyst surface that block the active sites [13].

12

4.6

Biodiesel properties

13

The biodiesel synthesised in present study was also tested for various fuel properties,

14

viz., density, flash point, kinematic viscosity, cloud point, calorific value, oxidation stability

15

of the biodiesel using standard methods. Total conversion of triglycerides in present study

16

(~93%) is slightly smaller than the specified standards (conversion < 98% with limit of free

17

and total glycerine as 0.02% and 0.025%). The probable two causes leading to this effect: (1)

18

mass transfer limitations and relatively slow kinetics of transesterification reaction, and (2)

19

the mixture of non-edible oil, which after esterification has the acid value of 1.5 mg KOH/g

20

and that cannot be converted to methyl ester in presence of base catalyst. But, the results of

21

biodiesel properties (viz. density, flash point, kinematic viscosity, cloud point, calorific value,

22

oxidation stability) listed in Table 7 shows that these properties are as per standard norms of

23

ASTM D 6751 and EN 14214 for commercial application of biodiesel.

24 25

5.

Conclusion

22

1

The present study has demonstrated feasibility of mixed non–edible oil feedstock for

2

biodiesel synthesis with heterogeneous base catalyst. This process could be intensified with

3

application of ultrasound. Activation energy of the transesterification showed significant

4

reduction with application of ultrasound with concurrent rise in kinetics. However, the kinetic

5

analysis of the transesterification process using Eley–Rideal model has revealed strong mass

6

transfer influence on the process even with application of sonication. Methanol adsorption on

7

the solid catalyst has been revealed to be the rate limiting step of overall process. Moreover,

8

the synthesized KI/ZnO catalyst after regeneration was reusable up to 5 cycles. The physical

9

and fuel properties of biodiesel produced at optimized conditions were at par with ASTM and

10

EN standards. The kinetic model and analysis described in this study can be extended to other

11

similar heterogeneously catalysed reactions as well as transesterification process employing

12

different catalysts and feedstocks.

13 14 15

Conflicts of interest

16

There are no conflicts of interest to declare.

17 18

Acknowledgement

19

Authors acknowledge the analytical facilities, the surface area and pore analyser, FE–

20

SEM and NMR at Central Instrumental Facility, Indian Institute of Technology Guwahati.

21

Rheometer provided by Analytical Laboratory of Chemical Engineering Department, Indian

22

Institute of Technology Guwahati are gratefully acknowledged. Authors also acknowledge

23

the XPS facility provided by Surface Characterization Lab, Advanced Center for Material

24

Science, Indian Institute of Technology Kanpur, XRD facility provided by SAIF, Gauhati

25

University, Assam and GC–MS analysis provided by Biotech Park, Guwahati, Assam.

23

1

Authors are thankful to anonymous referees for their meticulous evaluation and constructive

2

criticism of the manuscript.

3 4

Supplementary Material

5

Supplementary material for this work can be found with e–version of this paper online.

6 7

References

8

1

V. S. Moholkar, H. A. Choudhury, S. Singh, S. Khanna, A. Ranjan, S. Chakma, J. B.

9

Bhasarkar, Physical and chemical mechanisms of ultrasound in biofuel synthesis, in:

10

Z. Fang, R. L. Smith, X. Qi (Eds.), Production of Biofuels and Chemical with

11

Ultrasound, Biofuels and Biorefineries series, Springer Science + Business Media,

12

Dordrecht, 2015, vol. 4, pp. 35–86.

13

2

I. Shancita, H. H. Masjuki, M. A. Kalam, S. S. Reham, S. A. Shahir, Comparative

14

Analysis on Property Improvement Using Fourier Transform Infrared Spectroscopy

15

(FT-IR) and Nuclear Magnetic Resonance (NMR) (1H and

16

Biodiesel Blended Fuels, Energy Fuels 30(6) (2016) 4790–805.

17

3

13C)

Spectra of Various

S. P. Singh, D. Singh, Biodiesel production through the use of different sources and

18

characterization of oils and their esters as the substitute of diesel: a review, Renew.

19

Sust. Energ. Rev. 14(1) (2010) 200–216.

20

4

M. Mofijur, A. E. Atabani, H. A. Masjuki, M. A. Kalam, B. M. Masum, A study on

21

the effects of promising edible and non-edible biodiesel feedstocks on engine

22

performance and emissions production: a comparative evaluation, Renew. Sust.

23

Energ. Rev. 23 (2013) 391-404.

