ISSN 1451 - 9372(Print) ISSN 2217 - 7434(Online) JULY-SEPTEMBER 2017 Vol.23, Number 4, 441-596

www.ache.org.rs/ciceq Journal of the Association of Chemical Engineers of Serbia, Belgrade, Serbia

EDITOR-In-Chief Vlada B. Veljković Faculty of Technology, University of Niš, Leskovac, Serbia E-mail: [email protected]

ASSOCIATE EDITORS Jonjaua Ranogajec Srđan Pejanović Milan Jakšić Faculty of Technology, University of Department of Chemical Engineering, ICEHT/FORTH, University of Patras, Novi Sad, Novi Sad, Serbia Faculty of Technology and Metallurgy, Patras, Greece University of Belgrade, Belgrade, Serbia

EDITORIAL BOARD (Serbia) Đorđe Janaćković, Sanja Podunavac-Kuzmanović, Viktor Nedović, Sandra Konstantinović, Ivanka Popović Siniša Dodić, Zoran Todorović, Olivera Stamenković, Marija Tasić, Jelena Avramović

ADVISORY BOARD (International) Dragomir Bukur Ljubisa Radovic Texas A&M University, Pen State University, College Station, TX, USA PA, USA Milorad Dudukovic Peter Raspor Washington University, University of Ljubljana, St. Luis, MO, USA Ljubljana, Slovenia Jiri Hanika Constantinos Vayenas Institute of Chemical Process Fundamentals, Academy of Sciences University of Patras, of the Czech Republic, Prague, Czech Republic Patras, Greece Maria Jose Cocero Xenophon Verykios University of Valladolid, University of Patras, Valladolid, Spain Patras, Greece Tajalli Keshavarz Ronnie Willaert University of Westminster, Vrije Universiteit, London, UK Brussel, Belgium Zeljko Knez Gordana Vunjak Novakovic University of Maribor, Columbia University, Maribor, Slovenia New York, USA Igor Lacik Dimitrios P. Tassios Polymer Intitute of the Slovak Academy of Sciences, National Technical University of Athens, Bratislava, Slovakia Athens, Greece Denis Poncelet Hui Liu ENITIAA, Nantes, France China University of Geosciences, Wuhan, China

FORMER EDITOR (2005-2007) Professor Dejan Skala University of Belgrade, Faculty of Technology and Metallurgy, Belgrade, Serbia Journal of the Association of Chemical Engineers of Serbia, Belgrade, Serbia

Vol. 23 Belgrade, October-December 2017 No. 4

Chemical Industry & Chemical Engineering CONTENTS Quarterly (ISSN 1451-9372) is published quarterly by the Association of Chemical Majid Aliabadi, Effect of electrospinning parameters on the Engineers of Serbia, Kneza Miloša 9/I, 11000 Belgrade, Serbia air filtration performance using electrospun polyamide-6 nanofibers ...... 441 Editor: Rahim Shojaat, Afzal Karimi, Naghi Saadatjoo, Soheil Aber, Vlada B. Veljković Dye removal from artificial wastewater using hetero- [email protected] geneous bio-Fenton system ...... 447

Editorial Office: Ana Belščak-Cvitanović, Viktor Nedović, Ana Salević, Saša Kneza Miloša 9/I, 11000 Belgrade, Serbia Despotović, Draženka Komes, Miomir Nikšić, Branko Phone/Fax: +381 (0)11 3240 018 Bugarski, Ida Leskošek Ćukalović, Modification of E-mail: [email protected] functional quality of beer by using microencapsulated www.ache.org.rs green tea (Camellia sinensis L.) and Ganoderma mushroom (Ganoderma lucidum L.) bioactive For publisher: compounds ...... 457 Tatijana Duduković Danica Savanović, Radoslav Grujić, Sladjana Rakita, Alek- Secretary of the Editorial Office: sandra Torbica, Ranko Bozičković, Melting and crys- Slavica Desnica tallization DSC profiles of different types of meat ...... 473 Ali Sakin, Irfan Karagoz, Numerical prediction of short-cut Marketing and advertising: flows in gas-solid reverse flow cyclone separators ...... 483 AChE Marketing Office Kneza Miloša 9/I, 11000 Belgrade, Serbia Larissa Nayhara Soares Santana Falleiros, Bruna Vieira Phone/Fax: +381 (0)11 3240 018 Cabral, Janaína Fischer, Carla Zanella Guidini, Vic- elma Luiz Cardoso, Miriam Maria de Resende, Eloízio Publication of this Journal is supported by the Júlio Ribeiro, Improvement of recovered activity and Ministry of Education, Science and β Technological Development of the Republic of stability of the Aspergillus oryzae -galactosidase Serbia immobilized on duolite A568 by combination of immobilization methods ...... 495 Subscription and advertisements make payable Dung Van Nguyen, Harifara Rabemanolontsoa, Shiro Saka, to the account of the Association of Chemical Fed-batch fermentation of nipa sap to acetic acid by Engineers of Serbia, Belgrade, No. 205-2172- Moorella thermoacetica (f. Clostridium thermoace- 71, Komercijalna banka a.d., Beograd ticum) ...... 507 Computer typeface and paging: Soraya Seankham, Senad Novalin, Suwattana Pruksasri, Vladimir Panić Kinetics and adsorption isotherm of lactic acid from fermentation broth onto activated charcoal ...... 515 Printed by: Faculty of Technology and Metallurgy, Wang Ran, Tao Jinliang, Ling Yanli, Wei Feng, Liu Jidong, Research and Development Centre of Printing Total spray tray (TST) for distillation columns: A new Technology, Karnegijeva 4, P.O. Box 3503, generation tray with lower pressure drop ...... 523 11120 Belgrade, Serbia Yanjun Guan, Xiuying Yao, Guiying Wu, Kai Zhang, CFD investigation of slugging behavior in a gas-solids Abstracting/Indexing: fluidized bed ...... 529 Articles published in this Journal are indexed in Thompson Reuters products: Science Citation Snezana Sevic, Branko Grubac, Simulation of temperature- Index - ExpandedTM - access via Web of –pressure profiles and wax deposition in gas-lift wells ...... 537 ® SM Science , part of ISI Web of Knowledge Ainhoa Rubio-Clemente, Edwin Chica, Gustavo A. Peñuela, Kinetic model describing the UV/H2O2 photodegradation of phenol from water ...... 547

CONTENTS Continued Nina Trupej, Maša Knez Hrnčič, Mojca Škerget, Željko Knez, Investigation of the thermodynamic properties of the binary system vitamin K3/carbon dioxide ...... 563 Sema Akyalcin, Kinetic study of the hydration of propylene oxide in the presence of heterogeneous catalyst ...... 573 Ilayda Acaroglu Degitz, Muge Sari Yilmaz, Sabriye Piskin, Recycle of gold mine tailings slurry into MCM-41 mesoporous silica with high specific surface area ...... 581 Contents: Vol. 23, Issues 1–4, 2017 ...... 589 Author Index, Vol. 23, 2017 ...... 593

Activities of the Association of Chemical Engineers of Serbia are supported by:

- Ministry of Education, Science and Technological Development, Republic of Serbia - Hemofarm Koncern AD, Vršac, Serbia - Faculty of Technology and Metallurgy, University of Belgrade, Belgrade, Serbia - Faculty of Technology, University of Novi Sad, Novi Sad, Serbia - Faculty of Technology, University of Niš, Leskovac, Serbia - Institute of Chemistry, Technology and Metallurgy, University of Belgrade, Belgrade, Serbia

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 441−446 (2017) CI&CEQ

MAJID ALIABADI EFFECT OF ELECTROSPINNING Chemical Engineering Department, PARAMETERS ON THE AIR FILTRATION Birjand Branch, Islamic Azad PERFORMANCE USING ELECTROSPUN University, Birjand, Iran POLYAMIDE-6 NANOFIBERS SCIENTIFIC PAPER Article Highlights UDC 66.074.9:546.217:678.675 • Nanofibers of polyamide (PA) were fabricated via electrospinning process • The average fiber diameter of PA nanofibers was found to be 220 nm • The solution concentration showed major influence on the performance of PA nanofibrous filters • The highest aerosol filtration efficiency was found to be 96% at optimum conditions

Abstract Due to their high filtration efficiency and low basis weight nanofibrous filters are suitable for filtration applications. The objective of this study was to investigate the effect of electrospinning parameters including polymer solution concentration (10-15 wt.%), applied voltage (15-25 kV) and tip-collector distance (7.5-12.5 cm) on the filtration efficiency of polyamide (PA) nanofibers. The morphology of the PA nanofibers was characterized using scanning electron microscope (SEM) analysis. The SEM image results indicated that the average fiber diameter of PA nanofibers was 220 nm at PA solution concentration of 12.5 wt.%, applied voltage of 20 kV, tip-collector distance of 10 cm, flow rate of 0.5 mL h-1, temperature of 25 °C and humidity of 40%. The obtained results showed that the highest quality factor and efficiency of 7.02×10-2 Pa-1 and 96% were optimal parameters for nanofibrous filters during 30 and 240 min of electrospinning time, respectively. The presented study showed that the morphology optimization of nanofibers is an effective method for improvement of filtration performance. Keywords: electrospinning, polyamide, nanofibers, parameter study, air filtration.

The removal of air pollution due to the increasing basis weight are used in the air purification processes industrial air emissions and automobile exhaust is a [3-6]. The electrospun nanofibers due to their unique very important factor of preserving human health and properties such as high specific surface area and high environmental protection [1]. The fine particles less porosity with fine pores have a higher potential for than 2.5 μm in diameter, have a negative effect on the filtration applications [7-9]. In the electrospinning respiratory and pulmonary organs [2]. Air filtration is technique, high voltage is applied between a nozzle one of the most widely applied methods for removing and a collector where an electrically charged jet of suspended particles from the air. In recent studies, polymer or composite solution creates. The solvent the nano-fibrous filters prepared by electrospinning evaporates before reaching the collector and the process due to the high filtration efficiency and low fibers with micro or nano-sized diameters assemble on the collector [10]. According to the filtration theory,

a reduction in fiber size leads to better filtration effi- Correspondence: Chemical Engineering Department, Birjand ciency [11]. The diameter of nano-fibrous filters can Branch, Islamic Azad University, Birjand, Iran. E-mail: [email protected] be controlled by various parameters such as polymer Paper received: 9 May, 2016 solution properties, process and ambient conditions Paper revised: 17 November, 2016 Paper accepted: 14 December, 2016 [12–14]. Solution parameters such as viscosity, con- ductivity, and surface tension; process parameters https://doi.org/10.2298/CICEQ160509059A

441 M. ALIABADI: EFFECT OF ELECTROSPINNING PARAMETERS… Chem. Ind. Chem. Eng. Q. 23 (4) 441−446 (2017) such as voltage, distance between needle tip and the Matulevicius et al. [24] compared the performance of collector, and geometry of the collector; also ambient PA-6 and PA-6/6 nanofibers for air filtration applic- parameters such as temperature and humidity affect ation. Kim et al. [25] examined the filtration efficiency the formation of the fibers and fiber morphologies. of the Nylon 6 nanofilters and the commercial HEPA The morphological structure of nano-fibers plays a filter media. The Nylon 6 nanofibrous filter had less pivotal role on the filtration performance. pressure drop performance compared to HEPA filter. In recent studies, various polymeric nanofibers In the present study, the PA-6 nanofibrous filters were developed by electrospinning process for air fil- were prepared by the electrospinning process, and tration applications [15]. Li et al. produced a thin layer the effects of parameters including polymer solution of polylactic acid (PLA) nanofibrous media. They indi- concentration, applied voltage and tip-collector dis- cated that the filtration efficiency increased due to tance on the filtration efficiency of PA-6 nanofibers enhancement in the thickness of PLA nanofibers [16]. were investigated. Additionally the performance of PA Qin and Wang produced the poly(vinyl alcohol) nano- nanofibers was tested as a function of quality factor at -fibrous filter [17]. Vitchuli et al. [18] found that dep- different electrospinning times. ositing ultrathin Nylon 6 nanofiber mats improve the filtration efficiency to 99.5% without sacrificing air per- EXPERIMENTAL meability. Yun et al. prepared the polyacrylonitrile nano-fibers with diameters in the range of 270-400 Materials nm by electrospinning for use as a filter media. They Polyamide-6 was provided from DSM Company found that the electrospun filters had higher efficiency (DSM, Netherlands). Formic acid (98 wt.%) was compared to commercial filters [19]. Shin produced obtained from Sigma-Aldrich (Sigma Aldrich, St. polystyrene fibers from recycled expanded polysty- Louis, MO, USA). The polyester based fabric was sel- rene through electrospinning. They found that adding ected as a substrate. nanofibers to conventional micron-sized fibrous filter Preparation of PA-6 nanofibers media improved the separation efficiency of the filter media from 61 to 91% [20]. Faccini et al. prepared The PA-6 solutions were prepared by dissolving polyamide-6 (PA-6) nanofibrous filter by electrospin- PA-6 in formic acid under constant stirring for 3 h ning process [21]. Marsano et al. showed that PA-6 before electrospinning process. The prepared sol- nanofibers of 100 to 600 nm in diameter were suc- utions were poured into the 5 mL plastic syringe cessfully prepared. The electrospun PA-6 nanofibers equipped with a syringe needle. The syringe was exhibited good mechanical properties, such as a high- placed into a KD programmable syringe pump in -tensile strength (12±0.2 MPa) and elongation (300± order to control the solution feeding rate. High voltage ±50%) [22]. Guibo et al. [23] showed that the elec- was applied between a needle and a cylindrical trospun PA-6 nanofiber membrane had good filtration collector resulting in polyamide nanofibers being pro- potential for the microparticles with 0.3 µm diameter. duced on the polyester substrate. Schematic diagram of the electrospinning process is shown in Figure 1.

Figure 1. Schematic diagram of electrospinning process.

442 M. ALIABADI: EFFECT OF ELECTROSPINNING PARAMETERS… Chem. Ind. Chem. Eng. Q. 23 (4) 441−446 (2017)

The electrospinning parameters which were scopy (SEM, MV2300) after gold coating. The aver- investigated include solution concentration, applied age diameter and the diameter distribution for nano- voltage and tip-collector distance. The solution con- -fibers were obtained by image analyzer (Image- centration was adjusted at 10-15 wt.% under the fixed Proplus, Media Cybrernetics). voltage of 20 kV and the fixed distance of 10 cm. The In order to estimate the average diameter, at voltage was varied in the range of 15–25 kV over a least 50 different fiber segments were randomly sel- fixed collection distance of 10 cm and optimum sol- ected and their diameters were measured out to gen- ution concentration. The fibers were collected at dis- erate an average fiber diameter. tance of 7.5–12.5 cm under optimum values of solution concentration and applied voltage. The nano- RESULTS AND DISCUSSION -fibrous media of different thickness were made by varying the electrospinning time in the range of 15-60 Effect of solution concentration on the filtration min. The flow rate and speed collector were 0.5 mL h-1 efficiency of PA-6 nanofibers and 1000 rpm, respectively. Thinner fibers have higher filtration efficiency. Solution concentration is considered to be among Filtration process most important parameters determining fiber mor- The schematic diagram of the filtration process phology and diameter. For lower concentrations, the is shown in Figure 2. A mono-dispersed electrospray lower viscosity of solution resulted in the formation of aerosol generator (TSI ESP-3480) was used to bead fibers [26]. By increasing concentration, the ensure uniform aerosol concentration. The particles fibers became homogeneous with thinner diameters. generated by the machine had a particle size ranging Increased stability of solution jet resulted in the fab- from 100 to 300 nm. The upstream and downstream rication of homogeneous fibers. The higher concen- aerosol concentrations were determined by a conden- trations lead to larger diameters. For the higher con- sation particle counter (Model 5.412, GRIMM Co). The centrations, the movement of the polymer chains is aerosol filtration efficiency (η) was defined as follows: restricted thus limiting the ability of the jet to stretch, CC− leading to the thicker fibers. Therefore, the optimiz- η(%)= 100 up down (1) C ation of solution concentration for fabrication of homo- up geneous thinner fibers is very important. The influ- where Cup and Cdown are the aerosol concentrations ence of PA solution concentration on the filtration before and after passing through the filtration media, efficiency of PA-6 nanofibrous filters was investigated respectively. The aerosol filtration efficiency was at different times of electrospinning process. The measured for the face velocity of 5 cm s-1. The pres- results are shown in Figure 3. As shown, the filtration sure drop was measured by a pressure sensor (model efficiency of the 300 nm in diameter atmospheric par- P300-5-in-D, Pace Scientific Inc., USA) before and ticle for the substrate was only 28%.At the same time, after filtration. Prior to filtration tests, all nanofibrous the electrospun PA nanofibers significantly increased filters were discharged by a filter charge neutralizer. the filtration efficiency. Each experiment was performed in triplicate and The results revealed that the concentration of the results were given as averages. 12.5% was appropriate for fabrication of homogeneous fibers with thinner fibers as well as higher filt- Characterization of nanofibrous filter ration efficiency for using in filtration process. The filt- The morphology of the optimized nanofibrous ration efficiency and pressure drop for 10, 12.5 and filter was determined using scanning electron micro- 15% PA concentrations were found to be 72.3, 86.3,

Figure 2. Schematic diagram of filtration process.

443 M. ALIABADI: EFFECT OF ELECTROSPINNING PARAMETERS… Chem. Ind. Chem. Eng. Q. 23 (4) 441−446 (2017) and 76.1%, and 25.9, 28.6 and 37.1 Pa, respectively, of the electrical field increased the instabilities of jet after 30 min of electrospinning time. The filtration solution which led to increase in the fiber diameters efficiency did not change significantly by further coat- and decrease in the filtration efficiency. Therefore, the ing of nanofibers. Furthermore, the higher thickness applied voltage of 20 kV was selected for further exp- of nanofibers produced the higher pressure drop eriments. which was not appropriate for filtration process. Although, the pressure drop for 10% PA was lower than that of 12.5% PA; the quality factor for nano- -fibrous filter 12.5% was at the maximum. Therefore, the 12.5% PA concentration for 30 min electrospinning time was selected for further experiments.

Figure 4. Effect of applied voltage on the filtration efficiency of PA nanofibrous filters.

Effect of tip-collector distance on the filtration efficiency of PA-6 nanofibers Distance between the needle tip and the collector affected the deposition time, evaporation rate of the solvent and instability interval [28]. For the very Figure 3. Effect of solution concentration on the performance of short distance, electrical field became very strong PA nanofibrous filters. which increased the instability of the jet solution and enabled multiple jets emerged from one nozzle Effect of applied voltage on the filtration efficiency of resulting in the formation of bead fibers. Increase in PA-6 nanofibers the tip–collector distance led to formation of the fibers The applied voltage during electrospinning pro- without presence of the beads. For the very long dis- cess is one of the most essential parameters that tance, the electrical field strength became very weak affect fibers morphology. When the applied voltage resulting in the diameter fiber increase [28]. There- increases, the electrical field strength also increases, fore, the optimization of the distance between tip which causes fiber diameters to decrease [27]. needle and collector had a very important role in the Applied higher voltage causes the acceleration of formation of homogeneous and fine fibers. The fine charge on the spinning solution which leads to inc- fibers resulted in the higher filtration efficiency. Figure rease in the fiber diameter. Therefore, the optimiz- 5 shows the influence of tip-collector distance on the ation of applied voltage for formation of thinner nano- PA filtration efficiency. As shown, the maximum filt- -fibers and subsequently the maximum filtration effi- ration efficiency for the formation of PA fibers was ciency is necessary. The applied voltage effect on the achieved at tip-collect distance of 10 cm. Therefore, performance of prepared PA nanofibrous filters was the distance of 10 cm was selected for the formation evaluated during filtration process. The results are of PA nanofibers as filter media. shown in Figure 4. As shown, the optimal filtration Quality factor efficiency was achieved at applied voltage of 20 kV The quality factor is often used to evaluate the for fabrication of PA nanofibers. For the voltage of 15 filtration performance, defined as: kV, the applied electrical force is not sufficient enough to form the perfect fibers, which resulted in the lower −−ln(1η ) q = (2) filtration efficiency compared to applied voltage of 20 f Δp kV. At voltage of 20 kV, the formation of homogen- η Δ eous thinner fibers resulted in the highest filtration where is the filtration efficiency and p is the efficiency. At the voltage of 25 kV, the higher strength pressure drop across the filter. The electrospinning

444 M. ALIABADI: EFFECT OF ELECTROSPINNING PARAMETERS… Chem. Ind. Chem. Eng. Q. 23 (4) 441−446 (2017) time is the main factor that influenced the nanofiber Filtration Pressure drop Quality factor Sample performance. In order to investigate the electrospin- efficiency, % Pa 10-2 Pa-1 ning time on the performance of nanofibers, the fiber PA/30 86.3±0.2 28.3±0.1 7.02 media collection time was set for 30, 60, 120, 180 and PA/60 88.9±0.1 32.4±0.1 6.78 240 min under the optimum conditions of electrospin- PA/120 92.0±0.3 37.9±0.2 6.66 ning process. The results are presented in Table 1. PA/180 94.1±0.2 45.2±0.1 6.26 The penetration of nanoparticles through the electro- PA/240 96±0.3 54.5±0.1 5.91 spun filter media was reduced by increasing the filter thickness. Although the pressure drop increased by SEM image of PA nanofibers increasing the electrospinning time and nanofibrous The SEM image and distribution diameter of PA filter thickness (Table 1), the highest filtration effi- electrospun nanofibers at solution concentration of ciency was found to be 96% for PA nanofibrous filter 12.5 wt.%, applied voltage of 20 kV, tip-collector dis- during 240 min of electrospinning time. However, the tance of 10 cm and flow rate of 0.5 mL h-1 were shown higher pressure drop led to decrease in the quality in Figure 6. As shown, the homogeneous fibers with factor of PA/240 nanofibers compared to PA/30 nano- average diameter of 220 nm formed on the collector. -fibers. Therefore, the PA nanofibers for 30 min electrospinning time were more suitable for filtration CONCLUSION applications. In the presented work, the homogeneous PA nano-fibers were fabricated via electrospinning process for application to the filtration media. The investigation of the electrospinning parameters including solution concentration, applied voltage and tip-collector distance on the filtration efficiency of PA nano- -fibers indicated that the solution concentration showed great influence on the performance of PA nano-fibrous filters. Based on the results, the optimal conditions of electrospinning process were found to be solution concentration of 12.5%, applied voltage of 20 kV and tip-collector distance of 10 cm. For the optimal conditions, the obtained maximum quality factor and effi- × -2 -1 Figure 5. Effect of tip-collector distance on the filtration ciency were 7.02 10 Pa and 96% for PA nano- efficiency of PA nanofibrous filters. -fibers at 30 and 240 min electrospinning time intervals. The SEM image showed that the homogeneous Table 1. Filtration performance of PA nanofibers at different nanofibers with average diameter of 220 nm were electrospinning times fabricated on the collector at optimal conditions of electrospinning process.

10 Frequency (%)

0 100 150 200 250 300 350 400 Fiber diameter (nm) Average diameter = 220 nm Figure 6. SEM image and fiber diameter distribution of PA nanofibers.

445 M. ALIABADI: EFFECT OF ELECTROSPINNING PARAMETERS… Chem. Ind. Chem. Eng. Q. 23 (4) 441−446 (2017)

REFERENCES [15] Y.C. Ahn, S.K. Park, G.T. Kim, Y.J. Hwang, C.G. Lee, H.S. Shin, J.K. Lee, Curr. Appl. Phys. 6 (2006) 1030- [1] G. Oberdörster, M.J. Utell, Environ. Health Perspect. 110 –1035 (2002) A440-A441 [16] J. Li, X. Shi, F. Gao, L. Liu, R. Chen, C.Y. Chen, Z. [2] N. Wang, Y. Si, N. Wang, G. Sun, M. El-Newehy, S.S. Al- Zhang, Sci. China Tech. Sci. 57 (2014) 239-243 -Deyab, B. Ding, Sep. Purif. Technol. 126 (2014) 44-51 [17] X. Qin S. Wang, J. Appl. Polym. Sci. 102 (2006) 1285- [3] Q. Zhang, J. Welch, H. Park, C.Y. Wu, W. Sigmund, –1290 J.C.M. Marijinissen, J. Aerosol. Sci. 41 (2010) 230-236 [18] N. Vitchuli, Q. Shi, J. Nowak, M. McCord, M. Bourham, X. [4] J. Wang, S.C. Kim, D.Y.H. Pui, J. Aerosol Sci. 39 (2008) Zhang, J. Appl. Polym. Sci. 116 (2010) 2181-2187 323-334 [19] K.M. Yun, C.J. Hogan, J.Y. Matsubayashi, M. Kawabe, F. [5] R.S. Barhate, R. Ramakrishna, J. Membr. Sci. 296 (2007) Iskandar, K. Okuyama, Chem. Eng. Sci. 62 (2007) 4751- 1-8 –4759 [6] R.S. Barhate, C.K. Loong, S. Ramakrishna, J. Membr. [20] C. Shin. J. Colloid Interf. Sci. 302 (2006) 267-271 Sci. 283 (2006) 209-218 [21] M. Faccini, D. Amantia, S Vazquez-Campos, C. Vaquero, [7] P. Heikkila, A. Taipale, M. Lehtimaki, A. Harlin, Polym. J. M. López de Ipiña, L. Aubouy, J. Phys. 304 (2011) 1-7 Eng. Sci. 48 (2008) 1168-1176 [22] E. Marsano, L. Francis, F. Giunco, J. Appl. Polym. Sci. [8] L. Shi, X. Zhuang, X. Tao, B. Cheng, W. Kang, Fiber. 117 (2010) 1754-1765 Polym. 14 (2013) 1485-1490 [23] Y. Guibo, Z. Qing, Z. Yahong, Y. Yin, Y. Yumin. J. Appl. [9] X. Mao, Y. Si, Y. Chen, L. Yang, F. Zhao, B. Ding, J. Yu, Polym. Sci. 128 (2013) 1061-1069 RSC Adv. 2 (2012) 12216-12223 [24] J. Matulevicius, L. Kliucininkas, D. Martuzevicius, E. [10] S.A. Tehron, E. Zussman, A.L. Yarin, Polymer 5 (2004) Krugly, M. Tichonovas, J. Baltrusaitis. J Nanomater 2014 2017-2030 (2014) 14 [11] K.M. Yun, A.B. Suryamas, F. Iskandar, L. Bao, H. [25] G.T. Kim, Y.C. Ahn, J.K. Lee. Korean J. Chem. Eng. 25 Niinuma, K. Okuyama, Sep. Purif. Technol. 75 (2010) (2008) 368-372 340-345 [26] J.M. Deitzel, J. Kleinmeyer, D. Harris, N.C. Beck Tan, [12] P. Heikkila, A. Harlin, Eur. Polym. J. 44 (2008) 3067-3079 Polymer 42 (2001) 261-72 [13] U.S. Sajeev, K.A. Anand, D. Menon, S. Nasir, Bull. Mater. [27] M. Aliabadi, M. Irani, J. Ismaeili, S. Najafzadeh, J. Taiwan Sci. 31 (2008) 343-351 Inst. Chem. Eng. 45 (2014) 518-526 [14] S. Koushkbaghi, P Jafari, J. Rabiei, M. Irani, M. Aliabadi. [28] T. Subbiah, G.S. Bhat, R.W. Tock, S. Parameswaran, Chem. Eng. J. 301 (2016) 42-50 S.S. Ramkumar, J. Appl. Polym. Sci. 96 (2005) 557-569.

MAJID ALIABADI UTICAJ PARAMETARA ELEKTROPREDENJA NA Chemical Engineering Department, PERFORMANSE FILTRACIJE VAZDUHA POMOĆU Birjand Branch, Islamic Azad NANOVLAKANA POLIAMIDA-6 University, Birjand, Iran

Nanovlaknasti filtri, zbog njihove visoke efikasnosti filtracije i male težine, su pogodni za filtraciju. Cilj ovog rada je bilo istraživanje uticaja parametara elektropredenja uključujući NAUČNI RAD koncentraciju rastvora polimera (10-15 %), napon (15-25 kV) i rastojanje vrh-kolektor (7,5-12,5 cm) na efikasnost filtracije poliamidnih nanovlakana. Morfologija poliamidnih nanovlakana je okarakterisana pomoću skenirajuće elektronske mikroskopije (SEM). Rezultati SEM analiza pokazuju da je prosečan prečnih poliamidnih nanovlakana 220 nm pri koncentraciji rastvora poliamida 12,5 %, napona 20 kV, rastojanja vrh-kolektor 10 cm, protoka 0,5 ml/h, temperature 25 ˚C i vlažnosti 40 %. Dobijeni rezultati su pokazali da su pri optimalnim uslovima parametara postižu najveći faktor kvaliteta i efikasnost 7,02 × 10-2 Pa-1 i 96 % u toku 30 i 240 min elektropredenja, redom. Ovaj rad je pokazao da je optimizacija morfologije nanovlakana efikasna metoda za poboljšanje performansi filtracije.

Ključne reči: elektropredenje, poliamid, nanovlakna, istraživanje parametara, filtracija vazduha

446 Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 447−456 (2017) CI&CEQ

RAHIM SHOJAAT1 DYE REMOVAL FROM ARTIFICIAL AFZAL KARIMI2 1 WASTEWATER USING HETEROGENEOUS NAGHI SAADATJOO BIO-FENTON SYSTEM SOHEIL ABER3

1Department of Applied Chemistry, Article Highlights Faculty of Chemistry, Semnan • The heterogeneous bio-Fenton catalyst (GOx/MnFe2O4/calcium alginate) was prepared University, Semnan, Iran • Various kinetic models for adsorption and decolorization of AR14 by using catalyst were 2Department of Biotechnology, studied • Faculty of Advanced technologies The AR14 dye removal by bio-Fenton catalyst was more than 60% for 3 h • The obtained value of the Michaelis-Menten kinetic constant for decolorization of in medicine, Iran University of AR14 dye was 114 μM Medical Sciences, Tehran, Iran 3 Research Laboratory of Abstract Environmental Protection In the present study, GOx/MnFe O /calcium alginate nano-composite was Technology, Department of Applied 2 4 Chemistry, Faculty of Chemistry, prepared by the trapping enzyme/nanoparticles in calcium alginate. The pre- University of Tabriz, Tabriz, Iran pared absorbent was applied for decolorization of artificial dye wastewater of acid red 14 (AR14) by heterogeneous bio-Fenton system. Kinetic and isotherm SCIENTIFIC PAPER studies were carried out. The decolorization of acid red 14 followed the Micha- elis-Menten, pseudo-first order and pseudo-second order kinetic models. Good UDC 628.3:66:544.4 correlation coefficients were obtained by fitting the experimental data to Micha- elis-Menten and pseudo-second order kinetic models. The adsorption isotherms were described by Langmuir, Freundlich and Temkin isotherms. Among the three isotherm models, the Freundlich model was fitted with the equilibrium data

obtained from adsorption of AR14 onto MnFe2O4/calcium alginate; while Temkin isotherm gave the best correlation for adsorption on MnFe2O4 nanoparticles. The effect of various parameters such as initial pH of solution, initial dye concen-

tration, and contact time on the adsorption of AR14 on MnFe2O4 and MnFe2O4/ /calcium alginate as well as dye enzymatic decomposition was studied. The decolorization of AR14 with initial concentration of 10 mg.L−1 by using GOx/

/MnFe2O4/calcium alginate was 60.17%. Keywords: adsorbents, decolorization, enzyme, isotherm, kinetic.

Most industries use dyes and pigments to color structure, are known to be one of the most important their products. Dyes are aromatic organic colorants, groups of synthetic colorants that are extensively which have wide applications in textile, plastic, rub- used in various industries [2-4]. The effluents dis- ber, paper and food industries. [1]. Azo dyes, which charge from dyeing industries is highly colored with a contain at least one –N=N– group in their chemical large amount of suspended organic solids [5]. There are many studies on techniques for removal of dyes from wastewater [6-9]. Adsorption of pollutants is Correspondence: A. Karimi, Department of Biotechnology, superior to other methods for wastewater treatment in Faculty of Advanced technologies in medicine, Iran University of terms of low cost of raw materials, simplicity of des- Medical Sciences, Tehran, Iran; N. Saadatjoo, Department of Applied Chemistry, Faculty of Chemistry, Semnan University, ign, and ease of operation [10-14]. In the recent dec- Semnan, Iran. ades, magnetic particles such as iron oxide nanopar- E-mail: [email protected] (A. Karimi); ticles, especially magnetite, hematite, and spinel fer- [email protected] (N. Saadatjoo) Paper received: 21 June, 2016 rite are used for this purpose since they are easy to Paper revised: 23 October, 2016 collect and separate from the aqueous environments Paper accepted: 15 December, 2017 because of their magnetic property [15,16]. Biological https://doi.org/10.2298/CICEQ160621058S

447 R. SHOJAAT et al.: DYE REMOVAL FROM ARTIFICIAL WASTEWATER… Chem. Ind. Chem. Eng. Q. 23 (4) 447−456 (2017) adsorbents are another type of adsorbents that have erate OH radical species based on a heterogeneous low preparation cost, simple design, and are easy to Fenton system [27]. The possible mechanism for 2+ operate with biological materials such as chitosan and H2O2 decomposition by Mn is outlined below. Man- alginate [17-21]. Alginate is a polysaccharide, which ganese exhibit the redox pair Mn2+/Mn3+, which could is widely used for removal of heavy metals and org- produce radicals according to the following reactions: anic dyes from wastewater [22,23]. Enzymatic pro- 2+ 3+• - Mn + H22 O → Mn + HO + OH (1) cesses are frequently used to decolorize wastewater surf surf due to their cost effectiveness. The enzymatic decol- •• HO22 + HO → HO 2 + HOO (2) orization of some azo dyes cause the formation of 3+• → 2+ + non-specific free radicals without direct cleavage of Mnsurf + HOO Mn surf + H + O2 (3) the azo bonds thus avoiding the production of highly 3+• 2+ + Fe + HOO → Fe + H + O2 (4) toxic aromatic amines [24]. In addition, dye degrad- surf surf ation using enzymes has received great attention, . . The above generated HOO , HO and O2 react because it has high efficiency to decolorize dyes [25]. with organic dye (RH): Glucose oxidase (GOx) is one of the most widely • used available enzymes, which is known as oxido- R-H + HOO → X + H22 O + CO (5) reductase. It catalyzes the oxidation of β-D-glucose to Where X is degraded organic product. The reduction gluconic acid by utilizing molecular oxygen as an of Mn3+ by Fe2+ is thermodynamically favorable and electron acceptor with simultaneous production of produces Mn2+ ions which again continue the catalytic hydrogen peroxide (H2O2). Immobilization of enzymes cycle of H2O2 decomposition: on insoluble support provides a possibility for their 2+ 3+→ 3+ 2+ continuous or multiple uses. The immobilization of Fesurf + Mn surf Fe surf + Mn surf (6) enzymes on porous-supports through adsorption is a The GOx causes the production of in-situ hydro- simple method, and it is one of the first enzyme immo- gen peroxide during electron transfer reaction from bilization techniques [26]. The ease of immobilization, substrate to oxygen. The role of MnFe O in dye deg- lack of chemical modification and enhancement in 2 4 radation is due to the presence of Mn2+ in this mat- stability are some of the advantages offered by the erial’s structure which strongly favors the peroxide adsorption procedure. However, this technique has decomposition [27,28]. disadvantages, such as low linking energy between The present study reports the preparation of the enzyme and the support, which may cause enz- GOx/MnFe O /calcium alginate nanocomposite for yme desorption in presence of the substrate or when 2 4 simultaneous adsorption and degradation of dye polit is exposed to variations in the temperature, pH and lutant acid red 14 (AR14) from aqueous solutions by ionic strength. Nevertheless, the immobilization by heterogeneous bio-Fenton reactions. The enzymatic adsorption is one of the most used techniques in the decolorization followed the Michaelis-Menten kinetic attainment of insoluble support. The Fenton reagent, 2+ model. The effect of various parameters such as ini- a mixture of Fe and H2O2,is one of the most active tial pH of solution, initial dye concentration, and con- systems for oxidation of organics in water. Fenton tact time on the adsorption of AR14 on MnFe O and and Fenton-like technology have been extensively 2 4 MnFe O /calcium alginate was studied. Also, the ads- studied as one of the best options for the degradation 2 4 orption kinetics and equilibrium characteristics of the of various recalcitrant organic pollutants in water. adsorption on MnFe O and MnFeO /calcium algi- However, there is a non-ignorable drawback of the 2 4 2 4 nate have been studied. traditional homogeneous Fenton and that is the diffic- ulty to remove the iron sludge after the treatment, which MATERIALS AND METHODS may cause secondary pollution. Recently, heterogeneous Fenton-like systems using solid catalysts have Materials been developed. Most studies of heterogeneous Fen- ton-like catalysts have been focused on iron sup- GOx (EC.1.1.3.4) Aspergillusniger, glucose, ⋅ ⋅ ported materials, such as iron oxide minerals. The NaOH (99%), HCl, FeCl3 6H2O, MnCl2 4H2O and development of active heterogeneous Fenton sys- CaCl2 were obtained from Merck. Alginic acid sodium tems is of considerable interest since it could offer salt was obtained from Sigma-Aldrich and AR14 were some advantages, such as no need of acid or base, purchased from BuyakhSazan Company. and no sludge generation. The Mn2+, like Fe2+, played an active role in the decomposition of H2O2 to gen-

448 R. SHOJAAT et al.: DYE REMOVAL FROM ARTIFICIAL WASTEWATER… Chem. Ind. Chem. Eng. Q. 23 (4) 447−456 (2017)

Preparation of nanocomposites of aqueous solution of AR14 was measured at wavelength of 514 nm, which has been determined as the MnFe2O4 nanoparticles were prepared by co-pre- wavelength of maximum dye absorption, for a spe- cipitation phase inversion method from FeCl3⋅6H2O cified interval using UV–Vis spectrophotometer (1700 and MnCl2⋅4H2O as raw materials with molar ratio of Fe3:Mn2+ = 2:1. 100 mL of the solution containing 0.1 UV–Vis Shimadzu, Japan). The percent of the decol- M Mn2+ and 0.2 M Fe3+ were added drop wise to 100 orization was calculated by Eq. (7): mL of 3 M NaOH and the obtained solution was stir- A0 − At ° Decolorization (%) = 100 (7) red for 2 h at 95 C on a magnetic stirrer. The A0 obtained powder was washed several times with distilled water and dried for 12 h at 60 °C [29]. For each where A0 and At are the absorbance of the sample of the adsorption tests, 0.2 g of the obtained powder solution at times 0 and t, respectively. was used. MnFe2O4 particles were dispersed in water Kinetic and isotherm studies and cast onto a copper grid to study their size and The kinetic parameters of the enzyme for the morphology by transmission electron microscopy dye were obtained after determination of the velocity (TEM) using a Philips–CM300–150 kV microscope. for different concentrations of the dye. Afterwards, a To prepare MnFe O /calcium alginate, 0.2 g of 2 4 Michaelis-Menten curve was plotted as the obtained MnFe O in 10 mL of distilled water was mixed with 2 4 velocity (V) vs. dye concentrations (S). Then, the 20 mL of 2 wt.% sodium alginate solution. The mix- Michaelis-Menten constant (k ) and maximal velocity ture was transferred to 2 wt.% solution of calcium M (V ) were calculated using Lineweavere-Burk trans- chloride drop wise to obtain insoluble granular cal- max formation of the Michaelise-Menten equation [34]. cium salt. To prepare the enzymatic nanocomposite, The kinetics of adsorption without enzyme was the same process was followed except that the GOx determined by analyzing the adsorption of the dye enzyme was immobilized on MnFe O . 2 4 from aqueous solution at different time intervals.

