Ultrasonics Sonochemistry 22 (2015) 463–473

Contents lists available at ScienceDirect

Ultrasonics Sonochemistry journal homepage: www.elsevier.com/locate/ultson

Mechanistic analysis of cavitation assisted transesterification on biodiesel characteristics Baharak Sajjadi, A.R. Abdul Aziz ⇑, Shaliza Ibrahim Department of Chemical Engineering, Faculty of Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia

a r t i c l e

i n f o

Article history: Received 6 December 2013 Received in revised form 2 June 2014 Accepted 8 June 2014 Available online 20 June 2014 Keywords: Biodiesel Ultrasound Transesterification Mechanical stirring Biodiesel properties Cavitaion

a b s t r a c t The influence of sonoluminescence transesterification on biodiesel physicochemical properties was investigated and the results were compared to those of traditional mechanical stirring. This study was conducted to identify the mechanistic features of ultrasonication by coupling statistical analysis of the experiments into the simulation of cavitation bubble. Different combinations of operational variables were employed for alkali-catalysis transesterification of palm oil. The experimental results showed that transesterification with ultrasound irradiation could change the biodiesel density by about 0.3 kg/m3; the viscosity by 0.12 mm2/s; the pour point by about 1–2 °C and the flash point by 5 °C compared to the traditional method. Furthermore, 93.84% of yield with alcohol to oil molar ratio of 6:1 could be achieved through ultrasound assisted transesterification within only 20 min. However, only 89.09% of reaction yield was obtained by traditional macro mixing/heating under the same condition. Based on the simulated oscillation velocity value, the cavitation phenomenon significantly contributed to generation of fine micro emulsion and was able to overcome mass transfer restriction. It was found that the sonoluminescence bubbles reached the temperature of 758–713 K, pressure of 235.5–159.55 bar, oscillation velocity of 3.5–6.5 cm/s, and equilibrium radius of 17.9–13.7 times greater than its initial size under the ambient temperature of 50–64 °C at the moment of collapse. This showed that the sonoluminescence bubbles were in the condition in which the decomposition phenomena were activated and the reaction rate was accelerated together with a change in the biodiesel properties. Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction Among the biofuels currently in use or researched, Free Fatty Acid Methyl Esters (FAME), which is also known as biodiesel is considered an ideal alternative to diesel due to their similarities and advantages [1]. Biodiesel is commonly generated through a three-step, consecutive and reversible reaction called ‘‘transesterification’’. Low mass transfer due to immiscible nature of reactants is the main weakness of transesterification [2,3]. Recently, ultrasound assisted transesterification has been confirmed as a green synthesis method that is fast and energy-efficient [4]. It is due to the ultrasound ability to enhance mass transfer between the immiscible reactants. In other words, ultrasound waves are sinusoidal mechanistic waves consisting of both expansion (negative) and compression (positive) pressure waves. Hence, irradiation of ultrasound waves generates vacuum micro-regions in the liquid called ‘‘sonoluminescence bubble’’ that are filled with reactants

⇑ Corresponding author. Tel.: +60 379675300; fax: +60 379675319. E-mail address: [email protected] (A.R. Abdul Aziz). http://dx.doi.org/10.1016/j.ultsonch.2014.06.004 1350-4177/Ó 2014 Elsevier B.V. All rights reserved.

vapors. Mass transfer goes efficiently inside the micro-fine bubbles. ‘‘Cavitation’’ happens when bubbles grow (expansion) and collapse (compression) intensively [5]. This phenomenon assists the system to generate fine micro-emulsion through generation of micro streams, micro turbulent eddies and shock waves. Besides, the collapse is extremely energetic, resulting in generation of highly pressurized and over-heated regions called ‘‘hot spots’’ which induce the reaction [6]. Many authors have reported that low frequency ultrasound accelerates reaction rate in which higher conversion is achieved in transesterification within shorter reaction time compared to the other approaches [7–9]. Thermal decomposition can also be carried out in parallel with transesterification within these bubbles. Highly volatile hydrophobic molecules can easily and directly decompose in hot spots. Generally, biodiesel starts to decompose upon thermal stressing at 275 °C and above. The decomposition mainly involves (i) dimerization or polymerization reactions that form higher molecular weight components; (ii) isomerization reactions that transfer unsaturated cis-type FAMEs to trans-type FAMEs and (iii) pyrolysis reactions that break down FAMEs to form lower molecular weight FAMEs and hydrocarbons. These reactions occur

464

B. Sajjadi et al. / Ultrasonics Sonochemistry 22 (2015) 463–473

at 300–425 °C, 275–400 °C and >350 °C, respectively [10]. Bruno and co-workers [11] observed that the polymerization and cracking of unsaturated FAMEs significantly influenced the volatility of biodiesel fuel. Imahara et al. [12] found that the cold flow properties of biodiesel produced under high temperature and pressure (350 °C and 43 MPa) increased slightly due to cis–trans isomerization. Lin et al. [10] investigated the influence of thermal stressing on biodiesel viscosity and cold flow properties at 250–425 °C. They demonstrated that cis–trans isomerization reactions had a minimal effect on biodiesel characteristics. Meanwhile, pyrolysis reactions reduced both viscosity and the crystallization onset temperature but polymerization showed the opposite effects. Xin et al. [13] observed that the oxidation stability of biodiesel made from waste cooking oils with high peroxide values was improved at temperature of 270 °C and pressure of 17 MPa compared to biodiesel produced from the same feedstock by conventional method at lower temperature and pressure. Based on a detailed review done by Veljkovic´ et al. [6], despite a large quantity of researches on intensification of transesterification under irradiation of ultrasound energy, the quality of biodiesel produced in such conditions have not been well addressed. These characteristics influence flow injection and atomization of fuel as well as energy content of engine and its safety. In the current study, the authors tried to identify the physical mechanism of cavitation and provide a deep insight into the sonoluminescence transesterification that form conditions that may activate the thermal decomposition phenomena and influence the biodiesel characteristics. The principle aim is to find a link between the influence of thermal stressing and quality of biodiesel produced from alkaline transesterification of palm oil under ultrasound irradiation. 2. Methodology