24 25

5

N. N. A. N. Yusuf, S.K. Kamarudin, Z. Yaakub, Overview on the current trends in biodiesel production, Energy Convers. Manage. 52(7) (2011) 2741–2751.

24

1

6

H. A. Choudhury, R. S. Malani, V. S. Moholkar, Acid catalyzed biodiesel synthesis

2

from Jatropha oil: mechanistic aspects of ultrasonic intensification, Chem. Eng. J. 231

3

(2013) 262–272.

4

7

(1999) 1-15.

5 6

F. Ma, M. A. Hanna, Biodiesel production: a review, Bioresour. Technol. 70(1)

8

L. Bournay, D. Casanave, B. Delfort, G. Hillion, J. A. Chodorge, New heterogeneous

7

process for biodiesel production: a way to improve the quality and the value of the

8

crude glycerin produced by biodiesel plants, Catal. Today 106(1–4) (2005) 190–192.

9

9

T. Sakai, A. Kawashima, T. Koshikawa, Economic assessment of batch biodiesel

10

production processes using homogeneous and heterogeneous alkali catalysts,

11

Bioresour. Technol. 100(13) (2009) 3268–3276.

12

10

H. A. Choudhury, P. P. Goswami, R. S. Malani, V. S. Moholkar, Ultrasonic biodiesel

13

synthesis from crude Jatropha curcas oil with heterogeneous base catalyst:

14

mechanistic insight and statistical optimization, Ultrason. Sonochem. 21(3) (2014)

15

1050–1064.

16

11

E. F. Aransiola, T. V. Ojumu, O. O. Oyekola, T. F. Madzimbamuto, D. I. O. Ikhu–

17

Omoregbe, A review of current technology for biodiesel production: State of the art

18

Biomass Bioenergy 61 (2014) 276–297.

19

12

KF/ZnO catalyst, Catal. Lett. 107(1–2) (2006) 53–59.

20 21

W. Xie, X. Huang, Synthesis of biodiesel from soybean oil using heterogeneous

13

A. S. Chouhan, A. K. Sarma, Modern heterogeneous catalysts for biodiesel

22

production: A comprehensive review. Renew. Sust. Energ. Rev. 15(9) (2011) 4378-

23

4399.

25

1

14

A. P. Vyas, N. Subrahmanyam, P. A. Patel, Production of biodiesel through

2

transesterification of Jatropha oil using KNO3/Al2O3 solid catalyst, Fuel 88(4) (2009)

3

625–628.

4

15

K2CO3/Al-O-Si aerogel catalysts. J. Serb. Chem. Soc. 75(6) (2010) 789-801.

5 6

I. Lukić, J. Krstić, S. Glišić, D. Jovanović, D. Skala, Biodiesel synthesis using

16

S. Baroutian, M. K. Aroua, A. A. Raman, N. M. Sulaiman, Potassium hydroxide

7

catalyst supported on palm shell activated carbon for transesterification of palm oil,

8

Fuel Process. Technol. 91(11) (2010) 1378–1385.

9

17

transesterification of palm oil to biodiesel, Appl. Clay Sci. 53(2) (2011) 341–346.

10 11

F. E. Soetaredjo, A. Ayucitra, S. Ismadji, A. L. Maukar, KOH/bentonite catalysts for

18

J. X. Wang, K. T. Chen, J. S. Wu, P. H. Wang, S. T. Huang, C. C. Chen, Production

12

of biodiesel through transesterification of soybean oil using lithium orthosilicate solid

13

catalyst, Fuel Process. Technol. 104 (2012) 167–1673.

14

19

M. Takase, M. Zhang, W. Feng, Y. Chen, T. Zhao, S. J. Cobbina, L. Yang, X. Wu,

15

Application of zirconia modified with KOH as heterogeneous solid base catalyst to

16

new non-edible oil for biodiesel, Energy Convers. Manage. 80 (2014) 117–125.

17

20

G. Tao, Z. Hua, Z. Gao, Y. Zhu, Y. Chen, Z. Shu, L. Zhang, J. Shi, KF-loaded

18

mesoporous Mg–Fe bi-metal oxides: high performance transesterification catalysts for

19

biodiesel production, Chem. Commun. 49(73) (2013) 8006–8008.

20

21

Y. M. Dai, K. T. Chen, P. H. Wang, C. C. Chen, Solid-base catalysts for biodiesel

21

production by using silica in agricultural wastes and lithium carbonate, Adv. Powder

22

Technol. 27(6) (2016) 2432–2438.