Immobilization of enzyme on MnFe2O4 and assay of For determination of the adsorption isotherms, activities dye solutions of different concentrations were agitated To immobilize the enzyme on the support, 0.2 g with known amount of adsorbents till the equilibrium was reached at temperature 30 °C. The amount of the MnFe2O4 was added to 10 mL GOx solution (882 U/mL) in 0.1 M acetate buffer (pH 5.5). Then, it was adsorbed dye was calculated using Eq. (8): ° stirred in the incubator shaker at 120 rpm and 30 C (CCV0e - ) qe = (8) for 60 min [30]. The immobilization amount was cal- m culated by measuring the difference between the initial activity of the solution and the supernatant. The where qe, C0, Ce, V and m are the amount of dye −1 activity of GOx enzyme was measured by the method adsorbed per unit mass of the adsorbent (mg g ) at of Shoaebargh et al. [31]. The amount of GOx immo- equilibrium, the initial, equilibrium concentrations of the dye (mg L-1), solution volume (L) and adsorbent bilized in GOx/MnFe2O4/calcium alginate nanocom- weight (g), respectively. posite was calculated as 5850 U/gMnFe2O4.

Point of zero charge of the MnFe2O4 nanoparticles RESULTS AND DISCUSSION The determination of the point of zero charge pH Characterization of the nanoparticles (pHpzc) of MnFe2O4 nanoparticles at 30 °C in the 0.1

M NaNO3 solutions was conducted by using the Meh- Figures 1 a-c show the TEM images of the dizadeh et.al. method with a slight change [32]. The obtained nanomaterials synthesized by co-precipit- value of pHpzc ≈7.1 was obtained according to our pre- ation phase inversion method. Figure 1a shows that vious work [33]. the obtained material was composed of particles structures. It was found that the obtained material had Decolorization process cubic morphology (Figure 1b). Besides, the average The experiments were conducted in 250 mL size of each flank was measured to be about 39 nm Erlenmeyer flasks, containing 50 mL dye solution with (Figure 1c). The other characterizations of MnFe2O4 given concentration and a specific amount of different nanoparticles synthesized have been presented in ° nanocomposites at the set temperature of 30 C and our previous works [33]. 120 rpm. To the catalyst (GOx/MnFe2O4/calcium alginate) containing GOx enzyme, 80 mM glucose was added as a substrate to the solution. The absorption

449 R. SHOJAAT et al.: DYE REMOVAL FROM ARTIFICIAL WASTEWATER… Chem. Ind. Chem. Eng. Q. 23 (4) 447−456 (2017)

Effect of pH The adsorption capacity of the adsorbents were studied by setting the initial pH in the range of 3–7 by HCl (0.1 M) and NaOH (0.1 M) by using a pH-meter at an adsorbent dose of 0.2 g, 50 mL AR14 solution, initial concentration 50 mg.L−1 and temperature 30 °C. Figure 2 shows the removal of AR14 by using

MnFe2O4, MnFe2O4/calcium alginate and GOx/

/MnFe2O4/calcium alginate at different pHs. The graph indicated clearly that the adsorption process was pH dependent and the maximum uptake of the dye occured at pH 3. Thus, the pH 3 was selected as optimum pH for the following study. The effect of the initial pH depends on the type of pollutants and pzc of

the adsorbent. At pH below pHpzc ≈7.1, the electrostatic interaction between the anions of AR14 and adsorbent positive surface increased, which led to increase of the decolorization of anionic AR14 on

MnFe2O4 with decrease in the pH. The reduction of AR14 removal in higher pH could be contributed to the increase of repulsive forces between the negative

charged surface MnFe2O4 and the anionic AR14 dye. Also, the results showed that the removal of anionic

AR14 dye by GOx/MnFe2O4/calcium alginate was

higher than the removal by using MnFe2O4/calcium

alginate and MnFe2O4 due to the enzymatic degradation of the dye. Effect of contact time The effect of contact time on the percentage of color removal was studied at different time intervals from 5 to 60 min at an adsorbent dose 0.2 g, 50 mL AR14 solution, initial concentration 50 mg.L−1, pH 3 and temperature 30 °C. The results in terms of percentage decolorization are presented in Figure 3. It was observed that the adsorption capacity of AR14 was increased with increase in the contact time. As Figure1. TEM images of MnFe O synthesized by 2 4 could be seen from Figure 3, the removal efficiency of co-precipitation method.

Figure 2. Effect of solution pH on the adsorption of AR14 (50 mL AR14 solution, initial concentration 50 mg. L−1, 0.2 g adsorbent, 120 rpm, 1 h).

450 R. SHOJAAT et al.: DYE REMOVAL FROM ARTIFICIAL WASTEWATER… Chem. Ind. Chem. Eng. Q. 23 (4) 447−456 (2017)

Figure 3. Effect of contact time on the adsorption of AR14 (50 mL AR14 solution, initial concentration 50 mg. L−1, 0.2 g adsorbent, initial pH 3.0, 30 °C ,120 rpm).

AR14 onto adsorbents by adsorption was initially rapid Also, the results showed that the removal of anionic and then slowed down gradually and the amount of AR14 dye by using GOx/MnFe2O4/calcium alginate dye removed by GOx/MnFe2O4/calcium alginate was was higher than the one obtained by using MnFe2O4/ more than that removed by MnFe2O4/calcium alginate /calcium alginate and MnFe2O4, due to the enzymatic and MnFe2O4 at the same conditions. degradation of the dye. The difference between the decolorization amount of GOx/MnFe O /calcium alg- Effect of initial dye concentration 2 4 inate and MnFe2O4/calcium alginate nanocomposites Different concentration of AR14 (10, 25, 50 and confirmed the existence of simultaneous adsorption- 75 mg L−1) was selected to investigate the effect of -degradation of organic dye by GOx/MnFe2O4/calcium initial dye concentration on the MnFe2O4, MnFe2O4/ alginate nanocomposite. For example, the decoloriz- −1 /calcium alginate and GOx/MnFe2O4/calcium alginate ation of AR14 at initial concentration of 10 mg L by adsorbents. Figure 4 shows the effect of initial dye GOx/MnFe2O4/calcium alginate and MnFe2O4/calcium concentrations on the adsorption process. The results alginate was 60.17 and 23.41%, respectively (Figure 4). showed that the percentage of adsorption efficiency decreased with increasing of the initial dye concen- Kinetic of adsorption of AR14 tration in the solution reflecting the decreasing ads- Enzyme kinetics orption with increasing of the initial dye concentration. The experiments were conducted in 250 mL This could be attributed to the large number of vacant Erlenmeyer flasks containing 50 mL dye solution of surface sites that were available for adsorption at the GOx/MnFe2O4/calcium alginate adsorbent and 80 mM beginning of the process, and which after the initial glucose as a substrate at temperature 30 °C, initial pH stage decreased and were hard to be occupied due to 3.0, time 3 h, 120 rpm and with various initial concen- the repulsive forces between AR14 adsorbed on the trations 10, 25, 50 and 75 mg L-1. The kinetic constant surface of the composite and the solution phase. -1 Michaelise-Menten (kM / mg L ) and the maximum

Figure 4. Decolorization (%) as a function of the initial concentration of AR14 (0.2 g adsorbent, initial pH 3, 30 °C, 3 h, 120 rpm).

451 R. SHOJAAT et al.: DYE REMOVAL FROM ARTIFICIAL WASTEWATER… Chem. Ind. Chem. Eng. Q. 23 (4) 447−456 (2017)

-1 -1 decolorization rate (Vmax / mg L min ) of GOx/

/MnFe2O4/calcium alginate were determined for AR14 by linear regression and Lineweavere-Burk plots and are obtained, respectively: 57.1 and 5.64, with R2 = = 0.995.

kM corresponds to the substrate concentration for which the reaction rate is half of the maximum rate, being a measure of the affinity of the enzyme for a particular substrate. The lesser value of kM, the larger the affinity of the enzyme to the substrate [35].

A. Michniewicz [36] has reported the kM of the enzymatic decolorization of acid blue 62 by laccase (II and crude enzyme) from Cerrena unicolor to be 131 and 306 μM, respectively. Thus, kM values of laccase II were lower than that of Crude enzyme, which indicated a higher affinity of laccase II to acid blue 62 dyes. In this study, the kM value decolorization of

AR14 dye by GOx/MnFe2O4/calcium alginate was found to be 114 μM (57.51 mg. L-1). Kinetics of adsorption without enzyme Batch experiments were carried out using 0.2 g adsorbent at temperature 30 °C, initial pH 3.0, time 3 Figure 5. Pseudo-second order kinetic model for adsorption of h, 120 rpm and various initial concentrations 10, 25, AR14 on adsorbents: a) MnFe2O4 and b) MnFe2O4/calcium 50 and 75 mg L-1. alginate (50 mL AR14 solution, 0.2 g adsorbent, initial pH 3.0, ° In order to predict adsorption kinetic models of 30 C, 3 h, 120 rpm).

AR14 solutions on MnFe2O4/calcium alginate and 2 R values for adsorption on MnFe2O4/calcium alginate MnFe2O4, pseudo-first order and pseudo-second-order kinetic models were applied to the experimental data. and MnFe2O4 ranged (between 0.735-0.936) and A linear form of pseudo-first-order model was des- (between 0.722-0.887), respectively. The absolute deviation of the experimental adsorption capacity cribed by Lagergren [37]: -1 (qDev / mg g ) was calculated using the following Ln (qee1 -qqktt ) = Ln - (9) equation: where qe and qt are the amounts of dye adsorbed per qqqDev.=- e,exp e,cal. (11) unite mass of the adsorbent (mg.g−1) at equilibrium and time t, respectively and k1 is the rate constant of where qe,exp. denotes the experimental value of the −1 adsorption (min ). When Ln (qe–qt) was plotted adsorption capacity and qe,cal. denotes the calculated against time, a straight line should be obtained with a value of the adsorption capacity. Pseudo-first order slope of k1, if the first order kinetics is valid. The qDev. values for AR14 adsorption on MnFe2O4/calcium -1 pseudo-second order model as developed by Ho and alginate (0.3160–1.2950 mg g ) and MnFe2O4 -1 McKay [38] has the following form: (0.4553–1.5781 mg g ) were found to be high. On the tt1 other hand, qDev. values of the pseudo second order = 2 + (10) model for all dye concentrations were distributed qt kq2 e qe within the range of (0.0075–0.1283 mg g-1) and -1 where k2 is the rate constant of the pseudo-second (0.0019–0.0404 mg g ), respectively. These results −1 −1 2 order equation (g mg min ). A plot of t/qt versus t (high R and low qDev.) suggested the adequacy of the would yield a line with a slope of 1/qe and an intercept pseudo second order model to explain the kinetics of 2 of 1/(k2qe ), if the second order model is a suitable AR14 adsorption onto MnFe2O4/calcium alginate and expression. MnFe2O4, and this showed that the overall rate of the Figure 5 and Table 1 show that the pseudo-sec- dye adsorption process appeared to be controlled by ond order fit well to the experimental data for AR14 a chemisorption process. It can also be seen in Table 2 adsorption on MnFe2O4/calcium alginate (R > 0.997) 1 that, with decrease in the initial dye concentration, 2 and MnFe2O4 (R > 0.999), while the pseudo-first order the rate constant of adsorption k2 increased, as the

452 R. SHOJAAT et al.: DYE REMOVAL FROM ARTIFICIAL WASTEWATER… Chem. Ind. Chem. Eng. Q. 23 (4) 447−456 (2017)

Table 1. Pseudo-first order and pseudo-second-order reaction kinetic constants for AR14 adsorption onto MnFe2O4 and

MnFe2O4/Calcium alginate

Pseudo-first order Pseudo-second-order Adsorbent C0 q e,exp. 2 qDev. 2 qDev. qe,cal. k1 R qe,cal k2 R

MnFe2O4 10 0.5657 0.1114 0.0237 0.722 0.4553 0.5676 1.1542 0.999 0.0019

25 1.3273 0.3011 0.0396 0.884 1.0262 1.3378 0.5771 1.000 0.0105 50 2.0304 0.4523 0.0446 0.887 1.5781 2.0479 0.3819 1.000 0.0175 75 2.2685 0.7345 0.0246 0.733 1.5340 2.3089 0.1089 0.999 0.0404

MnFe2O4/calcium alginate 10 0.5853 0.1681 0.0375 0.903 0.4172 0.5928 0.8634 1.000 0.0075 25 1.2626 0.1816 0.0421 0.735 1.0810 1.2735 0.6755 0.999 0.0109 50 2.1957 1.8797 0.0472 0.926 0.3160 2.2914 0.0599 0.999 0.0957 75 2.5469 1.2519 0.0465 0.936 1.2950 2.6752 0.0542 0.997 0.1283

reaction rate constant was highest for the initial where qe is the solid phase equilibrium concentration -1 −1 concentration of 10 mg L in respect to the initial con- (mg g ); Ce is the liquid equilibrium concentration of -1 −1 centrations of 25, 50 and 75 mg L . The kinetics of dye in solution (mg L ); KL is the equilibrium ads- the adsorption of many dye species onto various adsorption constant related to the affinity of binding sites −1 orbents was also found to be of second order in the (L mg ); qm is the maximum amount of dye per unit literature [1]. weight of adsorbent for complete monolayer coverage −1 Adsorption isotherm of AR14 (mg g ). As shown in Table 2, the maximum adsorption capacity of AR14 onto MnFe O /calcium algin- The adsorption of AR14 was performed by 2 4 ate and MnFe O was found to be 3.696 and 4.775 shaking 0.2 g of adsorbent in 50 mL AR14 with var- 2 4 mg g-1, respectively. The essential characteristics of ious initial concentrations of 10, 25, 50 and 75 mg L-1 the Langmuir equation could be expressed in terms of at 30 °C and pH 3 for 3 h. The equilibrium data iso- a dimensionless separation factor R , which is def- therm analysis for AR14 adsorption onto MnFe O / L 2 4 ined by the following relationship (13): /calcium alginate and MnFe2O4 could be expressed by adsorption isotherm models. Langmuir, Freundlich 1 RL= (13) and Temkin isotherm models were used to describe 1 + KCL0 the adsorption equilibrium. The objective was to evaluate the most appropriate correlations for the equilib- where C0 is the highest initial solute concentration, KL rium curves and then to optimize the design of the is the Langmuir adsorption constant (L/mg). The RL adsorption system. The isotherm parameters of Lang- value implies the adsorption to be unfavorable (RL > muir, Freundlich and Temkin models for AR14 ads- 1), linear (RL = 1), favorable (0 < RL < 1) or irreversible (RL = 0). The values of RL listed in Table 2 orption onto MnFe2O4/calcium alginate and MnFe2O4 are shown in Table 2. confirmed that MnFe2O4/calcium alginate (RL = 0.419) and MnFe2O4 (RL = 0.339) were favorable for ads- Langmuir isotherm model orption of AR14 dye under the conditions used in this The Langmuir isotherm assumes monolayer study. adsorption on a uniform surface with a finite number Freundlich adsorption model of adsorption sites. The Langmuir isotherm equilibrium constants could be determined from the linear The Freundlich adsorption isotherm model, which is an empirical equation, is used to describe plot of Ce/qe versus Ce (Figure 6a) by using the following equation: heterogeneous adsorption systems. The Freundlich equilibrium constants could be determined from the CCee1 = + (12) linear plot of Ln qe versus Ln Ce (Figure 6b) by using qemLmqK q the following equation:

Table 2. Langmuir and Freundlich isotherm parameters for the adsorption of AR14 onto MnFe2O4/calcium alginate and MnFe2O4

Adsorbent Temkin model Langmuir model RL Freundlich model 2 2 2 KT B1 R qm KL R 1/n KF R

MnFe2O4 0.2619 0.815 0.994 3.696 0.0260 0.972 0.339 0.658 0.1618 0.959

MnFe2O4/Calcium alginate 0.2247 0.949 0.984 4.775 0.0185 0.978 0.419 0.709 0.1442 0.985

453 R. SHOJAAT et al.: DYE REMOVAL FROM ARTIFICIAL WASTEWATER… Chem. Ind. Chem. Eng. Q. 23 (4) 447−456 (2017)

MnFe2O4/calcium alginate (1/n = 0.709) and MnFe2O4

(1/n = 0.658) were chemical processes. Temkin model The Temkin model considered the effects of indirect adsorbent/adsorbate interactions on the adsorption isotherm. The heat of adsorption of all molecules in the layer would decrease linearly with the coverage due to the adsorbent/adsorbate interactions. Also, the adsorption is characterized by a uniform distribution of binding energies, up to some maximum binding energy. The following form of the Temkin isotherm model has been used:

qe1= BK Ln T + BC 1 Ln e (15)

A plot of qe versus ln Ce enables the deter-

mination of the isotherm constants B1 and KT from the

slope and the intercept, respectively (Figure 6c). KT is the equilibrium binding constant (L/mg) corresponding

to the maximum binding energy and constant B1 is

related to the heat of adsorption. When KT value is larger, this means that the adsorbent/adsorbate inter-

action is also larger. The constants KT and B1 together with the R2 value are listed in Table 2. The

KT value for MnFe2O4 (0.2619) to MnFe2O4/calcium alginate (0.2247) was high. The high values of R2

(=0.994) for MnFe2O4 showed that the adsorption of AR14 dye on this adsorbent was described by the Temkin isotherm. Finally, among the three isotherm models, the Temkin isotherm gave the best correlation for AR14 adsorption onto MnFe2O4. While the Figure 6. a) Langmuir plot for AR14 adsorption onto MnFe2O4; Freundlich model gave the best fit to the equilibrium b) Freundlich plot for AR14 adsorption onto MnFe2O4/calcium

alginate; c) Temkin plot for AR14 adsorption onto MnFe2O4 data of AR14 adsorption onto MnFe2O4/calcium alg- (50 mL AR14 solution, initial concentrations 10, 25, 50 and 75 inate. mg L−1, 0.2 g adsorbent, initial pH 3.0, 30 °C, 3 h, 120 rpm). CONCLUSION 1 Ln qeF = Ln KC + Ln e (14) n The present work suggested that GOx/MnFe2O4/ /calcium alginate, owing to its simple preparation, where qe is the solid phase equilibrium concentration −1 better efficiency and easy for separation, could be (mg g ); Ce is the liquid equilibrium concentration of −1 applied as adsorbent for organic pollutants removal dye in the solution (mg L ); KF is the Freundlich from aqueous solutions based on a heterogeneous constant representing the adsorption capacity (mg −1 Fenton system. The obtained results showed that the g ) and n is the heterogeneity factor describing the enzymatic decolorization of AR14 onto GOx/ adsorption intensity. The slope (value of 1/n) ranges /MnFe2O4/calcium alginate could be described by the between 0 and 1 is a measure of the adsorption int- Michaelis-Menten kinetic model. The Vmax and Km ensity or surface heterogeneity. A value of the slope were found to be 5.64 mg L-1 min-1 and 57.51 mg L-1, close to zero means that the process is heterogen- respectively, using Lineweavere-Burk transformation eous, while a value below unity implies that the ads- of the model. The kinetic data correlated better with orption is a chemical process. A value of the slope pseudo-second order kinetic model than with pseudo above 1 indicates that the adsorption is a physical first-order kinetic model. Among the three isotherm process [40]. The values of the slope (1/n) listed in models, the Freundlich model was fitted with the Table 2 indicated that the AR14 adsorption onto equilibrium isotherm for the AR14 adsorption onto

MnFe2O4/calcium alginate, while Temkin isotherm

454 R. SHOJAAT et al.: DYE REMOVAL FROM ARTIFICIAL WASTEWATER… Chem. Ind. Chem. Eng. Q. 23 (4) 447−456 (2017)

gave the best correlation for adsorption onto MnFe2O4 [18] X. Yu, D. Kang, Y. Hu, S. Tong, M. Ge, C. Cao, W. Song, nanoparticles. RSC Adv. 4 (2014) 31362-31369 [19] A. Córdoba, I. Magario, M.L. Ferreira, J. Chem. Technol. Acknowledgements Biotechnol. 90 (2015) 1665-1676 The authors truly appreciate Semnan University [20] V. Grudic, J. Scepanovic, I. Boskovic, Chem. Ind. Chem. and University of Tabriz for providing all of the sup- Eng. Q. 21 (2015) 285-293 ports for research. [21] J. Tan, X. Wei, Y. Ouyang, R. Liu, P. Sun, J. Fan, Chem. Ind. Chem. Eng. Q. 21 (2015) 465-476 REFERENCES [22] J. Xu, X. Liu, J. He, L. Hu, B. Dai, B. Wu, J. Chem. Technol. Biotechnol. 90 (2015) 57-63 [1] M. Iram, C. Guo, Y. Guan, A. Ishfaq, H. Liu, J. Hazard. [23] S.K. Papageorgiou, F.K. Katsaros, E.P. Kouvelos, N.K. Mater. 181 (2010) 1039-1050 Kanellopoulos, J. Hazard. Mater. 162 (2009) 1347-1354 [2] K. Nadafi, M. Vosoughi, A. Asadi, M.O. Borna, M. [24] R. Khan, P. Bhawana, M.H. Fulekar, Rev. Environ. Sci. Shirmardi, J. Water Chem. Technol. 36 (2014) 125-133 Biotechnol. 12 (2013) 75-97 [3] B.Y. Chen, J. Hong, I.S. Ng, Y.M. Wang, S.Q. Liu, B. Lin, [25] A. Kunamneni, I. Ghazi, S. Camarero, A. Ballesteros, F.J. C. Ni, J. Taiwan Inst. Chem., E 44 (2013) 446-453 Plou, M. Alcalde, Process Biochem. 43 (2008) 169-178 [4] Z. Asadgol, H. Forootanfar, S. Rezaei, A. Mahvi, M. [26] R.O. Cristóvão, A.P.M. Tavares, A.I. Brígida, J.M. Loure- Faramarzi, J. Environ. Health Sci. Eng. 12 (2014) 1-5 iro, R.A.R. Boaventura, E.A. Macedo, M.A.Z. Coelho, J. [5] V. Dulman, S.M. CucuMan, J. Hazard. Mater. 162 (2009) Mol. Catal., B: Enzym. 72 (2011) 6-12 1457-1464 [27] R.C.C. Costa, M.F.F. Lelis, L.C.A. Oliveira, J.D. Fabris, [6] H. Zangeneh, A.A.L. Zinatizadeh, M. Feyzi, CLEAN–Soil, J.D. Ardisson, R.R.V.A. Rios, C.N. Silva, R.M. Lago, J. Air, Water 44 (2016) 78-88 Hazard. Mater. 129 (2006) 171-178 [7] V. Elhami, A. Karimi, M. Aghbolaghy, J. Taiwan Inst. [28] P. Baldrian, V. Merhautová, J. Gabriel, F. Nerud, P. Chem., E 56 (2015) 154-159 Stopka, M. Hrubý, M.J. Beneš, Appl. Catal., B: Environ. [8] W. Li, D. Wan, G. Wang, K. Chen, Q. Hu, L. Lu, Korean J. 66 (2006) 258-264 Chem. Eng. 33 (2016) 1557-1564 [29] A. Han, J. Liao, M. Ye, Y. Li, X. Peng, Chinese J. Chem. [9] R. Rajeshkannan, M. Rajasimman, N. Rajamohan, Eng. 19 (2011) 1047-1051 Chem. Ind. Chem. Eng. Q. 19 (2013) 461-470 [30] A. Karimi, F. Mahdizadeh, D. Salari, A. Niaei, Desalin- [10] M.T. Mihajlovic, S.S. Lazarevic, I.M. Jankovic-Castvan, ation 275 (2011) 148-153 B.M. Jokic, D.T. Janackovic, R.D. Petrovic, Chem. Ind. [31] S. Shoaebargh, A. Karimi, G. Dehghan, J. Taiwan Inst. Chem. Eng. Q. 20 (2014) 283-293 Chem., E 45 (2014) 1677-1684 [11] Y. He, Z. Xu, F. Wu, Q. Yang, J. Zhang, J. Chem. [32] F. Mahdizadeh, S. Aber, A. Karimi, J. Taiwan Inst. Technol. Biotechnol. 90 (2015) 2257-2264 Chem., E 49 (2015) 212-219 [12] S. Sen-Gupta, K.G. Bhattacharyya, RSC Adv. 4 (2014) [33] R. Shojaat, N. Saadatjoo, A. Karimi, S. Aber, J. Environ. 28537-28586 Chem. Eng. 4 (2016) 1722-1730 [13] Y. Song, S. Ding, S. Chen, H. Xu, Y. Mei, J. Ren, Korean [34] S.D. Ashrafi, S. Rezaei, H. Forootanfar, A.H. Mahvi, M.A. J. Chem. Eng. 32 (2015) 2443-2448 Faramarzi, Int. Biodeter. Biodegr. 85 (2013) 173-181 [14] A.A. Babaei, A. Khataee, E. Ahmadpour, M. Sheydaei, B. [35] N.M. Mahmoodi, M. Arabloo, J. Abdi, Water Res. 67 Kakavandi, Z. Alaee, Korean J. Chem. Eng. 33 (2016) (2014) 216-226 1352-1361 [36] A. Michniewicz, S. Ledakowicz, R. Ullrich, M. Hofrichter, [15] M. Ozmen, K. Can, G. Arslan, A. Tor, Y. Cengeloglu, M. Dyes Pigments 77 (2008) 295-302 Ersoz, Desalination 254 (2010) 162-169 [37] H. Yuh-Shan, Scientometrics 59 (2004) 171-177 [16] L. Deng, Z. Shi, X. Peng, RSC Adv. 5 (2015) 49791- [38] Y.S. Ho, J. Hazard. Mater. 136 (2006) 681-689 49801 [39] R. Slimani, A. Anouzla, Y. Abrouki, Y. Ramli, S. El Antri, [17] H. Momenzadeh, A.R. Tehrani-Bagha, A. Khosravi, K. R. Mamouni, S. Lazar, M. El Haddad, J. Mater. Environ. Gharanjig, K. Holmberg, Desalination 271 (2011) 225- Sci. 2 (2011) 77-87 –230 [40] M. Matouq, N. Jildeh, M. Qtaishat, M. Hindiyeh, M.Q. Al- -Syouf, J. Environ. Chem. Eng. 3 (2015) 775-784.

455 R. SHOJAAT et al.: DYE REMOVAL FROM ARTIFICIAL WASTEWATER… Chem. Ind. Chem. Eng. Q. 23 (4) 447−456 (2017)

RAHIM SHOJAAT1 OBEZBOJAVANJE VEŠTAČKE OTPADNE VODE AFZAL KARIMI2 POMOĆU HETEROGENOG BIO-FENTONOVOG 1 NAGHI SAADATJOO SISTEMA SOHEIL ABER3

1 Department of Applied Chemistry, U ovom radu je napravljen nanokompozit GOx/ MnFe2O4/kalcijum-alginat zarobljava- Faculty of Chemistry, Semnan njem enzim/nanočestica u kalcijum alginatu. Dobijeni adsorbent je primenjen za obez- University, Semnan, Iran bojavanje veštačke otpadne vode sa bojom kiselo crveno 14 korišćenjem heterogenog 2 Department of Biotechnology, Faculty bio-Fentonovog procesom. Proučavane su kinetika i ravnoteža procesa. Dobre vred- of Advanced technologies in medicine, nosti koeficijentq korelacije su utvrđene za Mihaelis-Mentenov i model i pseudo-drugog Iran University of Medical Sciences, reda. Obezbojavanje je pratilo kinetičke modele Mihaelis-Mentena, kao i pseudo-prvog i Tehran, Iran pseudo-drugog reda. Adsorpciona ravnoteža je opisana Langmirovom, Frojndliovomh i 3Research Laboratory of Environmental Temkinovom izotermom. Frijndliov model je dobro fitovao ravnotežne podatke za ads- Protection Technology, Department of Applied Chemistry, Faculty of orpciju boje na MnFe2O4/kalcijum-alginatu, dok je Temkinova izoterma dala najbolju Chemistry, University of Tabriz, Tabriz, korelaciju za adsorpciju boje nanočesticama MnFe2O4. U radu je analiziran efekat raz- Iran ličitih parametara, kao što su pH rastvora, početna koncentracija boje i vreme kontakta,

na adsorpciju boje na MnFe2O4 i MnFe2O4/kalcijum-alginat, kao i enzimska razgradnja NAUČNI RAD boje. Stepen obezbojavanja kiselog crvenog 14 sa nanokompozitom GOx/MnFe2O4/ /kalcijum-alginat je bio 60,17% pri početnoj koncentraciji od 10 mg/l.

Ključne reči: adsorbenti, obezbojavanje, enzim, izoterma, kinetika.

456 Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 457−471 (2017) CI&CEQ

ANA BELŠČAK-CVITANOVIĆ1 MODIFICATION OF FUNCTIONAL QUALITY VIKTOR NEDOVIĆ2 2 OF BEER BY USING MICROENCAPSULATED ANA SALEVIĆ GREEN TEA (Camellia sinensis L.) AND SAŠA DESPOTOVIĆ2 DRAŽENKA KOMES1 GANODERMA MUSHROOM (Ganoderma MIOMIR NIKŠIĆ2 lucidum L.) BIOACTIVE COMPOUNDS BRANKO BUGARSKI3 IDA LESKOŠEK ĆUKALOVIĆ2 Article Highlights • Alginate and pectin microbeads were produced by electrostatic extrusion and spray 1Department of Food Engineering, drying Faculty of Food Technology and • Chitosan reinforcement provided smaller beads but prolonged release of polyphenols • Biotechnology, University of Pilsner beer with Ganoderma and radler with green tea encapsulates were preferred • The use of spray dried microbeads enabled the most pronounced augmentation of Zagreb, Zagreb, Croatia 2 TPC of enriched beer Department of Food Technology • Stability of enriched beer’s phenolic content during one month-storage was and Biochemistry, Faculty of established Agriculture, University of Belgrade, Belgrade-Zemun, Serbia Abstract 3 Department of Chemical Increasing interest in production of frequently consumed functional food products Engineering, Faculty of Technology has focused the present study on implementation of microencapsulated Gano- and Metallurgy, University of derma mushroom and green tea bioactive compounds in beer production. Belgrade, Belgrade, Serbia Electrostatic extrusion assisted microencapsulation of green tea and Ganoderma SCIENTIFIC PAPER extracts enabled production of particles ranging from 490 to 1000 µm in size, with up to 75% of entrapped total polyphenols. Dried, powdered extracts, as well UDC 663.4:635.8:66:60:54 as microparticles encapsulating Ganoderma and green tea extracts that exhibited the best morphological properties and retarded release of polyphenols (alginate and alginate-chitosan coated, as well as chitosan coated pectin microbeads) were implemented in beer production. The addition of Ganoderma microbeads to pilsner beer did not augment its polyphenolic concentration (TPC), as opposed to the addition of green tea encapsulating microbeads to radler, while adding dried Ganoderma and spray dried green tea extracts enabled to increase the TPC for up to 3-fold higher values. Ganoderma dried extract-enriched pilsner beer and spray dried green tea extract-enriched radler were preferred in terms of sensory properties, due to the lowest bitterness intensity and most pronounced herbal aroma of the added adjuncts. Refrigerated storage of Ganoderma hydrogel microbeads-enriched pilsner beer revealed fluctuations of TPC, while green tea hydrogel microbeads-enriched radler exhibited better stability. The established methodology provides a procedure suitable for microencapsulate-enrichment of drink and food products, thus setting a reliable basis for future functional food production by microencapsulate implementation strategies. Keywords: beer, Ganoderma, green tea, microencapsulation, polyphenolic antioxidants.