time, in the range of 20–60 min. Each parameter was fixed and coded into levels: 1 (Minimum level), 0 (middle level) and +1 (Maximum level). Finally, 50 experimental runs including 10 axial points, 32 factorial points, and 1 replicable central point were designed. The central point was repeated 8 times in order to identify the experimental error. The employed design matrix with the related yield and biodiesel properties are reported in Tables 2 and 3 respectively. 2.3. Biodiesel synthesis A 300 ml stainless steel and a 400 ml glass flask were used as the reaction vessels. Both were equipped with a water bath to control the temperature; a condenser to inhibit the exhaust of evaporated methanol and a thermometer to record the reaction temperature. The sonication was carried out by a 24 kHz-ultrasonic processor, UP400S (Hielscher Ultrasonics). The mechanical stirring was done by an overhead stirrer (13516 IKA Eurostar 60 Digital). Both reactors were connected to a 220 V voltage regulator. Since ultrasonication generates heat, the temperature in the ultrasonic bath was controlled at about 2–10 °C below the reaction temperature. The reactors were initially charged with favorable amount of palm oil and then heated to reaction temperature by the surrounding water. In the next step, a pre-provided and pre-heated solution of methanol and catalyst was added to the oil. Then, the stirring/ sonication of the reaction started immediately and continued with the specified mixing intensity/sonication power and reaction time. After the stirring/sonification, the mixture was transferred to a separatory funnel for gravitational separation for 48 h. The methyl ester was then washed with diluted hydrochloric acid to neutralize the remaining alkaline catalyst before being washed by distilled water twice. Finally, the excessive methanol and water in the product were removed by rotary evaporation at 70 °C for 120 min.

2.1. Material and method 2.4. Biodiesel analysis and characterization RBD (Refined, Bleached and Deodorized) palm oil was employed as the source of triglyceride. Its chemical composition and physical properties can be found in Table 1. Absolute methanol (99.99%) and hydrochloric acid (30 wt.%) for the pre-treatment step were purchased from Sigma Aldrich while potassium hydroxide pellets (99.99 wt.%) were purchased from Merck Companies, Malaysia. A mixture of fatty acid including palmitic, stearic, oleic, linolenic and linoleic methyl esters (called MSTFA) which was used as the HPLC standards in this study, was supplied by Supelco Company. 2.2. Experimental design Central composite design (CCD) was employed to specify the effects of operational variables on conversion of oil to FAME. This type of design is desirable for sequential experiments to obtain proper information for examining lack of fit without a large number of design points [14]. The operational parameters included (i) temperature, in the range of 50–64 °C; (ii) MeOH:oil molar ratio, in the range of 6–12; (iii) concentration of catalyst, in the range of 1–2%; (iv) sonication power, in the range of 0–400 W; (v) mechanical stirring, in the range of 400–800 RPM and (vi) reaction Table 1 Properties of used palm oil.

The chemical compositions of biodiesel product and yield of reaction were determined by a gas chromatography (Agilent Technology gas chromatograph model GC 6890) equipped with cool-on column (DB-23, 60 mL, 0.250 mm ID, 0.15 lm film thickness). The results are summarised in Table 2. Helium was used as a carrier gas at the flow rate of 1.0 mL/min; nitrogen was used as a make-up gas at the flow rate of 25 mL/min and hydrogen was used as a fuel gas at the flow rate of 40 mL/min. The start-up temperature was 323 K, which was increased to 448 K at 25 K/min and further increased to 553 K at 5 K/min. 553 K was maintained for 5 min. The standard test methods according to ASTM Standard are applied to evaluate the quality of biodiesel [15]. The density of the biodiesel produced was analyzed (based on ASTM D405296) by a fully automatic density meter (Mettler Toledo model DM.40, Switzerland). The viscosity (based on ASTM D445-06) was measured by a fully automatic viscometer (Brookfield DV-II+PRO Viscometer). The pour and cloud point (based on ASTM D97-93 and D2500) were determined by a pour/cloud point tester (Normalab, model NTE 450, France) while flash point (based on ASTM D93-07) was determined by a Pensky–Martens Closed Cup Tester. 3. Mathematical model

Property

Value

Density at 15 °C Viscosity at 40 °C Moisture content (%) Acid value Iodine value

791.8 kg/m3 33.60 mm2 s1 400  106 ppm 51.8 ppm 0.151 mg/g

The dynamic of sonoluminescence bubble along with the phenomena of liquid vaporization, diffusion, heat transfer and chemical reactions within the cavitation bubbles were simulated using Matlab to get a quantitative insight of the influence of ultrasound irradiation on a system (Mathworks Inc., USA).

465

B. Sajjadi et al. / Ultrasonics Sonochemistry 22 (2015) 463–473 Table 2 Experimental design matrix and the values of related Macro/Micro reaction yield. Run

Type

Temp (°C)

Oil Molar ratio

Cat wt.%

Mixing Watt (Micro)

Intensity RPM (Macro)

Time Minutes

Yield %(Micro)

Yield %(Macro)

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 37 38 39 40 41 42 43 44 45 46 47 48 49 50

Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Factorial Axial Axial Axial Axial Axial Axial Axial Axial Axial Axial Center Center Center Center Center Center Center Center

50 64 50 64 50 64 50 64 50 64 50 64 50 64 50 64 50 64 50 64 50 64 50 64 50 64 50 64 50 64 50 64 50 64 57 57 57 57 57 57 57 57 57 57 57 57 57 57 57 57

6 6 12 12 6 6 12 12 6 6 12 12 6 6 12 12 6 6 12 12 6 6 12 12 6 6 12 12 6 6 12 12 9 9 6 12 9 9 9 9 9 9 9 9 9 9 9 9 9 9