23

22

H. Liu, H. shuang Guo, X. jing Wang, J. zhong Jiang, H. Lin, S. Han, S. peng Pei,

24

Mixed and ground KBr-impregnated calcined snail shell and kaolin as solid base

25

catalysts for biodiesel production, Renew. Energ. 93 (2016) 648–657.

26

1

23

M. Yadav, V. Singh, Y. C. Sharma, Methyl transesterification of waste cooking oil

2

using a laboratory synthesized reusable heterogeneous base catalyst: Process

3

optimization and homogeneity study of catalyst, Energy Convers. Manage.148 (2017)

4

1438–1452.

5

24

M. C. Albuquerque, I. Jiménez–Urbistondo, J. Santamaría–González, J. M. Mérida–

6

Robles, R. Moreno–Tost, E. Rodríguez–Castellón, A. Jiménez–López, D. C.

7

Azevedo, C. L. Cavalcante Jr, P. Maireles–Torres, CaO supported on mesoporous

8

silicas as basic catalysts for transesterification reactions, Appl. Catal., A 334(1–2)

9

(2008) 35–43.

10

25

A. C. Alba–Rubio, J. Santamaría–González, J. M. Mérida–Robles, R. Moreno–Tost,

11

D.

12

transesterification processes by using CaO supported on zinc oxide as basic catalysts

13

Catal. Today 149(3–4) (2010) 281–287.

14

26

Martín–Alonso,

A.

Jiménez–López,

P.

Maireles–Torres,

Heterogeneous

P. D. Luu, N. Takenaka, B. van Luu, L. N. Pham, K. Imamura, Y. Maeda, Co-solvent

15

method produce biodiesel form waste cooking oil with small pilot plant, Energy

16

Procedia, 61 (2014) 2822–2832.

17

27

synthesis: review and analysis, RSC Adv. 6(70) (2016) 65541–66562.

18 19

A. Ranjan, S. Singh, R. S. Malani, V. S. Moholkar, Ultrasound-assisted bioalcohol

28

R. S. Malani, A. Goyal, V. S. Moholkar, Ultrasound-Assisted Biodiesel Synthesis: A

20

Mechanistic Insight, in: A. K. Agrawal, R. A. Agarwal, T. Gupta, B. R. Gurjar (Eds.),

21

Biofuels, Springer, Singapore, 2017, pp. 103–135.

22

29

application: Optimization of parameters, Fuel 150 (2015) 636–644.

23 24 25

A. S. Reshad, P. Tiwari, V. V. Goud, Extraction of oil from rubber seeds for biodiesel

30

W. Xie, H. Li, Alumina-supported potassium iodide as a heterogeneous catalyst for biodiesel production from soybean oil, J. Mol. Catal. A: Chem. 255(1–2) (2006) 1–9.

27

1

31

biodiesel by a two-step process, Energy Convers. Manage. 51(12) (2010) 2802–2807.

2 3

X. Deng, Z. Fang, Y. H. Liu, Ultrasonic transesterification of Jatropha curcas L. oil to

32

R. S. Malani, S. Singh, A. Goyal, V. S. Moholkar, Ultrasound-Assisted Biodiesel

4

Production Using KI-Impregnated Zinc Oxide (ZnO) as Heterogeneous Catalyst: A

5

Mechanistic Approach, in: S. Kumar, R. K. Sani, Y. K. Yadav (Eds.) Conference

6

Proceedings of the Second International Conference on Recent Advances in

7

Bioenergy Research, Springer, Singapore, 2018, pp. 67–81.

8

33

materials used for potassium argon dating, Geochim. Cosmochim. Acta 27(5) (1963)

9

525–546.

10 11

J. A. Cooper, The flame photometric determination of potassium in geological

34

DOTD

Designation

531–11,

Available

online

at

http://wwwsp.dotd.la.gov/

12

Inside_LaDOTD/Divisions/Engineering/Materials_Lab/TPM_Vol_II_Part_V/tr_531–

13

11.pdf (Access on 11–06–18).

14

35

Plenum Press, New York, 1999.

15 16

36

S. Chakma, V. S. Moholkar, Mechanistic features of ultrasonic desorption of aromatic pollutants. Chem. Eng. J. 175 (2011) 356-367.

17 18

Y. T. Shah YT, A. B. Pandit, V. S. Moholkar, Cavitation Reaction Engineering,

37

H. A. Choudhury, S. Chakma, V. S. Moholkar, Mechanistic insight into sonochemical

19

biodiesel synthesis using heterogeneous base catalyst. Ultrason. Sonochem. 21(1)

20

(2014) 169-181.