Correspondence: V. Nedović, Department of Food Technology and Biochemistry, Faculty of Agriculture, University of Belgrade, Nemanjina 6, 11081 Belgrade-Zemun, Serbia. E-mail: [email protected] Paper received: 22 July, 2016 Paper revised: 14 Novembar, 2016 Paper accepted: 16 December, 2016 https://doi.org/10.2298/CICEQ160722060B

457 A. BELŠČAK-CVITANOVIĆ et al.: MODIFICATION OF FUNCTIONAL QUALITY… Chem. Ind. Chem. Eng. Q. 23 (4) 457−471 (2017)

Beer is the most popular alcoholic beverage Following the principle of using widely available, worldwide and the second most consumed alcoholic economically acceptable sources of functional ingre- beverage in Europe, accounting for 37% of the total dients for production of novel food products, instead EU alcohol consumption [1]. Beer contains numerous of using plain, isolated and expensive bioactive com- ingredients that exert physiological effects upon con- pounds, green tea (Camellia sinensis L.) and Gano- sumption, which can be attributed to its diverse com- derma (Ganoderma lucidum) powder were selected position (water, malt, hops, brewer’s yeast, non-malted as the active ingredients in the present study. Green cereals, starch or starch syrups as adjuncts). Charac- tea and medicinal mushrooms such as Ganoderma terized by a high carbohydrate concentration and pre- are today among the most well-known and popular sence of proteins, amino acids, vitamins, organic functional ingredients, owing to their scientifically acids, over 30 different micro and macro elements proven beneficial health effects attributed to a high and a range of antioxidants, beer is considered as a concentration of polyphenols and polysaccharides nutritively valuable product [2]. However, in recent [12,13]. Although polyphenolic compounds are not the years, food industry efforts are focused on fitting the main active ingredients of Ganoderma mushroom, in consumer trends towards more natural products, order to achieve innovative and sensory appealing obtained by either using natural raw materials for food beer, Ganoderma was chosen as the functional sub- production or enriching conventional products with strate to be implemented in different beer products. diverse natural substrates or extracts. With regard to Namely, a previous study by Leskošek-Čukalović et the beer segment, the mixtures of high-quality beer al. [4] revealed that Ganoderma-extract enriched beer and lemonade (popularly called radler) have become exhibited significantly improved sensory properties in the largest volume innovation on beer markets glob- comparison to plain beer, especially with regard to ally, while interest has arisen in specialty types of bitterness which was scored much higher and more beer enriched with other bioactive compounds. Vari- harmonic and finer. The reason for this may be the ous natural sources of bioactive compounds, such as combination of all specific bioactive compounds of medicinal plant extracts (Melissa officinalis and Thy- Ganoderma, comprising terpens and polyphenols mus vulgaris) [3], medicinal mushrooms (Ganoderma which contributed to that aspect. For the purpose of lucidum) [4], grape must [5] and honey [6] have been encapsulation of green tea polyphenols, different used so far for enriching beer and formulating new techniques, such as internal gelation [14], spray dry- taste and aroma attributes. ing [15] or liposome entrapment [16] have been When consumed in the form of a functional food applied so far, while limited number of studies such product, the effectiveness of naturally derived bioact- as the ones from broken Ganoderma lucidum spores ive compounds in providing therapeutic or physiolog- [17] or Ganoderma mushroom oil [18] reported the ical benefits is usually relieved by their low concen- microencapsulation of Ganoderma bioactive com- tration and bioavailability, or the effects of chemical, pounds (polysaccharides and ganoderic acids). physical and physiological factors during food pro- Therefore, in the present study, electrostatic extrusion duction and consumption [7]. Encapsulation enables and spray drying microencapsulation of polyphenolic to improve the bioactive effectiveness of beneficial compounds from Ganoderma mushroom and green compounds by protecting unstable substances of vari- tea extracts were employed, in alginate and pectin or ous external influences, as well as by enhancing their chitosan-reinforced hydrogel microparticles, aiming to stability and activity during production and storage [8]. augment the retention performance of entrapped In recent years, encapsulation has been used exten- bioactives. The effect of encapsulates upon their sively for immobilization of polyphenolic compounds incorporation into a real food product was examined from different plant substrates such as medicinal plant by using microencapsulated polyphenolic compounds extracts [9], cocoa extract [10] or blueberry anthocya- from Ganoderma mushroom and green tea extracts nins [11], etc. Encapsulation of polyphenolic com- for the formulation of innovative beers with improved pounds as highly potent antioxidants, but charac- bioactive and sensory properties. terized with a pronounced instability and reactivity, contributes to their preservation, increasing their bio- EXPERIMENTAL availability, masking of unpleasant tastes, such as polyphenol-imparted bitterness and astringency and Chemicals enabling the formation of new dosage forms for pro- All chemicals used for the analytical procedures duction of innovative food products [8,9]. were of analytical grade, while the natural biopoly-

458 A. BELŠČAK-CVITANOVIĆ et al.: MODIFICATION OF FUNCTIONAL QUALITY… Chem. Ind. Chem. Eng. Q. 23 (4) 457−471 (2017) mers used as carriers for encapsulation purposes centration – TPC determined using the Folin-Ciocalteu were appropriate for food purposes. Sodium alginate assay [20]) of the extracts amounted to 820 mg·L-1 was purchased from Sigma-Aldrich (St. Louis, MI, expressed as mg of gallic acid equivalents (GAE) per USA). Chitosan (molecular weight 100000–300000 L and 3472 mg·L-1 GAE, respectively. g·mol-1) was obtained from Acros organic (Geel, Bel- Microencapsulation by electrostatic extrusion in gium) and pectin (low methoxyl pectin, molecular different carrier materials weight 70–140 kDa) from Cargill (Zagreb, Croatia). Green tea Darjeeling (Müller, Abtswind, Germany) in The preparation of microbeads by electrostatic loose leaf form and powdered Ganoderma mushroom extrusion was carried out by ionic gelation technique extract (Fujian Xianzhilou Biological Science & Tech- following the method adopted by Belščak-Cvitanović -1 -1 nology, Fuzhou, China) were purchased at a local et al. [9] Alginate (2 g·L ) and pectin (2 g·L ) sol- organic market in Croatia. utions were prepared by dissolving sodium alginate or pectin in the previously prepared mushroom or green Preparation of green tea and Ganoderma mushroom tea extracts. Dissolved solutions were homogenized extracts and microencapsulated using the electrostatic droplet Green tea (ground to a fine powder using a ball generation method [9], at a constant flow rate (25.2 mill) was repeatedly extracted (2-fold extraction), by mL·h-1) of encapsulant solution and varying the volt- pouring 100 ml of distilled water heated to 80 °C over age between 4.6 and 5.2 kV, depending on the form- 8 g of plant material and stirring on a magnetic stirrer ulation (Table 1). The cross-linking solution consisted at 80 °C for 15 min. After each extraction, the extracts of 2 g·L-1 of calcium chloride prepared in the green tea were centrifuged at 5310g for 5 min and the combined or Ganoderma mushroom extracts. For chitosan coat- supernatants filled up to 200 ml. ing of alginate and pectin microparticles, the encap- In case of Ganoderma mushroom extract, 5 g of sulant solutions were extruded into the collecting sol- powdered extract was resuspended in 100 ml of dis- ution consisting of 1 g·L-1 of chitosan dissolved in the tilled water on a magnetic stirrer during 24 h, centri- previously prepared mushroom or green tea extract fuged at 5310g for 5 min and filled up to 100 ml. In containing 2 g·L-1 of citric acid (pH 2.35) and 2 g·L-1 of order to determine the content of soluble (extractable) calcium chloride. All obtained microparticles were solids present in the obtained extracts, dry matter washed several times with the extract after production content of prepared Ganoderma and green tea ext- and stored in the original extracts containing 2 g·L-1 racts was determined according to the AOAC method calcium chloride in the dark at 4 °C. By preparing and 920.151 [19] and exhibited 32.3 and 15.1 g·kg-1, res- storing the formulated microbeads in the correspond- pectively. The polyphenolic yield (total phenol con- ing extracts, isotonic concentrations (the same con-

Table 1. Encapsulation parameters, loading efficiency and particle size distribution of microparticles encapsulating Ganoderma mushroom and green tea polyphenols; The values in each column superscripted with the same alphabeth letter (a-h) are not significant (p > 0.05); CH – chitosan coating; d(0.1), d(0.5), d(0.9) – particle diameters indicating that 10, 50 and 90% of the particle size distribution is below these values, respectively; encapsulation efficiency and retained antioxidant capacity are expressed as percentages of total polyphenols concentration and antioxidant capacity of formulated particles and initial encapsulant solution

Needle Retained Particle size, μm Particle Extrusion Encapsulation Material dimension antioxidant type voltage, kV efficiency, % d (0.1) d (0.5) d (0.9) mm capacity, % Hydrogel Ganoderma mushroom Alginate 4.8 0.60 71.8±4.3abc 72.7±1.4a 428.96±0.18a 621.20±7.27a 826.19±13.70 Pectin 4.9 0.84 66.6±2.4adg 67.2±2.3b 147.53±2.36 682.79±6.81 907.94±12.47a Alginate CH 5.0 0.60 75.2±3.8be 70.2±3.5ab 310.45±0.42 490.64±4.53 881.68±12.68a Pectin CH 5.2 0.84 61.6±4.2df 71.3±3.3ab 393.71±1.47 616.42±5.35a 955.04±15.78 Green tea Alginate 4.9 0.60 62.6±3.8d 56.8±1.7cd 791.20±39.61 987.27±40.48 1250.28±75.02b Pectin 4.7 0.41 52.2±1.7h 52.2±2.8c 425.17±21.14a 814.23±36.78 1237.15±35.75b Alginate CH 5.2 0.41 51.4±2.6h 56.0±3.4c 725.87±14.52 905.55±54.33 1156.88±62.47 Pectin CH 4.6 0.41 47.6±1.4h 59.4±2.4d 546.41±21.86 773.17±18.56 1047.30±29.32 Spray dried Plain extract - - 66.9±4.7cdf 69.1±3.4ab 1.45±0.12 3.71±0.15 7.80±0.06 green tea

459 A. BELŠČAK-CVITANOVIĆ et al.: MODIFICATION OF FUNCTIONAL QUALITY… Chem. Ind. Chem. Eng. Q. 23 (4) 457−471 (2017) centration of polyphenolic compounds in the formul- (distilled water). The samples were submitted to con- ated microbeads and outside) during all stages of tinuous agitation on an orbital shaker (New Brunswick microbeads preparation (carrier, cross-linking and Scientific, Edison, New Jersey, USA) operating at storage solutions) were maintained. In that way, the 1.67 Hz. At defined time intervals, an aliquot of the concentration gradient between the beads and sur- supernatant was taken for analysis and replaced by rounding solution was maintained and the driving the same amount of fresh medium. force for diffusion-related release of polyphenols was Formulation of beers enriched with microencap- eliminated, ensuring equal polyphenol content of the sulated Ganoderma or green tea extract polyphenols beads at all times. Green tea extract was also spray dried using a In order to improve the functional properties of Büchi mini B-290 spray dryer (Büchi Labortechnik, beer, pilsner beer and lemon radler were enriched Flawil, Switzerland) equipped with a 0.7 mm standard with prepared liquid extracts, microencapsulated diameter nozzle. The inlet and outlet temperature Ganoderma mushroom and green tea extracts in hyd- were 130±3 and 56±2 °C, respectively. The spraying rogel and dried (powdered) form (according to the air flow rate, rate of liquid feed, atomisation pressure schematic display on Figure 1). With the purpose of and pump speed were 536 L·h-1, 8 mL min-1, 41 kPa neutralizing the interferences of naturally present anti- and 30%, respectively. A free-flowing powder used for oxidants in lemon radler (ascorbic acid) which can enriching beer was obtained. cause interferences and overestimation of total polyphenols and antioxidant capacity during analytical Characterization of physico-chemical and bioactive measurements, the authors wanted to „equalize“ both properties of formulated particles types of beer samples (pilsner and radler), so lemon Randomly selected hydrogel beads were visual- juice was added to pilsner beer. Preliminary sensory ized using an optical microscope, whereas the evaluations (data not shown) also revealed that in appearance and surface morphology of spray dried that way the taste attributes of pilsner beer were imp- beads were examined by scanning electronic micro- roved in comparison to plain beer without lemon juice. scopy (SEM) performed by using a Mira3 microscope For the preparation of samples,1 g·L-1 of Ganoderma (Tescan, Brno – Kohoutovice, Czech Republic). The mushroom hydrogel microbeads or 1.5 g·L-1 of green samples were attached to stubs, coated with a layer tea extract hydrogel microbeads and 0.5 g·L-1 of of gold (50 nm) and examined using an acceleration lemon juice were added to pilsner beer, while 1.5 g·L-1 voltage of 4-5 kV. Particle size of the obtained micro- of green tea extract hydrogel microbeads or 0.5 g L-1 particles was determined using a Mastersizer 2000 of spray dried green tea extract were added to radler. (Malvern Instruments, Malvern, United Kingdom) The described microencapsulates and plain non- equipped with a Hydro 2000S dispersion unit in case encapsulated green tea or Ganoderma mushroom of hydrogel beads and a Scirocco 2000 unit in case of extracts (in concentration equal as the added micro- spray dried green tea extract. beads, representing control sample) were added The concentration of entrapped total polyphe- aseptically to commercially produced bottled beers or nols and antioxidant capacity of produced hydrogel radler, and the bottles immediately closed to mature microbeads was estimated by dissolving a known at 5 °C for one day. amount of microparticles in 2 g·L-1 of sodium citrate In the formulated enriched beers, the concen- (or in distilled water in case of spray dried micro- tration of total polyphenols and antioxidant capacity beads). The concentration of total polyphenols was were determined as previously described and their determined using the Folin-Ciocalteu assay [20], and stability was evaluated during one month of refri- the antioxidant capacity using the DPPH radical gerated storage (4 °C), directly upon production and scavenging assay [21]. The percentage of loading each 10th day of storage. The basic chemical para- efficiency of hydrogel beads was calculated as the meters of enriched beers such as: alcohol concen- ratio between the total polyphenols concentration in tration, real extract, apparent extract and energy the citrate solution of dissolved beads and their res- value, were also determined during storage using pective concentration in the initial solution. Alcolyzer Beer ME analyzing system (Anton Paar, The release kinetics of total polyphenols and Graz, Austria). antioxidant capacity profile was determined from the Sensory evaluation of enriched beers obtained hydrogel microbeads in distilled water. For the analysis a known amount of microparticles (of The enriched beers were evaluated using quan- about 3 g) was suspended in 50 mL of the medium titative descriptive analysis procedure excerpt from the literature data on sensory evaluation of beer [5]

460 A. BELŠČAK-CVITANOVIĆ et al.: MODIFICATION OF FUNCTIONAL QUALITY… Chem. Ind. Chem. Eng. Q. 23 (4) 457−471 (2017)

PILSNER BEER

+ lemon juice

Ganoderma adjuncts Green tea adjuncts

C CE A A-CH P-CH DE C CE A A-CH P-CH SDE

RADLER BEER

Ganoderma adjuncts Green tea adjuncts

C CE A A-CH P-CH DE C CE A A-CH P-CH SDE

C - Control plain beer A - Beer with alginate microbeads DE - Beer with dried Ganoderma extract CE - Beer with liquid extract A-CH - Beer with chitosan-alginate microbeads SDE - Beer with spray dried green P-CH - Beer with chitosan-pectin microbeads tea extract Figure 1. Schematic display of sample setup during preparation of enriched beers. and according to ISO standard [22]. The internal sen- aftertaste and acidity, F = 0.41 for taste and bitter- sory panel of the Faculty of Agriculture in Belgrade, ness, while F = 0.5 for remaining attributes (appear- Serbia comprised 25 people, 7 female and 18 male ance and overall acceptability). The number of points members, who had undergone extensive sensory describing each of the attributes was multiplied by the training and had previous experience in the assess- importance factor, and the average point number was ment of beer. The experimental samples were eval- calculated. uated during two sessions (one during which only Statistical analysis Ganoderma-enriched drinks were evaluated and the second one evaluating only green tea-enriched drinks), The results were analyzed statistically using the with conventional beer (or radler) produced with the Statistica AXA ver.12 ACADEMIC software (StatSoft, addition of plain Ganoderma or green tea extract Tulsa, Oklahoma, USA) to determine the average serving as the controls. For the sensory analysis, 50 value and standard error. Variance analysis, with a mL of sample adjusted to 12 °C was served in trans- significance level of α = 0.05%, was performed to det- parent plastic cups encoded with random alphabet ermine the influence of adding different types of sub- letters. Water was provided for rinsing between strates and microbeads on the evaluated beer para- samples. The sensory properties were presented on a meters. 5-point scale (according to which 5 meant extremely desired quality and 1 was for a defective product). RESULTS AND DISCUSSION Eight attributes were evaluated for all beers Microencapsulation of green tea and Ganoderma (appearance, colour, transparency, odour, taste, mouth- mushroom polyphenolic extracts feel, aftertaste, acidity, bitterness and overall acceptability). For each assessed attribute an importance Results of the microencapsulation experiments factor (F = 0.05 to F = 0.5) was considered, based on aimed at producing encapsulated particulates as the the guidelines determined by MEBAK [23], where delivery systems of green tea and Ganoderma bioact- F = 0.05 was used for the less important attributes ives revealed that the particle size and morphology, (colour, transparency), F = 0.29 was used for odour, encapsulation yields and release profiles as the main

461 A. BELŠČAK-CVITANOVIĆ et al.: MODIFICATION OF FUNCTIONAL QUALITY… Chem. Ind. Chem. Eng. Q. 23 (4) 457−471 (2017) attributes of formulated microparticles were in dep- bits a very broad particle size distribution (PSD) endence on both the employed carrier material (or width, suggesting uneven and non-homogenous mic- combination thereof) as the encapsulants and the roparticle formation, in comparison to alginate par- type of substrate (green tea or Ganoderma). ticles. The same situation was observed in case of pectin particles encapsulating green tea extract, indi- Particle size and morphological properties of cating more non-homogenous particles than alginate formulated microparticles ones. These results imply that alginate hydrogel par- Marked differences can be observed in the particles may be preferred over pectin ones according to ticle size distribution parameters of microbeads the PSD and size homogenicity (Figure 2), since for encapsulating Ganoderma extract, which were sig- the purpose of food and beverage applications visual nificantly (p < 0.05) smaller in comparison to the ones appearance is a crucial factor for the product deve- encapsulating green tea extract (Table 1). According lopment. to the median diameter (d 0.5) alginate particles The differences observed in the particle size of encapsulating Ganoderma were smaller than the pec- alginate and pectin microparticles reveal the effect of tin particles, but an opposite ranking can be observed pH influenced by the substrate (Ganoderma or green in case of green tea encapsulating microbeads. tea) composition that can mediate during ionic cross- Although pectin particles encapsulating Ganoderma -linking of alginate and pectin hydrogels and result exhibited higher d 0.5 in comparison to alginate mic- with possible modifications of the hydrogel properties. roparticles, they were characterized with a very low d Namely, the gelation kinetics and the mechanism 0.1 = 147.53 µm (particle diameter at which 10% of involved in the ionotropic gel formation are strictly the sample is comprised of smaller particles), indi- dependent upon the pH of alginate and pectin sol- cating that a low number of very small microbeads utions. pH regulates the dissociation and availability (smaller than 147.53 µm) is formed during the micro- for ionic complexation of guluronic and carboxylic particles production. If taking into account the high d groups of alginate and pectin, affecting the final gel 0.9 (particle diameter at which 90% of the sample is properties [24]. At lower pH, the ionization of galactur- comprised of smaller particles) of pectin particles onate carboxylic (pectin) groups is repressed by the (907.94 µm), the results imply that this sample exhi-

A B C

D E H

F G

Figure 2. Optical micrographs of plain alginate (A, D), plain pectin (B, E), chitosan coated alginate (C, F) and chitosan coated pectin (G) hydrogel microbeads encapsulating Ganoderma (A, B, C) and green tea (D-G) extracts, as well as SEM micrograph of spray dried green tea extract (H); the enhancements of -10× for optical micrographs and 3000× for SEM imaging was used.

462 A. BELŠČAK-CVITANOVIĆ et al.: MODIFICATION OF FUNCTIONAL QUALITY… Chem. Ind. Chem. Eng. Q. 23 (4) 457−471 (2017) higher concentration of hydrogen ions, thus producing Encapsulation efficiencies and release profiles of less negatively charged groups (such as COO-) that polyphenolic compounds 2+ can form intermolecular junction zones, via Ca - The loading capacities of polyphenols as the bridges [25]. Since the pH of Ganoderma extract target active ingredients revealed a fluctuating trend, (4.12) was lower than the one of green tea extract depending on the encapsulated substrate and emp- (5.15), less Ca-crosslinking could have occurred while loyed carrier material. According to the obtained encapsulating Ganoderma bioactive compounds, thus results, the substrate and the composition of its active producing larger pectin particles with a very broad ingredients, rather than their concentration is the pre- PSD width. Correspondingly, in case of green tea bio- dominant factor influencing the encapsulation per- actives encapsulation in pectin hydrogel, higher pH formance. This is confirmed by the fact that a much favored more pronounced ionization of carboxylate higher TPC of green tea extract (3472 mg·l-1 GAE) in groups that could form the cross-linked junction zones comparison to low TPC of Ganoderma extract (820 2+ with Ca , producing better cross-linked and smaller mg·l-1 GAE), did not account for higher encapsulation particles. efficiency of polyphenolic compounds from green tea. According to the d 0.5, chitosan coated micro- As can be seen in Table 1, hydrogel microparticles beads were slightly smaller in comparison to the non- encapsulating green tea generally exhibited lower coated particles. Similar results were already pre- encapsulation efficiency of total polyphenols in com- viously observed in case of chitosan coating form- parison to the particles encapsulating Ganoderma. ation on alginate particles [9]. This may have occur- The same was also applied to the percentage of red due to the hypertonic concentration of chitosan retained antioxidant capacity, which was evidenced coating medium in relation to the hydrogel beads, by a high correlation (r = 0.79) established between which facilitated the hypertonic-driven diffusion of the total polyphenolics loading capacity and percent- water from the beads, resulting with smaller particles. age of retained antioxidant capacity. Lower total poly- Another possible explanation is that since alginate phenolics loading capacity and percentage of retained and pectin as anionic polysaccharides form polyelec- antioxidant capacity in case of green tea encapsul- trolyte complex with the oppositely charged (cationic) ating microbeads can be attributed to the composition chitosan, the interactions between the carrier mater- of green tea and its active constituents, which influ- ials and coating resulted with additional cross-linking ence their retention in the hydrogel microbeads struc- of particle matrix, reducing the water concentration ture. Namely, at the pH of green tea extract (pH 5.1), and shrinking of the formed particles [26]. Spray dried flavan-3-ol catechin molecules as the representative green tea extract was expectedly in micro-sized range active compounds of green tea [12] are dissociated to (< 4 µm, according to d 0.5). some extent (deprotonated) [28], and produce more With regard to the morphological properties of electrostatic repulsion between both anionic alginate produced microparticles (Figure 2), electrostatic ext- or pectin charged groups. This electrostatic repulsion rusion of both Ganoderma and green tea extracts and of deprotonated catechins [28] and alginate or pectin plain alginate as the encapsulants provided spheric- functional groups may contribute to the poor retention ally shaped particles. of small molecular weight catechins in the porous Employing pectin as the carrier material, as well matrix of alginate and pectin microbeads. Since the as chitosan as the coating revealed more irregul- predominant active ingredients of Ganoderma are arities and size-inhomogeneous particles with pro- soluble polysaccharides [13], rather than polyphenolic nounced structural differences. Both plain pectin and compounds, the fair encapsulation efficiency of poly- chitosan coated alginate and pectin particles were phenols achieved in this type of microbeads may be characterized with more conically-shaped (“tear-like”) the consequence of the higher molecular weight and structure, which has previously also been observed size of polysaccharides. Moreover, the neutral char- for similar delivery formulations [27]. SEM micrograph acter of soluble Ganoderma polysaccharides [29] may of spray dried green tea extract revealed a polydis- have facilitated the retention of Ganoderma extract persed system, and large aggregates formed by the constituents in the hydrogel matrix and induced pos- particles. The obtained particles had a pronounced, sible interactions with the carrier materials, resulting irregular shape with morphologically dented and dis- in better retention and loading yield of polyphenolic turbed surfaces, which occurs as the consequence of compounds. Chitosan reinforcement of alginate and water leakage and shrinking of particles during spray pectin particles of both Ganoderma and green tea drying [15]. encapsulating microbeads produced no marked effect on the loading capacity of total polyphenols and anti-

463 A. BELŠČAK-CVITANOVIĆ et al.: MODIFICATION OF FUNCTIONAL QUALITY… Chem. Ind. Chem. Eng. Q. 23 (4) 457−471 (2017) oxidant capacity (Table 1). With regard to the spray ison to plain, uncoated particles), it enabled an imp- dried green tea extract, employing lower spray drying roved, prolongued release pattern of total polyphenols temperatures (130 °C) in order to minimize the neg- (Figure 3A) and antioxidant capacity (Figure 3B) from ative temperature-related degradation of green tea the particles in water. Green tea encapsulating micro- extract polyphenols, enabled to entrap a high pro- particles exhibited significantly higher polyphenolic portion of these compounds in the produced powder. concentration and antioxidant capacity in relation to The spray dried green tea extract even exhibited sig- the ones encapsulating Ganoderma. Due to the pro- nificantly (p < 0.05) higher contents of total polyphe- nounced porosity of ionically cross-linked alginate gel, nols loading capacity and retained antioxidant cap- a rapid release (within the first 10 min) of encap- acity in comparison to hydrogel microencapsulated sulated compounds occurs. Chitosan coating on both green tea, implying that the use of this dosage form of alginate and pectin microbeads enabled to retard the bioactive compounds would be preferred in terms of release even up to 60 min in case of total polyphenols augmenting the bioactive content of an enriched food released from chitosan coated alginate particles, or product. antioxidant capacity exhibited by chitosan coated pec- Although chitosan coating provided structurally tin particles. In this way, longer stability and protection poorer particles and no marked differences in the of the encapsulated bioactives and their functional pro- loading capacities of active compounds (in compar- perties may be achieved when incorporated in beer.

A 10

8 ] -1 6

4 TPC [g · [g TPC kg

0 0 2 5 10152030601200 2 5 1015203060120

Ganoderma mushroom Green tea min A A-CH P P-CH B 600 ] -1

450

300

150

Antioxidant capacity [mmol kg · capacity [mmol Antioxidant 0 2 5 10152030601200 2 5 1015203060120

Ganoderma mushroom Green tea min A A-CH P P-CH Figure 3. Release profiles of: A) total polyphenolic compounds (mg GAE/g beads) and B) antioxidant capacity (µmol trolox/g beads) from alginate, pectin and corresponding chitosan coated microbeads with encapsulated Ganoderma mushroom and green tea extracts in water. The total polyphenolic content is expressed as g of gallic acid equivalents per kg of beads ± standard deviation. Antioxidant capacity is expressed as mmol of trolox equivalents per kg of beads ± standard deviation. A - alginate microbeads; P – pectin microbeads; A-CH – chitosan-coated alginate microbeads; P-CH – chitosan-coated pectin microbeads.

464 A. BELŠČAK-CVITANOVIĆ et al.: MODIFICATION OF FUNCTIONAL QUALITY… Chem. Ind. Chem. Eng. Q. 23 (4) 457−471 (2017)

A correspondence of the sample ranking Formulation of beers and radlers enriched with green according to the content of released total polyphenols tea and Ganoderma microencapsulates and dried and antioxidant capacity (Figure 3) and the total poly- powders phenols loading capacity and retained antioxidant For the purpose of enriching beers with different capacity was observed (Table 1). Namely, it can be types of microencapsulated bioactive compounds, observed that the samples exhibiting the highest load- only morphologically regular and visually acceptable ing capacity of total polyphenols and retained antioxi- microparticles enabling the most retarded release of dant capacity also released the highest contents of polyphenols were implemented in the production of the corresponding compounds. The discrepancy of pilsner or radler beers; alginate and alginate-chitosan the sample ranking based on the released TPC and coated, as well as chitosan coated pectin microbeads antioxidant capacity (lack of the same trend) may be encapsulating Ganoderma and green tea extract. the consequence of the non-selectivity and method mechanisms involved in the assays employed for the Polyphenolic content of microencapsulates-enriched release determinations, since it is well known that the beers presence of some substances such as the citric acid The addition of hydrogel microbeads encapsul- in chitosan-coated particles may quenche the ABTS ating either Ganoderma (Figure 4A) or green tea ext- radical thus providing smaller discrepancies between ract (Figure 4B) to pilsner beer or radler, regardless of the sample ranking based on their TP and antioxidant the employed carrier material, did not markedly capacity release profiles. The release of polyphenols modify the TPC of enriched drinks when compared to from the spray dried green tea extract was not control beers, so the addition of dried Ganoderma evaluated since thus obtained powders dissolve in extract and spray dried green tea extract were also water and do not display classic diffusion- or swell- evaluated. mediated release. Generally, there was no significant (p > 0.05) increase in TPC after implementation of Ganoderma

A 1200

900 ] -1

600

ac ad ae cde de cde cbde 300 a ab TPC [mg · l · [mg TPC

0 CE A A-CH P-CH DE CE A A-CH P-CH DE

ControlPilsner Light beer beer + + additions additions Radler + additions

B 1200 900 ] -1 600 bb aaaa 300 TPC [mg · l · [mg TPC 0 CE A A-CH P-CH SDE CE A A-CH P-CH SDE

ControlPilsner Light beer beer + +additions additions Radlers + additions Figure 4. The content of total polyphenolic compounds (mg GAE/L) of beers and radler enriched with hydrogel encapsulates and dried extracts of: A) Ganoderma and B) green tea polyphenols. Control – plain pilsner beer; CE – control pilsner beer or radler + liquid (non- -encapsulated) extract; A - alginate microbeads; A-CH – chitosan-coated alginate microbeads; P-CH – chitosan-coated pectin microbeads; DE –dried Ganoderma extract; SDE – spray dried green tea extract. The total polyphenolic concentration is expressed as mg of gallic acid equivalents per L of sample ± standard deviation. The values for each sample superscripted with the same alphabeth letter (a-e) are not significant (p > 0.05).

465 A. BELŠČAK-CVITANOVIĆ et al.: MODIFICATION OF FUNCTIONAL QUALITY… Chem. Ind. Chem. Eng. Q. 23 (4) 457−471 (2017) encapsulating hydrogel microbeads to either pilsner beer. With regard to the type of beer, primarily the or radler beer. Only the addition of dried Ganoderma combinations of pilsner beer and Ganoderma addi- extract enabled to significantly (p < 0.05) augment the tions, as well as radler and green tea were estab- TPC of enriched beer and radler, even up to 3-fold in lished as the preferred ones, which can be observed comparison to control drinks (Figure 3a). The addition as the much lower sensory scores of all properties for of chitosan coated alginate and pectin hydrogel beads radler with Ganoderma additions (Figure 5B) and encapsulating green tea bioactives to pilsner beer pilsner beer with green tea (Figure 5C). enabled a marked increase of TPC (Figure 4B) in In the category of pilsner beers, Ganoderma comparison to both plain control beer (sample C) and enriched samples were preferred by the panelists the control beer enriched only with liquid green tea over the green tea-enriched beers with respect to extract (sample CE). In case of chitosan coated alg- odour, taste, mouthfeel and overall acceptability. inate particles encapsulating green tea bioactives, up Although higher bitterness intensity was perceived in to 40% increase in the TPC of formulated pilsner beer Ganoderma-enriched beers in comparison to green was achieved. Likewise, in case of radlers, a signific- tea, this did not affect the overall preference of con- ant (p < 0.05) increase of 32% of TPC was obtained sumers towards the beer. Figure 5A reveals that upon the addition of chitosan coated alginate and Ganoderma hydrogel microbeads enriched pilsner pectin particles to radler. Slightly higher TPC that can beer exhibited equal or slightly higher bitterness int- be observed for radler in comparison to pilsner beer ensity, which is attributed to the bitter-tasting triter- may be attributed to the presence of lemon juice and penoids characteristic for Ganoderma [31]. However, thus ascorbic acid in radler drinks influencing the in case of Ganoderma enriched radler even lower overall polyphenolic concentration. The addition of bitterness intensity than the control beer was noted, spray dried green tea extract significantly (p < 0.05) which confirms the off-taste masking effect of micro- augmented the TPC of enriched samples, in compar- encapsulation approach. Namely it has been well est- ison to plain pilsner beer or radler (∼3.2-fold higher ablished that microencapsulation in different polymers TPC than control radler). With regard to different car- can mask the bitter taste of numerous pharmaceutic- rier materials, chitosan coated particles enabled to ally active compounds, as well as to retard the rel- deliver a higher TPC to pilsner and radler drinks, ease of bitter tasting active compound [32]. Also, which may indicate that a higher amount of polyphen- higher sweetness and sourness in radler drinks, olics is released from those particles in comparison to which may have masked and covered the bitterness alginate ones, when incorporated in beer as a med- of Ganoderma, can be attributed for the lower bitter- ium. These findings point to the fact that the release ness intensity of those beverages. of polyphenols may be triggered in beer due to its The application of Ganoderma in beer product- complex physicochemical properties. The release of ion was already evaluated [4] and confirmed to pro- catechins as the representative polyphenols of green vide sensory viable and attractive products, while tea is known to be more pronounced at lower pH [30], according to our knowledge no scientific studies are which corresponds to the lower pH of chitosan coated available adressing the incorporation of green tea particles in beer. Insignificant differences (p > 0.05) of extract in beer and its functional effects. The taste TPC between the alginate and pectin microencap- attributes of green tea microencapsulate-enriched sulate-enriched beers were established. The obtained radlers did not differ or were slightly higher than the results suggest that dried polyphenolic extracts and control indicating the favourable effect of enriching substrates are more appropriate for increasing the radler with green tea encapsulating microbeads (Fig- bioactive concentration of beers, rather than hydrogel ure 5D). Adding dried powders of Ganoderma or beads. However, since this is only valid in case of spray dried green tea was preferred only in case of compositional/bioactive purposes, other product pro- taste attributes of pilsner beer with Ganoderma pow- perties, such as the sensory properties, especially der. The addition of spray dried green tea powder to visual appearance and the stability of the product, both pilsner beer and radler markedly enhanced the need to be taken into account. aftertaste and bitterness attributes, providing generally lower overall acceptability of enriched drinks. Sensory properties of microencapsulates-enriched Regarding the visual appearance of enriched beers beers, no haze or turbidity development was noticed Sensory properties of microencapsulate-enriched following the addition of hydrogel microencapsulated beers differed markedly depending on the type of ext- Ganoderma or green tea extracts to beer. The most ract (Ganoderma or green tea extract) and the type of common cause of haze formation in beer production

466 A. BELŠČAK-CVITANOVIĆ et al.: MODIFICATION OF FUNCTIONAL QUALITY… Chem. Ind. Chem. Eng. Q. 23 (4) 457−471 (2017)

Colour Colour A 2 B 2,5 Appearance Transparency 1,5 Appearance 2 Odour 1,5 1 1 Overall acceptability Odour 0,5 Overall acceptability 0,5 Tast e 0 0

Acidity Tast e Acidity After taste

Bitterness After taste Bitterness Mouthfeel Mouthfeel

C CE A A-CH P-CH DE C CE A A-CH P-CH DE

Colour Colour C 2,5 D 2,5 cdAppearance 2 Transparency Appearance 2 Odour 1,5 1,5 1 1 Overall acceptability Odour 0,5 Overall acceptability 0,5 Tast e 0 0

Acidity Tast e Acidity After taste

Bitterness After taste Bitterness Mouthfeel Mouthfeel

C CE A A-CH P-CH SDE C CE A A-CH P-CH SDE Figure 5. Average scores of sensory quality indicators for: A) pilsner beer with Ganoderma additions; B) radler with Ganoderma additions; C) pilsner beer with green tea additions; D) radler with green tea additions. are the interactions between proteins and polyphen- ences in the visual appearance of microencapsulate ols, while polysaccharides do not participate in the enriched beers was observed. haze formation mechanism, but are instead simply Due to the natural turbidity of radler, the addition incorporated as haze particles [33]. Being composed of hydrogel microbeads encapsulating green tea ext- of polysaccharide biopolymers (alginate and pectin), ract did not visually modify the appearance of the microencapsulates in this study clearly did not enriched radler. Conversely, the addition of dried affect the postproduction haze formation in beer, rev- Ganoderma and spray dried green tea powders ealing their suitability for this purpose. The addition of caused significant modifications of their visual appear- hydrogel microbeads enabled to produce visually ance with a pronounced turbidity development (figure attractive products, in case of Ganoderma encapsul- not shown), which may be due to the Ganoderma ating microbeads due to their more pronounced, polysaccharides precipitation initiation by the pre- darker colour, while in case of green tea encapsul- sence of alcohol in the medium [34]. Namely, beers ating microbeads due to their larger size (d 0.5 < also contain a number of constituents, such as alco- 1000 µm) which evoked a specific consumer interest, hol and the hydrogen ion (pH), which can influence especially in case of morphologically regular particles. the interactions between the main constituents and The larger size of green tea encapsulating micro- protein–polyphenol haze formation [35]. beads contributed to the sensory preference of rad- Since visual appearance can be regarded as lers containing green tea microbeads exhibited by the one of the most important quality parameters of food, panelists (in comparison to control radler), and form- directly related to consumer perception and purchase ulation of an innovative, consumer-appealing product. decisions, which was in both pilsner beer and radlers With respect to the type of microparticles, chitosan achieved by using hydrogel microbeads, while the coated alginate particles encapsulating Ganoderma combinations of pilsner beer and Ganoderma addi- dispersed in pilsner beer (Figure 5A) provided visually tions, as well as radler plus green tea additions exhi- the most acceptable product, while in case of green bited better taste properties, only those combinations tea encapsulating microbeads no significant differ- with the highest overall acceptability were selected for additional beer quality control and stability analysis.