1 1 1 1 2 2 2 2 1 1 1 1 2 2 2 2 1 1 1 1 2 2 2 2 1 1 1 1 2 2 2 2 1.5 1.5 1.5 1.5 1 2 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5

300 300 300 300 300 300 300 300 400 400 400 400 400 400 400 400 300 300 300 300 300 300 300 300 400 400 400 400 400 400 400 400 350 350 350 350 350 350 300 400 350 350 350 350 350 350 350 350 350 350

400 400 400 400 400 400 400 400 800 800 800 800 800 800 800 800 400 400 400 400 400 400 400 400 800 800 800 800 800 800 800 800 600 600 600 600 600 600 400 800 600 600 600 600 600 600 600 600 600 600

20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 40 40 40 40 40 40 40 40 20 60 40 40 40 40 40 40 40 40

86.99 88.17 88.94 90.28 91.44 92.63 91.92 92.57 89.20 91.26 90.35 91.26 91.19 93.84 93.75 93.05 89.92 91.91 91.61 93.13 91.88 95.81 91.26 93.70 90.89 93.39 92.27 93.20 89.74 95.44 92.81 95.71 90.15 95.15 89.66 92.67 90.93 92.26 90.95 91.96 89.39 91.95 91.11 91.44 91.90 90.88 91.53 90.18 90.80 91.47

81.60 84.27 84.93 86.33 83.49 86.85 86.41 89.47 82.82 87.90 85.96 89.12 86.89 89.09 87.34 89.16 87.12 88.11 88.74 90.03 87.99 90.50 88.72 92.43 90.24 91.75 90.67 93.65 91.01 92.70 91.31 94.01 86.35 88.12 87.04 88.62 87.18 88.04 85.79 88.49 83.64 89.89 87.08 86.98 87.18 87.08 87.13 87.62 86.49 87.15

3.1. Bubble dynamics The radial motion of the cavitation bubble is described by a variant of Rayleigh–Plesset equation [16], which takes into account the liquid compressibility:

  2   2 dR=dt d R 3 dR=dt dR R 2þ 1 1 c 3c dt dt 2   dR=dt 1  R dPg dR=dt 2r ¼ 1 P  Pa  P0 þ  4m  c R qL g qL c R qL R ð1Þ where q, r, l are the density, surface tension coefficient and viscosity of the liquid. c denotes the sound speed in the liquid; P0 is the static pressure and Pa is the acoustic driving pressure that can be explained by:

Pa ¼ PA sin ðxtÞ

ð2Þ

pffiffiffiffiffiffiffiffiffiffiffiffiffiffi where, P A ¼ 2qIUS c and x are the pressure amplitude and angular velocity of the acoustic wave. pg also denotes the internal pressure of the bubble which is modeled by van der Waals equation as shown in [17]:

Ntot ðtÞkT i pg ¼ h  3 4p R3 ðtÞ  h 3

ð3Þ

In Eq. (3), Ntot is the total number of molecules in the bubble including TG (Triglyceride), DG (DiGlyceride), MG (MonoGlyceride), G (Glycerol), FAME (Free fatty Acid Methyl Ester) and MeOH (Methanol), which varies according to evaporation, diffusion, condensation, reaction, temperature and time. k and h also represent the Boltzmann constant and van der Waals hard-core radius of various species in the bubble. 3.2. Energy balance Temperature is determined by the energy balance over the cavitation bubble. The energy balance is approximated by considering

466

B. Sajjadi et al. / Ultrasonics Sonochemistry 22 (2015) 463–473

Table 3 Run numbers and biodiesel specifications. Run

Micro mixing

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 37 38 39 40 41 42 43

Macro mixing

q

l

PP °C

FP °C

l

mm2/sec

CP °C

q

kg/m3

kg/m3

mm2/s

CP °C

PP °C

FP °C

875.8 875.8 875.7 875.5 875.8 875.5 875.6 875.5 875.7 875.6 875.7 875.5 875.8 875.5 875.5 875.5 875.7 875.6 875.6 875.4 875.8 875.4 875.6 875.5 875.6 875.5 875.6 875.4 875.8 875.4 875.5 875.4 875.7 875.4 875.7 875.5 875.6 875.6 875.6 875.6 875.7 875.6 875.6

4.467 4.467 4.391 3.833 4.244 3.955 4.148 3.987 4.367 4.167 4.256 3.765 4.172 3.883 3.782 3.833 4.283 4.148 4.130 3.833 4.132 3.819 4.171 3.922 4.153 4.008 4.121 3.783 4.257 3.783 4.003 3.783 4.257 3.783 4.322 4.057 4.181 4.130 4.162 4.119 4.357 4.098 4.193

14 14 14 13.5 14 13.5 13.5 13 14 13.5 14 13 13.5 13 13.5 13 13.5 13.5 13.5 13 13.5 13 13.5 13 13 13 13 12.5 13 13 13 12 13.5 13.5 13.5 13.5 13.5 13.5 14 13.5 14 13 13.5

10 10 10 8.5 9.5 9 9 9 10 9.5 9.5 8.5 9.5 8.5 8.5 9 9.5 9 9 8.5 9 8.5 9.5 8.5 9.5 9 9 8.5 9.5 8.5 9 8.5 9.5 8.5 9.5 9 9.5 9 9.5 9 10 9 9.5

204.5 202.5 196.5 194.5 192.5 188.5 190.5 188.5 196.5 192.5 194.5 192.5 192.5 188.5 188.5 188.5 194.5 190.5 190.5 188.5 190.5 186.5 192.5 188.5 192.5 188.5 190.5 188.5 194.5 186.5 188.5 186.5 194.5 186.5 194.5 188.5 192.5 190.5 192.5 190.5 196.5 190.5 192.5