21

38

R. S. Malani, S. Patil, K. Roy, S. Chakma, A. Goyal, V. S. Moholkar, Mechanistic

22

analysis of ultrasound-assisted biodiesel synthesis with Cu2O catalyst and mixed oil

23

feedstock using continuous (packed bed) and batch (slurry) reactors, Chem. Eng. Sci.

24

170 (2017) 743–755.

28

1

39

G. Gelbard, O. Bres, R. M. Vargas, F. Vielfaure, U. F. Schuchardt, 1H nuclear

2

magnetic resonance determination of the yield of the transesterification of rapeseed

3

oil with methanol, J. Am. Oil Chem. Soc. 72(10) (1995) 1239–1241.

4

40

biodiesel, T. ASAE 44(2) (2001) 193–200.

5 6

G. Knothe, Analytical methods used in the production and fuel quality assessment of

41

A. Kapil, K. Wilson, A. F. Lee, J. Sadhukhan, Kinetic modeling studies of

7

heterogeneously catalyzed biodiesel synthesis reactions, Ind. Eng. Chem. Res. 50(9)

8

(2011) 4818–4830.

9

42

V. R. Midathana, V. S. Moholkar, Mechanistic studies in ultrasound-assisted

10

adsorption for removal of aromatic pollutants, Ind. Eng. Chem. Res. 48(15) (2009)

11

7368–7377.

12

43

R. S. Malani, H. Sardar, Y. Malviya, A. Goyal, V. S. Moholkar, Ultrasound

13

Intensified Biodiesel Production from Mixed Non-Edible Oil Feedstock Using

14

Heterogeneous Acid Catalyst Supported on Rubber De-Oiled Cake. Ind. Eng. Chem.

15

Res. 57(44) (2018) 14926–14938.

16

44

R. Bhoi, D. Singh, S. Mahajani, Investigation of mass transfer limitations in

17

simultaneous

esterification

and

transesterification

18

heterogeneous catalyst, React. Chem. Eng. 2(5) (2017) 740–753.

19

29

of

triglycerides

using

a

Table 1: Experimental sets for feed optimization experiments

1 2

Sr. No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

Castor 0.05 0.275 0.05 0.05 0.2 0.2 0.095 0.1625 0.05 0.05 0.05 0.095 0.05 0.1625 0.1625 0.05 0.2 0.275 0.05 0.2 0.32 0.095 0.05 0.2 0.275 0.275 0.05 0.1625 0.05 0.05 0.2 0.05 0.05 0.14 0.5 0.095

Volumetric ratio of oils Jatropha Rubber Waste cooking 0.05 0.275 0.2 0.05 0.05 0.425 0.05 0.05 0.65 0.275 0.05 0.2 0.2 0.2 0.2 0.05 0.2 0.2 0.32 0.095 0.245 0.05 0.1625 0.3125 0.2 0.05 0.35 0.05 0.05 0.2 0.05 0.5 0.2 0.095 0.095 0.47 0.5 0.05 0.2 0.1625 0.1625 0.3125 0.1625 0.05 0.3125 0.2 0.2 0.2 0.05 0.05 0.35 0.05 0.275 0.2 0.275 0.275 0.2 0.05 0.2 0.35 0.095 0.095 0.245 0.095 0.32 0.245 0.05 0.05 0.425 0.2 0.05 0.2 0.05 0.05 0.2 0.275 0.05 0.2 0.275 0.05 0.425 0.1625 0.1625 0.2 0.05 0.275 0.425 0.2 0.2 0.35 0.2 0.05 0.35 0.05 0.2 0.35 0.1625 0.1625 0.3125 0.14 0.14 0.29 0.05 0.05 0.2 0.095 0.095 0.245