467 A. BELŠČAK-CVITANOVIĆ et al.: MODIFICATION OF FUNCTIONAL QUALITY… Chem. Ind. Chem. Eng. Q. 23 (4) 457−471 (2017)

Stability of the polyphenolic content and quality occur during storage. Namely, it is well established parameters during refrigerated storage of beer that flavonoid polyphenols are very prone to polymer- Beer quality is known to be affected by the sto- ization, or formation of complexes with proteins or rage conditions, such as temperature and light, which carbohydrates. In case of those compositional mod- is in the brewing process often solved by the addition ifications, overestimation of total polyphenolic con- of antioxidants (e.g., glutathione [36]), aiming to imp- centration may occur as already previously estab- rove beer stability. The approach of using micro- lished in numerous studies [37]. Although previous encapsulated polyphenolic antioxidants, especially studies on enriched beers [6] evidenced an achieved from green tea, may provide a natural alternative to increase in the polyphenolic concentration in compar- the improvement of beer stability during storage. In ison to control (beer enriched with honey), the stability the present study, fluctuations of TPC of hydrogel of enriched beers and its bioactive concentration was microbeads enriched beers during the storage period not evaluated. The results obtained in this study sug- were revealed, since during storage period a slightly gest that although no marked increase in the poly- higher TPC was revealed in both green tea and phenolic concentration of Ganoderma and green tea Ganoderma enriched beers (Figure 6A and B), in microbeads enriched beers and radler in comparison comparison to initial control sample (upon product- to plain, control beers was obtained, the stability of ion), while after one month of storage even a dec- polyphenolic concentration in the enriched beers was rease of TPC occurred (Figure 6A). achieved, pointing to a beneficial effect of functional The fluctuations were however not so obvious ingredients added for production of enriched beers for the dried extracts enriched beer and radler (DE from the preservation aspect. Ganoderma and spray dried green tea extract), since Quality parameters of finally selected enriched no significant differences between the TPC of those beers and radlers (Table 2) such as alcohol concen- samples even after one month of storage were deter- tration, real extract, apparent extract and energy mined. The observed fluctuations may be an indicator value revealed that the addition of either plain ext- of polyphenolic compositional changes that may racts or microencapsulates of Ganoderma and green tea (hydrogel and dried extracts) resulted with no sig-

A 800

600 ] -1 400

200 TPC [mg · l · [mg TPC 0 C C C C A A A A CE CE CE CE DE DE DE DE A-CH A-CH A-CH A-CH 1st day 10th day 20th day 30th day B 1200

900 ] -1 600

300 TPC [mg · l · [mg TPC

0 C C C C A A A A CE CE CE CE SDE SDE SDE SDE A-CH A-CH A-CH A-CH

1st day 10th day 20th day 30th day Figure 6. Storage stability of total polyphenolic compounds (mg GAE/L) in A) pilsner beer with Ganoderma mushroom additions and B) radler with green tea additions during one month. The total polyphenolic concentration is expressed as mg of gallic acid equivalents per L of product ± standard deviation

468 A. BELŠČAK-CVITANOVIĆ et al.: MODIFICATION OF FUNCTIONAL QUALITY… Chem. Ind. Chem. Eng. Q. 23 (4) 457−471 (2017)

Table 2. The concentration of alcohol, real and apparent extract and energy value of enriched beers

Enriched 1st day 30th day beer samples Alcohol con- Real extract Apparent Energy, kJ Alcohol con- Real extract Apparent Energy, kJ centration, mg·ml-1 % extract, % centration, mg·ml-1 % extract, % Pilsner beer enriched with Ganoderma additions C 4.6 3.9 2.2 165.6 4.6 4.1 2.5 164.3 CE 4.7 3.9 2.2 165.7 4.6 4.0 2.3 166.2 A 4.7 3.8 2.1 164.4 4.5 4.1 2.5 164.8 A-CH 4.6 4.0 2.3 165.6 4.3 4.1 2.5 161.3 DE 4.3 4.2 2.4 164.8 4.3 4.2 2.5 165.2 Radler enriched with green tea additions C 2.2 7.6 6.6 172.1 2.2 7.6 6.8 164.3 CE 2.4 7.8 6.9 173.3 2.2 7.7 6.9 169.8 A 1.9 7.8 7.1 164.1 1.9 7.9 7.2 164.5 A-CH 1.8 7.4 6.8 156.2 2.1 7.9 7.2 168.6 SDE 2.5 7.8 6.9 176.2 2.5 7.9 6.9 175.2

nificant differences in these parameters compared to radler did not provide a significant enhancement of the control drinks. their TPC, which was instead achieved by adding The values obtained for each of the parameters dried or spray dried extracts. The combinations of do not differ notably between the control beer pilsner beer enriched with Ganoderma and radler with enriched with plain Ganoderma extract and beers green tea hydrogel microbeads and spray dried ext- enriched with Ganoderma hydrogel microparticles, ract were preferred according to their sensory pro- which is also applicable to beers enriched with plain files, especially lower bitterness, astringency and green tea extract and beers containing green tea more intense herbal flavour, respectively. hydrogel microparticles. Although radlers were char- Acknowledgements acterized by higher values of real and apparent extract (due to the higher saccharides concentration) no This work was performed within the framework significant effect of enrichment with different micro- of the research projects 46001, 46010 and 31020, encapsulated green tea delivery systems was obs- supported by the Ministry of Education, Science and erved, except of a slight increase of real and apparent Technological Development, Republic of Serbia. extract in the beer and radler with dried and spray dried extracts. The four-week refrigerated storage had REFERENCES no significant effect on the quality parameters of [1] P. Quifer-Rada, A. Vallverdú-Queralt, M. Martínez-Huél- enriched beers, which supports the notion that the addi- amo, G. Chiva-Blanch, O. Jáuregui, R. Estruch, R. tion of microparticles with polyphenolic compounds Lamuela-Raventós, Food Chem. 169 (2015) 336-343 does not affect the basic parameters of beer quality. [2] H. Zhao, H. Li, G. Sun, B. Yang, M. Zhao, J. Sci. Food Agric. 93 (2013) 910–917 CONCLUSIONS [3] A. Kalušević, G. Uzelac, S. Despotović, I. Leskošek-Ču- kalović, M. Nikšić, in Proceedings of the XV Advising on In this study, the potential and suitability of imp- biotechnology, Serbia, 2010, pp. 743-749 lementing microencapsulated bioactive compounds [4] I. Leskošek-Čukalović, S. Despotović, N. Lakić, M. Nikšić, from Ganoderma mushroom and green tea extracts V. Nedović, V. Tešević, Food Res. Int. 43 (2010) 2262- for the formulation of innovative, attractive beers with –2269 improved functional properties was revealed. Electro- [5] M. Veljovic, R. Djordjevic, I. Leskosek-Cukalovic, N. static extrusion assisted microencapsulation of poly- Lakic, S. Despotovic, S. Pecic, V. Nedovic, J. Inst. Brew. phenolic compounds in alginate and pectin particles, 116 (2010) 440-444 mediated by ionic gelling, enabled to entrap up to [6] A. Kalušević, G. Uzelac, M. Veljović, S. Despotović, M. ∼70% of polyphenolic compounds from both exam- Milutinović, I. Leskošek-Čukalović, V. Nedović, in Pro- ceeding of 11th International Congress of Engineering and ined substrates, with up to ∼72% of their antioxidant food, Greece, 2011, pp. 2057-2058 capacity recovered after encapsulation. The incorpor- [7] Y. Ting, Y. Jiang, C.T. Ho, Q. Huang, J. Funct. Foods 7 ation of formulated hydrogel microparticles to beer or (2014) 112-128

469 A. BELŠČAK-CVITANOVIĆ et al.: MODIFICATION OF FUNCTIONAL QUALITY… Chem. Ind. Chem. Eng. Q. 23 (4) 457−471 (2017)

[8] V. Nedović, A. Kalušević, V. Manojlović, S. Lević, B. [24] H. Moreira, F. Munarin, R. Gentilini, L. Visai, P.L. Granja, Bugarski, Procedia Food Sci. 1 (2011) 1806-1815 M.C. Tanzi, P. Petrini, Carbohydr. Polym. 103 (2014) [9] A. Belščak-Cvitanović, R. Stojanović, V. Manojlović, D. 339-347 Komes, I. Juranović-Cindrić, V. Nedović, B. Bugarski, [25] B.M. Yapo, D. Gnakri, in Polysaccharides, Bioactivity and Food Res. Int. 44 (2011) 1094-1101 Biotechnology, K.G. Ramawat, J.M. Mérillon, Eds., [10] B. Lupo, A. Maestro, M. Porras, J.M. Gutierrez, C. Gon- Springer International Publishing, Cambridge, 2014, pp. zales, Food Hydrocolloids 38 (2014) 56-65 1-18 [11] F.P. Flores, R.K. Singh, W.L. Kerr, R.B. Pegg, F. Kong, [26] A. Belščak-Cvitanović, V. Đorđević, S. Karlović, V. Pav- Food Chem. 153 (2014) 272–278 lović, D. Komes, D. Ježek, B. Bugarski, V. Nedović, Food Hydrocolloids 51 (2015) 361-374 [12] S.M. Chacko, P.T. Thambi, R. Kuttan, I. Nishigaki, Chin. Med. 5 (2010) 13-22 [27] A. Belščak-Cvitanović, D. Komes, S. Karlović, S. Dja- ković, I. Špoljarić, G. Mršić, D. Ježek, Food Chem. 167 [13] J. Xie, J. Zhao, D.J. Hu, J.A. Duan, Y.P. Tang, S.P. Li, Molecules 17 (2012) 740-752 (2015) 378-386 [28] M.M. Ramos-Tejada, J.D.G. Duran, A. Ontiveros-Ortega, [14] J. Lee, E. Kim, D. Ching, H.G. Lee, Colloids Surfaces, B 74 (2009) 17-22 M. Espinosa-Jimenez, R. Perea-Carpio, E. Chibowski, Colloids Surfaces, B 24 (2002) 309-320 [15] N. Fu, Z. Zhou, T.B. Jones, T.T.Y. Tan, W.D. Wu, S.X. [29] S. Huang, J. Li, Y. Li, Z. Wang, Int. J. Biol. Macromol. 48 Lin, X.D. Chen, P.P.Y. Chan, Int. J. Pharm. 413 (2011) (2011) 165–169 155-166 [30] H. Pool, D. Quintanar, J. de Dios Figueroa, C.M. Mano, J. [16] A. Rashidinejad, E.J. Birch, D. Sun-Waterhouse, D.W. Etelvino, H. Bechara, L.A. Godínez, S. Mendoza, J. Everett, Food Chem. 156 (2014) 176–183 Nanomater. 2012 (2012) Article ID 145380 [17] D. Zhao, Doctoral thesis, UCL (University College [31] T. Nishitoba, H. Sato, S. Sakamura, Agric. Biol. Chem. 52 London), 2014 (1988) 1791-1795 [18] Y.‐J. Gao, B. Yuan, W.‐J. Yang, Y. Fang, N. Ma, Q.‐H. [32] V. Sharma, H. Chopra, Int. J. Pharm. Pharm. Sci. 2 Hu, Mycosystema 33 (2014) 483-492 (2010) 14-18 [19] Association of Analytical Communities, AOAC official [33] E. Steiner, T. Becker, M. Gastl, J. Inst. Brew. 116 (2010) method 920.151. in AOAC International Official Methods th 360-368 of Analysis, 16 ed, Arlington, 1999, p. 5 [20] J. Lachman, V. Hosnedl, V. Pivec, M. Orsák, in Pro- [34] J. Habijanič, M. Berovič, B. Wraber, D. Hodzar, B. Boh, Food Technol. Biotechnol. 39 (2001) 327-331 ceedings of Conference Cereals for Human Health and Preventive Nutrition, Czech Republic, 1998, pp. 118-125 [35] M.M.L. Rovaletti, E.I. Benítez, N.M.J. Martinez Amezaga, N.M. Peruchena, G.L. Sosa, E.J. Lozano, Food Res. Int. [21] W. Brand-Williams, M.E. Cuvelier, C. Berset, LWT-Food Sci. Technol. 28 (1995) 25-30 62 (2014) 779-785 [36] L. Gijs, P. Perpète, A. Timmermans, C. Guyot-Declerck, [22] International Organization for Standardization , ISO 8586- 2/2008, Geneva, 2008 P. Delince, S. Collin, J. Am. Soc. Brew. Chem. 62 (2004) 97-102 [23] Mitteleuropäische Brautechnische Analysenkommission [37] D. Komes, D. Horžić, A. Belščak, K. Kovačević Ganić, I. (MEBAK) Method Collection of the Mitteleuropäische Vulić, Food Res. Int. 43 (2010) 167-176. Brautechnische Analysenkommission, 2014

470 A. BELŠČAK-CVITANOVIĆ et al.: MODIFICATION OF FUNCTIONAL QUALITY… Chem. Ind. Chem. Eng. Q. 23 (4) 457471 (2017)

ANA BELŠČAK-CVITANOVIĆ1 MODIFIKACIJA FUNKCIONALNOG KVALITETA VIKTOR NEDOVIĆ2 PIVA KORIŠĆENJEM BIOAKTIVNIH JEDINJENJA 2 ANA SALEVIĆ MIKROENCAPSULIRANOG ZELENOG ČAJA SAŠA DESPOTOVIĆ2 1 (Camellia sinensis L.) I HRASTOVE SJAJNICE DRAŽENKA KOMES MIOMIR NIKŠIĆ2 (Ganoderma lucidum L.) BRANKO BUGARSKI3 2 Sve veći interes za proizvodnju često konzumiranih funkcionalnih prehrambenih proiz- IDA LESKOŠEK ĆUKALOVIĆ voda usredsredilaa je istraživanja primeni bioaktivnih jedinjenja mikroencapsulisane 1 Department of Food Engineering, Faculty of gljive hrastova sjajnica i zelenog čaja u proizvodnji piva. Mikrokapsulacija zelenog čaja i Food Technology and Biotechnology, ekstrakta gljive pomoću elektrostatičke ekstruzije daje čestice u rasponu od 490 mm do University of Zagreb, Zagreb, Croatia 1000 mm, sa sadržajem ukupnih polifenola do 75%. U proizvodnji piva su korišćeni osu- 2Katedra za prehrambenu tehnologiju i šeni ekstrakti u prahu, kao i mikročestice sa kapsulama ekstrakta gljive i zelenog čaja biohemiju, Poljoprivredni fakultet, Univerzitet koje su pokazale najbolje morfološke osobine i odloženo oslobađanje polifenola (pektin- u Begradu, Nemanjina 6, 11081 Begrad- ske mikrosfere obložene alginatom, alginatom i hitozanom ili hitozanom). Dodavanje Zemun, Srbija 3Katedra za Hemijsko inženjerstvo, mikrosfera sa gljivom u pilsner pivo nije povećalo koncentraciju polifenola (TPC), za raz- Tehnološko-metalurški fakultet, Univerzitet u liku od dodavanja mikrosfera sa zelenim čajem na radler pivo, dok je dodavanje suvog Begradu, Karnegijeva 4, 11120 Begrad, ekstrakta gljive ekstrakte zelenog čaja sušenog u spreju povećalo TPC do tri puta. Pils- Srbija ner pivo obogaćeno suvim ekstraktom gljive i radler pivo obogaćeno ekstraktom zelenog čaja sušenog u spreju imaju poboljšana senzorna svojstva, zbog najnižeg intenziteta NAUČNI RAD gorčine i najizraženije biljne arome dodatih ekstrakata. Kod pilsner piva obogaćenom hidrogelnim mikrosferama gljive zapažene su fluktuacije TPC, dok je radler pivo oboga- ćeno hidrogelnim mikrosferama zelenog čaja pokazao bolju stabilnost. Razvijena meto- dologija obezbeđuje proceduru pogodnu za mikroencapsulaciju - obogaćenje pića i prehrambenih proizvoda, čime se postavlja pouzdana osnova za buduću proizvodnju funk- cionalne hrane primenom strategija mikroenkapsulacije.

Ključne reči: pivo, hrastova sjajnica. zeleni čaj. mikroenkapsulacija, polifenolni anti- oksidanti.

471

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 473−481 (2017) CI&CEQ

DANICA SAVANOVIĆ MELTING AND CRYSTALLIZATION DSC RADOSLAV GRUJIĆ2 PROFILES OF DIFFERENT TYPES OF MEAT SLADJANA RAKITA3 3 ALEKSANDRA TORBICA Article Highlights 4 RANKO BOZIČKOVIĆ • Freezing rate influences the thermo-physical properties of frozen meat • 1 Types of meat influence the thermal-physical properties of frozen meat Faculty of Technology, University • The temperatures of crystallization and melting of meat depend on freezing rate of Banja Luka, Banja Luka, • The temperature of crystallization is different from the temperature of melting of some Republic of Srpska, Bosnia and type of meat Herzegovina • Freezing rate affects the ratio of freezable/unfreezable water in the frozen meat 2Faculty of Technology, University of East Sarajevo, East Sarajevo, Abstract Republic of Srpska, Bosnia and The aim of this study was to test the influence of scanning rate and meat type on Herzegovina the thermo-physical properties of meat and content of the freezable water in 3 Institute of Food Technology, frozen meat, using differential scanning calorimetry (DSC). In this study, three University of Novi Sad, Novi Sad, types of meat were investigated: beef (M. Longissimus dorsi), pork (M. Serbia Longissimus dorsi), and chicken meat (Pectoralis major). The cooling rate 4Faculty of Mechanical affected the onset (T on), peak (T ) and end (T end) temperatures of crystal- engineering, University of East c c c lization process of beef meat (p < 0.05). Decreasing cooling rate from 20 to 2 Sarajevo, East Sarajevo, Republic ° of Srpska, Bosnia and Herzegovina C/min resulted in significant (p < 0.05) change of the crystallization enthalpy (ΔHc) of beef meat, from -220.17 to -168.20 J/g, respectively. Reduction of the SCIENTIFIC PAPER heating rate caused significant (p < 0.05) decrease in enthalpy of melting (ΔHm) for beef meat, from 228.87 to 161.13 J/g. The heating rate affected the peak (Tm) UDC and end temperatures (Tmend) of melting process of beef meat (p < 0.05). The type of meat did not have effect on ΔHc and ΔHm as well as temperature of crystallization (Tcon, Tc and Tcend) and temperature of melting (Tm and Tmend) in meat. Significant (p < 0.05) change in freezable water content were recorded between heating rate 20 °C/min and other heating rates, for all three meat types. Keywords: DSC, meat, crystallization, melting.

Differential scanning calorimetry (DSC) is a the energy released or absorbed in the form of latent thermoanalytical technique used to study the thermal heat during phase transition of that sample [2,7,8]. behaviors of food and pharmaceutical systems [1-6]. The most commonly occurring phase transitions This technique has some advantages over other clas- in food are crystallization of water during freezing and sical detection methods, as it is rapid and does not melting of ice during thawing of food [3,9-12]. These require additional sample preparation or solvent util- processes are extremely important, because physico- ization and, therefore, it is an environmentally friendly chemical and sensory characteristics of foods may be technique. A DSC generated thermogram, in which altered after freezing and thawing [13-17]. the rate of heat transfer is plotted versus temperature Freezing and thawing processes play an import- for a small amount of sample, provides information on ant role in food processing [9,18]. Freezing is the most widely used technique for food preservation and Correspondence: D. Savanović, University of Banja Luka, especially if large quantities of meats are frozen Faculty of Technology, Vojvode Stepe Stepanovica 73, 78000 before use [6,14,19]. To optimize these processes Banja Luka, Republic of Srpska, Bosnia and Herzegovina. and the quality of the final product, it is very useful to E-mail: [email protected] Paper received: 7 July, 2016 analyze the phase transitions that occur in the pro- Paper revised: 5 October, 2016 duct [2,20]. However, mathematical description of Paper accepted: 11 January, 2017 phase transition is difficult because of the heterogen- https://doi.org/10.2298/CICEQ160707001S

473 D. SAVANOVIĆ et al.: MELTING AND CRYSTALLIZATION DSC PROFILES… Chem. Ind. Chem. Eng. Q. 23 (4) 473−481 (2017) eous nature of food [2,21-23]. Therefore, it is import- crystallization process. The following temperatures ant to determine analytical techniques for monitoring were analyzed for the melting process: temperature of melting and crystallization processes in food. These onset melting (Tmon), peak melting (Tm) and end techniques are based on the measurements of ther- melting (Tmend). The melting temperature interval mo-physical properties of food, which change during was computed as the width of the melting peak (ΔTm a phase transition [2,23,24]. = Tmend - Tmon) and the crystallization temperature Thermo-physical properties of food depend to a interval was computed as the width of the crystal- great extent on food composition and previous treat- lization peak (ΔTc = Tcon - Tcend). Enthalpy of melting ment, but attention must be paid to the use of different ΔHm (J/g) was determined as the area, limited by the measuring techniques because they can also influ- melting curve and baseline. Proteus software, version ence the experimental results [23]. Calorimetry is one 6.1.0 (NETZSCH–Gerätebau GmbH, Germany) was of the most powerful tools to study thermo-physical used to analyze the DSC thermograms and evaluate properties of food and phase transition phenomena the thermo-physical properties of meat. [25] and it can be used in the analysis and charac- Influence of scanning rate on thermo-physical terization of oils and fats, crystallization and melting properties of beef meat profiles, enthalpy of transitions and polymorphic forms [1,5,8,26,27]. The procedure of determining enthalpy Samples of beef meat were cooled and heated in foods, the freezing and melting temperature and at five rates (2, 5, 10, 15, 20 °C/min). In each scan, the content of unfreezable water by differential scan- the sample was equilibrated at 20 °C for 5 min and ning calorimetry (DSC) is fast and allows systematic then cooled to -40 °C at a preset rate; after an iso- data acquisition and processing to be obtained by util- thermal holding stage (-40 °C) of 5 min, the sample izing specific software [23,24]. DSC analysis can be was heated at same rate to 20 °C. All the measure- performed with different scanning rates, however, ments were performed in triplicate. there are little studies in the literature which compare Influence of meat type on thermo-physical properties DSC thermograms in case of different cooling and of meat heating rates of food. Samples of beef, pork and chicken meat were The aim of this study was to test the influence of cooled and heated at 10 °C/min. This scan rate was selected factors, such as scanning rate and meat chosen because it is the most commonly used scan- type, on the thermo-physical properties of meat and ning rate for DSC analysis. In each scan the sample content of the freezable water in frozen meat. was equilibrated at 20 °C for 5 min and then cooled to ° ° MATERIALS AND METHODS -40 C at a scan rate of 10 C/min; after an isothermal holding stage (-40 °C) of 5 min, the sample was Standard procedure heated at same rate to 20 °C. All the measurements were performed in triplicate for each meat sample. In this study, three types of meat were investigated: beef (M. Longissimusdorsi), pork (M. Longissi- Determination of the freezable water mus dorsi) and chicken meat (Pectoralis major). In all The percent of freezable water (FW) was cal- experiments, we used fresh post-rigor meat. Meat culated using the area of the integrated endothermic was purchased at the slaughter house. peak at the range of 0 °C. The endothermic peak Differential scanning calorimetry thermograms arises from the phase transition of ice into water. Eq. were obtained using a differential scanning calori- (1) was used in order for the freezable water of the meter DSC (204 F1 Phoenix, NETZSCH–Gerätebau meat to be estimated [19]: GmbH, Germany). Samples (14±2 mg) were weighed 100Q into 25 µl capacity aluminum pans. After that, pans FW (%) = (1) were hermetically sealed. An identical empty pan was Hmfs used as a reference sample during the experiments. where FW is the percent of freezable water, Q is the The calibration of the cell was made following the enthalpy of melting (J/g), H represents the heat of DSC manufacturer’s recommendation. Flow rate of f fusion ice–water equal with 333.50 J/g of ice/water, purge nitrogen atmosphere was 20 ml/min. and m is the mass of the sample. The percent of The temperature of onset crystallization (T on), s c unfreezable water (UFW) was calculated by subtract- peak crystallization (T ), end crystallization (T end) c c ing the percent of freezable water from the percent of and enthalpy of crystallization ΔH (J/g) were mea- c total water [28]. The total water content of meat was sured from the cooling curves and analyzed for the

474 D. SAVANOVIĆ et al.: MELTING AND CRYSTALLIZATION DSC PROFILES… Chem. Ind. Chem. Eng. Q. 23 (4) 473−481 (2017) determined by the method of drying at 105±2 °C to ing rates (2, 5, 10, 15 and 20 °C/min). The tempe- constant mass [29]. rature at which begins the process of crystallization (T on) in beef meat significantly (p < 0.05) changed Statistical analysis c with changing cooling rate (Table 1). Therefore, the The results of this study were presented as the mean values of Tcon of beef meat were: -14.93, mean values accompanied with their standard devi- -19.40 and -18.10 °C, for the rate of 2, 10 and 20 ations of three measurements. One factor analysis of °C/min, respectively (Table 1). The mean values of variance (ANOVA) was performed using IBM SPSS Tcon did not show difference for scanning rate from 5 Statistics for Windows, version 22.0 (Armonk, NY, to 20 °C/min (p ˃ 0.05), and Tcon for 2 °C/min was USA). Where significant differences (p < 0.05) were only different from 10 °C/min (p < 0.05). detected, Tukey’s multiple comparison was used to Theoretically, the onset values for pure sub- compare treatment means and create statistically stances should be always the same, regardless of the homogeneous groups. scanning rate. For the tested meat samples, various cooling rates caused different courses of crystalliz- RESULTS AND DISCUSSION ation, which was manifested in different shapes of peaks, their sizes, as well as different temperatures The effect of scanning rate on the crystallization and [8]. A food material is a complex biochemical system melting processes in beef meat where multiple interactions occur during any type of DSC is a suitable method to characterize phase process [25], thus differing from pure substances. transitions that require the intake or release of ther- Marini et al. [30] reported that chemical com- mal enthalpy, such as crystallization and melting [1]. position, especially water content, presents a great Figure 1 presents crystallization curves of beef meat influence in thermo-physical properties of meat pro- obtained using different cooling rates (2, 5, 10, 15 and ducts. The soluble components such as various 20 °C/min). The shape of the curves, as well as the sugars, ions and acids and soluble proteins, will con- width of crystallization peaks changed depending on tribute to the freezing point depression, while the ins- the cooling rate (curves are presented in the same oluble components such as fat and the insoluble pro- scale). As seen in Figure 1, the width of crystallization teins will not. Furthermore, some components, such as peaks decreased with decreasing cooling rate, which protein and starch, have such high molecular masses resulted in significant (p < 0.05) change of the crys- that their contribution to the mole fraction of solutes is Δ tallization enthalpy ( Hc) of beef meat, from -220.17 negligible; it is the small molecules such as sugars, J/g for the rate of 20 °C/min to -168.20 J/g for the rate ions and acids, which contribute appreciably [31]. of 2 °C/min (Table 1). Tomaszewska-Gras [8] also Similarly, as in the case of onset crystallization observed the decrease in the width of crystallization temperature (Tcon), the mean values of peak tempe- peaks of milk fat with a decreasing cooling rate. ratures (Tc) and temperature of end crystallization

Different transition temperatures of crystalliz- (Tcend) for beef meat significantly (p<0.05) changed ation for beef meat were recorded with various cool- with the increase in the cooling rate (Table 1). The

Figure 1. DSC crystallization curves of beef meat at different cooling rates (2, 5, 10, 15 and 20 °C/min).

475 D. SAVANOVIĆ et al.: MELTING AND CRYSTALLIZATION DSC PROFILES… Chem. Ind. Chem. Eng. Q. 23 (4) 473−481 (2017)

Table 1. Temperatures and enthalpies of crystallization and melting process for beef meat, in relation to different cooling and heating rates; data are expressed as mean ± standard deviation; mean values in the same column followed by different letters indicate significant difference (p < 0.05); Tcon - temperature of onset crystallization; Tc - temperature of peak crystallization; Tcend - temperature of end crystallization; ΔTc - width of crystallization peak; ΔHc - enthalpy of crystallization; Tmon - temperature of onset melting; Tm - temperature of peak melting; Tmend - temperature of end melting; ΔTm - width of melting peak; ΔHm - enthalpy of melting

Cooling/heating rate Crystallization Melting °C/min Temperature, °C Entalphy Temperature, °C Entalphy Δ –1 Δ –1 Tcon Tc Tcend ΔTc Hc / J g Tmon Tm Tmend ΔTm Hm / J g 2 -14.93b -14.93b -15.60c 0.67a -168.20b -2.87a 0.23a 0.60a 3.47a 161.13a ±1.07 ± 1.07 ± 1.13 ± 0.06 ± 6.30 ± 0.21 ± 0.06 ± 0.17 ± 0.00 ± 9.05 5 -16.23ab -16.40ab -17.77bc 1.53b -171.00b -1.60a 1.90b 2.93b 4.53a 169.77a ± 2.31 ± 2.31 ± 2.10 ± 0.23 ± 4.10 ± 1.25 ± 0.46 ± 0.32 ± 1.36 ± 16.60 10 -19.40a -19.63a -21.93a 2.53c -179.70b -1.43a 3.73c 5.27c 6.70b 178.93a ± 1.00 ± 1.05 ± 0.85 ± 0.15 ± 11.14 ± 0.06 ± 0.06 ± 0.15 ± 0.17 ± 10.66 15 -18.03ab -18.33ab -20.73ab 2.70c -182.97b -2.63a 5.10d 7.13d 9.77c 183.07a ± 0.75 ± 0.75 ± 0.55 ± 0.20 ± 6.29 ± 0.06 ± 0.30 ± 0.25 ± 0.25 ± 6.29 20 -18.10ab -18.60ab -22.00a 3.90d -220.17a -2.63a 7.60e 10.60e 13.23d 228.87b ± 1.30 ± 1.20 ± 0.90 ± 0.40 ± 2.55 ± 0.06 ± 0.10 ± 0.10 ± 0.11 ± 1.21

mean values of Tc of beef meat were: -14.93, -19.63 cess (of ice) in beef meat. Similarly, as in the process and -18.60 °C for the rate of 2, 10 and 20 °C/min, of crystallization, the shape of the curves, the position respectively. The mean values of Tc did not show and the size of peaks changed with various heating difference for scanning rate from 5 to 20 °C/min (p ˃ rates.

0.05), and Tc for 2 °C/min was statistically different The crystallization curves exhibited narrower from 10 °C/min (p < 0.05). The mean value of Tcend peaks when compared to the curves of melting. This of beef meat was -15.60 °C for the rate of 2 °C/min indicated that during thawing, phase change occurred and -22.00 °C for the rate of 20 °C/min (Table 1). over a wider temperature range and more gradually Moreover, it was observed that the cooling rate had when compared to freezing, as was previously obs- significant (p < 0.05) effect on the crystallization tem- erved in the literature [2]. Analogously, as in the case perature interval (ΔTc, Table 1). of crystallization, a reduction of the heating rate This study was also conducted on the process of caused significant (p < 0.05) decrease in width of melting at different scanning rates. Figure 2 presents melting peak (ΔTm) for beef meat. In this study, sig- the melting curves of beef meat obtained at different nificant (p > 0.05) differences were not found between heating rates (2, 5, 10, 15 and 20 °C/min). As it is the values of onset melting temperature (Tmon) in the presented in Figure 2, different heating rates (from 2 range of heating rates from 2 to 20 °C/min. On the to 20 °C/min) influenced changes in the melting pro- other hand, significant (p < 0.05) differences between

Figure 2. DSC melting curves of beef meat at different heating rates (2, 5, 10, 15 and 20 °C/min).

476 D. SAVANOVIĆ et al.: MELTING AND CRYSTALLIZATION DSC PROFILES… Chem. Ind. Chem. Eng. Q. 23 (4) 473−481 (2017)

mean values of peak melting (Tm) and end melting place, increased with increasing heating rate. For temperature (Tmend) of beef meat were observed for faster thawing procedures, enthalpy of melting inc- all the heating rates. The values of enthalpy of melt- reased during the melting of meat samples (Figure 4). ing did not differ (p ˃ 0.05) for heating rates from 2 to Influence of meat type on thermo-physical properties 15 °C/min, while significantly (p < 0.05) different value of meat was recorded for heating rate of 20 °C/min (Table 1). Enthalpy is used for calculating the total heat to The DSC method can be used to obtain the free- be removed and to determine the rate removal during zing point, heat of fusion (ΔH) and apparent specific refrigeration and freezing of food products [32]. The heat [33]. Figures 5 and 6 present the crystallization peak area represents the melting latent heat of ice in and melting DSC curves of beef, pork and chicken the tested frozen meat sample. meat for the scanning rate of 10 °C/min. Obtained The influence of the cooling rate on the tempe- results were analyzed statistically, and no significant ratures and enthalpy of crystallization for beef meat is (p ˃ 0.05) effect of meat type on temperatures (Tcon, presented in Figure 3, and the influence of the heat- Tc and Tcend) and enthalpy of crystallization (ΔHc) ing rate on the temperatures and enthalpy of melting were observed (Table 2). for beef meat is demonstrated in Figure 4. In the melting process, no significant (p ˃ 0.05) The obtained parameters which indicate the differences were found between meat type and temperature of peak melting (T ), end melting (T end) crystallization temperatures of the meat (Tcon, Tc and m m and enthalpy of melting (ΔH ) at the scanning rate of Tcend) showed that these temperatures are lower m when meat undergo freezing at higher freezing rate. 10 °C/min. In this study, it was observed that the The width of the crystallization temperature interval crystallization temperature interval (ΔTc), melting temperature interval (ΔT ) and temperature of onset melt- (ΔTc), in which the crystallization process of water in m meat takes place, increased with increasing cooling ing (Tmon) of beef meat were significantly (p < 0.05) dif- rate. In addition, the enthalpy of crystallization of meat ferent from those of pork and chicken meat (Table 2). samples gradually decreased with increasing freezing The DSC method can be used to obtain the free- rate from 2 to 15 °C/min, while the reduction was zing point, heat of fusion (ΔH) and apparent specific more pronounced at freezing rate of 20 °C/min (Fig- heat [33]. The onset of melting was considered as the freezing point (Tf). Freezing is a safe and a commonly ure 3). Similarly, the width of melting peak (ΔTm), in which the melting process of water in meat takes used preservation method for meat and fish products in general [30,33].

Figure 3. Influence of cooling rate on the temperatures and enthalpy of beef meat crystallization (Tcon - temperature of onset

crystallization; Tc - temperature of peak crystallization; Tcend - temperature of end crystallization; ΔTc - width of crystallization peak;

ΔHc - enthalpy of crystallization).

477 D. SAVANOVIĆ et al.: MELTING AND CRYSTALLIZATION DSC PROFILES… Chem. Ind. Chem. Eng. Q. 23 (4) 473−481 (2017)

Figure 4. Influence of heating rate on the temperature and enthalpy of beef meat (Tmon - temperature of onset melting; Tm - temperature

of peak melting; Tmend - temperature of end melting; ΔTm - width of melting peak; ΔHm - enthalpy of melting).

Figure 5. DSC crystallization curves of three types of meat at cooling rate of 10 °C/min.

Figure 6. DSC melting curves of three types of meat at heating rate of 10 °C/min.

478 D. SAVANOVIĆ et al.: MELTING AND CRYSTALLIZATION DSC PROFILES… Chem. Ind. Chem. Eng. Q. 23 (4) 473−481 (2017)

Table 2. Temperatures and enthalpies of crystallization and melting process in beef, pork and chicken meat for the rate of 10 °C/min

Crystallization Melting Sample Temperature, °C Entalphy Temperature, °C Entalphy Δ –1 Δ –1 Tcon Tc Tcend ΔTc Hc / J g Tmon Tm Tmend ΔTm Hm / J g Beef meat -14.93b -14.93b -15.60c 0.67a -168.20b -2.87a 0.23a 0.60a 3.47a 161.13a ±1.07 ± 1.07 ± 1.13 ± 0.06 ± 6.30 ± 0.21 ± 0.06 ± 0.17 ± 0.00 ± 9.05 Pork meat -16.23ab -16.40ab -17.77bc 1.53b -171.00b -1.60a 1.90b 2.93b 4.53a 169.77a ± 2.31 ± 2.31 ± 2.10 ± 0.23 ± 4.10 ± 1.25 ± 0.46 ± 0.32 ± 1.36 ± 16.60 Chicken meat -19.40a -19.63a -21.93a 2.53c -179.70b -1.43a 3.73c 5.27c 6.70b 178.93a ± 1.00 ± 1.05 ± 0.85 ± 0.15 ± 11.14 ± 0.06 ± 0.06 ± 0.15 ± 0.17 ± 10.66

The initial freezing point is defined as the tem- Unfreezable water is the amount of water unav- perature at which the first ice crystals start to form. ailable for freezing in a food product at reference tem- Marini et al. [30] reported that chicken sausages perature of -40 °C [9,32,35]. Unfreezable water is (Frankfurter type), mortadela (Bologna type) and bound water, whereas freezable water fraction ref- mechanically deboned chicken meat (MDCM) had lects the fraction of free water of the total water in a very different initial freezing temperature and end product. point of freezing. Initial freezing temperature and end The mean values of freezable water and unfree- point of freezing for mortadela (bologna) were -4.46 zable water content in three meat types are presented and -10.14 °C, respectively. MDCM presented an ini- in Table 3. Analogously, as enthalpy of melting, the tial freezing temperature of -0.43 °C and end point of mean values of freezable water contents and unfree- freezing of -4.46 °C, and for frankfurter sausages these zable water contents did not differ (p > 0.05) for heat- temperatures were -2.49 and -9.71 °C, respectively. ing rates from 2 to 15 °C/min. However, significant (p < 0.05) increase in freezable water content and the The effect of scanning rate on the content of decrease in unfreezable water content were recorded freezable and unfreezable water in meat between heating rates 2 and 20 °C/min for all three Different type of food has different water content meat types (Table 3). that may undergo the phase transition, and hence on Tolstorebrov et al. [3] were investigated phase the basis of sample mass only partial water–ice trans- transitions for Atlantic Salmon, Cod, Herring, Mack- formations can be expected [25]. Water present in erel and Rainbow Trout. The amount of unfreezable food can be classified into two categories according water was calculated by the DSC melting endotherm to its reaction to freezing process: freezable and integration for scaning rate 5 °C/min and was in the unfreezable water. During freezing, only freezable range between 5.1 and 8.6% for all investigated water crystallizes into ice, whereas unfreezable water samples. undergoes no changes [34].