876.1 876 876 875.9 876 875.9 875.9 875.7 876.1 875.8 875.9 875.7 875.9 875.7 875.8 875.7 875.8 875.8 875.8 875.7 875.8 875.7 875.8 875.6 875.7 875.6 875.7 875.5 875.6 875.6 875.6 875.5 875.9 875.8 875.8 875.8 875.8 875.8 875.9 875.8 876 875.7 875.8

4.56 4.37 4.37 4.31 4.44 4.30 4.30 4.14 4.47 4.33 4.35 4.15 4.30 4.15 4.24 4.15 4.24 4.21 4.16 4.09 4.23 4.09 4.17 4.01 4.09 4.04 4.09 4.05 4.07 4.03 4.07 3.94 4.30 4.20 4.28 4.17 4.16 4.21 4.35 4.18 4.41 4.14 4.19

14.5 14.5 14.5 13.0 14.0 13.5 13.5 13.5 14.5 14.0 14.0 13.0 14.0 13.0 13.0 13.5 14.0 13.5 13.5 13.0 13.5 13.0 14.0 13.0 14.0 13.5 13.5 13.0 14.0 13.0 13.5 13.0 14.0 13.0 14.0 13.5 14.0 13.5 14.0 13.5 14.5 13.5 14.0

11 11 11 10.5 11 10.5 10.5 10 11 10.5 11 10 10.5 10 10.5 10 10.5 10.5 10.5 10 10.5 10 10.5 10 10 10 10 9.5 10 10 10 9 10.5 10.5 10.5 10.5 10.5 10.5 11 10.5 11 10 10.5

206.5 202.5 202.5 198.5 204.5 198.5 198.5 194.5 206.5 198.5 200.5 194.5 200.5 194.5 198.5 194.5 198.5 196.5 196.5 192.5 198.5 192.5 196.5 188.5 192.5 190.5 192.5 188.5 190.5 188.5 190.5 186.5 200.5 196.5 198.5 196.5 198.5 196.5 202.5 196.5 204.5 194.5 198.5

the heat transfer across the bubble (dQ/dt); the rate of work done by the bubble (PidV) which is reduced during the expansion phase and the varying rate of number of molecules within the bubble (dNS/dt). Therefore, the energy balance is introduced as:

C m;mix

dT dQ dNs ¼  Pi dV þ ðhs  U s Þ dt dt dt

ð4Þ

where hs and Us are the molecular enthalpy and internal energy of each compound. Cm,mix represents the specific heat of the mixture that can be obtained by the molecular specific heat of each comN P pound and the number of its molecules C m;i N i :

Table 4 Vibrational temperatures of various species. Species

Characteristic wave number

Triglyceride Diglyceride Monoglyceride Gelycerol

1713 1713 1713 919 1565 722 1464

FAME

1745 1559 1733 1040 1656 1117 1743

1734 2915 1745 1120 2914 1159 2852

1410 3025 1378 2920

1474 3308

i¼1

fi X ðhi =TÞ expðhi =TÞ þ 2 ðexpðexpðhi =TÞ  1Þ2 Þ 2

C m;i ¼ Ni k

! ð5Þ

In Eq. (5), hi denotes the vibrational temperature of each compound which provides information on energy transfer between the electrons and heavy particles, especially molecules, and can be obtained by:

hm hi ¼ kB

ð6Þ

m represents the wave number and the related values are given in Table 4

3.3. Heat transfer Conductive heat transfer across the bubble wall is determined analogously with mass transfer by Eqs. (7) and (10) respectively:

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi !   dQ T0  T Rv R 2 ; lth ¼ min ¼ 4pR j ; dt lth jdR=dtj p

ð7Þ

where T and T0 are the temperature of bubble contents and bubble wall, respectively. j is the thermal conductivity of the bubble contents that can be obtained by the following equation:

467

B. Sajjadi et al. / Ultrasonics Sonochemistry 22 (2015) 463–473

jmix ¼

X i

nj Xi i ; nj Ui;j

ð8Þ

j

2 !1=2   32  1=2 1=4 1 mi g mj i 41 þ 5 Ui;j ¼ pffiffiffi 1 þ mj gj mi 8

ð9Þ

where mi, gi, ni are the molecular mass, viscosity and the mole fraction of species i. The thermal diffusivity is eventually obtained from v = jmix/cp. lth is also the thermal diffusion thickness. 3.4. Mass transfer The mass-transfer effect which is critically influenced by Ntot, is obtained by employing the bubble dynamic and diffusion equations along with the reaction. Generally, the vapor transport is a two-step process, consisting of diffusion to the wall and condensation. However, the transport phenomena layer is more controlled by diffusion rather than condensation [18]. Therefore, the diffusion-limited model of the boundary layer approach proposed by Toegel et al. [19] was used in this study:

  dNd;i @n ni;0  ni ; ld ¼ 4pR2 D  4pR2 D @r r¼R dt ld sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ! RD R ¼ min ; jdR=dtj p

ð10Þ

where ni,0 and ni are the equilibrium and instantaneous concentration of molecules of compound i, respectively. D represents the binary diffusion coefficient and ld represents the instantaneous diffusive penetration depth. Diffusion coefficients are also approximated for each species by employing the model of Wilke and Chang [18]. Since methanol was used excessively in the reaction mixture in this study, the diffusivity of all other compounds in methanol was considered:

Di;MeOH ¼ 7:4  108

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi UMeOH MMeOH T

gMeOH m0:6 i

ð11Þ

where, UMeOH, MMeOH and gMeOH represent the association factor, molecular weight (g mol1) and viscosity of methanol (cP) respectively. mi also denotes the molar volume of species i at its normal boiling temperature (cm3 mol1).

Fig. 1. Perturbation graph of reaction yeild at the temperature of (a) 57 °C, (b) 64 °C A: temperature, (°C), B: MeOH:oil (molar ratio), C: catalyst concentration (wt.%), D: sonication power (W), E: time (min): left: US, right: CS.