3 4

30

Average Palm % Conversion 0.425 43.18 ± 0.28 0.2 15.89 ± 0.52 0.2 20.18 ± 0.35 0.425 22.18 ± 0.44 0.2 26.08 ± 0.41 0.35 38.17 ± 1.09 0.245 20.73 ± 0.62 0.3125 30.89 ± 0.75 0.35 27.31 ± 0.82 0.65 43.17 ± 0.78 0.2 46.18 ± 0.51 0.245 26.17 ± 0.54 0.2 21.76 ± 0.83 0.2 26.02 ± 0.33 0.3125 27.53 ± 0.51 0.35 28.68 ± 0.57 0.35 27.14 ± 0.95 0.2 30.11 ± 0.35 0.2 22.16 ± 0.27 0.2 26.38 ± 0.83 0.245 19.92 ± 0.52 0.245 36.28 ± 0.86 0.425 33.28 ± 0.72 0.35 27.76 ± 1.46 0.425 30.83 ± 0.92 0.2 17.98 ± 0.83 0.2 26.91 ± 1.13 0.3125 29.89 ± 1.20 0.2 32.27 ± 1.29 0.2 24.18 ± 1.00 0.2 19.94 ± 0.41 0.35 36.11 ± 1.39 0.3125 27.19 ± 0.83 0.29 26.15 ± 0.37 0.2 8.18 ± 0.55 0.47 35.16 ± 0.38

Table 2 (A): Box–Behnken experimental range and level of independent variables

1 2

Variables Symbol Level of factors coded values (Actual values) Catalyst loading (wt% oil) C –1 (3) 0 (6) +1 (9) Temperature (oC) T –1 (45) 0 (55) +1 (65) Alcohol : oil molar ratio (M) M –1 (5:1) 0 (10:1) +1 (15:1) 3 4 5

Table 2 (B): Experimental sets of Box–Behnken design with triglyceride conversion Sr. No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Catalyst Loading (% w/w) 3 6 9 9 3 6 3 6 9 6 6 6 9 3 6

Molar Temperature Ratio (oC) 10 45 10 55 15 55 10 65 5 55 10 55 15 55 15 45 5 55 15 65 5 45 5 65 10 45 10 65 10 55

6 7

31

Experimental % Conversion 28.78 ± 1.13 86.78 ± 2.14 71.98 ± 0.99 73.98 ± 0.57 37.69 ± 0.95 86.02 ± 0.83 41.28 ± 0.79 40.82 ± 0.58 41.79 ± 1.41 71.28 ± 0.89 24.28 ± 0.55 52.18 ± 0.51 27.18 ± 0.41 39.28 ± 0.72 86.22 ± 0.72

Model Predicted % Conversion 28.44 86.34 72.00 74.33 37.67 86.34 41.73 40.72 41.35 70.92 24.65 52.28 27.26 39.20 86.34

1 2 3

Table 3: Elementary reaction steps and corresponding forward kinetic rate expressions in transesterification process based on Eley–Rideal (ER) mechanism Step in the mechanism

Chemical equation

Rate expression

1. Methanol adsorption

 CH 3OH     CH 3OH  

r1  k1  f CH 3OH 

 D  F  CH 3OH   T  

r2  k2 T  CH 3OH  

M  F  CH 3OH   D  

r3  k3  D  CH 3OH  

 G  F  CH 3OH   M  

r4  k4  M  CH 3OH  

 D  D  

r5  k5  D 

 M  M   

r6  k6  M  

G    G   

r7  k7 G  

2. Transesterification reactions

3. Desorption of adsorbed species 4 5 6 7

Symbols: * – free catalyst active site, T – triglyceride, D – diglyceride, M – monoglyceride, G – glycerol, F – fatty acid methyl ester (biodiesel) (Reproduced with permission from Malani et al.34, Copyright © 2018, Elsevier)

32

Table 4: Final kinetic rate expressions in transesterification process based on Eley–Rideal (ER) mechanism Rate of consumption of triglyceride

Rate of consumption of diglyceride

Rate of consumption of monoglyceride

Rate of consumption of methanol

Rate of formation of glycerol

Rate of formation of FAME or biodiesel

rT 

d T   dt

rD 

d D  dt

rM 

d M   dt

rCH3OH 

k 2 T  k T  k 3  D  k 4  M  k 2 T   k 3  D   k 4  M  1 2    k5 k6 k7 k1 CH 3OH 

k3  D   k2 T  k T  k3  D  k4  M  k2 T   k3  D   k4  M  1 2    k5 k6 k7 k1 CH 3OH 

k 4  M   k3  D  k T  k3  D  k4  M  k2 T   k3  D   k4  M  1 2    k5 k6 k7 k1 CH 3OH 

d CH 3OH 

rG 

rF 

dt d G  dt



d F   dt



k2 T   k3  D   k4  M  k T  k3  D  k4  M  k2 T   k3  D   k4  M     1 2 k1 CH 3OH  k7 k6 k5

k4  M  k T  k3  D  k4  M  k2 T   k3  D   k4  M     1 2 k5 k6 k7 k1 CH 3OH 

( k2 T   k3  D   k4  M ) k T  k3  D  k4  M  k2 T   k3  D   k4  M  1 2    k5 k6 k7 k1 CH 3OH 