Table 3. Content of freezable and unfreezable water (%) in relation to different type of meat and different heating rate; FW - freezable water; UFW - unfreezable water; data are expressed as mean ± standard deviation. Mean values in the same column followed by different letters indicate significant difference (p < 0.05)

Heating rate Beef meat Pork meat Chicken meat ° C/min FW UFW FW UFW FW UFW 2 48.32a 26.52b 57.94a 15.64b 52.07a 22.69b ± 2.71 ± 2.71 ±1.19 ± 1.19 ± 3.00 ± 3.00 5 50.90a 23.94b 58.69a 14.89b 55.82ab 18.94ab ± 4.98 ± 4.98 ± 2.61 ± 2.61 ± 1.25 ± 1.25 10 53.65a 21.19b 59.02a 14.56b 56.53ab 18.23ab ± 3.20 ± 3.20 ± 0.02 ± 0.02 ± 2.39 ± 2.39 15 54.89a 19.95b 59.19a 14.39b 57.52ab 17.24ab ± 1.89 ±1.89 ± 1.89 ± 1.89 ± 1.28 ± 1.28 20 68.63b 6.21a 64.05b 9.53a 61.53b 13.23a ± 0.36 ± 0.36 ± 0.34 ± 0.34 ± 3.04 ± 3.04

479 D. SAVANOVIĆ et al.: MELTING AND CRYSTALLIZATION DSC PROFILES… Chem. Ind. Chem. Eng. Q. 23 (4) 473−481 (2017)

CONCLUSIONS Pezo, N. Mišljenović, Chem. Ind. Chem. Eng. Q. 20 (2014) 155-162 The shape of the DSC cooling and heating [12] D.Vujadinović, R.Grujić, V.Tomović, A.Torbica, Chem. curves of meat depended on the scanning rate. Ind. Chem. Eng. Q. 20 (2014) 407-416 Therefore, it was observed that crystallization and [13] Lj. Petrović, R.Grujić, M. Petrović, Meat Science 33 melting peaks changed with an increasing scanning (1993) 319-331 rate. The cooling rate effected the onset (Tcon), peak [14] M. Bueno, V.C. Resconi, M. Mar Campo, J. Cacho, V. Ferreira, A. Escudero, Food Res. Int. 54 (2013) 772–780 (Tc) and end temperatures (Tcend) of crystallization process for beef meat. Enthalpy of crystallization [15] Z. Pietrasik, J.A.M. Janz, Meat Science 81 (2009) 523– –532 (ΔHc) for beef meat changed with a changing cooling rate. The effect of heating rate on enthalpy of melting [16] S. Meziani, M. Kaci, M. Jacquot, J. Jasniewski, P. Ribotta, J.-M. Muller, M. Ghoul, S. Desobry, J. Food Eng. (ΔH ), peak (T ) and end temperatures (T end) of m m m 111 (2012) 336–342 melting process for beef meat was reported. How- [17] M. Holzwarth, S. Korhummel, R. Carle, D.R. Kammerer, ever, the type of meat did not influence temperatures Food Res. Int. 48 (2012) 241–248 of crystallization (T on, T and T end), enthalpy of c c c [18] S. Da-Wen, Handbook of frozen food packaging and crystallization (ΔHc), temperatures of melting (Tm and processing, CRC Press Taylor & Francis Group, Boca Tmend) and enthalpy of melting (ΔHm). In addition, the Raton, FL, 2006, pp. 3-141 increase in freezable water content and the decrease [19] E. Xanthakis, M. Havet, S. Chevallier, J. Abadie, A. Le- in unfreezable water content were recorded between Bail, Innovative Food Sci. Emerging Technol. 20 (2013) heating rates 2 and 20 °C/min for all three meat types. 115–120 The information presented can be helpful for simple [20] M. Castro-Giráldez, N. Balaguer, E. Hinarejos, P.J. Fito, and rapid engineering calculations and for imple- Innov. Food Sci. Emerging Technol. 23 (2014) 138–145 mentation in complex mathematical models of heat [21] R. Grujic, T. Vasiljevic, Meat Technol. 37 (1996) 133-136 transfer. [22] R. Grujić, D. Vujadinović, G. Tadić, V. Tomović, in Proceedings of 6th Central European Congress on Food, REFERENCES CEFood2012, Novi Sad, Serbia, 2012, pp. 719-725 [23] A.M. Tocci, R.H. Mascheroni, LWT Food Sci. Technol. 31 [1] O. Dahimi, A.A. Rahim, S.M. Abdulkarim, M.S. Hassan, (1998) 418–426 S.B.T. ZamHashari, A.S. Mashitoh, S. Saadi, Food [24] G. Chen, C. Öhgren, M. Langton, K.F. Lustrup, M. Nydén, Chem. 158 (2014) 132–138 J. Swenson, J. Cereal Sci. 57 (2013) 120-124 [2] J.S. Karthikeyan, K.M. Desai, D. Salvi, R. Bruins, M.V. [25] S. Zhu, A. Le Bail, H.S. Ramaswamy, J. Food Eng. 75 Karwe, Food Res. Int. 76 (2015) 595–604 (2006) 215–222 [3] I. Tolstorebrov, T.M. Eikevik, M. Bantle, Food Res. Int. 55 [26] L.I. Najafabadi, A. Le-Bail, N. Hamdami, J.-Y Monteau, J. (2014) 303–310 Keramat, J. Cereal Sci. 60 (2014) 151-156 [4] I.S.M. Zaidul, N. Absar, S.-J. Kim, T. Suzuki, A.A. Karim, [27] D. Micić, S. Ostojić, M. Simonović, B.R. Simonović, J. H. Yamauchi, T. Noda, J. Food Eng. 86 (2008) 68–73 Process. Energy Agric.18 (2014) 204-206 [5] A. Fini, C. Cavallari, F. Ospitali, Pharmaceutics 2 (2010) [28] A.L. Simmons, K.B. Smith, Y. Vodovotz, J. Cereal Sci. 56 136–158 (2012) 232-238 [6] A. Soyer, B. Ozalp, U. Dalmıs, V. Bilgin, Food Chem. 120 [29] AOAC (2006) Official Method, 950.46 (2010) 1025–1030 [30] G.A. Marini, E.M. Bainy, M.K. Lenzi, M.L. Corazza. Acta [7] M. Reading, D. Hourston, Modulated-temperature Scientiarum Technol. 36 (2014) 361-368 differential scanning calorimetry, Springer, Dordrecht, [31] C.A. Miles, Z. Mayer, M.J. Morley, M. Houska, Int. J. 2006, pp. 38-49 Food Sci. Technol. 32 (1997) 389–400 [8] J. Tomaszewska-Gras, J. Therm. Anal. Calorim. 113 [32] O. Fasina, J. Agric. Sci. Techol., B 2 (2012) 1287-1292 (2013) 199–208 [33] E.M. Bainy, M.L. Corazza, M.K. Lenzi, J. Food Eng. 161 [9] R. Grujić, Lj. Petrović, B. Pikula, Lj. Amidžić, Meat (2015) 82–86 Science 33 (1993) 301-318 [34] X. Ding, H. Zhang, L. Wang, H. Qian, X. Qi, J. Xiao, Food [10] H. Kiani, D.-W. Sun, Trends Food Sci. Tech. 22 (2011) Hydrocolloids 47 (2015) 32-40 407-426 [35] R.G.M. van der Sman, E. Boer, J. Food Eng. 66 (2005) [11] V. Pitschmann, Z. Kobliha, I. Tušarová, L. Bártová, 469–475. D.Vetchý, P.V.Filipović, L. Lević, B. Ćurčić, M. Nićetin, L.

480 D. SAVANOVIĆ et al.: MELTING AND CRYSTALLIZATION DSC PROFILES… Chem. Ind. Chem. Eng. Q. 23 (4) 473481 (2017)

DANICA SAVANOVIĆ DSC PROFILI TOPLJENJA I KRISTALIZACIJE RADOSLAV GRUJIĆ2 RAZLIČITIH VRSTA MESA SLADJANA RAKITA3 3 ALEKSANDRA TORBICA Cilj ovog rada bio je ispitati uticaj brzine skeniranja i vrste mesa, na termo-fizička svojs- 4 RANKO BOZIČKOVIĆ tva mesa i sadržaj smrznute vode u smrznutom mesu, koristeći metodu diferencijalne

1 skenirajuće kalorimetrije (DSC). U ovom istraživanju ispitivane su tri vrste mesa: goveđe Tehnološki fakultet, Univerzitet u Banjoj Luci meso (M. Longissimus dorsi), svinjsko meso (M. Longissimus dorsi) i pileće meso (Pec- 2Tehnološki fakultet, Univerzitet u toralis major). Brzina hlađenja je uticala na temperaturu početka procesa kristalizacije Istočnom Sarajevu (Tcon), temperaturu pika krive kristalizacije (Tc) i temperaturu završetka procesa kristali- 3 Institut za prehrambene tehnologije, zacije (Tcend) u goveđem mesu (p < 0,05). Smanjenje brzine hlađenja od 20 do 2

Univerzitet u Novom Sadu C/min rezultiralo je značajnom (p < 0,05) promjenom entalpije kristalizacije (Hc) 4 Mašinski fakultet, Univerzitet u goveđeg mesa, od -220,17 do -168,20 J/g. Smanjenje brzine zagrijavanja uzrokovalo je Istočnom Sarajevu značajano (p < 0,05) smanjenje entalpije topljenja (Hm) goveđeg mesa od 228,87 do 161,13 J/g. Brzina zagrijavanja uticala je na temperaturu pika krive topljenja (T ) i tem- NAUČNI RAD m peraturu završetka procesa topljenja (Tmend) u goveđem mesu (p < 0,05). Vrsta mesa

nije imala uticaj na vrijednosti Hc i Hm kao i na temperature kristalizacije (Tcon, Tc i

Tcend) i temperature topljenja (Tm i Tmend) u mesu. Značajna (p < 0,05) promjena sadr- žaja smrznute vode zabilježena je između brzine zagrijavanja 20 C/min i ostalih brzina zagrijavanja, za sve tri vrste mesa.

Ključne reči: DSC, meso, kristalizacija, topljenje.

481

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 483−493 (2017) CI&CEQ

ALI SAKIN1 NUMERICAL PREDICTION OF SHORT-CUT IRFAN KARAGOZ2 FLOWS IN GAS-SOLID REVERSE FLOW 1TOFAS-FIAT R&D Department, CYCLONE SEPARATORS Istanbul Caddesi, Osmangazi, Bursa, Turkey 2 Article Highlights Department of Mechanical • Short-cut flows are mainly affected by vortex finder diameter and insertion length Engineering, Uludag University, • Insertion length has an optimum value for minimum short-cut flow Nilufer, Bursa, Turkey • Decrease in vortex finder diameter increases short-cut flow and pressure drop • Short-cut flow increases with the cone tip diameter SCIENTIFIC PAPER Abstract UDC 519.876.5:66:66.07 The effect of operational and geometrical parameters on the short-cut flow in cyclone separators has been investigated computationally using the Reynolds stress model (RSM). The motion of solid particles in the flow field was simulated using the Eulerian-Lagrangian approach with one way discrete phase method (DPM). Eleven cyclones with different cone tip diameters, vortex finder lengths and diameters were studied and the simulation results were analyzed in terms of velocity fields, pressure drops, cut-off diameters and short-cut flows. The numerical simulation was verified with the published experimental results. The results obtained demonstrate that all three parameters, particularly, vortex finder diameter, have significant effects on the cut-off diameter (collection efficiency), the short-cut flow and the pressure drop. Keywords: CFD, cut-off diameter, pressure drop, separation efficiency, swirl flow.

A cyclone separator is a simple tool that leads comprehensive [5,6] for the prediction of a cyclone the incoming solid-gas flow into a spiral motion and performance. separates the solid particles from the main stream Several studies of small scale reverse flow cyc- under the influence of centrifugal and gravity forces. lones have been undertaken by examining the effects Robust in structure, relatively inexpensive to construct of cyclone design and air flow rate on collection effi- and operate with little maintenance, cyclones are ciency and pressure drop [7-10]. used for removal of harmful or nuisance air-borne Optimization of cyclone performance by increas- particulates or liquid droplets. There are many ver- ing cyclone separator collection efficiency and red- sions of cyclone separators with various geometries ucing the pressure drop was mainly derived from exp- and designs. eriments rather than theoretical studies because the The major performance parameters of a cyclone fluid motion in cyclones are complicated, including separator are pressure drop and cut-off diameter highly swirling turbulent characteristics and restrict- values. Numerous theoretical and experimental stu- ions in the usage of empirical models [8,9,11-16]. dies carried out on this tough subject provide semi- Advances in computer capabilities, software and empirical models varying from basic [1-4] to more numerical methods provide more opportunities to computational calculation of cyclone separators [17-34].

Elsayed and Lacor performed numerical cyclone Correspondence: A. Sakin, TOFAS-FIAT R&D Department, Istanbul Caddesi, No:574 16369 Osmangazi, Bursa, Turkey. studies for the effect of cone tip diameter, dust bin geo- E-mail: [email protected] metry, vortex finder and inlet dimensions [28,31,35,36]. Paper received: 9 October, 2016 They reported that for the inlet dimensions effect of Paper revised: 4 January, 2017 Paper accepted: 13 January, 2017 changing the inlet width on the cut-off diameter is more considerable than inlet height and optimum ratio https://doi.org/10.2298/CICEQ161009002S

483 A. SAKIN, I. KARAGOZ: NUMERICAL PREDICTION OF SHORT-CUT FLOWS… Chem. Ind. Chem. Eng. Q. 23 (4) 483−493 (2017) of the inlet width to the inlet height is from 0.5 to 0.7 540 L/min. They stated that the capability of CFD [36]. Regarding the cone tip diameter change, they accurately predicts performance indicators of pres- reported that there is no significant effect on flow sure drop and collection efficiency. pattern and performance, and that decreasing the Taking advantage of the increased availability of cone tip diameter increases the pressure drop [28]. adjoint solvers in many commercial CFD codes, Els- Another study based on vortex finder dimensions, ayed [42] performed CFD analysis to find optimum reported that reducing the vortex finder diameter by shape of vortex finder. Resultant optimum vortex finder 40% leads to 175% increase in the dimensionless geometry provided reduction of 34% pressure drop pressure drop (Euler number) and 50% reduction in and 83% cutoff diameter. the Stokes number [35]. In contrast with previous stu- Song et al. [43] performed CFD analysis for D = dies, they performed numerical cyclone analysis with = 290 mm and they investigated the influence of par- Reynolds stress model (RSM) for investigation of dust ticle forces on separation efficiency. They analyzed outlet shape and reported that lack of dust outlet pro- effect of drag force, pressure gradient force, added vides saving of large computational cost and it can mass force and Saffmann lift force. They stated that affect the calculations in an error margin about 10% for pressure gradient and added mass forces are at least the Euler number and 35% for the cut-off diameter [31]. 3 orders of magnitude smaller than the drag force and Kaya and Karagoz [37] performed numerical the Saffmann lift force is at least 15 orders of magni- studies to investigate performance characteristics of a tude smaller. Therefore these forces can be neg- cyclone prolonged with a dipleg and they reported lected and the main driver of separation forces are that particle collection efficiency can be enhanced by drag and centrifugal forces. applying a dipleg, which results in a variation in flow The present study aims to investigate the effect pattern for tangential inlet short cyclones. Increasing of short-cut flow on the pressure drop and collection the dipleg length surface of the wall friction is inc- efficiency of reverse flow cyclone by changing cyc- reased which should lead to moderately lower tan- lone tip diameter and vortex finder dimensions. gential velocity. Kaya et al. [38] investigated effects of surface MATHEMATICAL MODEL AND NUMERICAL roughness parameters on the performance of tan- SIMULATION gential inlet cyclones. Body diameter of 31 mm cyclone was utilized for numerical simulations by assign- Calculations were carried out on three cyclones ing different roughness values for inlet velocities 8 with different cone tip diameters (fixed S and Dx and 16 m/s, respectively. They concluded that the values), four cyclones with different vortex finder tangential velocity in the cyclone is decreased by inc- lengths (fixed Bc and Dx values) and four cyclones reasing wall roughness, due to an increase in flow with different vortex finder diameters (fixed S and Bc resistance and weakening of swirl so particle col- values). The dimensional layout and cyclone config- lection efficiency is deteriorated by greater surface urations are given in Figure 1 and Table 1. roughness values. El-Batsh [39] performed CFD stu- Numerical method dies to improve performance of cyclone separator by The Eulerian-Lagrangian approach was used for appropriate selection of vortex finder dimensions. It numerical calculations of gas-solid flow in this study was found that pressure drop is decreased with inc- and flow is assumed as unsteady, incompressible and reasing vortex finder diameter and length of exit pipe turbulent. Continuity and momentum equations are is not a significant parameter on the cyclone perform- solved in order to obtain velocity field. The turbulent ance. flow was represented by the Reynolds stress model Houben et al. [40] performed CFD simulations of (RSM) and particle trajectories and collection effi- cyclone separator with central vortex stabilization rod ciencies were calculated by Lagrangian approach. to suppress the vortex core precession and results Numerous particles that are injected from the inlet were compared with the experimental data. They surface were followed through the flow field by solving stated that the vortex is kept in position by using a the particle force balance equation of motion. stabilizer and maximum tangential velocity is found to The flow field was defined as three-dimensional, be larger which has a positive effect on collection effi- the particle phase was assumed as dilute, and the ciency of small particles. particle loading level is low so interaction between the Haig et al. [41] performed CFD analysis on small particles are insignificant and particles have no scale cyclones with the body diameter of ranging from influence on the main stream flow. Therefore, the 29 to 52 mm and inlet flow rates ranging from 60 to problem was assumed as one-way coupling [44]. By

484 A. SAKIN, I. KARAGOZ: NUMERICAL PREDICTION OF SHORT-CUT FLOWS… Chem. Ind. Chem. Eng. Q. 23 (4) 483−493 (2017) this assumption, the source terms in Navier-Stokes ∂     ()ρρvvvp+∇ ( ) =−∇ +∇ () τρ + gF + (2) equations due to two-way coupling strategy are not ∂t considered. where ρ is the fluid density, v is the fluid velocity, p is the static pressure, τ is the stress tensor  2 ( =∇+∇−∇μ[(vvT ) vI ] ), μ is the molecular visco- 3 sity, I is the unit tensor, ρg is the gravitational body force, and F is the external body force. The RSM was used to modify the viscous turbulent flow in reverse flow cyclone. The turbulent transport equation for RSM is [45]: ∂∂ ()ρρuu′′+= ( u uu ′′ ) ∂∂ij kij txk ∂ =−ρδδuuu′′′ + p() u ′ + u ′ + ∂ ijk kji ikj x k ∂∂ ∂u ∂u  +−+−μρ()uu′′ uu ′′j uu ′′ i (3) ∂∂ij ij ∂ jk ∂ xxkk  x k x k  ∂u ′ ∂u ′ −+++−ρβ()gu′′ θ gu θ pi j u ′ ij ji∂∂xx j Figure 1. Schematic diagram of cyclone geometry and ji coordinate system definition. ∂uu′′ −+2μ ij S ∂∂xx Transport equations kk ′ The three-dimensional flow field in reverse flow where t is the time, ui is the fluctuating velocity to =− cyclone was simulated using CFD. The conservation direction i ()uuim, ui is the velocity direction i, um ′′ equations for mass and momentum in an incompres- is the mean velocity to direction i, uui j is the sible Newtonian flow are as follows [45]: Reynolds stress tensor, β is the thermal expansion, μ ∂ρ  p is the pressure, is the eddy viscosity, and S is the +∇()ρvS = (1) source term. ∂t The two terms on the left-hand side of Eq. (3) indicate local time derivative and convection, res-

Table 1. Geometrical dimensions of cyclones; body diameter, D = 31 mm. Inlet length, Li = 1.5D from the cyclone center, outlet length above cylindrical surface of cyclone is Le = D

Dimension Cyclonea Dimension, D Inlet height, a – 0.4 Inlet width, b – 0.16 Cylinder height, h – 1.0

Total cyclone height, Ht – 2.5

Cone tip diameter, Bc S = 0.5D CY1 0.625 Dx = 0.5D CY2 0.5 CY3 0.375

Vortex finder length, S Bc = 0.375D CY4 0.4 Dx = 0.5D CY5 0.5 CY6 0.875 CY7 1.0

Vortex finder diameter, Dx Bc = 0.375D CY8 0.3 S = 0.5D CY9 0.4 CY10 0.5 CY11 0.645 aCyclone CY3, CY5 and CY10 are identical

485 A. SAKIN, I. KARAGOZ: NUMERICAL PREDICTION OF SHORT-CUT FLOWS… Chem. Ind. Chem. Eng. Q. 23 (4) 483−493 (2017) pectively from left to right. The terms on the right- Numerical scheme -hand side of Eq. (3) indicate turbulent diffusion, mol- The selection of spatial and temporal discret- ecular diffusion, stress production, buoyancy product- ization schemes has an important effect on the CFD ion, pressure strain and dissipation respectively from results and fluent provides various solution strategies left to right. The terms for turbulent diffusion, buoy- for pressure velocity coupling, pressure, momentum, ancy production, pressure strain and dissipation are kinetic energy, rate of kinetic energy dissipation dis- needed for modeling [46]. cretization. Both Kaya and Karagoz [25] and Shukla Discrete phase modeling et al. [30] analyzed various solution strategies in steady and unsteady simulations of cyclone separ- The motion of solid particles in a flow field was ators. Based on the study of Kaya and Karagoz [25], simulated using the Eulerian-Lagrangian approach governing equations of the three-dimensional, incom- with a discrete phase method (DPM). Gas phase was pressible flow inside the cyclone, Eqs. (1) and (2), treated as continuum by Navier-Stokes equations and and the RSM turbulence equations were discretized the solid phase is calculated. After flow field variables over the computational cells and iteratively solved by are obtained, particle tracking calculations are per- using ANSYS Fluent commercial CFD software. formed using Lagrangian approach. The volume fract- SIMPLEC algorithm was used for pressure velocity ion of the dispersed phase did not exceed 10%, so coupling in this study and pressure staggering option particle-particle interaction can be neglected and par- (PRESTO) scheme was chosen for the pressure inter- ticle phase did not affect flow field (one-way coup- polation as it was shown to be well suited for steep ling). A stochastic tracking method was used for pressure gradient involved in complex swirling flows. modeling turbulent dispersion of particles. The traject- QUICK scheme was chosen for discretization of ory of a particle can be calculated by integrating the momentum, second order upwind scheme was chosen force balance in a Lagrangian reference frame. This for turbulent kinetic energy and turbulent dissipation force balance equates the particle inertia with the forces rate and first order upwind scheme was used for turb- acting on the particle, and can be written as [45]:   ulence stresses. d()ug ρρ−  pp=−+Fuu() + F (4) Simulations started with steady solution, when Dp ρ dt p convergence plot becomes nearly horizontal and  there was no significant change, temporal discretizat- where F is an additional acceleration (force/unit par-  ion was switched to unsteady solution with the time ticle mass) term, Fuu()− is the drag force per unit Dp step of 0.0001 s. Residence time values of cyclone particle mass and: configurations depending on inlet velocity are given in 18μC Re Table 2. Residence time exceeded for all simulations F = Dp (5) D ρ 2 to prevent inconsistent results of flow field. 24 ppd Boundary conditions and other settings where u is the air velocity in a separator, u p is the particle velocity, μ is the viscosity of air, ρ is the air Velocity inlet boundary condition is employed at ρ density, p is the particle density, d p is the particle inlet, outflow at gas exit and wall (no-slip condition) at diameter, Re is the relative Reynolds number, CD is all other boundaries. Volumetric flow rate equals to 30 the drag coefficient and C1 to C3 are the constants and 60 L/h, corresponding to air inlet velocity 8 and that depends on the range of Re [47]: 16 m/s, respectively, dynamic viscosity of 2.11×10-5 ⋅ 3 C C Pa s and air density 1.0 kg/m . The turbulence inten- CC=+2 +3 (6) D 1 Re Re sity and the hydraulic diameter were defined as 5% and 0.007 m, respectively. For the near wall treat- ρdu− u ment, scalable wall function was selected. Re = pp (7) μ

Table 2. Residence time and cell numbers

Parameter CY1 CY2 CY3, CY4, CY5, CY6, CY7, CY10 CY8 CY9 CY11 Cyclone volume×103, m3 0.0552 0.0521 0.0493 0.0456 0.0472 0.0532 tres / s (Uin = 8 m/s) 0.1122 0.1059 0.1003 0.0927 0.0960 0.1082 tres / s (Uin = 16 m/s) 0.0561 0.0529 0.0501 0.0463 0.0480 0.0541 Cell number 98582 95631 93756 92138 92556 96806

486 A. SAKIN, I. KARAGOZ: NUMERICAL PREDICTION OF SHORT-CUT FLOWS… Chem. Ind. Chem. Eng. Q. 23 (4) 483−493 (2017)

104 particles were released from the inlet sur- RESULTS face with the air inlet velocity and a mass flow rate  Model validation mp of 0.001 kg/s and particle diameter varies from 0.5 to 10 μm. The density of released particles is 860 Stairmand cyclones are widely used for indus- 3 5 kg/m and the maximum time steps number was 10 trial purposes so they became a significant research steps for each injection. The restitution coefficient area from past to today. Hoekstra [48] studied barrel was set as 0.8 for both normal and tangential dir- diameter of 290 mm and measured velocity profiles ection. Trap DPM boundary condition was applied by LDA. This study is the newest one that can be bottom surface of cyclone to calculate particle collect- taken as a reference for model validation. Xiang et al. ion efficiency, escape was applied both inlet and out- [9] studied barrel diameter of 31 mm for various pur- let. The solution was assumed to be converged at poses and most of the studies, especially based on each time step when preset scaled residuals reached CFD, used this geometry due to small geometric -5 10 as convergence criteria for all variables. The cal- domain and lower computational effort. It is also used ® culations were carried out on an Intel Core i7- for comprehensive optimization studies. 2630QM 2.0 GHZ with 8 GB of RAM. Calculated results were compared with the LDA Solution of grid dependency velocity measurements of Hoekstra [48], measured by using laser Doppler anemometry (LDA) system. Fig- The flow volume was divided into a number of ure 2 shows the comparison between RSM solution blocks by using Icem CFD 15.0 software to generate and the measured axial and tangential velocity pro- unstructured topology among the domain. However, files at axial station z = 2.5D (Plane 3). As can be fine mesh structure was used in the core region seen from Figure 2b, very good agreement was where strong gradients in the flow parameters were obtained for tangential velocity profiles at all planes. present. Grid independency was examined for CY3 The axial velocity profiles do not agree well with the cyclone. Three different grid domains containing experimental values (Figure 2a). Considering comp- 25.330, 93.756 and 349.222 cells were compared. lexity of the flow in cyclones, the agreement between The calculated grade efficiency curve with mesh the calculations and measurements is considered design 349.222 cells is very close to mesh design highly sufficient. with 93.756 cells and there is no appreciable difference in the prediction and mesh design with 93.756 Calculation of short-cut flow rate cells is provided for computational solution. Grid inde- A small rate of mass flow directly moves toward pendent cell numbers of other cyclone configurations the vortex finder by entering upward the vortex were given in Table 2. region. The short-cut flow rate strongly depends on cyclone separator parameters. To predict the short-

Figure 2. Comparison of axial (a) and tangential (b) velocity profiles between CFD results and LDA measurements (cyclone barrel diameter for validation study, D = 290 mm).

487 A. SAKIN, I. KARAGOZ: NUMERICAL PREDICTION OF SHORT-CUT FLOWS… Chem. Ind. Chem. Eng. Q. 23 (4) 483−493 (2017)

-cut flow rate, the downward flow rate at an axial Present CFD results of 8 m/s inlet velocity are in station should be calculated instantly [49]. If the good agreement with LES [28] and for 16 m/s results downward axial velocity is integrated between cyc- are close to experimental data (Figure 4). Calculated lone wall and the locus of zero at the axial station, the cut off diameter differs from experimental data 41, 33 total downward flow rate is obtained. Since we create and 31% for CY1, CY2 and CY3, respectively. Els- a new surface 1 mm below bottom of the vortex finder ayed and Lacor [31] observed that cyclone para- and clip axial velocity by locus of the zero and dif- meters are predicted without dust bin, an error margin ference between downward and total flow rate in the around 10% in the calculations of Euler number and cyclone will give the short-cut flow rate. 35% in the calculation of cut-off diameter should be In this study, short-cut flow effect on cyclone considered for CFD calculations. performance is explained with tangential velocity pro- Maximum collection efficiency is obtained for files and short-cut flow rate. Tangential and axial vel- smaller cone tip diameter of 11.625 mm (CY3). Total ocity components have more significant effect on vel- friction surface area rates based on CY3 are 9.15% ocity field rather than radial velocity component. Tan- and 13.52% larger for CY2 and CY1 respectively, so gential velocity component is more significant because maximum tangential velocity profiles of CY3 for both it leads to centrifugal force for particle separation inlet velocities are higher than CY1 and CY2. This mechanism [49]. The axial flow has effect on down- results stronger swirl intense and more centrifugal ward and upward flows. To represent all velocity pro- force for particle separation. files for both inlet velocities and all five sections will The short-cut flow increases with increasing create more pages so to prevent exceeding page cone tip diameter and this results in decreasing swirl number, velocity profiles at 1.75D section is pre- intensity. While short-cut flow rate is increased, dec- sented for comparison of axial and tangential velo- reasing separation efficiency is expected but down- cities for both inlet values. ward flow migration increases with increasing cone tip diameter so there is no significant change in the sep- Effect of cone tip diameter aration efficiency due to the change in cone tip dia- Velocity profiles are given in Figure 3 represent meter. In this case, maximum particle collection effi- the tangential and axial velocity distribution at ciency is obtained for maximum tangential velocity z = 1.75D distance. The tangential velocity profiles and minimum short-cut flow percentage. show axis-symmetrical distribution for both inlet vel- The pressure drop values are in good agree- ocity values and inner region profiles are identical. ment with experimental data and LES [28] calcul- The axial flows indicate existing of two flow streams: ations. Pressure drop increases with decreasing cone downward (positive axial velocity) and upward (neg- tip diameter as expected. Contrary to cut-off diameter, ative axial velocity) directed to vortex finder exit.

Figure 3. Comparison of axial (a) and tangential (b) velocity of CY1, CY2 and CY3 for Uin = 8 and 16 m/s at z = 1.75D section.

488 A. SAKIN, I. KARAGOZ: NUMERICAL PREDICTION OF SHORT-CUT FLOWS… Chem. Ind. Chem. Eng. Q. 23 (4) 483−493 (2017)

Figure 4. Cone tip diameter effects on cut off diameter, shortcut flow (a), pressure drop and downward flow (b). The results are compared to “Xiang and Lee [49]” literature [9]. pressure drop decreases with increasing short-cut increased short-cut flow so swirl intensity and pres- flow due to weaker swirl in the flow field. sure drop decrease concurrently. For this case, the inlet height (a) and vortex finder length (s) are equal Effect of vortex finder length and portion of inlet flow joins upward inner flow easily In Figure 5, the tangential velocity profiles are and short-cut flow rate is greater than CY5. quite similar. Maximum tangential velocity occurs at Only for small vortex finder length (CY4), axial CY5 configuration (s = 0.5D) and maximum collection velocity and downward flow through the cone is inc- efficiency for both inlet velocities (Figure 6) is obtained reased but at the same time separation efficiency is for this configuration due to domination of tangential decreased due to increasing short-cut flow rate. CY5 velocity on centrifugal force. has the optimum vortex finder length and results mini- Shortest vortex finder (CY4, s = 0.4D) results mum short-cut flow and maximum efficiency.

Figure 5. Comparison of axial (a) and tangential (b) velocity of CY4, CY5, CY6 and CY7 for Uin = 8 and 16 m/s at z = 1.75D section.

489 A. SAKIN, I. KARAGOZ: NUMERICAL PREDICTION OF SHORT-CUT FLOWS… Chem. Ind. Chem. Eng. Q. 23 (4) 483−493 (2017)

Figure 6. Vortex finder length effects on cut off diameter, shortcut flow (a), pressure drop and downward flow (b).

From the viewpoint of velocity profiles, CY5 has ration is decreased by increasing vortex finder length. the maximum tangential velocity and pressure drop. Sharp decrease between CY4 and CY5 occurs and Vortex finder of CY6 and CY7, extended into the cone this can be explained by axial velocity profiles. For section of cyclone, where the wall was contracting CY6 and CY7 both axial velocity profiles and down- and flow coming from the cylindrical section, will be ward flow rates are quite similar however swirl int- influenced by the contracting wall and easily join into ensity is decreased with increasing friction surface. the upward flow. The trend of short-cut flow rate is in Effect of vortex finder diameter good agreement with the explanation of Xiang and Lee [49]. The separation efficiency is decreased with It can be clearly seen that maximum tangential decreasing tangential velocity and increasing short- velocity is decreased with increasing vortex finder cut flow rate for CY6 and CY7. diameter (Figure 7). From the viewpoint of tangential The pressure drop is increases with increasing velocity profiles, cut off diameter is increased and tangential velocity and CY5 has the maximum pres- pressure drop is decreased with increasing vortex sure drop for both inlet values. Downward flow mig- finder diameter. Increased vortex finder diameter

Figure 7. Comparison of axial (a) and tangential (b) velocity of CY8, CY9, CY10 and CY11 for Uin = 8 and 16 m/s at z = 1.75D section.

490 A. SAKIN, I. KARAGOZ: NUMERICAL PREDICTION OF SHORT-CUT FLOWS… Chem. Ind. Chem. Eng. Q. 23 (4) 483−493 (2017) causes larger friction surface inside cyclone body so CONCLUSIONS vortex strength (also tangential velocity which is the most dominant component for particle collection Cyclones of different cone tip diameters, vortex efficiency) is decreased. finder lengths and diameters were numerically analyzed using Reynold stress model (RSM). The follow- Smaller vortex finder diameter (Dx) results the maximum separation efficiency although maximum ing conclusions were obtained. short-cut flow rate occurs at the same time. The axial • Short-cut flow curve trend of cone tip diameter velocity profile demonstrates reversed W profile configurations at both inlet velocities are similar. Mini- everywhere, but for small vortex finder diameter pro- mum short-cut flow rate is obtained for minimum cone file initially exhibits reversed V and becoming W at tip diameter on contrary to maximum tangential velo- downward sections. Similar behavior was also rep- city and pressure drop obtained with this cyclone con- orted by Elsayed and Lacor [35] and Hoekstra et al. figuration. Decreasing cone tip diameter increases [18]. Elsayed and Lacor [35] reported that this change separation efficiency, pressure drop, short-cut and results 73% increase in the maximum axial velocity. downward flow concurrently. Hoekstra et al. [18] explained this situation; adverse • Vortex finder length has no significant effect on pressure gradient leads to decay of swirl due to frict- tangential velocity profile but shortest length causes ion losses in the vortex finder. Swirl intensity inc- increasing short-cut flow. S = 0.5D is the optimum reases for narrow vortex finder diameters, so that the value for minimum short-cut flow and greater values adverse pressure gradient can be overcome. result increasing short-cut flow due to contraction of Pressure drop is monotonically decreased for conic cyclone walls. Downward flow migration is dec- both inlet velocities as shown Figure 8. For smaller reased with increasing vortex finder length due to inc- vortex finder diameters, pressure drop is increased reasing friction surface area. due to increase in velocity and swirl intensity as exp- • Vortex finder diameter has an important effect lained previously and for larger vortex finder dia- on cut-off diameter and tangential velocity. Increased meters pressure drop decreases with weaker swirl vortex finder diameter results larger friction surface in intensity. For CY8, downward flow rate is small in cyclone body, weaker swirl intense and decreased comparison with the others and this can be explained pressure drop. Dx = 0.3D is the smallest vortex dia- by the shape change of velocity profile. Downward meter configuration and maximum short-cut flow and flow rate is decreasing monotonically for CY9, CY10 pressure drop is obtained due to velocity profile shape and CY11 due to increased friction surface and change from inverted “W” to inverted “V” shape. weaker swirl intensity. From the viewpoint of short-cut flow rate, vortex finder dimensions (diameter and length) have more significant effect than cone tip diameter. Vortex finder

Figure 8. Vortex finder diameter effects on cut off diameter, shortcut flow (a), pressure drop and downward flow (b).

