Modeling of a Bioreactor-Extractor for the Production of Biodiesel Part 1: Dynamic Shortcut Model

Transcription

1 Contemporary Engineering Sciences, Vol. 11, 2018, no. 77, HIKARI Ltd, Modeling of a Bioreactor-Extractor for the Production of Biodiesel Part 1: Dynamic Shortcut Model Juan B. Restrepo Programa de Ingeniería Química, Facultad de Ingeniería Universidad del Atlántico, Cra 30 No Puerto Colombia, Atlántico Ramiro Betancourt G Departamento de Ingeniería Química Universidad Nacional de Colombia - sede Manizales Cra. 27 No , Manizales, Colombia Gerard Olivar Department of Electrical and Electronics Engineering and Computer Science Universidad Nacional de Colombia - sede Manizales Cra. 27 No , Manizales, Colombia Copyright c 2018 Juan B. Restrepo, Ramiro Betancourt G, and Gerard Olivar. This article is distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract As an emergent technology, the simultaneous reaction and extraction in the recent literature is scarce and mostly if not totally experimental. Therefore, this work proposes a shortcut model that takes advantage of as much simplifications as possible and the already available experimental data in the scientific literature to portray the dynamic behavior of a reactive-extractor for the enzymatic production of Biodiesel. The main purpose of this model is the theoretical design and assessment of this process intensification technology that can solve inherent problems to the enzymatic biodiesel transesterification such as inhibition by

2 3808 Juan B. Restrepo et al. methanol and glycerol, as also can reduce the process flowsheet in the industrial production of biodiesel. Keywords: Simultaneous Processes, simultaneous bioreactor-extractor systemm, Dynamical Model, Bifurcation Analysis. 1 Introduction The transformation of fats and vegetable oils (triglycerides) into long chain fatty acid esters by the reaction with an alcohol generally methanol or ethanol in presence of a catalyst (see Figure 1) is the base of industrial production of biodiesel and it is called transesterification [4]. CH2-OOC-R1 Catalyst R-OOC-R1 CH2-OH CH-OOC-R2 + 3R-OH R-OOC-R2 + CH-OH CH2-OOC-R3 R-OOC-R3 CH2-OH Figure 1: Transesterification reaction A successful biodiesel production process must assure a complete removal of the catalysts, alcohol, glycerol and a complete esterification of fatty acids. The flowchart for biodiesel production as any other chemical industrial process comprises three stages: the pretreatment stage, the reaction stage and the separation stage. In the first stage raw materials are processed (i.e. palm, rapeseed, sunflowers) to obtain vegetable oils. In the second stage the transesterification reaction, in presence of a catalyst, occurs producing a mixture of biodiesel and glycerol, which is separated in the third stage. The most relevant steps are: the reaction, where the different technologies and catalyst are used to obtain biodiesel with different degrees of efficiency, and the separation, where different technologies are used to purify biodiesel to the required degree. There are three different types of catalyst for biodiesel production: alkali, acid and enzymatic(lipases), For a basic catalyst production, either diluted sodium hydroxide (NaOH) or diluted Potassium hydroxide (KOH) are added as catalysts, the acid catalyst process uses an acid such as sulfuric or sulphonic acid [10], enzymatic biodiesel production, if immobilized enzymes are used [6], solves some of the inherent problems of alkali and acid catalysts, such as the formation of by-products and the catalyst removal. As an emergent technology, the simultaneous reaction and extraction in the recent literature is scarce and mostly if not totally experimental [3, 8]. Its main advantage is that the coupling of the reaction and the product separation allows continuous operation and withstands undesirable effects such as inhibition by

3 Modeling of a bioreactor-extractor for the production of biodiesel 3809 methanol and glycerol. This work proposes a theoretical shortcut model to address the study case of the biodiesel production from sunflower oil. 2 Shortcut Dynamic Model Development The theoretical assessment of new a technology such as the solvent free reactive extraction system should be, at least in the initial stage of a viability analysis, addressed by a shortcut model based only on the available data and the correlated physical properties found in the scientific literature. 2.1 Study Case One of the most studied feedstocks for enzymatic transesterification is the Sunflower oil [11, 2], that confers low temperature propertiesto the produced biodiesel [5]. The sunflower oil was chosen to test the developed model. 2.2 Initial Simplifications To approximate the sunflower oil and the biodiesel produced from the sunflower oil to one reactant and one product, the composition of sunflower oil must be analyzed. The sunflower oil main triglycerides are: LLL (33.8 mol %), OLL (29.3 mol %), PLL (10.6 mol %), OOL (8.4 mol %), SLL (7.4 mol %), POL (5.1 mol %), OOO (1.4 mol %) and SOL (2.4 mol %) as reported in Stolyhwo (1985) [13], where the main substituents of the glycerol molecule in the triglyceride are L for Linoleyl, O for Oleyl, P for Palmityl and S for Stearyl radicals. If sunflower oil is transesterified with methanol, the main component of the produced biodiesel will be methyl linoleate (72 mol %). The most accurate assumption under the current model simplifications would be the use of Trilinolein as a reactant and methyl linoleate as a product. Also, in a liquid mixture of Sunflower oil, biodiesel, glycerol and methanol, the molecular interaction of the sunflower oil and biodiesel with the other mixture components will be dominated by the linoleyl since it is abundant. 2.3 Kinetic Model Industrially biodiesel is produced produced from vegetable oils and during the reaction monoglycerides and diglycerides are formed as intermediate products, glycerol and biodiesel are the final products. For that reason Al-Zuhair [1] based on the mechanism showed in Figure 2 proposed a ping-pong model with inhibition by the alcohol:

4 3810 Juan B. Restrepo et al. Oil6Substrate6(S) Alcohol6(A) Glycerol6(G) Ester6(Bd) E E.A k 1 k -1 k 3 k -3 k 5 k 6 k 7 k -7 k E.S 2 k E.F E.F.A 4 E.Bd.G E.Bd Alcohol6(A) k -4 E Figure 2: Graphical representation of the mechanism steps of triglyceride esterbond transesterification from [1] υ = V max [S] 1 + K IS [S] + K S [S] [ ] 1 + [A] K IA + K A [A] where υ is the transesterification reaction rate, V max is the maximum reaction rate, K S is the reaction constant for the substrate, K A is the reaction constant for the alcohol, and K IA is the inhibition constant for the alcohol, S denotes the substrate (ester bond on the triglyceride), A denotes the alcohol and the brackets denote molar concentration (mol cm 3 ). The results obtained experimentally in [1] for immobilized Rhizomucor miehei (RM) lipase and immobilized Thermomyces lanuginosa (TL) lipase, using Sunflower oil and methanol as reactants, were fitted for equation 1 and are presented in the Table 1. (1) Table 1: Kinetic parameters Obtained in [1] for the Ping Pong kinetics V max K IS K S K A K AI (min 1 ) (cm 3 mol 1 ) (molcm 3 ) (molcm 3 ) (molcm 3 ) RM Lipase E TL Lipase E The correlated parameters of the kinetic model in [1] were obtained in a batch processes with the silica gel immobilized lipase, n-hexane as organic solvent, the reaction temperature was 40 o C and it was agitated with a magnetic stirrer at 200 rpm. 2.4 Start Stage The start of the reactive extraction system is performed with high stirring and in batch operation. As a result of the high stirring and the presence of the enzymes the reaction takes place converting the reactants into products, leading the global composition of the mixture into the immiscibility gap. While this stage takes place, the stirring is high enough to assume that the whole

5 Modeling of a bioreactor-extractor for the production of biodiesel 3811 system is behaving as a single phase system, no matter if the global composition is within the immiscibility gap. The reactor can be modeled by material balances. N T dz ROH = υ ROH V T, N T dz T rig = υ T rig V T, N T dz BioD = υ BioD V T (2) 1 z ROH z T rig z BioD z Glyc = 0 (3) Where N T is the total number of moles inside the reactor, V T is the reactor volume, z is the mole fraction inside the reactor and υ is the chemical reaction rate based on the specific reaction kinetics. The component mass balance equations for alcohol, biodiesel, triglyceride and glycerol are represented by the subindexes ROH, BioD, T rig and Glyc, respectively. 2.5 Operation Stage Once a successful starting stage is carried out, the stirring velocity is slowed down. The new stirring velocity will allow the formation of two liquid phases but it will allow the reaction to take place. Also the input and output valves of the reactor will be opened, allowing the continuous input of reactants and the continuous flow of products. In this stage the process can be modeled by material balances over the now open control volume of the stirred tank(see Figure 3) Reaction Phase II V II c Phase I V I Control Volume c Figure 3: Biodiesel Simultaneous Bioeactor-Extractor system for the production of N T d(z ROH) N T d(z T rig) N T d(z BioD) = N 0 R R + 1 (N I x I ROH + N II x II ROH) υ ROH V II (4) = N 0 1 R + 1 (N I x I T rig + N II x II T rig) υ T rig V II (5) = (N I x I BioD + N II x II BioD) + υ BioD V II (6)

6 3812 Juan B. Restrepo et al. 0 = 1 z ROH z T rig z BioD z Glyc (7) 0 = N 0 (N I + N II ) (8) Where N 0 is the molar flow entering the reactor, z, x I and x II are the global, phase one and phase two mole fractions respectively, N I and N II are the outgoing molar flows of phase one and phase two respectively and R is the molar relation of alcohol to triglyceride in the reactor inlet flow. Physical equilibrium will be assumed for simplicity. Therefore, there will be a relation between the global composition (z) and the composition in each phase(x I,x II ) for every component (i), the variable φ is introduced as the number of moles of phase one inside the system (N(in) I ) divided by the total moles inside the system (N T ) z i = φ x I i + (1 φ) x II i (9) (10) To represent the physical equilibrium, thermodynamic equilibrium relations are added: x I i γi I = x II i γi II (11) (12) where γ I and γ II are the activity coefficients of phase one and phase two respectively. Activity coefficients are highly non-linear functions that depend on both phases mole fractions and overall temperature. Empirical and fitted data to NRTL[12] activity coefficient model would be ideal; nevertheless, it is not always available therefore a viable yet not so accurate alternative are group contribution models, and specially their modification to improve the prediction of triglycerides and methanol mixtures ([?, 9]). The volume of the phase where the reaction occurs is unknown, but the total system volume is fixed. Some algebraic volume restrictions are needed to complete the model. Reaction volume, in this case the volume of phase two is dependent of the number of moles in phase two (N T (1 φ)) divided by the molar density of phase two (ρ II ), which can be approximated as the weighted average of the molar densities (ρ) of each one of the components of phase two. V II = N T (1 φ) ρ II (13) (14)