33

1 2

Table 5: Kinetic rate constants (min–1) for different steps of transesterification process at 59oC (332 K) Rate constants Control experiment Test experiment (min–1) (with mechanical agitation) (with sonication) k1 7.1910–3 4.4210–3 k2 4.3610–2 5.6010–2 k3 4.9210–2 6.2410–2 –2 k4 5.2610 7.3810–2 k5 6.6410–2 8.1710–2 k6 8.4810–2 1.0210–1 –1 k7 2.4210 2.7810–1 Cumulative error 2.8510–2 4.0510–2

3 4

34

1 2 3 4

Table 6: Arrhenius analysis of transesterification process: kinetic rate constants (min–1) and activation energies (kJ/mol) for the three steps and overall reaction of transesterification (A)

With Mechanical agitation

Transesterification reaction (MA) Step 1 (rate expression = r2, kinetic constant = k2) Step 2 (rate expression = r3, kinetic constant = k3) Step 3 (rate expression = r4, kinetic constant = k4) Overall transesterification reaction (k)

59oC (332 K)

54oC (327 K)

49oC (322 K)

Activation energy (kJ/mol)

R2

4.3610–2

3.8610–2

3.1410–2

29.09

0.98

4.9210–2

4.2510–2

3.8510–2

21.80

0.99

5.2610–2

4.7910–2

4.5010–2

13.80

0.98

1.9710–2

1.0710–2

4.3010–3

135.40

0.99

59oC (332 K)

54oC (327 K)

49oC (322 K)

Activation energy (kJ/mol)

5.6010–2

4.8510–2

4.3510–2

22.40

0.99

6.2410–2

5.5910–2

5.2810–2

14.81

0.96

7.3810–2

6.9010–2

6.6410–2

9.42

0.97

3.9110–2

1.5110–2

9.7010–3

123.65

0.95

5 6

(B)

With Ultrasound system

Transesterification reaction (US) Step 1 (rate expression = r2, kinetic constant = k2) Step 2 (rate expression = r3, kinetic constant = k3) Step 3 (rate expression = r4, kinetic constant = k4) Overall transesterification reaction (k) 7 8

35

R2

Table 7: Properties of Biodiesel

1 2

Property Density at 15 oC (kg/m3) Flash point (oC) Kinematic viscosity at 40 oC (mm2/s) Cloud point (oC) Oxidation stability at 110 oC (h) Calorific value (MJ/Kg)

Present study 887 128 4.08 2 5.58 36.23

3 4

36

ASTM D 6751 standard 860–900 Min. 93 1.9–6.0 – Min. 3 –

EN 14214 standard 860–900 Min. 101 3.5–5.0 – Min. 6 –

(a)

(b)

(c)

1 2 3

Figure 1: XPS spectra of (a) Potassium; (b) Oxygen and (c) Zinc molecule in KI/ZnO catalyst

4

37

1

(a) (b) Figure 2: FE–SEM micrographs of (a) ZnO particle and (b) KI/ZnO catalyst particles

2

38

1 2

Figure 3: N2 adsorption–desorption isotherm for ZnO support and KI/ZnO catalyst

3

39

1 2

(a)

3 4

(b)

5 6

Figure 4: 1H NMR spectra of mixed oil transesterification reaction at optimum conditions (a) with mechanical agitation and (b) with ultrasound treatment 40

(a)

(b) 1 2 3

Figure 5: Fitting of experimental and model predicted data for biodiesel yield (a) with mechanical agitation and (b) with ultrasound treatment

4

41

1

(a)

(b) 2 3 4

Figure 6: Arrhenius plots of individual reaction steps and overall transesterification process (a) with mechanical agitation; (b) with ultrasound system

5

42

1

2

Figure 7: Performance of KI/ZnO catalyst in the reusability study

3 4 5

6 7

RESEARCH HIGHLIGHTS

8 9 10 11 12

   

Screening of various non-edible oils and their blend in optimum ratio. Statistical optimization of transesterification with lab synthesized KI/ZnO catalyst Ultrasound boosts the reaction kinetics, but lowers methanol adsorption on catalyst Application of ultrasound lowers activation energy, but increases mass transfer barrier

43

1 2 3



Activation energy contributed by mass transfer, is higher than sum of three reaction steps

44