491 A. SAKIN, I. KARAGOZ: NUMERICAL PREDICTION OF SHORT-CUT FLOWS… Chem. Ind. Chem. Eng. Q. 23 (4) 483−493 (2017) length of 0.5D is an optimum point for short-cut flow Subscript rate and for greater values of vortex finder length i Cartesian coordinate downward flow is joining upward flow due to con- j Cartesian coordinate traction wall between cylindrical and conical section. z Cartesian coordinate Smaller vortex finder diameter causes change of vel- ax Axial ocity profile from “W” to “V” and short-cut flow per- in Inlet centage is increased. tan Tangential As a recommendation of future work, numerical calculations can be carried out by modifying the cyl- Abbreviations inder and total height of the cyclone in order to inv- CFD Computational Fluid Dynamics estigate the effect of short-cut flow on pressure drop DPM Discrete Phase Method and particle collection efficiency. LDA Laser Doppler Anemometry LES Large Eddy Simulation Nomenclature PRESTO Pressure Staggering Option a Inlet height of the cyclone RSM Reynolds Stress Model b Inlet width of the cyclone SIMPLEC Semi-Implicit Method for Pressure Linked BC Cone tip diameter Equations-Consistent C1 Constant in Equation (6) QUICK Quadratic Upwind Interpolation C2 Constant in Equation (6) C3 Constant in Equation (6) REFERENCES CD Particle drag coefficient D Cyclone body diameter [1] G.B. Shepherd, C.E. Lapple, J. Ind. Eng. Chem. 31 (1939) 972-984 d p Particle diameter [2] R.M. Alexander, Proc. - Australas. Inst. Min. Metall. 203 DX Vortex finder diameter F External body force in Equation (2) (1949) NS 152–153:203-228 F Additional acceleration term in Equation (4) [3] W. Barth, Brennst.-Waerme-Kraft 8 (1956) 1–9 [4] W. Barth, L. Leineweber, Staub - Reinhalt. Luft 24 (1964) FD Drag force g Acceleration due to gravity 41–55 h Cylinder height [5] A. Avci, I. Karagoz, J. Aerosol Sci. 34 (2003) 937–955 [6] I. Karagoz, A. Avci, Aerosol Sci. Technol. 39 (2005) 857– Ht Total cyclone height I Unit tensor 865 L Outlet length above cylindrical surface of cyc- [7] J.C. Kim, K.W. Lee, Aerosol Sci. Technol. 12 (1990) e 1003–1015 lone [8] M.E. Moore, A.F. McFarland, Environ. Sci. Technol. 27 L Inlet length from the cyclone center i (1993) 1842–1848 m Dust mass flow rate P [9] R. Xiang, S.H. Park, K.W. Lee, J. Aerosol Sci. 32 (2001) p Static pressure 549–561 Re Reynolds number based on the relative par- p [10] Y. Zhu, K.W. Lee, J. Aerosol Sci. 30 (1999) 1303–1315 ticle velocity [11] L.C. Kenny, R.A. Gussman, J. Aerosol Sci. 28 (1997) s Vortex finder length 677–688 S Source term in Equations (1) and (3) [12] C.J. Stairmand, Trans. Inst. Chem. Eng. 29 (1951) 356– t Time 383 tres Residence time [13] C.E. Lapple, Chem. Eng. 58 (1951) 144–151 u Flow velocity component in i direction [14] J. Dirgo, D. Leith, Aerosol Sci. Technol. 4 (1985) 401-415 u ′ Fluctuating velocity component in i direction [15] C. Köning, H. Büttner, F. Ebert, Part. Part. Syst. Charact. u p Particle velocity 8 (1991) 301–307 v Fluid velocity [16] S.L. Upton, D. Mark, W.D. Griffiths, J. Aerosol Sci. 25 z Section position in z direction (1994) 1493–1501 Greek [17] J. Gimbun, T.G. Chuah, A. Fakhru’l-Razi, S.Y. Choong Thomas, Chem. Eng. Process. 44 (2005) 7–12 ρ Fluid density [18] A.J. Hoekstra, J.J. Derksen, H.E.A. Van Den Akker, ρ Particle density p Chem. Eng. Sci. 54 (1999) 2055–2065 μ Molecular viscosity [19] A.L. Gong, L.Z. Wang, Aerosol Sci. Technol. 38 (2004) τ Stress tensor 506–512

492 A. SAKIN, I. KARAGOZ: NUMERICAL PREDICTION OF SHORT-CUT FLOWS… Chem. Ind. Chem. Eng. Q. 23 (4) 483−493 (2017)

[20] T.G. Chuah, J. Gimbun, S.Y. Choong, Powder Technol. [35] K. Elsayed, C. Lacor, Comput. Fluids 71 (2013) 224–239 162 (2006) 126–132 [36] K. Elsayed, C. Lacor, Appl. Math. Model. 35 (2011) 1952– [21] L. Ma, D.B. Ingham, X. Wen, J. Aerosol Sci. 31 (2000) –1968 1097–1119 [37] F. Kaya, I. Karagoz, Chem. Eng. J. 151 (2009) 39–45 [22] D.B. Ingham, L. Ma, Int. J. Energy Res. 26 (2002) 633– [38] F. Kaya, I. Karagoz, A. Avci, Aerosol Sci. Technol. 45 –652 (2011) 988-995 [23] I. Karagoz, F. Kaya, Int. Commun. Heat Mass Transfer 34 [39] H.M. El-Batsh, Appl. Math. Model. 37 (2013) 5286–5303 (2007) 1119–1126 [40] J.J.H. Houben, C. Weiss, E. Brunnmair, S. Pirker, J. Appl. [24] J.J. Derksen, H.E.A. Van den Akker, S. Sundaresan, Fluid Mech. 9 (2016) 487-499 AIChE J. 54 (2008) 872–885 [41] C.W. Haig, A. Hursthouse, D. Sykes, S. Mcilwain, Appl. [25] F. Kaya, I. Karagoz, Curr. Sci. 94 (2008) 1273–1278 Math. Model. 40 (2016) 6082-6104 [26] K. Elsayed, C. Lacor, Chem. Eng. Sci. 65 (2010) 6048– [42] K. Elsayed, Sep. Purif. Technol. 142 (2015) 274-286 –6058 [43] C. Song, B. Pei, M. Jiang, B. Wang, D. Xu, Y. Chen, [27] K. Elsayed, C. Lacor, Powder Technol. 212 (2011) 115– Powder Technol. 294 (2016) 437-448 –133 [44] M. Sommerfeld, Theoretical and experimental modeling [28] K. Elsayed, C. Lacor, Comput. Fluids 51 (2011) 48–59 of particulate flow, Lecture Series 2000-6, von Karman [29] J. You-hai, J. Guang-qin, C. Qing-yun, W. Jian-jun, Institute for Fluid Dynamics, 2000 Zhongguo Shiyou Daxue Xuebao, Ziran Kexueban 32 [45] ANSYS Inc., ANSYS FLUENT 15.0 theory guide, (2008) 109–112 Canonsburg, 2013 [30] S.K. Shukla, P. Shukla, P. Ghosh, Adv. Powder Technol. [46] S.K. Shukla, P. Shukla, P. Ghosh, Appl. Math. Model. 37 22 (2011) 209–219 (2013) 5774-5789 [31] K. Elsayed, C. Lacor, Comput. Fluids 68 (2012) 134–147 [47] S.A. Morsi, A.J. Alexander, J. Fluid Mech. 55 (1972) 193- [32] F. Qian, Z. Huang, G. Chen, M. Zhang, Comput. Chem. –208 Eng. 31 (2007) 1111-1122 [48] A.J. Hoekstra, Gas flow field and collection efficiency of [33] H. Shalaby, K. Wozniak, G. Wozniak, Eng. Appl. Comput. cyclone separators (Ph.D. thesis), Technical University Fluid Mech. 2 (2008) 382-392 Delft, 2000 [34] S.K. Shukla, P. Shukla, P. Ghosh, Eng. Appl. Comput. [49] R.B. Xiang, K.W. Lee, Chem. Eng. Process. 44 (2005) Fluid Mech. 5 (2011) 235–246 877–883.

ALI SAKIN1 NUMERIČKA PREDVIĐANJE STRUJNIH PREČICA IRFAN KARAGOZ2 U CIKLONSKIM SEPARATORIMA GASNO-ČVRSTO SA REVERSNIM TOKOM

NAUČNI RAD Efekat operativnih i geometrijskih parametara na strujne prečice u ciklonskim separatorima istražen je pomoću modela Rejnoldsovih napona (RSM). Kretanje čvrstih čestica u strujnom polju simulirano je korišćenjem Ojler-Lagranžovog pristupa metodom dis- kretne faze u jednom smeru (DPM). Ispitivana su jedanaest ciklona sa različitim prečni- cima konusa, dužinama i prečnicima vorteksa, a rezultati simulacije su analizirani u pogledu polja brzine, pada pritiska, kritičnog prečnika i strujnih prečica. Numerička simulacija je potvrđena već publikovanim eksperimentalnim rezultatima. Dobijeni rezultati pokazuju da sva tri parametra, a pre svega prečnik vorteksa, imaju značajne efekte na kritični prečnik (efikasnost sakupljanja), strujne prečice i pad pritiska.

Ključne reči: CFD, kritični prečnik, pad pritiska, efikasnost razdvajanja, kružni tok.

493

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 495−506 (2017) CI&CEQ

LARISSA NAYHARA SOARES IMPROVEMENT OF RECOVERED ACTIVITY SANTANA FALLEIROS AND STABILITY OF THE Aspergillus oryzae BRUNA VIEIRA CABRAL β-GALACTOSIDASE IMMOBILIZED ON JANAÍNA FISCHER ® CARLA ZANELLA GUIDINI DUOLITE A568 BY COMBINATION OF VICELMA LUIZ CARDOSO IMMOBILIZATION METHODS MIRIAM MARIA DE RESENDE ELOÍZIO JÚLIO RIBEIRO Article Highlights • β-Galactosidase from Aspergillus oryzae was immobilized on Duolite® A568 Faculty of Chemical Engineering, • Three steps in the immobilization process allows a higher catalytic activity Uberlândia Federal University, • Physical adsorption, covalent multipoint and cross-linking process were investigated • Campus Santa Mônica, The biocatalyst obtained was highly stable over the entire pH range studied • The first-order model described well the kinetics of thermal deactivation Uberlândia-MG, Brazil

Abstract SCIENTIFIC PAPER The immobilization and stabilization of Aspergillus oryzae β-galactosidase on ® UDC 582.282.123.4:66.081.3:544:577 Duolite A568 was achieved using a combination of physical adsorption, incubation step in buffer at pH 9.0 and cross-linking with glutaraldehyde and in this sequence promoted a 44% increase in enzymatic activity as compared with the biocatalyst obtained after a two-step immobilization process (adsorption and cross-linking). The stability of the biocatalyst obtained by three-step immobilization process (adsorption, incubation in buffer at pH 9.0 and cross-linking) was higher than that obtained by two-steps (adsorption and cross-linking) and for free enzyme in relation to pH, storage and reusability. The immobilized biocatalyst was characterized with respect to thermal stability in the range 55-65 °C. The kinetics of thermal deactivation was well described by the first-order model, which resulted in the immobilized biocatalyst activation energy of thermal deactivation of 71.03 kcal/mol and 5.48 h half-life at 55.0 °C. Keywords: β-galactosidase, Duolite® A-568, immobilization, incubation in buffer at pH 9.0.

The β-galactosidases obtained from filamentous reason major fields of its applications include pro- fungi have shown to be a promising alternative for cessing of lactose-containing products, allowing the use in the lactose hydrolysis of cheese whey, since improvement of digestibility and technological and these enzymes present optimum pH in the acidic sensorial characteristics of sweetened condensed region and high thermal stability and thus have and frozen dairy products, as well as solving the sub- potential applications in food processing [1-3]. stantial problem of whey utilization and appropriate The β-galactosidase primary biochemical func- disposal [4,5]. However, the use of this enzyme to tion is associated with lactose hydrolysis, for this catalyze the kinetically controlled reaction of transgal- actosylation has received great attention recently in the synthesis of physiologically active galactosides, Correspondence: L.N.S.S. Falleiros, Faculty of Chemical such as the galacto-oligosacharides (GOS) [6-8]. Engineering, Uberlândia Federal University, P.O. Box 593, Av. β João Naves de Ávila, 2121, Campus Santa Mônica - Bloco 1K - -Galactosidase has been isolated from different 38400-902 – Uberlândia-MG, Brazil. microbial species, plants and animals, however, E-mail: [email protected] according to its source, the properties of the enzyme Paper received: 12 September, 2016 Paper revised: 10 January, 2017 vary markedly. The sources of the enzymes that are Paper accepted: 25 January, 2017 used for the hydrolysis of milk and its derivatives are https://doi.org/10.2298/CICEQ160912010F Aspergillus sp. and Lactobacillus bulgaricus because

495 L.N.S.S. FALLEIROS et al.: IMPROVEMENT OF RECOVERED ACTIVITY… Chem. Ind. Chem. Eng. Q. 23 (4) 495−506 (2017) the enzymes obtained from these microorganisms are [19,20]. However, adsorption of the enzyme on a sup- more stable and active at an acidic pH [3,9]. port is not always advantageous. The main disadv- The β-galactosidase from A. oryzae is indus- antage of these procedures is that the enzyme can be trially important. This enzyme is composed of 985 released from the support during the reaction in func- amino acid residues, presents monomeric structure tion of the pH value change or increase of ionic [10], having molecular weight of 105 kDa and max- strength, so to minimize the enzyme desorption is imum activity at pH 4.5 with o-NPG (ortho-nitro- recommended a cross-linking step, generally using phenyl-β-galactoside) and at pH 4.8 with lactose as glutaraldehyde as reticulant agent [16]. substrate. The optimum temperature is 50 °C and it is Chemical enzyme modification can be per- stable in the pH range from 4.0 to 9.0 [11,12]. formed to further upgrade enzyme stability (e.g., via For most applications of enzymes as biocat- chemical cross-linking), and may be also utilized as alysts an immobilization process is required, since it another tool to modify enzyme catalytic character- is the most common method for improving enzyme istics [21]. Glutaraldehyde is one of the most emp- stability [13]. A proper immobilization process, in addi- loyed reagents in enzyme cross-linking and immobil- tion to improving the enzyme stability, may also greatly ization [22]. enhance other enzyme properties, altering of the In addition to the numerous studies on the enzyme environment, or promoting protein rigidific- immobilization conditions and techniques that are ation [14]. Besides enhanced stability, immobilization available in the literature, there is great interest in the of enzymes permits efficient recovery of biocatalysts stabilization of the biocatalyst immobilized through from the reaction medium as well as allows their use multipoint covalent immobilization (it is not the unique in continuous operations, consequently reducing the methodology for this aim), which comprises all com- enzyme and product costs significantly [15-17]. binations of support functional groups and amino acid Furthermore, the terms immobilization and stab- residues that provide formation of multiple covalent ility are closely associated; however, if the immobil- bonds. Thus, this multipoint covalent immobilization ization procedure is not well designed, permitting promotes stiffening of the immobilized enzyme struc- uncontrolled enzyme-support interactions, immobil- ture and reduces the conformational changes invol- ized enzymes can be even less stable than soluble ved in its inactivation [23,24]. enzymes [18]. On the other hand, if an adequate According Gürdaş et al. [25] the anion exchange ® technique is utilized the stability of the enzyme may resin Duolite A568 has a highly porous matrix, is be also increased by immobilization process [18]. granular, weakly basic and based on a crosslinked Various immobilization methods for β-galactos- phenol-formaldehyde polycondensate with tertiary idase, including entrapment, cross-linking, adsorption amine functional groups that attach the enzyme through and covalent binding have already been suggested a combination of the ion exchange bonding between [8]. Nevertheless, immobilization by physical adsorp- the tertiary amine groups presents in the resin and the tion onto solid carriers can be a compelling strategy to carboxylic acid groups of the enzyme and the simple improve the operational stability and reusability, thus adsorption by hydrogen bonds and van der Waals ensuring process cost effectiveness, since that its forces. These authors studied a simple adsorption employment was ensured through a simple protocol, mechanism to immobilize A. oryzae β-galactosidase ® mild immobilization conditions, minimal damage of in Duolite A568. This resin has a hydrophilic struc- catalytic activity, and a wide variety of inexpensive ture and controlled pore size distribution that make it materials that can serve as enzyme carriers with the suitable for employment as a support for enzyme in a possibility of simple regeneration [8]. Physical adsorp- variety of bioprocessing applications. tion is a standard technique constituted by binding Abuquerque et al. [20] studied the effect of the between enzyme and solid support based on van der immobilization pH via adsorption on the final perform- Waals, ion exchange, and/or hydrophobic interact- ance of the β-galactosidase from A. oryzae. These ions. In the case of using this technique to immobilize authors concluded that the enzyme immobilization in proteins, those that do not become adsorbed on the ion exchangers may be more versatile than sup- support will be discarded. posed. If the protein has specially enriched areas in Enzyme immobilization by ion exchange is a anion groups, the change in the immobilization pH multipoint process, since the enzyme is fixed to the during ion exchange may alter the immobilized enz- support just when a high sufficiently number of ionic yme functional properties, and the simplest explan- bridges is established between the protein and the ation for this is that the enzyme orientation regarding support to compensate the ionic strength of the media the support may be altered by using different pH value.

496 L.N.S.S. FALLEIROS et al.: IMPROVEMENT OF RECOVERED ACTIVITY… Chem. Ind. Chem. Eng. Q. 23 (4) 495−506 (2017)

The present work investigated the immobil- pw is the total protein content remained in solution ization procedure of β-galactosidase from A. oryzae after immobilization process; αsi is the total specific ® on Duolite A568 combining the immobilization activity of immobilized biocatalyst and αsf is the total methods: physical adsorption, incubation in buffer pH specific activity of the initial enzyme solution. 9.0 to permit multi interaction and cross-linking pro- α cess, in order to aggregate the advantages of each Immobilization efficiency (%)= 100 i (1) α method. The influence of the buffer composition used f in the incubation step and the effect of the presence p − p Immobilization yield (%)= 100 fw (2) of the substrate during the immobilization process p were studied. It was evaluated the stability of the bio- f catalyst obtained in relation to pH, temperature, stor- α Immobilization specific efficiency (%)= 100 si (3) age and reusability. α sf

MATERIALS AND METHODS The specific activity (U/mgprotein) was determined as the quantity of β-galactosidase that hydrolyzes 1 Materials μmol of lactose per min per mg of protein at pH 4.5 ° The β-galactosidase (E.C. 3.2.1.23) from A. and 35 C, employing lactose 50 g/L as substrate. oryzae was purchase from Sigma. The enzyme act- Immobilization procedure ivity determined experimentally was 4 U per mg solid To promote the formation of multipoint links using lactose as a substrate. The unit of enzymatic between the enzyme and the support, the immobil- activity (U) was defined as the amount of β-galac- ization process was composed into three steps: tosidase that hydrolyzes 1 μmol of lactose per min at immobilization by adsorption to the resin, incubation pH 4.5 and 35 °C, employing lactose 50 g/L as sub- at pH 9.0 and cross-linking with glutaraldehyde. The strate. The protein concentration of the commercial procedures employed for each step are outlined enzyme, determined using the method reported by below. Lowry et al. [26], was 79.4 mg per g of solid. The ion ® Immobilization. The immobilization procedure exchange resin employed in this work (Duolite A568) involved the adsorption of the enzyme onto the ion was kindly donated by Dow Chemical Brazil S.A. All ® exchange resin Duolite A568 which was previously of the other reagents used in the study were of ana- activated conform described by Guidini et al. [28]. lytical purity. Samples of 0.5 g of this resin were mixed with 10 mL Determination of the activity of the immobilized of the enzyme solution (at the indicated concentration enzyme for each experiment) in 0.1 M acetate buffer, pH 4.5, ° The catalytic activity of the immobilized β-galac- and incubated at 25±1 C for 15 h in a shaking inc- tosidase for the catalysis of the lactose hydrolysis ubator with an agitation of 150 rpm. The immobiliz- reaction was measured by the initial rate method, and ation conditions used were those employed by Guidini the glucose produced in the reaction was determined et al. [28]. The strength of the buffer used was chosen by the glucose oxidase method [27]. based on the work of Husain et al. [31], Ansari et al. The hydrolysis reactions were performed in a [32] and Urrutia et al. [33]. reactor containing 100 mL of a 50 g/L lactose solution Incubation in buffer at pH 9.0. The incubation prepared with 0.1 M acetate buffer pH 4.5, at 35±1 °C step was performed by immersion the resin or the under magnetic agitation, according Guidini et al. [28]. enzyme derivative in 10 mL of 0.1 M phosphate buffer, pH 9.0, in a shaking incubator with an agitation of The unit of activity (U/gsupport) was defined as the ° amount of β-galactosidase that hydrolyzes 1 μmol of 150 rpm at 25±1 C for 24 h [34,35]. lactose per min per g of biocatalyst immobilized at pH Cross-linking. In this step, glutaraldehyde was 4.5 and 35 °C, employing lactose 50 g/L as substrate. used as the cross-linking agent at a ratio of 1 g of All of the experiments were conducted in triplicate. resin to 10 mL of 3.5 g/L glutaraldehyde. A volume of The immobilization efficiency and yield were 5 mL of the 3.5 g/L glutaraldehyde solution H 3.86 measured according to Zhang et al. [29] and Facin et was added to the biocatalyst immobilized, and the reaction system was maintained at a stirring of 150 al. [30] by Eqs. (1)-(3), where αi is the total enzymatic rpm at 25±1 °C for 1.5 h. The employed variables (the activity of immobilized biocatalyst and αf is the total glutaraldehyde concentration and the cross-linking activity of the initial enzyme solution; pf is the total protein content of the initial enzyme preparation and time) were defined by Guidini et al. [28].

497 L.N.S.S. FALLEIROS et al.: IMPROVEMENT OF RECOVERED ACTIVITY… Chem. Ind. Chem. Eng. Q. 23 (4) 495−506 (2017)

Influence of the buffer composition used during the Effect of the addition of substrate during the incubation step at pH 9.0 of the immobilized immobilization process biocatalyst To verify the influence of the substrate (lactose) To evaluate the effect of the buffer composition during the immobilization of the enzyme, six assays employed in the incubation step, three samples were were performed using the immobilization by ads- prepared using the three steps, immobilization by orption and the cross-linking steps as described pre- adsorption, with enzyme concentration of 16 g/L, fol- viously. The amount of substrate added in each step lowed by cross-linking with glutaraldehyde and finally used to obtain the biocatalyst was calculated for a lac- incubation step. Three different buffers (borate, Tris tose concentration of 50 g/L. In these experiments, the (hydroxymethyl) and phosphate) 0.1 M, pH 9.0, were concentration of the enzyme solution used was 16 g/L. used. The pH 9.0 phosphate buffer was prepared by pH Stability of the immobilized β-galactosidase 955.0 mL of 0.1 M disodium hydrogen phosphate solution (14.2 g/L) with 45.0 mL of 0.1 M HCl, accord- Samples of 0.5 g of the immobilized enzyme ing to the procedure reported by Farag et al. [36]. obtained after the three-step (immobilization by adsorption using an enzyme solution of 5 g/L, incubation Influence of the sequence of the steps on the in buffer pH 9.0 and cross-linking) immobilization pro- β immobilized -galactosidase cess, according to the procedure described pre- To study the influence of the step sequence in viously, were incubated in 10 mL of a buffer solution the preparation of the biocatalyst, experiments were with a pH value in the range of 1.5 to 9.0 at 25±1 °C performed using different methods as described pre- for 24 h. After the incubation, the biocatalysts were viously. The first experiment evaluated the influence washed with 0.1 M acetate buffer, pH 4.5. The resi- of the incubation in buffer at pH 9.0 of the resin before dual enzymatic activity was calculated in relation to the immobilization step, as described by Letca et al. the initial activity. The following buffers were used: ® [35]. The sample of 0.5 g of Duolite A568, previously hydrochloric acid-potassium chloride (0.1 M) at pH activated, was mixed with 10 mL of 0.1 M phosphate 1.5; 0.1 M citrate at pH 3.0, 0.1 M acetate for the pH buffer, pH 9.0, in a shaking incubator at 150 rpm at values between 4.0 to 5.0, and 0.1 M phosphate at 25±1 °C for 24 h. After this step the equilibrated resin the pH range of 6.0 to 9.0. For pH range of 6.0 to 8.0 was submitted to the immobilization by adsorption, were used stocks solutions of monobasic sodium followed by cross-linking with glutaraldehyde. phosphate and dibasic sodium phosphate. The pH In two other experiments, was studied the influ- 9.0 phosphate buffer was prepared as mentioned pre- ence of performing the incubation in buffer pH 9.0 of viously. the immobilized enzyme before and after the cross- Storage stability -linking step. The relative activity (ratio of the activity of the biocatalyst obtained after the three steps to the The biocatalysts obtained after the immobiliz- activity of the biocatalyst obtained only after ads- ation by adsorption, incubation in buffer pH 9.0 and orption step and cross-linking with glutaraldehyde) for cross-linking steps, using an enzyme solution of 5 each biocatalyst was measured by the method of ini- g/L, were stored in buffer solution pH 4.5 (0.1 M tial rates. In all of the experiments performed in this sodium acetate) at 4 °C in for 98 days. The stability of part of the study, the concentration of the enzyme sol- two different samples was studied: one of the samples ution used in the immobilization was 16 g/L. was obtained by washing the resin between each Two experiments were conducted using the enz- immobilization step, and the other was obtained without yme obtained after the three steps (immobilization by washing the resin. The residual enzymatic activity in adsorption, incubation in buffer 9.0 and cross-linking): comparison to the initial activity was determined by the resin was washed with acetate buffer pH 4.5 the method of initial rates. between each step in the first experiment, whereas Reusability of immobilized β-galactosidase the resin was not washed in the second experiment, The immobilized biocatalyst obtained by the and the cross-linking step carried out at pH 9.0. three-step (immobilization by adsorption, incubation To analyze the effect of the washing step, the in buffer pH 9.0 and cross-linking) process using an protein concentration in the supernatant was deter- enzyme solution of 5 g/L, was used in 10 consecutive mined before and after the immobilization step using lactose hydrolyses assays of 300 min each. The glu- the method related by Lowry et al. [26]. In these exp- cose produced in the reaction was determined by the eriments, the concentration of the enzyme solution glucose oxidase method [27]. Between each hydro- used in the immobilization step was 5 g/L. lysis reaction, the biocatalyst was washed with a 0.1

498 L.N.S.S. FALLEIROS et al.: IMPROVEMENT OF RECOVERED ACTIVITY… Chem. Ind. Chem. Eng. Q. 23 (4) 495−506 (2017)

M sodium acetate buffer, pH 4.5 and stored at 4 °C in The analysis of the results presented in Table 1 the same buffer until the next hydrolysis reaction. The show that the choice of buffer significantly affects the conditions utilized for all hydrolyses experiments were enzymatic activity of the biocatalyst. The use of bor- temperature of 35±1.0 °C, 0.1 M sodium acetate buf- ate decreases the activity almost by 70% when com- fer, pH 4.5 and initial lactose concentration of 50 g/L. pared with the activity of the derivative obtained employing phosphate buffer. Thermal stability of the immobilized β-galactosidase A blank assay was performed employing enz- With the purpose of studying the thermal stab- yme solution 5 g/L prepared with 0.1 M borate buffer ility of the biocatalyst immobilized obtained after the pH 9.0 and then determined the activity under the three-step (immobilization by adsorption, incubation same conditions employed by immobilized enzyme. in buffer pH 9.0 and cross-linking) immobilization pro- The enzymatic activity for soluble enzyme achieved cess, the samples of β-galactosidase immobilized using borate buffer was 13.33 U, which corresponds obtained using an enzyme solution of 5 g/L, were to about 7% of the soluble enzyme activity prepared incubated into 50 mL of pH 4.5 acetate buffer in a in acetate buffer (194.44 U). thermostatic bath and the temperature ranged between This result can be explained by the fact of borate 55.0 to 65.0 °C. For each studied temperature, the buffer contains boric acid in its composition, that samples of immobilized biocatalyst were removed of behaves like a Lewis acid (electron acceptor), allow- the thermostatic bath at specific intervals of time (5 ing their complexation with the electron pair of the min for 65.0±1 °C, 7 min for 62.5±1 °C, 10 min for amino groups of the enzyme, causing a decrease in 60.0±1 °C, 15 min for 57.5±1 °C, 20 min for 55.0±1 enzyme activity [39,40]. °C) and were cooled fast in an ice bath, and the resi- Shubhada and Sundaram [41] have studied the dual activity was measured by the initial rate of react- stability of β-galactosidase from A. oryzae in water- ion using 50 g/L of initial lactose concentration, pH -miscible organic solvents at different pH values in 4.5 at 35±1 °C. The residual enzymatic activities various buffers. These authors noted that in the stu- obtained as a function of incubation time for each tem- died pH values, the loss of stability was greater in perature were adjusted to first order thermal deactiv- sodium borate buffer than in sodium citrate-sodium ation model and in series thermal deactivation model phosphate or sodium phosphate buffers. These authors in a single step [38] by the numerical method of Lev- state that the greater loss of stability observed when ® enberg–Marquardt [37] using the Statistica 7.0 soft- used sodium borate buffer, in comparison with other ware to obtain the kinetic parameters for the best fit [38]. studied solvents, is related to the fact that borate ions are known to complex with the carbohydrate part of RESULTS AND DISCUSSION glycoproteins. Since β-galactosidase is a glycopro- tein, borate complexation may further assist the access Influence of the buffer composition used during the of organic solvents, possibly through a change of con- incubation step at pH 9.0 of the immobilized formation. biocatalyst The use of phosphate and Tris buffers produced The results of the relative enzymatic activity immobilized biocatalysts with similar catalytic acti- (ratio of the activity of the biocatalyst obtained using vities, with a lower cost for the phosphate buffer. different buffers in relation to the activity of the bio- Thus, phosphate buffer was chosen to prepare the catalyst obtained using phosphate buffer) of the samples of biocatalyst in the work sequence. immobilized enzymes obtained using different buffers Influence of the sequence of the steps on the during the incubation step are presented in Table 1. immobilized β-galactosidase Table 1. Results of the influence of the buffer composition used The influence of the sequence of the immobil- in the incubation step in the immobilization process ization by adsorption, cross-linking and incubation (in

a buffer pH 9.0) steps on the immobilized enzyme act- Buffer composition Activity , U/gsupport Relative activity, % ivity is shown in Table 2, that demonstrate that the Borate 155.56±6.67 32 incubation step (immersion in 0.1 M phosphate buffer, Phosphate 486.67±12.78 100 pH 9.0) increased the enzyme activity, regardless of Tris 462.22±11.11 95 a where in the sequence this step was performed. In The unit of activity (U/gsupport) was defined as the amount of β-galactosidase that hydrolyzes 1 μmol of lactose per min per g of biocatalyst addition, an increase in enzyme activity of 44% was immobilized at pH 4.5 and 35 °C, employing lactose 50 g/L as substrate achieved when the enzyme was subjected to immobil-

499 L.N.S.S. FALLEIROS et al.: IMPROVEMENT OF RECOVERED ACTIVITY… Chem. Ind. Chem. Eng. Q. 23 (4) 495−506 (2017)

Table 2. Results of the enzymatic activity of the biocatalyst obtained by combination of steps in the immobilization process compared with the enzymatic activity achieved by the immobilized biocatalyst obtained by adsorption and subsequent cross-linking

a Steps for obtaining the biocatalyst Activity , U/gsupport Relative activity, % Immobilization 401.11 ± 18.33 106 Immobilization + Crosslinking 377.78 ± 10.56 100 Immobilization + Incubation at pH 9.0 462.22 ± 11.11 124 Incubation at pH 9.0 + Immobilization + Crosslinking 451.11 ± 10.00 119 Immobilization + Incubation at pH 9.0 + Crosslinking 543.33 ± 15.00 144 Immobilization + Crosslinking + Incubation at pH 9.0 486.67 ± 12.78 129 a The unit of activity (U/gsupport) was defined as the amount of β-galactosidase that hydrolyzes 1 μmol of lactose per min per g of biocatalyst immobilized at pH 4.5 and 35 °C, employing lactose 50 g/L as substrate ization by adsorption followed by incubation in buffer that glutaraldehyde’s contribution could probably be pH 9.0 and then exposed to the cross-linking agent. ascribed predominantly to cross-linking between enz- The improved in enzyme activity can be conse- yme molecules. quence of conformational changes that drive to a final An important aspect of pH influence on immobil- conformation with more activity since in fact, the results ization by adsorption can be explained by the fact that shows that the loss of activity caused by the immobil- enzymes are more susceptible to interact with positi- ization step was prevented. vely charged functional groups if they are negatively One of the experimental conditions that can alter charged, which occurs at pH values above their pI, the balance between cationic and anionic groups on a which is around 4.6 for β-galactosidase from A. protein surface is the pH. It is well known and accepted oryzae [6]. that this variable determines if an enzyme is adsorbed Though, overall enzyme charge is not the only or not during anion exchange [19-20]. However, it has factor that impact adsorption via ionic interactions, shown that the immobilization on an ion exchanger once distribution of charged residues on the surface support may involve different areas depending on the of enzyme also influences the efficiency of immobil- immobilization pH value and how this may affect the ization by directing the orientation of adsorbed mole- final enzyme properties, e.g., stability, activity or sel- cules. Since surface of the support is positively ectivity [20]. charged throughout examined pH range due to prim- Albuquerque et al. [20] affirm that if the protein ary amino groups, it can be assumed that distribution has specially enriched areas in anion groups, the vari- of negatively charged residues and properties of resi- ation in the immobilization pH during ion exchange dues in their proximity cause large discrepancies in may modify the immobilized enzyme functional pro- efficiency of immobilization at different pH values. The perties, and the simplest explanation for this is that effect of pH on the efficiency of immobilization could the enzyme orientation in relation to the support may be further explained by distribution of amino acid resi- be altered by using different pH value. According to dues relevant for adsorption on the surface of β-gal- the authors, their work was the first to probe this actosidase from A. oryzae [6]. statement for the β-galactosidase from A. oryzae Banjanac et al. [6] studied the immobilization of (same enzyme studied in the present work). β-galactosidase from A. oryzae onto unmodified, The catalytic activities of the enzymes obtained amino modified and cyanuric chloride functionalized after each step (immobilization by adsorption, stabil- amino modified nonporous fumed nano-silica par- ization and cross-linking) were determined. As shown ticles (FNS, AFNS and CCAFNS, respectively) and it in Table 2, a gradual increase in the activity was can be concluded that immobilization of β-galacto- obtained after each step. sidase is more favorable via carboxyl residues than In the first step, immobilization by adsorption, via amino acid residues present on enzyme surface the immobilization of the enzyme occurs through int- which is in agreement with previously study reported eractions between support and the enzyme. The by Carevic et al. [8]. incubation of the enzyme immobilized by adsorption The influence of washing between each step in in a buffer at pH 9.0 can permit to achieve an intense the process used to obtain the biocatalyst was stu- multi interaction enzyme-support. In the last step the died, and the specific activities and yield obtained addition of glutaraldehyde was responsible for the after the three steps of the process with and without appearance of new links between the enzyme mole- the washing step are presented in Table 3. cules linked in the support to other molecules still As shown in Table 3, the suppression of the available in the immobilization medium. In this way washing step between the preparation steps of the

500 L.N.S.S. FALLEIROS et al.: IMPROVEMENT OF RECOVERED ACTIVITY… Chem. Ind. Chem. Eng. Q. 23 (4) 495−506 (2017)

Table 3. Results of the enzymatic activity, specific activity, immobilization yield and efficiency in comparison with free enzyme

b a Specific activity , Immobilization Immobilization Immobilization specific Enzyme Enzyme activity , U U/mgprotein yield, % efficiency, % efficiency, % Soluble (Free) 194.44 32.03±1.71 - - - Immobilized without washing 166.66 31.32±0.96 88 86 98 Immobilized with washing 158.33 28.56±0.65 75 81 89 aThe unit of activity (U) refers to enzyme activity defined as the amount of β-galactosidase that hydrolyzes 1 μmol of lactose per minute at pH 4.5 and 35 °C, b employing lactose 50 g/L as substrate; the unit of activity (U/mgprotein) refers to specific activity defined as the amount of β-galactosidase that hydrolyzes 1 μmol of lactose per minute per mg of protein at pH 4.5 and 35 °C, employing lactose 50 g/L as substrate enzymatic derivative increased the immobilization The presented results show that the immobil- yield by approximately 17.3%, which resulted in an ization of β-galactosidase on anion exchange resin ® increase in the specific activity of 9.7%. (Duolite A568) does not affect the active site of the The analysis of the results of Table 3 shows enzyme, even in the absence of lactose. This finding similar catalytic activities exhibited by the soluble and is likely because the site is well protected inside the immobilized enzymes. Moreover, the experiences protein and does not significantly suffer from steric resulted in a high immobilization specific efficiency hindrance or obstructions during the immobilization (98%). A blank assay with the soluble enzyme was process. Thus, the addition of lactose as a protector performed, and after incubation in the same immobil- of active site did not change the activity of the immo- ization conditions there was no loss of activity. This bilized enzyme, which indicates that the active site is result indicates that nearly all of the catalytic activity not directly involved in the immobilization process and of the soluble enzyme was retained after the immobil- that the addition of the substrate during the immobilization procedure. ization process is unnecessary [42]. The results above are consistent with the work Effect of the addition of substrate during the of Aguiar-Oliveira [43], who studied the effect of the immobilization process addition of salts and substrate during the immobil- The biocatalyst obtained after three-steps has ization of fructosyltransferase on niobium and the effect presented high activity retention, nevertheless in the of these compounds on the catalytic properties of the present work was studied the possibility of increase resulting immobilized biocatalyst. The researcher the activity using the substrate as a protector of the found that the presence of the substrate does not active site. Table 4 shows the enzymatic activities affect the immobilized enzyme activity in the syn- obtained in the study of the effect of the presence of thesis of FOS (fructooligosaccharide). the substrate (lactose) during the immobilization pro- Based on the knowledge that the existence of cess. the substrate or a competitive inhibitor can preserve As shown in Table 4, the presence of lactose the active site during the cross-linking and immobil- does not significantly influence the immobilized bio- ization processes, Oosterom et al. [44] performed the catalyst under the conditions evaluated. The enzym- immobilization of the β-galactosidase from A. oryzae ® atic activities obtained when the enzyme was immo- on Duolite S761 in the presence of 0.1 M lactose. bilized by adsorption process in the presence and in The enzymatic activity of this immobilized enzyme the absence of substrate were similar. The analysis of increased around 30% in comparison with the immo- the cross-linking process showed that the enzymatic bilized biocatalyst obtained in the absence of lactose. activities of the enzymes obtained after the cross-link- However, the presence of lactose had no effect on the ing step performed in the presence and in the abs- performance of the immobilized biocatalyst when ence of the substrate were not significantly different.

Table 4. Results of the effect of the addition of substrate during the immobilization process; the unit of activity (U/gsupport) was defined as the amount of β-galactosidase that hydrolyzes 1 μmol of lactose per min per g of biocatalyst immobilized at pH 4.5 and 35 °C, employing lactose 50 g/L as substrate

Sample Steps used to obtain the biocatalyst Enzymatic activity (U/gsupport) 1 Immobilization by adsorption without lactose 401.11±18.33 2 Immobilization by adsorption with lactose 400.00±12.78 3 Immobilization by adsorption without lactose + cross-linking without lactose 377.78±10.56 4 Immobilization by adsorption without lactose + cross-linking with lactose 367.78±10.56 5 Immobilization by adsorption with lactose + cross-linking without lactose 326.67±6.11 6 Immobilization by adsorption with lactose + cross-linking with lactose 330.00±9.44

501 L.N.S.S. FALLEIROS et al.: IMPROVEMENT OF RECOVERED ACTIVITY… Chem. Ind. Chem. Eng. Q. 23 (4) 495−506 (2017) dimethyl adipimidate (DMA) was used as the cross- over a narrow pH range (from 4 to 5). However, after -linking agent. the addition of the cross-linking with glutaraldehyde step, these authors obtained the residual activity pH influence on the stability of the immobilized approximately 95% for a pH range of 2 to 7. biocatalyst Freitas et al. [45] present a study of kinetic pro- The reaction medium pH exerts a great inf- perties of β-galactosidase from A. oryzae soluble and luence on the catalytic stability of most enzymes. Fig- immobilized in alginate, gelatin and glutaraldehyde. ure 1 shows the residual activity, which is determined The enzyme maintain its total activity after 18 h of by the ratio of the activity after a 24-h incubation to incubation, for a pH extension of 4.0–7.0, in its free the initial activity. and immobilized forms, which coincides with the stability of the soluble enzyme, as recommended by the manufacturer and the literature [1,45,46]. Grazú et al. [47] immobilized penicillin G acyl- ase (PGA) at neutral pH for 24 h and then incubated the immobilized enzyme for 48 h at pH 10. These researchers found that more than 90% of the catalytic activity was recovered after the stiffening obtained during the immobilization process. The results obtained in present work are in accordance with the findings related by Albuquerque et al. [20] that explain, in first time, that the immobilization on an ion exchanger support may involve different areas depending on the immobilization pH value and this may affect the final enzyme properties, e.g., stability.