468

B. Sajjadi et al. / Ultrasonics Sonochemistry 22 (2015) 463–473

3.5. Chemical reactions

4. Result and discussion

The values of the reaction rate constants which exactly match this work were obtained from the related literature [20]. However, the kinetics of the reactions inside the cavitation bubble were several orders of magnitude higher than the time scale of bubble dynamics due to very high temperature within the bubbles [21]. Therefore, the well-known Arrhenius law was employed to describe the net reaction rates per unit volume within the cavitation bubble:

4.1. Experimental results

  Ef ;i r j ¼ kntot ni nj T  exp  kT

ð12Þ

By considering Eqs. (10) and (12) for diffusion and chemical reaction, the total changes in the number of molecules for each species is expressed by:

dNi dNd;i dNR;i ¼ þ dt dt dt

ð13Þ

The set of ODEs in the bubble dynamics model can be solved simultaneously using the well-known Runge–Kutta adaptive step size method. Generally, a single sonoluminescence bubble in the reactant mixture is described under the conditions of reaction. The mathematical analysis was carried out with the assumptions on compressible boundary conditions. The parameters and constants of the mixture were fixed accordingly based on Table 5. It worth noting that, such method of analyzing were employed by Kuppa et al. [22] and Choudhury et al. [23] firstly.

4.1.1. Reaction yield Table 2 and Fig. 1 present the experimental results of reaction yield for both ultrasound assisted transesterification and traditional macro mixing with conventional heating. Although investigation of reaction yield is not the major scope of this work, it is presented to clarify the other sections. As mentioned earlier, an ultrafine mixing is generated under ultrasound irradiation which leads to major mass transfer between the reactants. Therefore, the maximum yield of 93.84% was achieved in just about 20 min at MeOH:oil molar ratio of 6:1 (Run No. 14). Meanwhile, the reaction reached only 89.09% in the same condition with mechanical stirring and conventional heating. By further increasing the reaction time to 60 min under ultrasonication, the reaction reached 95.44% (Run No. 30) and 95.71% (Run No. 32) at the MeOH:oil molar ratio of 6:1 and 12:1 respectively. However, the maximum yield of 94.01% (Run No. 32) was achieved with mechanical stirring performed at the maximum values of all parameters. Generally, temperature was the most effective parameter in both methods (Fig. 1) while mixing intensity and time exerted more influence on the reaction under mechanical stirring (Figs. 1 and 2). In other words, a fine mico-emulsion was produced under ultrasound irradiation even with the lowest power; the reaction therefore slightly increased at higher sonication intensity. Another noteworthy point is presence of sufficient amount of alcohol had a noticeable impact on the reaction yield under traditional heating. The reaction yield

Fig. 2. Response surface of reaction yield versus temperature and (a) MeOH:oil molar (b) reactor power at the middle value of other parameters: left: US, right: CS.

B. Sajjadi et al. / Ultrasonics Sonochemistry 22 (2015) 463–473

increased by increasing MeOH:oil molar ratio up to 9 under ultrasound and it then reached a plateau by further increase in alcohol concentration. It dramatically reduced at higher MeOH:oil molar ratio and temperature (above 59 °C) (Fig. 2). The reason is increasing the methanol quantity till an optimum value leads to increase of cavitational intensity, formation of smaller drop sizes (higher emulsion quality) and provision of extra areas for mass transfer which eventually increases conversion. Beyond that, excessive methanol reduces the concentrations catalyst as well as interfacial area due to smaller volume of oil in reaction mixture. In general, the reaction yield produced under ultrasound irradiation was 3.69% higher than that produced by conventional heating under similar production conditions. 4.1.2. Biodiesel characterization Table 3 and Figs. 3 and 4 compare the physicochemical properties of biodiesel produced by sonoluminescence micro and stirring macro synthesis. Based on the table, the produced biodiesel by both systems fully conform to the range given in both ASTM and EN, except for 5 runs of traditional conventional heating, which did not conform to the pour point standard. In this study, the biodiesel samples produced by ultrasonication power had slightly lower densities and viscosities than those produced by traditional heating under the same operational conditions. The maximum density of 876.1 kg/m3 was observed in the case of CSR (Run No. 1), while the maximum density of 875.8 kg/m3 was

469

achieved in cases of UAR (Runs No. 1, 2, 5, 13, 21, 29, US). Furthermore, the viscosity of biodiesel produced under ultrasound irradiation was about 0.12 mm2/s lower than that produced under traditional conventional heating. The viscosity of biodiesel produced by ultrasound in this work was also lower compared to that produced by the other methods [24–27]. In terms of cold flow properties, the minimum and maximum cloud point observed were 12, the minimum an UAR) and 14.5 °C (Run No. 1, 2, 3) respectively while the minimum and maximum pour point observed were 8.5 °C (Runs with the greatest yield US) and 11 °C (Runs with the least reaction yield CON) respectively. Furthermore, ultrasound decreased pour point by nearly 1~2C. However, no significant improvement in cloud point was observed. Besides, the cold flow properties obtained under ultrasound irradiation in this work were nearly 5° lower than those produced by other conventional methods [11–16,18–21]. As summarized in Table 3, the flash point of all biodiesel samples was much greater than expected based on ASTM D93-07. However, the flash point of biodiesel produced by ultrasound waves was slightly lower than that produced by mechanical stirrer and conventional heating. The same observation in terms of biodiesel properties was also reported by Salamatinia et al. [28]. Analysis of the sample with the highest viscosity/density/flash point showed that it contained 32.86% of C16:0; 3.67% of C18:0; 35.06% of C18:1 and 9.67%. of C18:2. On the other hand, the average compositions of less viscous/dense samples were 33.75% of C16:0; 3.81% of 5C18:0; 36.48% of C18:1 and 10.16% of C18:2.

Fig. 3. Perturbation graph of (a) density, (b) viscosity A: temperture, (°C), B: MeOH:oil (molar ratio), C: catalyst concentration (wt.%), D: sonication power (W), E: time (min): left: US, right: CS.