7 Modeling of a bioreactor-extractor for the production of biodiesel 3813 It also can be assumed that the molar relation between the outgoing currents of both phases is equal to the molar relation of the phases formed inside the reactive extraction system. N I N II = φ 1 φ (15) Finally, the mole fraction restrictions or percent restrictions, must be added to the model, for every phase k and every component i: 1 4 x k i (16) i=1 3 Conclusions There is a great amount of physical and chemical data already available in the scientific literature about enzymatic production of Biodiesel and its properties, this abundance of models and experimental data must serve as a tool for a chemical engineer to predict, with an acceptable degree of accuracy, the performance of a given technology without the need of conducting any physical experiments, in other words, to perform a theoretical assessment of the viability of a technology. Nevertheless, due to the great amount of different models and approaches it is not a trivial labor to discern which are the most fit for a preliminary evaluation. In this case this labor was performed by using transient material balances (Ordinary Differential Equations) coupled to algebraic restrictions that arose from equilibrium and volumetric assumptions. This resulted in a piecewise continuous differential algebraic model that can portray the operation of the reactor under high and moderate stirring and the transition between this two operation modes. Some of the possible applications of the model are the simulation of pulsed reaction and extraction, the start up of a reactor-extractor and the continuous operation of the reactor extractor. References [1] S. Al-Zuhair, Production of biodiesel by lipase-catalyzed transesterification of vegetable oils: A kinetics study, Biotechnology Progress, 21 (2005), [2] J. Calero, C. Verdugo, D. Luna, E. D. Sancho, C. Luna, A. Posadillo, F. M. Bautista, A. A. Romero, Selective ethanolysis of sunflower oil with lipozyme rm im, an immobilized rhizomucor miehei lipase, to obtain a

8 3814 Juan B. Restrepo et al. biodiesellike biofuel, which avoids glycerol production through the monoglyceride formation, New Biotechnology, 31 (2014), [3] S. Chattopadhyay and R. Sen, Development of a novel integrated continuous reactor system for biocatalytic production of biodiesel, Bioresource Technology, 147 (2013), [4] A. Demirbas, Biodiesel, Springer-Verlag, New York, [5] A. Demirbas, Characterization of biodiesel fuels, Energy Sources, Part A: Recovery, Utilization, and Environmental Effects, 31 (2009), [6] K. Dussan, C. Cardona, O. Giraldo, L. Gutierrez, V. Perez, Analysis of a reactive extraction process for biodiesel production using a lipase immobilized on magnetic nanostructures, Bioresource Technology, 101 (2010), [7] E. Batista, S. Monnerat, L. Stragevitch, C. G. Pina, C. B. Goncalves, A. J. Meirelles, Prediction of liquid- liquid equilibrium for systems of vegetable oils, fatty acids, and ethanol Journal of Chemical & Engineering Data, 44 (1999), [8] M. Ilmi, A. Kloekhorst, J. Winkelman, G. Euverink, C. Hidayat, H. Heeres, Process intensification of catalytic liquid-liquid solid processes: Continuous biodiesel production using an immobilized lipase in a centrifugal contactor separator, Chemical Engineering Journal, 321 (2017), [9] M. Lanza,W. B. Neto, E. Batista, R. J. Poppi and A. Meirelles, Liquidliquid equilibrium data for reactional systems of ethanolysis at K, Journal of Chemical and Engineering Data, 53 (2008), [10] J. M. Marchetti, V. U. Miguel, and A. F. Errazu, Possible methods for biodiesel production, Renewable and Sustainable Energy Reviews, 11 (2007), [11] M. Michael, Lipase catalyzed alcoholysis of sunflower oil, Journal of American Oil chemical Society, 67 (1990),

9 Modeling of a bioreactor-extractor for the production of biodiesel 3815 [12] H. Renon, J. Prausnitz, Estimation of parameters for the nrtl equation for excess gibbs energies of strongly nonideal liquid mixtures, Industrial & Engineering Chemistry Process Design and Development, 8 (1969), [13] A. Stolyhwo, H. Colin, G. Guiochon, Analysis of triglycerides in oils and fats by liquid chromatography with the laser light scattering detector, Analytical Chemistry, 57 (1985), Received: August 21, 2018; Published: September 12, 2018