Figure 1. pH influence in the stability of the immobilized Storage stability β -galactosidase. The influence of the storage time on the activity retention of immobilized β-galactosidase was evaluated. The study of the pH stability of the immobilized The results show that the residual activity obtained for biocatalyst obtained by three-step immobilization pro- the samples with and without washing maintained cess (adsorptive immobilization, incubation at pH 9.0 their activities after 98 days of storage in 0.1 M and cross-linking with glutaraldehyde) shows that the sodium acetate buffer, pH 4.5, and at a temperature enzymatic derivative presents high stability at an ext- of 4±2 °C, as shown in Figure 2. These results indi- ensive range of pH values. The residual enzyme acti- cate that the immobilization procedure improved the vity was approximately 100% in relation to initial activity at all pH values studied, which indicates high stability. The maintenance of the activity of the biocatalyst after a 24-h incubation at different pH values may be further explained by the additional step of incubation at pH 9.0 realized between the immobilization by adsorption and the cross-linking with glutaraldehyde steps. The results observed by Guidini et al. [28], who studied the stabilities of the biocatalysts obtained after a one-step (immobilization by adsorption) and a two-step immobilization procedure (immobilization by adsorption followed cross-linking with glutaraldehyde), presented in Figure 1 for comparison, show less stability compared with that obtained in present work. According to the work of Guidini et al. [28], the Figure 2. Storage stability of immobilized β-galactosidase enzyme obtained after the one-step immobilization (sample 1 – the biocatalyst was washed between each step process (immobilization by adsorption) was stable used to obtain the same; sample 2 – without the washing step).

502 L.N.S.S. FALLEIROS et al.: IMPROVEMENT OF RECOVERED ACTIVITY… Chem. Ind. Chem. Eng. Q. 23 (4) 495−506 (2017) stability of the biocatalyst in relation to the storage, buffer pH 4.5) allowing their reuse for long and rep- since, according to Haider and Husain [1], the soluble eated times. A. oryzae β-galactosidase, in the same storage con- The results obtained in the present work were ditions, has maintained only 40±2.5% of the initial better than those obtained by Guidini et al. [28], who activity after 60 days. studied the reuse of β-galactosidase immobilized by These results are consistent with the work of adsorption on the same support followed by cross- Guidini et al. [28], who reported that the stability of the linking with glutaraldehyde, in which they found a loss β-galactosidase enzyme from A. oryzae immobilized of activity of 10% after the fifth 10 min assay in the ® on Duolite A568 by adsorption and then crosslinked same conditions. with glutaraldehyde was maintained after 90 days of Ansari and Husain [11] obtained for β-galacto- stocking in acetate buffer, pH 4.5, at 4±2 °C. sidase adsorbed on Con A-celite and adsorbed and Haider and Husain [1] immobilized the β-galac- crosslinked on Con A-celite, a residual activity of 64 tosidase from A. oryzae on calcium alginate and and 71% after 7th repeated use, respectively. crosslinked with concanavalin A. These researchers Thermal stability found that the resulting immobilized enzyme maintained an activity of 93% compared with the baseline Figure 4 shows the strong dependence of the after 2 months of storage at 4 °C. stability of the immobilized biocatalyst on tempera- Ansari and Husain [11] described the storage ture. It is observed that for 65.0 °C in 15 min there stability of free and immobilized β-galactosidase from was a decrease of 50% of the activity in relation to the A. oryzae at 4 °C in 0.1 M acetate buffer, pH 4.5 for initial activity. For the temperatures of 60.0, 57.5 and more than 2 months. The crosslinked enzyme pre- 55.0 °C the biocatalyst maintained the initial activity sented 78% activity after 2 months of storage while its until 30-min incubation, thereafter was observed the soluble form retained only 40% activity under similar thermal inactivation process, demonstrating that for storage conditions. the range of 55 to 60 °C, the immobilized enzyme is more resistant to thermal deactivation. For the tempe- β Reusability of immobilized -galactosidase rature of 55 °C, the biocatalyst retains almost all of its The immobilization of enzymes is an important initial activity after 60-min incubation and after 2.3 h, tool that improves the utilization of enzymatic cat- the biocatalyst maintained 80% of its initial activity. alysts increasing stability and allowing their reuse. The results of reuse of immobilized β-galactosidase have been presented in Figure 3. The biocatalyst was used during 10 consecutive hydrolyses assays of 300 min each and the conversion of lactose into glucose and galactose was the same for all assays. These results indicate the high stability of the biocatalyst in the reaction conditions (35 °C and 0.1 M acetate

Figure 4. Activity relative as a function of incubation time for the temperatures of 55.0–65.0 °C.

This maintenance of initial activity until 30 min incubation can be related to the insertion of the incubation in buffer pH 9.0 step in obtaining a biocatalyst, since this was not observed by Guidini et al. [48] that studied the thermal stability of the β-galactosidase Figure 3. Conversion of lactose using β-galactosidase immobilized by two-step immobilization process immobilized in relation to the number of uses. (immobilization by adsorption followed cross-linking)

503 L.N.S.S. FALLEIROS et al.: IMPROVEMENT OF RECOVERED ACTIVITY… Chem. Ind. Chem. Eng. Q. 23 (4) 495−506 (2017) and the results obtained showed a fall of 10% activity The authors Haider and Husain [1] evaluated after 60 min incubation at 55 °C. the thermal stability of calcium alginate entrapped Freitas [49] studied the thermal stability of β-gal- preparations of A. oryzae β-galactosidase and actosidase from A. oryzae free and was observed for obtained for entrapped soluble and insoluble concan- the temperature of 65 °C a rapid enzyme deactiv- avalin complex half-lifes of 45 and 60 min respect- ation, requiring only five minutes for the enzyme lost ively, at 60 °C in 0.1 M sodium acetate buffer (pH 80% of its catalytic activity. 4.5). An strong dependence on the thermal stability of The results obtained experimentally for the resi- the enzyme in function of method of immobilization dual enzymatic activity were adjusted to the model of was observed. thermal deactivation in the first order and in series in The kd values shown in Table 5, were adjusted a single stage [38] by the numerical method of to the Arrhenius model utilizing the Origin® 7.0 soft- Levenberg–Marquardt [37] employing the Statistica® ware. With a determination coefficient of 0.96 the 7.0 software. The significance analysis was realized linear fit was obtained. Therefore, the activation using the Student’s t-test test considering similar sig- energy determined from the process of thermal deact- nificant parameters as those with significance levels ivation was 71.03 kcal/mol, a value less than to the smaller than 10%. free enzyme obtained by Freitas [49], which was Analyzing the results (in supporting information, 88.14 kcal/mol for the same enzyme. These results available from the author upon request) of the indicate that the immobilized enzyme in the present modeling of thermal deactivation, it follows that for all work is more stable for an increment temperature in temperatures, the best adjust was achieved by apply- the lactose hydrolysis process in comparison to the ing first order deactivation kinetics. On the other same enzyme in the soluble form. hand, the model of deactivation in series in a single step was inappropriate for the fit in all temperatures. CONCLUSIONS With the experimental results of residual enzym- β atic activity on incubation time were calculated the The immobilization process of A. oryzae -galactosidase by adsorption, incubation in buffer pH 9 thermal deactivations constants, kd, for first order model for each incubation temperature, which were and cross-linking, in this sequence, employing ® used to calculate the half-life, as showed in Table 5. Duolite A568 as support, resulted in a biocatalyst Freitas [49] reported in your work, the half-life mea- with better activity and stability compared with the sured by the first order thermal deactivation model for biocatalyst obtained only by adsorption and cross- soluble form of β-galactosidase from A. oryzae at 65, -linking. The immobilized biocatalyst obtained pre- 63, 61, 57 and 55 °C were 2.40, 11.68, 13.59, 57.70 sented a good thermal stability at temperatures around ° ° and 177.70 min, respectively. 50 C, with a half-life of 5.48 h at 55 C and also These results clearly show the biocatalyst in the excellent pH stability in the range from 1.5 to 9.0. immobilized form has higher thermal stability in com- There was also an increase of stability in comparison parison to the soluble enzyme. Haider and Husain [1] to the storage time and in relation to the number of obtained for β-galactosidase from A. oryzae in the uses of the biocatalyst immobilized. Based on the free form a half-life of approximately 30 min for previously mentioned conclusions, the use of sequ- thermal stability at 60 °C and pH 4.6. Husain et al. ence of the steps of adsorption, followed by incub- [31] and Ansari et al. [32] have studied the thermal ation in buffer pH 9.0 and cross-linking with glutaraldehyde in the immobilization of β-galactosidase of A. denaturation of same enzyme in the soluble form at ® 60 °C in 0.1 M sodium acetate buffer (pH 4.5) and oryzae in Duolite A568 as support presents potential obtained the half-life of 30 to 40 min under these con- for industrial applications of this enzyme in the ditions. immobilized form.

Table 5. Half-life for each temperature using the model of thermal deactivation of the first order

-1 Temperature, °C kd / min R² t1/2 / min t1/2 / h 65.0 0.0546 0.92 12.70 0.21 62.5 0.0313 0.98 22.15 0.37 60.0 0.0211 0.95 62.85 1.05 57.5 0.0052 0.92 163.30 2.72 55.0 0.0024 0.94 328.81 5.48

504 L.N.S.S. FALLEIROS et al.: IMPROVEMENT OF RECOVERED ACTIVITY… Chem. Ind. Chem. Eng. Q. 23 (4) 495−506 (2017)

Acknowledgements [22] O. Barbosa, C. Ortiz, A. Berenguer-Murcia, R. Torres, R.C. Rodrigues, R. Fernandez-Lafuente, RSC Adv. 4 The authors wish to thank the CNPq, CAPES (2014) 1583-1600 and FAPEMIG (Brazil) for the financial support and [23] Y.A.R. Tapias, C.W. Rivero, F. López Gallego, J.M. Dow Chemical Brazil S.A for the resin donation. Guisán, J.A. Trelles, Food Chem. 208 (2016) 252–257 [24] B.C.C. Pessela, R. Torres, M. Fuentes, C. Mateo, M. REFERENCES Filho, A.V. Carrascosa, R. Fernandez-Lafuente, Biotechnol. Progress 20 (2004) 1507-1511 [1] T. Haider, Q. Husain, Int. J. Biol. Macromol. 41 (2007) 72- [25] S. Gürdaş, H.A. Güleç, M. Mutlu, Food Bioprocess –80 Technol. 5 (2012) 904-911 [2] R. Panesar, P.S. Panesar, R.S. Singh, J.F. Kennedy, J. [26] O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randal, J. Chem. Technol. Biotechnol. 86 (2011) 42-46 Biol. Chem. 193 (1951), pp. 193, 265-275 [3] I. Van De Voorde, K. Goiris, E. Syryn, C. Van Den [27] J. Raba, H.A. Mottola, Crit. Rev. Anal. Chem. 25 (1995) Bussche, G. Aerts, Process Biochem. 49 (2014) 2134- 1-42 –2140 [28] C.Z. Guidini, J. Fischer, L.N.S. Santana, V.L. Cardoso, [4] A. Nath, S. Mondal, S. Chakraborty, C. Bhattacharjee, R. E.J. Ribeiro, Biochem. Eng. J. 52 (2010) 137-143 Chowdhury, Asia-Pac. J. Chem. Eng. 9 (2014) 330–348 [29] Y. Zhang, M. Jeya, J. Lee, Appl. Microbiol. Biotechnol. 89 [5] M. Khan, Q. Husain, A. H. Naqvi, RSC Adv. 6 (2016) (2011) 1435-1442 53493–53503 [30] B.R. Facin, B. Moret, D. Baretta, L.A. Belfiore, A.T. [6] K. Banjanac, M. Carević, M. Ćorović, A. Milivojević, Paulino, Food Chem. 179 (2015) 44-51 N.Prlainović, A.Marinkovićc, D.Bezbradica, RSC Adv. 6 (2016) 97216–97225 [31] Q. Husain, S.A. Ansari, F. Alam, A. Azam, Int. J. Biol. Macromol. 49 (2011) 37-43 [7] A. Kamran, Z. Bibi, A. Amana, S.A.U. Qader, Biocatal. Agric. Biotechnol. 5 (2016) 99–103 [32] S.A. Ansari, R. Satar, F. Alam, M.H. Alqahtani, A.G. Chaudhary, M.I. Naseer, S. Karim, I.A. Sheikh, Process [8] M. Carević, M. Ćorović, M. Mihailović, K. Banjanac, A. Biochem. 47 (2012) 2427-2433 Milisavljević, D. Velicković, D. Bezbradica, Int. Dairy J. 54 (2016) 50–57 [33] P. Urrutia, C. Mateo, J.M. Guisan, L. Wilson, A. Illanes, Biochem. Eng. J. 77 (2013) 41-48 [9] V.Gékas, M.Lopez-Leiva, Process Biochem. 20 (1985) 2-12 [34] C. Mateo, O. Abian, R. Fernandez-Lafuente, J.M. Guisan, Enzyme Microb. Technol. 26 (2000) 509-515 [10] M.M. Maksimainen, A.Lampio, M. Mertanen, O.Turunen, J. Rouvinen, Int. J. Biol. Macromol. 60 (2013) 109-115 [35] D. Letca, C. Hemmerling, M. Walter, D. Wullbrand, K. Buchholz, Roman. Society Biol. Sci. 9 (2004) 1879-1886 [11] S.A. Ansari, Q. Husain, Food Bioprod. Process 90 (2012) 351-359 [36] R.S. Farag, M.S. Afifi, M.M. Abd-Rabow, Int. J. Pharm. Sci. Res. 2 (2011) 1197-1203 [12] Q. Husain, Critical Rev. Biotechnol.30 (2010) 41-46 [37] J.J. Moré, G.A. Watson, Lect. Notes Mathematics 630 [13] N. An, C.H. Zhou, X.Y. Zhuang, D.S. Tong, W.H. Yu, (1977) 105-116 Appl. Clay Sci. 114 (2015) 283–296 [38] J.P. Henley, A. Sadana, Enzyme Microb. Technol. 7 [14] F.F. Freitas, L.D.S. Marquez, G.P. Ribeiro, G.C. Brandão, (1985) 50-60 V.L. Cardoso, E.J. Ribeiro,Braz. J. Chem. Eng. 29 (2012) 15-24 [39] A.F.C. Alcântara, H.S. Barroso, D. Piló-Veloso, Quim. Nova 25 (2002) 300–311 [15] L.D.S. Marquez, B.V. Cabral, V.L. Cardoso, E.J. Ribeiro, J. Mol. Catal., B-Enzym. 51 (2008) 86-92 [40] J.P. Montalvo-Andia, L.A. Teixeira, PhD Thesis, Pontifical Catholic University of Rio de Janeiro (2009) [16] V. Stepankova, S. Bidmanova, T. Koudelakova, Z. Prokop, R. Chaloupkova, J. Damborsky, ACS Catal. 3 [41] S. Shubhada, P. Sundaram, Enzyme Microb. Technol. 15 (2013) 2823-2836 (1993) 881-886 [17] B. Das, A.P. Roy, S. Bhattacharjee, S. Chakraborty, C. [42] M. Wiltchek, T.J. Miron, Biochem. Bioph. Methods 55 Bhattacharjee, Ecotox. Environ. Safety 121 (2015) 244- (2003) 67-70 252 [43] E. Aguiar-Oliveira, PhD Thesis, Faculty of Food Eng- [18] O. Barbosa, R. Torres, C. Ortiz, A. Berenguer-Murcia, ineering, State University of Campinas (2012) R.C. Rodrigues, R. Fernandez-Lafuente, Biomacromole- [44] M.W. Oosterom, H.J.A. Belle, F. Rantwijl, R.A. Sheldon, cules 14 (2013) 2433-2462 J. Mol. Catal., A: Chem. 134 (1998) 267-274 [19] X.D. Tong, X.Y. Dong, Y. Sun, Biochem. Eng. J. 12 [45] F.F. Freitas, L.D.S. Marquez, G.P. Ribeiro, G.C. Brandão, (2002) 117–124 V.L. Cardoso, E.J. Ribeiro, Biochem. Eng. J. 58-59 [20] T.L. de Albuquerque, S. Peirce, N. Rueda, A. Marzo- (2011) 33-38 cchella, L.R.B. Gonçalves, M.V.P. Rocha, R. Fernandez- [46] A.S. Tanriseven, E. Dogan, Process Biochem. 38 (2002) -Lafuente, Process Biochem. 51 (2016) 875–880 27-30 [21] A. Díaz-Rodríguez, B.G. Davis, Curr. Opin. Chem. Biol. 15 (2011) 211–219

505 L.N.S.S. FALLEIROS et al.: IMPROVEMENT OF RECOVERED ACTIVITY… Chem. Ind. Chem. Eng. Q. 23 (4) 495−506 (2017)

[47] V. Grazú, F. López-Gallego, T. Montes, O. Abian, R. [48] C.Z. Guidini,J. Fischer, M.M. Resende, V.L. Cardoso, González, J.A. Hermoso, J.L. García,C. Mateo, J.M. E.J. Ribeiro, J. Mol. Cat., B-Enzym. 71 (2011) 139-145 Guisán, Process Biochem. 45 (2010) 390-398 [49] F.F. Freitas, PhD Thesis, Faculty of Chemical Eng- ineering, Federal University of Uberlândia (2007).

LARISSA NAYHARA SOARES POBOLJŠANJE OBNOVLJENE AKTIVNOSTI I SANTANA FALLEIROS STABILNOSTI β-GALACTOZIDAZE IZ Aspergillus BRUNA VIEIRA CABRAL oryzae IMOBILISANE NA DUOLITU A568 JANAÍNA FISCHER KOMBINACIJOM IMOBILIZACIONIH METODA CARLA ZANELLA GUIDINI VICELMA LUIZ CARDOSO Imobilizacija i stabilizacija β-galaktozidaze iz Aspergillus orizae na Duolitu A568 su izvr- MIRIAM MARIA DE RESENDE šene korišćenjem kombinacije fizičke adsorpcije, inkubacije u puferu pri pH 9,0 i umre- ELOÍZIO JÚLIO RIBEIRO žavanja sa glutaraldehidom, pri čemu je enzimska aktivnost povećana za 44% u odnosu Faculty of Chemical Engineering, na biokatalizator dobijen nakon procesa dvostepene imobilizacije (adsorpcija i umreža- Uberlândia Federal University, Campus vanje). Stabilnost biokatalizatora dobijenog u tristepenom procesu imobilizacije (adsorp- Santa Mônica, Uberlândia-MG, Brazil cija, inkubacija u puferu na pH 9,0 i umrežavanje) je veća od biokatalizatora dobijenog dvostepenim postupkom (adsorpcija i umrežavanje), kao i od slobodnog enzima u

odnosu na pH, skladištenje i ponovno korišćenje. Imobilizovani biokatalizator je okarak- NAUČNI RAD terisan u odnosu na termičku stabilnost u opsegu 55–65 °C. Kinetika termičke deakti- vacije opisana je modelom prvog reda, pri čemu je energija aktivacije termičke deakti- vacije imobilizovanog biokatalizatora 297,2 kJ/mol, a poluvremem života 5,48 h na 55,0 °C.

Ključne reči: β-galaktozidaza, Duolit A-568, imobilizacija, inkubacija u puferu pri pH 9,0.

506 Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 507−514 (2017) CI&CEQ

DUNG VAN NGUYEN FED-BATCH FERMENTATION OF NIPA SAP HARIFARA TO ACETIC ACID BY Moorella thermoacetica RABEMANOLONTSOA (f. Clostridium thermoaceticum) SHIRO SAKA

Department of Socio-Environ- Article Highlights mental Energy Science, Graduate • Acetic acid could be produced from the underutilized nipa sap using Moorella thermo- School of Energy Science, Kyoto acetica University, Yoshida-honmachi, • High substrate concentration in batch fermentation caused low conversion efficiency • Sakyo-ku, Kyoto, Japan Fed-batch fermentation provided over 42.6 g/L of acetic acid with high conversion efficiency of 87% • High feeding rate improved acetic acid productivity as compared with the low feeding SCIENTIFIC PAPER rate • The current method shows much higher efficiency than the traditional vinegar process UDC 633.8:58:663:547.292

Abstract An efficient process for conversion of nipa sap to acetic acid was developed. Nipa sap was hydrolyzed with invertase and provided glucose as well as fructose as main sugars. Batch fermentation of glucose and fructose was inadequate with increased substrate concentration. By contrast, fed-batch technique on hydrolyzed nipa sap with high feeding rate drastically increased acetic acid concentration and productivity to be 42.6 g/L and 0.18 g/(L/h), respectively. All the sugars in hydrolyzed nipa sap were consumed, with acetic acid yield of 0.87 g/g sugar. Overall, nipa sap as hydrolyzed with invertase was efficiently fermented to acetic acid, which is a valuable chemical and a potential biorefinery intermediate. Keywords: nipa sap, fed-batch fermentation, Moorella thermoacetica, acetic acid, biorefinery.

Acetic acid is an important chemical widely used ments such as coastal areas, river estuaries and as a solvent [1], food [2], or as a precursor for the mangrove forests with little care and investment synthesis of paints, plastics, fibers [3], deicers [4] and [11,12]. By tapping the stalk of nipa palm, sugar-rich pharmaceutical products [5]. Recently, acetic acid has sap can be obtained daily with high yield of around been explored for biodiesel [6] and for bioethanol 1.3 L/(palm/day) and average sugar content of 16.4% production after its hydrogenation over a metal [13]. Nipa palm can provide sap all year long and its catalyst [7-10]. Due to these increasing interests in annual productivity is estimated to be 169,000 biorefineries, acetic acid production through biological L/(ha/year) [13], equivalent to 181 t/(ha/year) [14], which processes has received considerable attention. Cost- is higher than sugarcane juice, 70 t/(ha/year) [15]. effective acetic acid production necessitates As reported by Tamunaidu and Saka [14], competitive and available feedstocks. chemical components of nipa sap include sucrose, Nipa (Nypa fruticans) is an abundant palm spe- glucose, fructose, and minor amounts of organic and cies that can naturally grow in brackish water environ- inorganic compounds. Although these constituents can be used for various purposes, the sap is collected

limitedly by local communities for their traditional use Correspondence: S. Saka, Department of Socio-Environmental Energy Science, Graduate School of Energy Science, Kyoto without any industrial applications [12,16]. The under- University, Yoshida-honmachi, Sakyo-ku, Kyoto 606-8501, Japan. utilized nipa sap could, thus, be a good raw material E-mail: [email protected] for acetic acid production. Paper received: 3 January, 2017 Paper revised: 26 January, 2017 Bio-derived acetic acid can be obtained from dif- Paper accepted: 29 January, 2017 ferent sugars through vinegar production using yeasts https://doi.org/10.2298/CICEQ170103003N and Acetobacters [17]. However, the conversion effi-

507 D. VAN NGUYEN et al.: FED-BATCH FERMENTATION OF NIPA SAP… Chem. Ind. Chem. Eng. Q. 23 (4) 507−514 (2017) ciency of sugars to acetic acid is as low as 27% [18] Hydrolysis of nipa sap due partly to CO2 emission [3] as can be seen in the Sucrose in nipa sap was hydrolyzed to glucose following equation: and fructose by 3.92 vol.% invertase solution at room

2CO2 2O2 temperature for 30 min. C H O 2C H OH 2CH COOH 6 12 6 2 5 3 Batch fermentation of standard sugars 2H2O Batch fermentations using glucose and fructose To avoid carbon loss during fermentation of nipa as carbon sources were first studied to investigate the sap, a novel process involving hydrolysis of the effects of high substrate concentration on free cells of sugars in nipa sap and subsequent fermentation with M. thermoacetica. For this purpose, a standard sugar Moorella thermoacetica has been developed [19]. The solution containing 100 g/L each of glucose and fruc- process does not release CO2 during fermentation tose was used. Fermentation was accomplished in and most carbon atoms in the sugar can be converted 500 mL DPC-2A fermenters. The composition and to acetic acid according to the following reaction preparation of nutrients in the fermentation medium [3,20]: are detailed by Rabemanolontsoa et al. [24]. The nutrients include 5 g/L yeast extract, 0.1 g/L cys- C6H12O6 3CH3COOH teine·HCl·H2O, 1 g/L (NH4)2SO4, 0.25 g/L MgSO4·7H2O, Initial application of this novel process on nipa 0.04 g/L Fe(NH4)2(SO4)2·6H2O, 0.24 mg/L NiCl2·6H2O, sap provided 9.9 g/L of acetic acid from 10.2 g/L 0.29 mg/L ZnSO4·7H2O, 0.017 mg/L Na2SeO3 and substrates, corresponding to 98% conversion effici- 1 mg/L resazurin. ency of the sugars to acetic acid [19]. To be more The fermenters as well as the substrate and effective for industrial-scale production, it is important nutrient solutions were sterilized in an autoclave at to improve the product concentration. For that matter, 121 °C for 20 min. Then, they were moved to a glove substrate concentration should be increased. How- box which provided an anaerobic environment by ever, many microorganisms can be inhibited by high filling with N2 gas. A working volume of 200 mL was substrate concentration [21,22], but evidence of such obtained by pouring 20 mL of M. thermoacetica inoc- effects on M. thermoacetica is scarce. Thus, the first ulum, 20-50 mL of standard sugar solution, adequate objective of this study is to investigate the effects of volumes of nutrient solution and water into each of high substrate concentration on batch fermentation by the fermenters. M. thermoacetica. Then, fed-batch technique was Fermentations were performed anaerobically at explored to improve acetic acid concentration from 60 °C with a stirring rate of 300 rpm. The pH was hydrolyzed nipa sap. controlled at 6.5±0.1 by automatic titration with oxygen-free 2 M NaOH. To correct for the acetic acid EXPERIMENTAL produced from the nutrients, inoculum and invertase, fermentations without substrates were accomplished Materials under the same conditions. Samples were collected Nipa sap was collected from Sarawak, Malaysia at specific time intervals and were immediately frozen and stored in concentrated and frozen state. To pre- at –31 °C until analysis with high-performance liquid pare the substrate for fermentation, 200 g of the con- chromatography (HPLC). centrated nipa sap was diluted with Milli-Q water to a total volume of 1 L in order to reach a sugar concen- Fed-batch fermentation of hydrolyzed nipa sap tration similar to that of the natural sap. Fed-batch fermentation was started with similar Invertase solution (EC No.3.2.1.26) with a mini- nutrient concentrations as in batch fermentation and mal activity of 4 units/mL was obtained from Wako approximately 20 g/L sugars from 28 mL hydrolyzed Pure Chemical Industries Ltd., Osaka, Japan. M. ther- nipa sap. In addition, oxygen-free feeding medium moacetica (formerly Clostridium thermoaceticum) was separately prepared by dissolving solid nutrients

ATCC 39073 was purchased from the American Type to hydrolyzed nipa sap under N2 atmosphere. Culture Collection (Manassas, VA, USA) in freeze- As total sugar concentration in the fermentation -dried state. The procedures for revival of the micro- broth decreased, the medium containing substrate organism and subsequent inoculum preparation were and nutrients was fed stepwise by a peristaltic pump. conducted as described by Nakamura et al. [23]. Total sugar concentration in the broth was increased by around 10 g/L after each feeding. The nutrient concentrations added for each feeding time, the ferment-

508 D. VAN NGUYEN et al.: FED-BATCH FERMENTATION OF NIPA SAP… Chem. Ind. Chem. Eng. Q. 23 (4) 507−514 (2017) ation temperature, stirring rate and pH control were pounds were detected. In fact, sucrose is regarded as similar to those for batch fermentation. Sampling was the primary sugar in original nipa sap [25]. During conducted just before and after feeding in order to collection and storage process, hexoses and lactic determine sugar and product concentrations. acid can be spontaneously formed by hydrolysis and fermentation [26]. Analyses

The concentrations of sucrose, glucose, fructose Table 1. Chemical composition of nipa sap before and after were determined by HPLC (Shimadzu, Kyoto, Japan) enzymatic hydrolysis; results are reported with standard devi- equipped with a RI detector and a Shodex sugar KS- ation: ≤ 0.2 g/L for sugars, ≤ 0.3 g/L for lactic acid, ≤ 0.4 g/L for 801 column (Showa Denko, Kanagawa, Japan). The ash and ≤ 1.3 g/L for total chemical compositions column was maintained at 80 °C and eluted with Milli- Concentration, g/L Chemical composition -Q water at a flow rate of 1 mL/min. The limit of det- Before hydrolysis After hydrolysis ection for this method is between 2.7 and 4.5 mg/L Sucrose 82.1 0.9 [24]. Ash content in the concentrated nipa sap was Glucose 29.8 72.2 determined by incineration at 600 °C for 2 h. Fructose 38.3 81.3 To analyze acid concentrations, an Aminex Lactic acid 1.6 1.6 HPX-87H column (Bio-rad, Hercules, CA, USA) was Ash 6.0 6.0 used at 45 °C with 5 mM aqueous H2SO4 solution as Total chemical 157.8 162.0 mobile phase, and the flow rate was set at 0.6 composition mL/min. Prior to injection into HPLC, samples were filtered through a 0.45 µm membrane to remove micro- In this process, after enzymatic treatment with organisms. The limit of detection of HPLC using this invertase, 99% of sucrose in nipa sap was hydro- column varied from 1.2 to 5.0 mg/L [24]. lyzed, increasing the concentrations of glucose and Growth of M. thermoacetica was determined by fructose to 72.2 and 81.3 g/L, respectively. The measuring the optical density (OD) using Shimazu obtained hydrolyzed nipa sap was used for acetic UV-Vis spectrophotometer (Kyoto, Japan) at 660 nm acid fermentation by M. thermoacetica. with water as reference. In order to evaluate the performance of the ferm- Batch fermentation entation techniques, conversion efficiency and acetic The capacity of batch fermentation to provide acid yield were respectively calculated according to high concentration of acetic acid was first explored Eqs. (1) and (2), while productivity was estimated with using glucose and fructose mixtures (1:1) as carbon Eq. (3): sources. Conversion efficiency () % = Figure 1a shows a comparison of acetic acid production from batch fermentations of 18.2, 28.9 and Acetic acid produced ()g /L (1) = 100 48.0 g/L of total sugar mixtures by M. thermoacetica. Total sugars added ()g /L In general, lag periods of about 24 h were observed, regardless of the substrate concentration. Acetic acid yield= Acetic acid production from 18.2 g/L sugars was Acetic acid produced () g = (2) faster than that from 28.9 and 48.0 g/L sugars Total sugars consumed () g between 24-72 h of fermentation. This might be explained by the fact that high sugar concentration can Acetic acid productivity = cause slow initial cell growth for M. thermoacetica Acetic acid produced ()g /L (3) = which might lead to low acetic acid production rate in Fermentation time (h) the first 30 h of fermentation, as demonstrated by Witjitra et al. [27]. RESULTS AND DISCUSSION At 96 h fermentation, acetic acid concentrations from fermentations of 18.2, 28.9 and 48.0 g/L sugars Chemical composition and hydrolysis of nipa sap were similar to be 15.9, 13.8, 16.1 g/L, respectively. Table 1 shows the chemical composition of nipa Then, the fermentation of 18.2 g/L sugars almost sap before and after hydrolysis by invertase. The raw stopped because no more substrate remained in the nipa sap contained 82.1 g/L of sucrose, 29.8 g/L of broth. In contrast, the fermentations of 28.9 and 48.0 glucose and 38.3 g/L fructose. In addition, 1.6 g/L of g/L continued for final acetic acid concentrations of lactic acid and 6.0 g/L of inorganics as minor com- 24.3 and 25.9 g/L, respectively.

509 D. VAN NGUYEN et al.: FED-BATCH FERMENTATION OF NIPA SAP… Chem. Ind. Chem. Eng. Q. 23 (4) 507−514 (2017)

reased. Many reports indicated that high concentration of the product might inhibit acetic acid production as well as cell growth [27-30]. Although Wang et al. [30] demonstrated that the maximum acetic acid tolerance for M. thermoacetica is around 48 g/L, other studies showed that acetic acid concentration as low as 10 g/L can cause a reduction in rates of both cell growth and acetic acid production [28]. Therefore, not only the substrate concentration, but also the formation of acetic acid in higher concentration could be a possible reason for the decrease in fermentation rate. Figure 1b compares the conversion efficiencies of the different sugar concentrations in function of time. Increasing substrate concentration sharply decreased conversion efficiency of sugars to acetic acid. This shows that high substrate concentration could not be effectively converted to acetic acid in batch fermentation. Fed-batch fermentation has potential to solve this problem. Choice of substrate concentration and feeding time for fed-batch fermentation Table 2 summarizes the results from batch fermentation of different substrate concentrations. Although 28.9 and 48.0 g/L provided higher acetic acid concentration than 18.2 g/L, their sugar con- Figure 1. Comparison of :a) sugar concentration (filled sumption could not reach 100%. In the fermentation symbols), acetic acid concentration (open symbols) and b) of 48.0 g/L sugars, 31% of the sugars still remained conversion efficiency during batch fermentations of glucose + after 240 h, 25.9 g/L of acetic acid was produced, fructose (1:1) in various concentrations by Moorella thermo- corresponding to a low conversion efficiency of 54%. acetica; square: 18.2 g/L; round: 28.9 g/L; triangle: 48.0 g/L of The conversion rate of sugars to acetic acid is total sugars (glucose + fructose). high with the lowest substrate concentration, as evidenced by the data in Figure 1a and b. In more detail, As shown in Figure 1a, acetic acid production similar conversion efficiencies of 90 and 84% were rates for both fermentations of 28.9 and 48.0 g/L dec- achieved respectively for 18.2 and 28.9 g/L after 168 h reased gradually when acetic acid concentration inc- of fermentation. However, high efficiency of 87% was

Table 2. Summary of batch and fed-batch fermentations of standard sugars and hydrolyzed nipa sap to acetic acid by Moorella thermoacetica; Results are reported with standard deviation ≤ 0.1 g/L for sugars and acetic acid

Substrate feeding, g/L Acetic acid Ferment– Time Sugar con- Substrate Concen– Average pro- Conversion Yield ation mode Initial Total h sumption, % tration, g/L ductivity, g/(L/h) efficiency, % g/g Hydrolyzed nipa sapa Batch 10.1 10.1 119 100 9.9 0.08 98 0.98 Glucose + fructose Batch 18.2 18.2 96 100 15.9 0.17 87 0.87 – – 168 100 16.3 - 90 0.90 28.9 28.9 168 97 24.3 0.15 84 0.87 48.0 48.0 168 56 21.9 - 46 0.81 – – 240 69 25.9 0.11 54 0.79 Hydrolyzed nipa sap Fed-batch 19.3 49.0 (low 408 100 42.3 0.10 86 0.86 feeding rate) 22.2 49.0 (high 240 100 42.6 0.18 87 0.87

feeding rate) aNguyen et al. [19]

510 D. VAN NGUYEN et al.: FED-BATCH FERMENTATION OF NIPA SAP… Chem. Ind. Chem. Eng. Q. 23 (4) 507−514 (2017) obtained much faster at 96 h for 18.2 g/L. Fur- one (named high feeding rate) would add the nutri- thermore, average productivities for the fermentations ents and substrates during the exponential and stat- of 18.2, 28.9 and 48.0 g/L sugars were 0.17, 0.15 and ionary phases between 24 and 72 h fermentation. 0.11 g/(L/h), respectively as seen in Table 2. Rapid These strategies were compared for fed-batch ferm- conversion and high productivity are desirable to red- entation of nipa sap using M. thermoacetica. uce energy consumption during fed-batch ferment- Fed-batch fermentation ation. For these reasons, sugar concentrations near 18.2 g/L would be ideal for fed-batch fermentation. Fed-batch fermentation of hydrolyzed nipa sap To determine the proper feeding time, batch fer- was performed at an initial substrate concentration of mentation of 18.2 g/L sugars was studied in detail. around 20 g/L. Total sugars after feeding were regul- Figure 2 shows the fermentation kinetics of 18.2 g/L ated to remain below this concentration. According to sugars. A lag phase was observed in the first 24 h Wang et al. [30], the maximum acetic acid tolerance when very little acetic acid was produced at low cell of M. thermoacetica was 0.8 M (48 g/L). Therefore, to density. In the next 24 h, M. thermoacetica grew consume all sugars and to achieve high conversion rapidly and reached a maximum optical density of 2.0. efficiency, total sugars were added until 50 g/L only. Subsequently, the cell density remained at maximum Figure 3 shows substrate consumption and the in the stationary growth phase between 48-72 h fer- obtained acetic acid concentration at (a) low and (b) mentation. Meanwhile, acetic acid was produced sig- high feeding rates during fed-batch fermentation of nificantly, with rapid substrate consumption. At 96 h, hydrolyzed nipa sap. For both experiments, the sec- no sugars remained and acetic acid concentration ond feeding was implemented at 72 h but in the case reached 15.9 g/L. These results demonstrated that of (a) low feeding rate, the third and fourth feedings fast acetic acid production occurred with increasing were conducted when most substrates were con- and high cell density. sumed with sugar concentration below 2.0 g/L.