470

B. Sajjadi et al. / Ultrasonics Sonochemistry 22 (2015) 463–473

Therefore, the content of triglycerides in the samples has more influence on the biodiesel properties compared to the content of saturated and unsaturated components in biodiesel. In other words, the samples with higher amount of unreacted triglyceride presented greater values for each property. The same flow pattern was also observed for cold flow properties. Therefore, the pour and cloud point increased in samples with higher amount of unreacted triglycerides. The reason is triglycerides (especially in palm oil) are huge molecules with three branches. At low temperature, the branches are stuck in each other and inhibit further molecular motions. This phenomenon is more prominent in palm oil biodiesel due to its long carbon chain (C16–C18) [29]. Hence, the samples with higher reaction yield presented lower values for each property (which is also more favorable). According to Analysis of Variance, a statistical analysis, all operational parameters except catalyst concentration showed positive influence on biodiesel properties in both cases. However, MeOH:oil molar ratio was more effective for the US samples. It is because greater alcohol concentration till an optimum value increases the cavitational intensity and formation of smaller drop sizes. This condition leads to higher emulsion quality, providing extra area for mass transfer and increasing the reaction yield subsequently.

Mixing intensity and time are more important for the CON samples. This is due to the necessity for a better mixing and contact time between the reactants in the mechanical stirring system to obtain better yield. Reaction temperature was also the most effective parameter in both cases. It is clear that the biodiesel properties were affected by reaction yield. Hence, the pattern of biodiesel properties changes and reaction yield were similar, based on comparison of Figs. 3 and 4 with Fig. 1. More details about the interaction and other influences of operational parameters are observed in Figs. 3 and 4. The other point that should be highlighted is related to decomposition reactions. As mentioned earlier, such reactions may be activated at extremely high reaction temperatures and pressures. The mathematical modeling may clarify the mechanistic features of sonoluminescence bubbles and develop a link between the bubbles properties and decomposition reactions.

4.2. Numerical analysis Figs. 5a–e present the simulation results of the sonoluminescence bubble characteristics, which included radius, temperature,

Fig. 4. Perturbation graph of (a) cloud point, (b) flash point A: temperture, (°C), B: MeOH:oil (molar ratio), C: catalyst concentration (wt.%), D: sonication power (W), E: time (min): left: US, right: CS.

B. Sajjadi et al. / Ultrasonics Sonochemistry 22 (2015) 463–473

, 64 °C

471

Fig. 5. Simulation results of cavitation bubble in first and second collapse at 50 °C oscillation velocity and (e) number of molecules.

, 57 °C

, (a) radius, (b) temperature, (c) pressure, (d)

pressure, oscillatory velocity and number of molecules within them, for three different reaction temperatures. As observed, the bubbles with the initial radius of 5 lm under ultrasound irradiation grew up to almost 17.9, 15.75 and 13.7 of its initial radius at the ambient temperatures of 323.15, 330.15, 337.15 K respectively. Besides, the cavitation cycles increased and generated finer microemulsion. This phenomenon is confirmed by the life span of bubbles which showed a progressive reduction in subsequent cycles. It is due to intensification of reaction temperature that improves the solubility of methanol in the oil and also elevates the evaporation and diffusion rate of methanol into the bubbles. Therefore, greater ambient temperature makes the bubbles grow (expansion phase) and collapse (compression phase) faster. Moreover, the maximum radius of cavitation bubbles in the next cycle decreased due to reduction of the collapse intensity. During the expansion phase, the molecules evaporate and are trapped within the cavitation bubbles. Then, the molecules start to react and undergo an energetic compression till the moment of collapse. The simulation results of this study showed that the bubbles reached the temperature and pressure of 758.9 K and 235.58 bar at the lowest ambient temperature. The bubble temperature and pressure slightly increased by further increasing the ambient temperature. As the ambient temperature rose till 64 °C, which is close to the boiling point of methanol, more methanol molecules evaporated into the sonoluminescence bubble. Hence, the heat capacity of the cavitation bubble increased by reducing

the temperature as much as 713.41 K at the moment of collapse. Methanol also cushioned the collapse, reducing the internal pressure of bubbles as much as 159.55 bar. The trapped molecules were subject to supercritical conditions (above 512.58 K and 80.96 bar). Such conditions confirm the acceleration of reaction within the cavitation bubble. At the same time, it may activate the decomposition reactions. Based on the previous studies, cis–trans isomerization, polymerization and pyrolysis reactions start at 275 °C. Unsaturated biodiesel molecules decompose completely at 425 °C after 8 min. Saturated molecules are more stable, but start to decompose at 400 °C as well [10,30]. Decomposition reactions are enhanced by increase in thermal stressing temperature or duration or both. The isomerization and polymerization reactions led to a slight increase in viscosity and cold flow properties, but the increase was negligible compared to the other decomposition reactions and their effects were neutralized by pyrolysis. Pyrolysis is the most effective reaction that becomes very significant at 350 °C and above. This reaction produces components of smaller molecular weight and reduces the viscosity and cold flow properties eventually. Therefore, it is suggested that the decomposition reaction, mainly pyrolysis, is responsible for the change of a part of biodiesel properties and unreacted reactants. The total number of the molecules within the sonoluminescence bubbles is illustrated in Fig. 5d. The evaporated molecules diffuse into bubbles as the bubbles expand. Therefore, the bubbles reach their maximum capacity at the moment of collapse. It was

472

B. Sajjadi et al. / Ultrasonics Sonochemistry 22 (2015) 463–473

Table 5 Summary of the simulation data. Temperature (°C)

Species MEOH

TG

DG

MG

G

FAME

Density (kg/m3) 50 57 64

764 757.7 751.4

875.1 871.67 868.286

863.7 859.465 855.393

852.3 847.26 842.5

1301.8 1196 1117.6

858.3 853.22 847.97

Viscosity (Pa s) 50 57 64

0.000392 0.000365 0.000337

0.0339 0.03082 0.02767

0.018824 0.017048 0.015275

0.003749 0.003276 0.002879

0.142 0.0813 0.0506

0.003013 0.002591 0.002278

Sound speed (m/s) 50 57 64

1036.6 1016 996

1350.6 1325.58 1301.65

1311.4 1286.49 1261.575

1272.2 1247.4 1221.5

1809 1782 1755.8

1291.55 1266.825 1242.5

Diffusion coefficient 50 57 64

3.093e9 5.157e9 8.23e9

4.463e10 6.55e10 9.65e10

5.62e10 7.73e10 10.86e9

0.800e9 2.82e9 5.83e9

2.04e9 4.082e9 7.13e9

3.09e9 2.87e9 5.89e9

Table 6 Summary of the simulation results. Simulations parameters, R0 = 5 lm T0 P0 MeOH:oil

323.15 350 6

330.15 350 6

337.15 350 6

Rmax/R0 Tmax Pmax Vmax Ntot

17.93 758.9 235.58 0.035–0.057 1.18e13

15.75 774.61 275.64 0.039–0.062 6.76e12

13.70 713.41 159.55 0.041–0.065 4.48e12

Species

Equilibrium composition of different species in the bubble

Triglyceride Diglyceride Monoglyceride Gelycerol FAME Methanol

6.77e10 2.89e10 9.20e9 2.54e11 7.69e11 1.07e13

observed from the study that the bubbles with the lowest ambient temperature had the greatest capacity due to their bigger size. In this situation, the bubbles contained 0.57% of triglycerides, 0.245% of diglycerides, 0.078% of monoglycerides, 2.1% of glycerol and 6.52% of FAME in the first lifecycle. However, this equilibrium was changed at the second lifecycle due to the reaction evolution within the bubbles. Furthermore, based on Table 6 and the simulation results, methanol had dominated the greatest capacity of sonoluminescence bubbles which depend on its higher vapor pressure, diffusion coefficient and methanol’s concentration within the media. It can be suggested that the reactions speed up drastically within these bubbles due to large evaporation of methanol into the bubbles that enhances the heat capacity of the bubble contents [31]. High pressure and high temperature, which increase the highly energetic collisions between the molecules in a gaseous atmosphere, may also contribute to the phenomenon. The last figure explains the oscillation velocity that generates microturbulence. As observed, the magnitudes of micro turbulence increased at higher reaction temperatures. Since the viscosity of oil and the total liquid bulk were smaller at higher temperatures, the intensity of the bubble motion increased. Furthermore, comparison of the micro turbulence velocity with the micro-streaming velocity (14.48 cm/s), revealed that the contribution of ultrasound to the

5.36e10 2.27e10 9.03e9 1.58e11 4.84e11 5.88e12

2.26e10 1.19e10 4.49e9 4.81e10 1.55e11 4.24e12

total convection in the medium was quite high. In other words, it confirmed that ultrasound plays a principal role in generation of a fine micro emulsion. The same observation was reported by Choudhury et al. [31].

5. Conclusion In the present work, the mechanistic features of sonoluminescence transesterification and its effects on the physical and chemical properties of biodiesel were discerned. The study was done by coupling the statistical analysis of the experimental results with the simulation of sonoluminescence/cavitation bubble dynamics. The results were further compared to that of macro mechanical stirring with traditional conventional heating. It was observed that ultrasound assisted transesterification could produce higher reaction yield in a shorter time in contrast to traditional macro mixing under the same condition. Furthermore, it could slightly decrease the density and viscosity of biodiesel compared to the mechanically stirred system. Moreover, the reaction results revealed that among the five operational parameters, which included reaction temperature, MeOH:oil molar ratio, catalyst concentration, sonication/mixing intensity and reaction

B. Sajjadi et al. / Ultrasonics Sonochemistry 22 (2015) 463–473

time; temperature had the main influence on the reaction yield and biodiesel properties. Some unclear aspects of cavitation phenomena in the system including transesterification reaction were also studied. It was observed that ultrasound waves resulted in generation of cavitation bubbles in the mentioned system. The bubbles expanded to about 13–18 times of their initial radius till the moment of collapse. After that, they experienced an energetic compression and collapse. In this phase, they reached the temperature and pressure of 758.9–713.41 K and 235.58–159.55 bar respectively. The reaction sped up within the sonoluminescence bubbles due to increasing number of energetic collisions within the bubble, which is similar to supercritical condition. However, decomposition reactions are activated easily in such conditions. Decomposition is one of the reasons of decreased physicochemical properties of biodiesel under ultrasound irradiation. The simulation results showed that sonication of reaction mixture does not always result in favorable effects. Another reason is lower amount of unreacted compounds in biodiesel samples produced by sonoluminescence transesterification. Finally, the oscillation velocity of these bubbles showed that cavitation plays the principal role in increasing micro stream, generating fine microemulsion and increasing mass transfer. Acknowledgments The authors are grateful to the University of Malaya High Impact Research Grant (HIR-MOHE-D000038-16001) from the Ministry of Higher Education and Bright Spark Unit (from University of Malaya, Malaysia) which financially supported this work. References [1] V.L. Gole, P.R. Gogate, A review on intensification of synthesis of biodiesel from sustainable feed stock using sonochemical reactors, Chem. Eng. Process. Intens. 53 (2012) 1–9. [2] B. Sajjadi, A.R. Abdul Aziz, S. Baroutian, S. Ibrahim, Investigation of convection and diffusion during biodiesel production in packed membrane reactor using 3D simulation, J. Ind. Eng. Chem. 20 (2014) 1493–1504. [3] B. Sajjadi, A.A. Abdul Raman, S. Baroutian, S. Ibrahim, R.S.S. Raja Ehsan Shah, 3D Simulation of fatty acid methyl ester production in a packed membrane reactor, Fuel Process. Technol. 118 (2014) 7–19. [4] S.M. Hingu, P.R. Gogate, V.K. Rathod, Synthesis of biodiesel from waste cooking oil using sonochemical reactors, Ultrason. Sonochem. 17 (2010) 827–832. [5] P.R. Gogate, Cavitational reactors for process intensification of chemical processing applications: a critical review, Chem. Eng. Process. Intens. 47 (2008) 515–527. [6] V.B. Veljkovic´, J.M. Avramovic´, O.S. Stamenkovic´, Biodiesel production by ultrasound-assisted transesterification: state of the art and the perspectives, Renew. Sust. Energy Rev. 16 (2012) 1193–1209. [7] P.R. Gogate, V.S. Sutkar, A.B. Pandit, Sonochemical reactors: important design and scale up considerations with a special emphasis on heterogeneous systems, Chem. Eng. J. 166 (2011) 1066–1082.

473

[8] N.P. Vichare, P.R. Gogate, V.Y. Dindore, A.B. Pandit, Mixing time analysis of a sonochemical reactor, Ultrason. Sonochem. 8 (2001) 23–33. [9] V.G. Deshmane, P.R. Gogate, A.B. Pandit, Ultrasound-assisted synthesis of biodiesel from palm fatty acid distillate, Ind. Eng. Chem. Res. 48 (2008) 7923– 7927. [10] R. Lin, Y. Zhu, L.L. Tavlarides, Effect of thermal decomposition on biodiesel viscosity and cold flow property, Fuel 117, Part B (2014) 981–988. [11] B.C. Windom, T.J. Bruno, Pressure-controlled advanced distillation curve analysis of biodiesel fuels: assessment of thermal decomposition, Energy Fuel 26 (2012) 2407–2415. [12] H. Imahara, E. Minami, S. Hari, S. Saka, Thermal stability of biodiesel in supercritical methanol, Fuel 87 (2008) 1–6. [13] J. Xin, H. Imahara, S. Saka, Oxidation stability of biodiesel fuel as prepared by supercritical methanol, Fuel 87 (2008) 1807–1813. [14] R.H. Myers, D.C. Montgomery, C.M. Anderson-Cook, Response Surface Methodology: Process and Product Optimization Using Designed Experiments, John Wiley, 2009. [15] A. D6584-07, Test method for determination of free and total glycerine in B100 biodiesel methyl esters by gas chromatography, in: ASTM International, 100 Bar Harbor Drive, West Conshohocken, PA, USA, 2003. [16] T.G. Leighton, U.o.S.I.o. Sound, V. Research, Derivation of the Rayleigh–Plesset Equation in Terms of Volume, University of Southampton, 2007. [17] R. Loftstedt, K. Weninger, S. Putterman, B.R. Barber, Sonoluminescing bubbles and mass diffusion, Phys. Rev. E 51 (1995) 4400. [18] C.R. Wilke, P. Chang, Correlation of diffusion coefficients in dilute solutions, AIChE Journal 1 (2) (1955) 264–270. [19] R. Toegel, B. Gompf, R. Pecha, D. Lohse, Does water vapor prevent upscaling sonoluminescence?, Phys Rev. Lett. 85 (2000) 3165–3168. [20] D. Darnoko, M. Cheryan, Kinetics of palm oil transesterification in a batch reactor, JAOCS 77 (2000). [21] J.S. Krishnan, P. Dwivedi, V.S. Moholkar, Numerical investigation into the chemistry induced by hydrodynamic cavitation, Ind. Eng. Chem. Res. 45 (2006) 1493–1504. [22] R. Kuppa, V.S. Moholkar, Physical features of ultrasound-enhanced heterogeneous permanganate oxidation, Ultrason. Sonochem. 17 (2010) 123–131. [23] H.A. Choudhury, S. Chakma, V.S. Moholkar, Mechanistic insight into sonochemical biodiesel synthesis using heterogeneous base catalyst, Ultrason. Sonochem. 21 (2014) 169–181. [24] H.C. Ong, T.M.I. Mahlia, H.H. Masjuki, R.S. Norhasyima, Comparison of palm oil, Jatropha curcas and Calophyllum inophyllum for biodiesel: a review, Renew. Sust. Energy Rev. 15 (2011) 3501–3515. [25] E.N. Ali, C.I. Tay, Characterization of biodiesel produced from palm oil via base catalyzed transesterification, Procedia Eng. 53 (2013) 7–12. [26] T. Muppaneni, H.K. Reddy, S. Ponnusamy, P.D. Patil, Y. Sun, P. Dailey, S. Deng, Optimization of biodiesel production from palm oil under supercritical ethanol conditions using hexane as co-solvent: a response surface methodology approach, Fuel 107 (2013) 633–640. [27] S.Y. Giraldo, L.A. Rios, N. Suárez, Comparison of glycerol ketals, glycerol acetates and branched alcohol-derived fatty esters as cold-flow improvers for palm biodiesel, Fuel 108 (2013) 709–714. [28] B. Salamatinia, A.Z. Abdullah, S. Bhatia, Quality evaluation of biodiesel produced through ultrasound-assisted heterogeneous catalytic system, Fuel Process. Technol. 97 (2012) 1–8. [29] M. Shahabuddin, M.A. Kalam, H.H. Masjuki, M.M.K. Bhuiya, M. Mofijur, An experimental investigation into biodiesel stability by means of oxidation and property determination, Energy 44 (2012) 616–622. [30] R. Lin, Y. Zhu, L.L. Tavlarides, Mechanism and kinetics of thermal decomposition of biodiesel fuel, Fuel 106 (2013) 593–604. [31] H.A. Choudhury, R.S. Malani, V.S. Moholkar, Acid catalyzed biodiesel synthesis from Jatropha oil: mechanistic aspects of ultrasonic intensification, Chem. Eng. J. 231 (2013) 262–272.

Mechanistic analysis of cavitation assisted transesterification on biodiesel characteristics.

The influence of sonoluminescence transesterification on biodiesel physicochemical properties was investigated and the results were compared to those ...
2MB Sizes 0 Downloads 3 Views