Figure 2. Batch fermentation profile of 9.0 g/L glucose + 9.2 g/L fructose by Moorella thermoacetica.

From 72 h to 96 h, the optical density of cells in the death phase showed 2.5-fold decrease from 2.0 to 0.8 when 1.5 g/L of sugars remained in the fermenter. As discussed in several reports [27,31-33], the lack of substrates and nutrients can cause cell death. There- fore, both substrate and nutrients would be added to the fermentation medium to maintain the activity of microorganism in fed-batch fermentation. Based on these fermentation kinetics, two stra- Figure 3. Effect of feeding rate on fed-batch fermentation of tegies are applicable for the feeding time. The first hydrolyzed nipa sap by Moorella thermoacetica. one (qualified as low feeding rate) would wait for most Short arrows indicate the feeding times. of the substrates to be consumed, while the second

511 D. VAN NGUYEN et al.: FED-BATCH FERMENTATION OF NIPA SAP… Chem. Ind. Chem. Eng. Q. 23 (4) 507−514 (2017)

As shown in Figure 3a, total 49.0 g/L sugars in could remain during extended fermentation with many hydrolyzed nipa sap were fermented completely in feeding cycles. low feeding rate and acetic acid concentration reached 42.3 g/L at 408 h. This corresponds to a conversion efficiency of 86%. However, when little sugar remained in the fermentation medium from 120 to 144 h and from 216 to 264 h, substrate consumption rate decal- erated. Even after the third and fourth feedings, substrate consumption rate remained slow for nearly 24 h, resulting into a prolonged fermentation time of 408 h. To increase the speed of fed-batch fermentation, high feeding rate was explored. Figure 3b presents the results from fed-batch fermentation of hydrolyzed nipa sap with high feeding rate. The time intervals for the third and fourth feedings were reduced to be lower than 72 h. As a result, substrate consumption and acetic acid production maintained their high rates until total substrates were consumed completely at 240 h. From total of 49.0 g/L sugars in nipa sap, 42.6 g/L acetic acid was obtained with a conversion efficiency of 87%. This conversion efficiency could not reach 100% because part of the sugars might have been used for cell growth [27-29]. High feeding rate and low feeding rate showed similar final acetic acid concentration and conversion efficiency. However, fed-batch fermentation with high feeding rate reduced the total fermentation time and therefore improved 1.7-fold acetic acid productivity from 0.10 to 0.18 g/L/h, as compared to that with low Figure 4. Comparison of: a) acetic acid yield and b) acetic acid feeding rate. productivity of each feeding cycle in fed-batch fermentations of Comparison of fermentation performance during hydrolyzed nipa sap by Moorella thermoacetica with low and high feeding rates. each feeding cycle Figure 4 compares acetic acid yield (a) and ace- Figure 4b compares acetic acid productivities for tic acid productivity (b) of each feeding cycle in fed- each feeding cycle in fed-batch fermentations. High batch fermentations of hydrolyzed nipa sap. In the and low feeding rates showed similar productivities in first 3 feedings, the yield obtained with high feeding the first batch cycle. In later batch cycles, high feed- rate was lesser than the one with low feeding rate ing rate expressed much superior productivity due to because the microorganisms might not have enough its shorter fermentation time. Additionally, the lack of time to finish the metabolism until the end-product, substrates and nutrients during low feeding rate with and part of the consumed sugars might be in forms of extended fermentation time could have caused cell intracellular intermediate compounds. Nevertheless, death and reduced acetic acid production [27,31-33]. th in the 4 feeding, metabolism could have been com- Thus, the fast feeding of substrate and nutrients might pleted and high feeding rate slightly surpassed low have maintained viable cells and bioconversion of feeding rate in terms of yield. sugars into acetic acid. Either for low or high feeding rate, the yield was In general, both fed-batch fermentations with lower in the first feeding as compared with the sub- low and high feeding rates showed a similar trend of sequent ones. This may be due to the fact that the acetic acid productivities. Compared with the first microorganisms actively grew in the first feeding and batch, higher acetic acid productivities were obtained made more cells than in the next ones [34]. Acetic in the second batch because the first batch might acid yield for all feedings were above 0.68 g/g sugar. have had lag and exponential phases while the sec- This reveals that high acetic acid yield from nipa sap ond feeding might have been conducted in stationary phase with already high cell density. For this reason,

512 D. VAN NGUYEN et al.: FED-BATCH FERMENTATION OF NIPA SAP… Chem. Ind. Chem. Eng. Q. 23 (4) 507−514 (2017) acetic acid production occurred rapidly in the second REFERENCES batch. However, acetic acid productivities declined after the second feeding cycle. Since acetic acid can [1] N. Yoneda, S. Kusano, M. Yasui, P. Pujado, S. Wilcher, Appl. Catal. A: Gen. 221 (2001) 253-265 inhibit cell growth and acetic acid production [30,34], high acetic acid concentration accumulated from the [2] E. Lück, G.-W. von Rymon Lipinski, in Ullmann's Encyc- lopedia of Industrial Chemistry, Wiley-VCH Verlag GmbH fermentation might have caused lower acetic acid & Co. KGaA, Weinheim, 2000, pp. 671-692 productivities in the third and fourth feeding cycles. [3] M. Cheryan, in Encyclopedia of Microbiology, M. Comparison with traditional acetic acid production Schaechter, Ed., Academic Press, Oxford, 2009, pp. 144- –149 In the traditional vinegar production from nipa [4] D.F. Vitaliano, J. Pol. Anal. Manag 11 (1992) 397-418 sap in the Philippines, around 4.5–5.5% acetic acid [5] H. Cheung, R.S. Tanke, G.P. Torrence, in Ullmann's was produced from 15–22% nipa sap [18]. The pre- Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag sent method achieved a similar acetic acid concen- GmbH & Co. KGaA, Weinheim, 2000, pp. 209-238 tration of 42.6 g/L. In addition, this study showed 87% [6] V. Béligon, L. Poughon, G. Christophe, A. Lebert, C. Lar- conversion efficiency which is 3.2-fold higher than the roche, P. Fontanille, Bioresour. Technol. 192 (2015) 582- one from the traditional method to be 27% [18]. In –591 other words, from 1 kg of nipa sap, the current pro- [7] S. Saka, H. Miyafuji, Y. Kohara, H. Kawamoto, (Shiro cess can provide 115 mL of acetic acid while only 36 Saka), US Patent 8409832 (2013) mL can be obtained from the traditional vinegar pro- [8] S. Saka, H. Kawamoto, M. Hisashi, Y. Kazuyoshi, M. duction. Consequently, this technology is ultra-effi- Shozo, N. Yousuke, S. Yutaka, N. Kenichi, N. Phaibo- cient to convert sugars in nipa sap to acetic acid. onsilpa, T. Shigeo, (Shiro Saka), Japan Patent 2010239913 (2010) CONCLUSIONS [9] H. Kawamoto, T. Fujii, Y. Ito, S. Saka, J. Jpn. Inst. Energy 95 (2016) 162-166 In this study, the conversion of high substrate [10] Y. Ito, H. Kawamoto, S. Saka, Fuel 178 (2016) 118-123 concentrations to acetic acid by M. thermoacetica [11] L. Hamilton, D. Murphy, Econ. Bot. 42 (1988) 206-213 was explored and showed that increasing sugar con- [12] P. Tamunaidu, N. Matsui, Y. Okimori, S. Saka, Biomass centration caused low conversion efficiency in batch Bioenerg. 52 (2013) 96-102 fermentation. Therefore, fed-batch fermentation tech- [13] A.E.A. Päivöke, Agric., Ecosyst. Environ. 13 (1985) 59-72 nique was investigated. As a result, fed-batch ferm- [14] P. Tamunaidu, S. Saka, in Zero-Carbon Energy Kyoto entation of total 49.0 g/L sugars in nipa sap could 2011, Springer, 2012, pp. 121-126 produce 42.6 g/L of acetic acid, which is similar to the [15] D. Rajagopal, S.E. Sexton, D. Roland-Holst, D. Zilber- vinegar production process but with much higher ace- man, Environ. Res. Lett. 2 (2007) 044004 tic acid yield of 0.87 g/g sugar. Thus, fed-batch ferm- [16] S. Udofia, E. Udo, Global J. Agric. Sci. 4 (2005) 33-40 entation of the underutilized nipa sap by M. thermo- [17] A. Paivoke, M.R. Adams, D.R. Twiddy, Process Biochem. acetica may be a promising route for efficient acetic 19 (1984) 84-87 acid production. [18] P. Lim-Castillo, in Authenticity in the Kitchen: Proceed- ings of the Oxford Symposium on Food and Cookery, R. Acknowledgments Hosking Ed., Oxford Symposium, UK, 2006, pp. 295-306 This work was supported by the Japan Science [19] D.V. Nguyen, P. Sethapokin, H. Rabemanolontsoa, E. Minami, H. Kawamoto, S. Saka, Int. J. Green Technol. 2 and Technology Agency (JST) under the Advanced (2016) 1-12 Low Carbon Technology Research and Development [20] H.L. Drake, A.S. Gößner, S.L. Daniel, Ann. N.Y. Acad. Program (ALCA), for which the authors are extremely Sci. 1125 (2008) 100-128 grateful. The first author would like to acknowledge [21] M.C. Reed, A. Lieb, H.F. Nijhout, Bioessays 32 (2010) the financial support received from JICA under the 422-429 AUN/SEED-Net Project for his PhD study. Further- [22] J. Hong, Biotechnol. Bioeng. 28 (1986) 1421-1431 more, the authors are thankful to Dr. Pramila Tamu- [23] Y. Nakamura, H. Miyafuji, H. Kawamoto, S. Saka, J. naidu from Malaysia-Japan International Institute of Wood Sci. 57 (2011) 331-337 Technology, Universiti Teknologi Malaysia for pro- [24] H. Rabemanolontsoa, K. Yoshimizu, S. Saka, J. Chem. viding the nipa sap samples used in this study. Technol. Biotechnol. 91 (2016) 1040-1047 [25] J. Van Die, P.M.L. Tammes, in Transport in Plants I: Phloem Transport, M.H. Zimmermann, J.A. Milburn Eds., Springer-Verlag, Berlin, 1975, pp. 196-222

513 D. VAN NGUYEN et al.: FED-BATCH FERMENTATION OF NIPA SAP… Chem. Ind. Chem. Eng. Q. 23 (4) 507−514 (2017)

[26] J. Santiago-Urbina, F. Ruíz-Terán, Int. Food Res. J. 21 [31] M.M. Shah, M. Cheryan, J. Ind. Microbiol. 15 (1995) 424- (2014) 1261-1269 –428 [27] K. Witjitra, M.M. Shah, M. Cheryan, Enzyme Microb. [32] S. Parekh, M. Cheryan, Biotechnol. Lett 12 (1990) 861- Technol. 19 (1996) 322-327 –864 [28] J.E. Brownell, J.P. Nakas, J. Ind. Microbiol. 7 (1991) 1-6 [33] S.R. Parekh, M. Cheryan, Biotechnol. Lett 16 (1994) 139- [29] K. Sugaya, D. Tusé, J.L. Jones, Biotechnol. Bioeng. 28 –142 (1986) 678-683 [34] M.M. Shah, M. Cheryan, Appl. Biochem. Biotechnol. 51 [30] G. Wang, D.I.C. Wang, Appl. Environ. Microbiol. 47 (1995) 413-422. (1984) 294-298

DUNG VAN NGUYEN DOLIVNA FERMENTACIJA SOKA NIPA PALME U HARIFARA SIRĆETNU KISELINU POMOĆU Moorella RABEMANOLONTSOA thermoacetica (f. Clostridium thermoaceticum) SHIRO SAKA

Department of Socio-Environ- Razvijen je efikasan proces za konverziju soka nipa palme u sirćetnu kiselinu. Ovaj biljni mental Energy Science, Graduate sok je hidrolizovan invertazom do glukoze i fruktozu. Šaržna fermentacija glukoze i fruk- School of Energy Science, Kyoto toze bila je neadekvatna pri povećanoj koncentraciji supstrata. Nasuprot tome, dolivna University, Yoshida-honmachi, fermentacija hidrolizovanog biljnog soka sa velikom brzinom dolivanja drastično je pove- Sakyo-ku, Kyoto, Japan ćala koncentraciju sirćetne kiseline i produktivnost na 42,6 g/l i 0,18 g/(l h), redom. Svi šećeri hidrolizovanog biljnog soka su potrošeni, sa prinosom sirćetne kiseline od 0,87 g/(g šećera). Sve u svemu, hidrolizovani sok nipa palme efikasno je fermentisan u sir- NAUČNI RAD ćetnu kiselinu, koja je vredna hemikalija i potencijalni biorafinerijski intermedijer.

Ključne reči: sok nipa palme, dolivna fermentacija, Moorella thermoacetica, sirćetna kiselina, biorefinerija.

514 Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 515−521 (2017) CI&CEQ

SORAYA SEANKHAM1 KINETICS AND ADSORPTION ISOTHERM OF SENAD NOVALIN2 1 LACTIC ACID FROM FERMENTATION BROTH SUWATTANA PRUKSASRI ONTO ACTIVATED CHARCOAL 1Department of Biotechnology, Faculty of Engineering and Article Highlights Industrial Technology, Silpakorn • Activated charcoal was used to remove lactic acid from fermentation broth University, Nakhon Pathom, • The isotherm was best fitted with the Freundlich isotherm Thailand • The kinetic data were described well by pseudo-second-order model 2Department of Food Science and • Maximum adsorption capacity of 57.06 mg/g was obtained Technology, University of Natural Resources and Life Sciences, Abstract Vienna, Austria Activated charcoal was applied for the recovery of lactic acid in undissociated form from fermentation broth. Lactic acid was obtained from the fermentation of SCIENTIFIC PAPER Lactobacillus casei TISTR 1340 using acid hydrolyzed Jerusalem artichoke as a carbon source. The equilibrium adsorption isotherm and kinetics for the lactic UDC 661.183.2:66.081.3:544:547.472.3 acid separation were investigated. The experimental data for lactic acid adsorption from fermentation broth were best described by the Freundlich isotherm and the pseudo-second order kinetics with R2 values of 0.99. The initial adsorption rate was 41.32 mg/g⋅min at the initial lactic acid concentration of 40 g/L. Keywords: lactic acid, activated charcoal, adsorption kinetics.

Lactic acid is an important organic acid that is mentation broth. Currently, many different adsorbents widely used in food, pharmaceutical and cosmetics such as Amberlite XAD1600, activated carbon, ads- industries. It can be produced by both chemical syn- orption resin AX-1 and Amberlite IRA-67 have been thesis and fermentation. Recently, many researchers investigated for lactic acid removal from fermentation have focused on lactic acid recovery from ferment- broth [1-6]. Therefore, activated charcoal is an inter- ation broth by calcium lactate precipitation. Its econ- esting resin for lactic acid adsorption due to its good omical cost is about 50% of the total capital cost but adsorbent characteristics, high porous and low cost. calcium sulphate was occurred as the waste from this The objectives of this work were to study the adsorpt- recovery process. So, the separation technologies for ion of lactic acid from fermentation broth on activated lactic acid separations such as solvent extraction, charcoal and to determine the adsorption isotherm electrodialysis, ion exchange chromatography and and adsorption kinetics. membrane separation etc. have been studied. These processes, however, have many disadvantages such MATERIALS AND METHODS as high purification cost, the more complex procedure, the usage of organic solvent, the fouling prob- Chemicals lem in case of membrane process and the short life- Activated charcoal with particle sizes in the time of expensive resins. As a result, adsorption tech- range of 20-60 mesh (0.84-2.4 mm) was purchased nique is of interest for lactic acid separation from fer- from Sigma-Aldrich and used in this study. Bacterial strains Correspondence: S. Pruksasri, Department of Biotechnology, Faculty of Engineering and Industrial Technology, Silpakorn Lactobacillus casei TISTR 1340 was purchased University, Nakhon Pathom 73000 Thailand. from The Thailand Institute of Scientific and Techno- E-mail: [email protected] logical Research, Bangkok, Thailand. Stock cultures Paper received: 11 May, 2016 ° Paper revised: 9 December, 2016 were maintained in MRS agar plates at 4 C and react- Paper accepted: 1 February, 2017 ivated every month. Cells were cultured in MRS broth https://doi.org/10.2298/CICEQ160511004S

515 S. SEANKHAM et al.: KINETICS AND ADSORPTION ISOTHERM OF LACTIC… Chem. Ind. Chem. Eng. Q. 23 (4) 515−521 (2017) at 37 °C for 18 h under static and anaerobic condit- The equilibrium data obtained from batch ads- ions. orption of lactic acid can be applied to predict the adsorption models and related theories, the Langmuir Lactic acid fermentation broth for adsorption and Freundlich isotherms were compared in this To prepare a fermentation broth for adsorption, work. The adsorption isotherms of lactic acid in the batch lactic acid fermentation was performed in a 5 L fermentation broth and in the pure lactic acid solution bioreactor (Biostat, B. Braun, Germany) using Jeru- at the concentration of 10 g/L were also evaluated salem artichoke hydrolysate as a carbon source. The because of the high adsorption efficiency of lactic acid Jerusalem artichoke hydrolysate was obtained from at this concentration. The pseudo-first order and the hydrolysis of Jerusalem artichoke tubers with 0.1 pseudo-second order equations were applied to model M H2SO4 at a ratio of tubers to conc. sulfuric acid of the kinetics of lactic acid adsorption onto activated 1:2 at 80 °C for 6 h. Then, the mixture was adjusted to charcoal. pH 6.5 with 4 M NaOH. The hydrolysate was passed through cloth bags to remove solid particles and ster- Analytical methods ilized at 121 °C for 15 min. The hydrolysate was Samples were taken periodically and centri- mixed with MRS medium and the L. casei TISTR fuged at 7000 rpm for 10 min to remove any particul- 1340 inoculum (10 vol.%) was added. Temperature ate solids. The supernatants were used for the ana- and agitation speed were controlled at 37 °C and 150 lysis of lactic acid and other components in the fer- rpm, respectively. The pH of the culture was kept at mentation broth. The lactic acid, organic acid and

6.0 by automatic addition of 25% Ca(OH)2. At the end sugar concentrations were measured with an HPLC of fermentation, cells were removed by centrifugation with a RI detector (Shimadzu, Japan). The separation at 4000 rpm for 10 min. The fermentation broth con- was performed on an Aminex HPX-87H column (Bio- tained approximately 40.6 g/L lactic acid, 1.05 g/L -Rad, CA) at 55 °C and the flow rate of 0.005 M sul- sucrose and 4.10 g/L acetic acid. furic acid as a mobile phase was set at 0.5 mL/min. Adsorption of lactic acid on activated charcoal RESULTS AND DISCUSSION Batch adsorption of lactic acid. The experiment was carried out in a 250 mL Erlenmeyer flask with Adsorption efficiency either 50 mL of fermentation broth or pure lactic acid Figure 1 shows the adsorption efficiency of acti- solution containing different initial lactic acid concen- vated charcoal for the removal of lactic acid from fer- trations in the range of 10-40 g/L. Then, 5 g of acti- mentation broth at 25 °C and pH 2. As can be seen, vated charcoal was added and incubated at 100 rpm, the maximum acid removal (∼45%) was obtained at ° ∼ 25 C until equilibrium ( 8 h). Samples were taken an initial concentration of 10 g/L, the increase of the periodically to determine the amount of adsorbed fer- lactic acid concentration from 10 to 40 g/L decreased mentation broth components at equilibrium. Lactic the adsorption efficiency from 45 to 20%. Therefore, a acid concentrations were determined by HPLC and comparison between pure lactic acid and lactic acid the amount of adsorbed lactic acid per g of activated from fermentation broth at 10 g/L was chosen and charcoal was calculated using Eq. (1): selected for further study on the adsorption isotherms ()CCV− and kinetics. The similar results were also presented q = 0e (1) e W where qe is the amount of lactic acid adsorbed by the adsorbent (g/L), Co is the initial lactic acid concentration (g/L), Ce is the lactic acid concentration at equilibrium (g/L), V is the initial solution volume (L) and W is the weight of activated charcoal. The adsorption efficiency of lactic acid was determined using Eq. (2): CC− Adsorption efficiency= 100 0f (2) C0 where Co and Cf are the initial and final concentration Figure 1. Adsorption efficiency of activated charcoal on the of lactic acid in solution (g/L), respectively. removal of lactic acid from fermentation broth under different initial lactic acid concentrations.

516 S. SEANKHAM et al.: KINETICS AND ADSORPTION ISOTHERM OF LACTIC… Chem. Ind. Chem. Eng. Q. 23 (4) 515−521 (2017) by Yousuf et al. [7] that the percentage of acid rem- The Langmuir and Freundlich isotherms were oval using activated carbon as an adsorbent was studied to find the equilibrium characteristics of ads- approximately 37% at pH 3. orption. The Langmuir adsorption model is based on the assumption that the solute molecules adsorbed Adsorption isotherms on the adsorbent surface in a monolayer without any The adsorption of lactic acid was performed at interactions between the adsorbed molecules. The pH 2 due to its undisscociated form at the pH lower linear form of Langmuir equation can be expressed by than the pKa value (pKa of lactic acid = 3.78). Even Eq. (3): though the physical properties of the used activated 1111 charcoal were not estimated, these physical proper- =+ (3) ties can be obtained from the work of Pyrzyńska and q eLmemKq C q Bystrzejewski [8] which used the same type of com- where q is the amount of lactic acid adsorbed at mercial activated charcoal. As reported in their work, e equilibrium (g/g ), q is the maximum adsorption the pH of the point of zero charge of this commercial AC m capacity (g/g ), C is lactic acid concentration at adsorbent was 4.44 with the methylene blue-relative AC e equilibrium (g/L) and K is equilibrium constant of the surface area of 118.7 m2/g. Therefore, the possibility L Langmuir constant (L/g). The values of q and K of using activated carbon for the adsorption of lactic m L were determined from the intercept and slope, res- acid from fermentation broth in a low pH region was pectively, and are presented in Table 1. studied and presented in Figure 2. The adsorption To determine whether the adsorption is favor- capacity of lactic acid from the fermentation broth by able or unfavorable, a dimensionless constant separ- activated charcoal was 57 mg/g adsorbent at an initial ation factor (R ) is presented and can be expressed lactic acid concentration of 40 g/L. As can be seen, L by Eq. (4): lactic acid is adsorbed on activated carbon whereas other components such as sugars still remained mainly 1 R = (4) L + in the solution which agrees with the previous work of 1 KCL0 Thang and Novalin [1]. where KL is the Langmuir constant and C0 is the initial

concentration of adsorbate. The RL value presents the

nature of adsorption as irreversible (RL = 0), favorable

(0 < RL < 1), linear (RL = 1) or unfavorable (RL > 1). In

this study, the calculated RL values were between 0 and 1 as listed in Table 1, indicating that the adsorption characteristic of lactic acid in fermentation broth on activated charcoal is favorable. The Freundlich isotherm is an empirical equation for heterogeneous systems and was used in this study as a second isotherm and the linear form of this model is expressed by Eq. (5): 1 logq =+logKClog (5) efn e

where Kf is a Freundlich constant, n indicates the Figure 2. Adsorption capacity of lactic acid, acetic acid and effect of concentration on the adsorption capacity and sucrose at equilibrium at different lactic acid concentrations. represents adsorption intensity. The values of Kf and

Table 1. The parameters of Langmuir and Freundlich isotherm of lactic acid adsorption

Lactic acid concentration, g/L Langmuir isotherm Freundlich isotherm –1 –1 2 –1 2 qm / g gAC KL / L g RL R Kf / L g n R 10 0.018 0.329 0.239 0.905 0.37 0.770 0.975 20 0.021 0.116 0.302 0.863 12.04 0.491 0.965 30 0.014 0.042 0.446 0.971 31282 0.241 0.993 40 0.011 0.008 0.766 0.962 331589 0.229 0.991 10 (pure) 0.008 0.250 0.320 0.905 1.28 0.455 0.975

517 S. SEANKHAM et al.: KINETICS AND ADSORPTION ISOTHERM OF LACTIC… Chem. Ind. Chem. Eng. Q. 23 (4) 515−521 (2017) n were determined from the intercept and slope, res- k log()q −=qq log −1 t (6) pectively, as shown in Table 1. eet 2.303 From Table 1, in case of Langmuir isotherm, the where q and q are the amount of lactic acid ads- value of q and K decreased with increasing initial e t m L orbed at equilibrium and time t (mg/g), respectively, lactic acid concentration indicates that the lactic acid and k is the rate constant of the first-order adsorption was adsorbed on the activated charcoal at lower con- 1 (min-1). centration. The Freundlich isotherm, the K value indi- f The plots of log (q - q ) versus t were used to cates the adsorption capacity of the adsorbent or the e t determine the rate constant (k ), q and correlation bonding energy which increased when increasing ini- 1 e coefficients (R2) for different lactic acid concentrations tial lactic acid concentration. Whereas n is the ads- as shown in Fig. 3. orption intensity, smaller n values at high lactic acid concentration implies that the adsorption is better with Pseudo-second order increasing lactic acid concentration [9]. The com- The pseudo-second order kinetic may be pre- parison between n value of pure lactic acid solution sented in a linear form as in Eq. (7): and the fermentation broth indicates that the lactic t 11 acid is adsorbed at higher intensity than that from fer- =+t (7) q 2 q mentation broth. However, as seen in Table 1, n value t kq2e e was not high enough for separation (n < 1). The coef- where k2 is the rate constant of second-order adsorp- ficient determination (R2) values in the Freundlich iso- tion (g/ mg⋅min) and qe and qt is amount of lactic acid therm of lactic acid adsorption were higher than those adsorbed at equilibrium and time t (mg/g). in the Langmuir isotherm. As a result, the Freundlich The plot of t/qt versus t should give a straight isotherm is more suitable to describe the adsorption line, and qe and k2 can be determined from the slope of lactic acid adsorption. and intercept, respectively, as shown in Fig. 4. Adsorption kinetics The pseudo-second order kinetic analysis presented that the values of the initial adsorption rates (h In this study, adsorption data are applied to the in the unit of mg/(g⋅min)) increased with increasing pseudo-first order, pseudo second-order kinetic models initial lactic acid concentrations as shown in Table 2. to find the rate constant of adsorption. It could be explained by the fact of the increase in Pseudo-first-order driving force at high lactic acid concentrations which A linear form of pseudo-first order model was is also in agreement with the previous literature work described by Lagergren as shown in Eq. (6): [10-13]. The equilibrium adsorption capacity (qe) also

Figure 3.The pseudo-first order kinetic plots for the adsorption of lactic acid on activated charcoal. At conditions: 100 rpm, pH 2 and temperature 25 °C. Different initial lactic acid concentrations: a) 10, b) 20, c) 30 and d) 10 g/L lactic acid solution.

518 S. SEANKHAM et al.: KINETICS AND ADSORPTION ISOTHERM OF LACTIC… Chem. Ind. Chem. Eng. Q. 23 (4) 515−521 (2017)

Figure 4. The pseudo-second order kinetics plots for the adsorption of lactic acid on activated charcoal. Different initial lactic acid concentrations at conditions: 100 rpm, pH 2 and temperature 25 °C.

Table 2. Comparison of pseudo-first order and pseudo-second order rate constants, normalized standard deviation (Δq) and correlation coefficients (R2) at different initial lactic acid concentrations

Pseudo-first order Pseudo-second order Lactic acid qe,exp Δq Δq –1 k1 qe,cal 2 h k2 qe,cal 2 concentration, g/L mg g -1 –1 R % –1 –1 –1 –1 –1 R % min mg g mg g min g mg ⋅min mg g 10 37.38 0.0044 6.104 0.434 83.67 3.758 0.0025 38.61 0.990 3.29 20 44.52 0.0042 7.12 0.464 84.01 9.124 0.0045 44.84 0.969 0.72 30 47.22 0.0001 4.571 0.0002 90.32 11.69 0.0052 47.39 0.984 0.36 40 57.06 - - - - 41.32 0.0093 66.67 0.997 16.84 10(pure) 22.99 0.0021 4.939 0.303 78.52 4.308 0.0052 28.65 0.986 24.62

increased with increasing initial lactic acid concentra- pared to the pseudo-first order (Fig. 3). The pseudo- tions. This may due to the large number of adsorbed -second order rate constants were in the range of lactic acid at the available adsorption sites. 0.0025 to 0.0093 g/(mg min). The theoretical values

The pseudo-first order and pseudo-second order of qe also agree very well with the experimental ones. rate constants are presented in Table 2 along with the These suggest that the sorption process could be corresponding correlation coefficients (R2) and the described using the pseudo second-order kinetic normalized standard deviation (Δq), which is model with higher correlation coefficients and lower calculated by the following equation: normalized standard deviation values than the first-

order kinetic model. The parameters k1, k2 and h from ()/qqq− 2 Δ=q exp cal exp (8) both model indicates the adsorption rate of activated N −1 charcoal that can be applied for further study. Inter- estingly, the q value of the pure lactic acid solution where the subscripts exp and cal are the experi- e was around 50% less than that of the fermentation mental and calculated data, respectively and N is the broth. This could be explained by the differences in number of data points [3]. From Table 2, it was obs- the chemical potentials and the activity coefficients of erved that the pseudo first-order model did not fit well the fermentation broth which may higher than those in because the calculated q values did not agree with e the pure solution due to the amount of other com- the experimental q values and the Δq was quite high. e ponents in the solution. That would lead to a higher This suggests that the adsorption of lactic acid did not chemical potentials and the adsorption capacity. follow the first-order kinetics. The pseudo-second order model (Fig. 4) gave good straight lines as com-

519 S. SEANKHAM et al.: KINETICS AND ADSORPTION ISOTHERM OF LACTIC… Chem. Ind. Chem. Eng. Q. 23 (4) 515−521 (2017)

In this present case, activated charcoal was more than the pseudo-first order kinetic model. The used as an adsorbent for the adsorption of low con- maximum adsorption capacity of 57 mg/g adsorbent centration of lactic acid from fermentation broth. The was achieved at pH 2. adsorption capacity of 37.4 mg/g activated charcoal Acknowledgement was obtained at pH 2, with an initial lactic acid concentration of 10 g/L in the fermentation broth. This This work was financially supported by The result was better than that reported by Yousuf et al. Research and Creativity Fund, Department of Bio- [7] which obtained the maximum adsorption capacity technology, Faculty of Engineering and Industrial of 18.63 mg lactic acid/g activated carbon from the Technology, Silpakorn University. mixture of acid system with an initial lactic acid concentration of 11.66 g/L, at pH 2, temperature 298 K REFERENCES and 5 g of adsorbent. This work presents a simple [1] V.H. Thang, S. Novalin, Bioresour. Technol. 99 (2008) process to recover lactic acid from the fermentation 4368-4379 broth at lower concentration and low pH value. The [2] S.a.S. Bayazit, I. Inci, H. Uslu, J. Chem. Eng. Data 56 efficiency of desorption of lactic acid is also another (2011) 1751-1754 important factor that needs to be considered. How- [3] J. Wu, Y. Hu, J. Zhou, W. Qian, X. Lin, Y. Chen, X. Chen, ever, this is outside of the scope of this study. It has J. Xie, J. Bai, H. Ying, Colloids Surfaces, A 407 (2012) been reported that acetone is an efficient solvent for 29-37 desorption of lactic acid from activated charcoal [14]. [4] X. Cao, H.S. Yun, Y.-M. Koo, Biochem. Eng. J. 11 (2002) With the ratio of acetone to activated charcoal at 4, 189-196 the recovery efficiency of lactic acid at around 80% [5] W.-Y. Tong, X.-Y. Fu, S.-M. Lee, J. Yu, J.-W. Liu, D.-Z. could be achieved. The acetone in lactic acid solution Wei, Y.-M. Koo, Biochem. Eng. J. 18 (2004) 89-96 can be subsequently removed under vacuum. In addi- [6] C.-C. Chen, L.-K. Ju, Sep. Sci. Technol. 33 (1998) 1423- tion, a long-term performance of the activated char- –1437 coal is another important factor that needs to be eval- [7] A. Yousuf, F. Bonk, J.-R. Bastidas-Oyanedel, J.E. uated in the further study. Schmidt, Bioresour. Technol. 217 (2016) 137-140 [8] K. Pyrzyńska, M. Bystrzejewski, Colloids Surfaces, A 362 CONCLUSIONS (2010) 102-109 [9] S. Bağcı, A.A. Ceyhan, Chem. Ind. & Chem. Eng. Q. 22 In this work, a possible approach to easily sep- (2016) 155-165 arate lactic acid from the fermentation broth in the [10] G. Crini, Bioresour. Technol. 97 (2006) 1061-1085 form of undissociated acids was presented. Lactic [11] F. Ferrero, J. Hazard. Mater. 142 (2007) 144-152 acid was much more adsorbed on to the adsorbent [12] I.A.W. Tan, B.H. Hameed, A.L. Ahmad, Chem. Eng. J. than other components such as acetic acid and suc- 127 (2007) 111-119 rose in the fermentation broth. Adsorption isotherm [13] B.H. Hameed, D.K. Mahmoud, A.L. Ahmad, J. Hazard. data of lactic acid could be well explained by the Mater. 158 (2008) 65-72 Freundlich model. The kinetic experimental data cor- [14] M.-T. Gao, T. Shimamura, N. Ishida, H. Takahashi, related well with the pseudo-second order kinetic Enzyme Microb. Technol. 48 (2011) 526-530.

520 S. SEANKHAM et al.: KINETICS AND ADSORPTION ISOTHERM OF LACTIC… Chem. Ind. Chem. Eng. Q. 23 (4) 515−521 (2017)

SORAYA SEANKHAM1 KINETIKA I RAVNOTEŽA ADSORPCIJE MLEČNE SENAD NOVALIN2 KISELINE IZ FERMENTACIONE TEČNOSTI NA 1 SUWATTANA PRUKSASRI AKTIVNOM UGLJU 1Department of Biotechnology, Faculty of Engineering and Industrial Aktivni ugalj je primenjen za izdvajanje mlečne kiseline u nedisociranom obliku iz fer- Technology, Silpakorn University, mentacione tečnosti. Mlečna kiselina je dobijena fermentacijom kiselog hidrolizata Jeru- Nakhon Pathom, Thailand salim artičoke, koji je izvor ugljenika, korišćenjem Lactobacillus casei TISTR1340. Anali- 2Department of Food Science and zirana je ravnotežna adsorpciona izoterma i kinetika izdvajanja mlečne kiseline. Eks- Technology, University of Natural perimentalni podaci za adsorpciju mlečne kiseline iz fermentacione tečnosti najbolje su Resources and Life Sciences, Vienna, opisani Frojndlihovom izotermom, a kinetika modelom pseudo-drugog reda sa vred- Austria nostima R2 od 0,99. Početna brzina adsorpcije je bila 41,32 mg/(g·min) pri početnoj kon- NAUČNI RAD centraciji mlečne kiseline od 40 g/L.

Ključne reči: mlečna kiselina, aktivni ugalj, kinetika adsorpcije.

521

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 523−527 (2017) CI&CEQ

WANG RAN TOTAL SPRAY TRAY (TST) FOR TAO JINLIANG DISTILLATION COLUMNS: A NEW LING YANLI GENERATION TRAY WITH LOWER WEI FENG LIU JIDONG PRESSURE DROP

School of chemical engineering, Article Highlights Hebei University of Technology, • The pressure drop of the tray is relatively lower than other type trays Tianjin, China • The plate has no independent downcomer • There is no flooding with TST SCIENTIFIC PAPER Abstract UDC 66.048.1:532.5:66 As a critical gas-liquid contact component, trays play an important role in the performance of distillation column. In this paper, the hydrodynamics of a new type of tray, the total spray tray (TST), were investigated. Taking air-water as the

working medium, the pressure drop with different F0 was tested under atmospheric pressure. The results indicate that there is no flooding with the tray and thus the wider operation flexibility can be provided by TST. Furthermore, the pressure drop of TST is much lower than that of other trays. The above properties of TST would benefit the production expansion and energy decreasing in distillation operation greatly. Keywords: distillation, tray, pressure drop, spray.

The plate distillation column is a typical type of However, regardless of the tray type, the downcomer, equipment that is widely used in chemical industries. as a separate part, is necessary because it is only Compared with packed columns, the trays have the through the downcomer that the upper plate liquid can advantages of low cost, easy installation and main- drop on the sub-plate. Hence, the flooding of the tenance, which made them more attractive in chem- downcomer under high vapor flux has become the ical, petroleum and energy industries. However, with bottleneck in the revamp for capacity increase. More- advancements in industrialization, new technologies, over, the specifications and layout of the downcomer processes and methods have been added to distil- also affect the plate efficiency, as well as the type and lation. Traditional trays may not meet the needs of the configuration of trays. For example, multi-down- energy saving and industrial scale-up. Thus, obtaining comers are often employed by large diameter col- a column tray with high capacity and low pressure umns to decrease liquid back-mixing and improve drop has become one of the main objectives for distil- mass transfer driving force. In some degree, the lation development. Thus, many types of trays have reasonable designs of downcomers may have the been produced, such as SNT [1], BVT [2], CTST-FI same key role with mass transfer element. Unfortun- [3], Triton [4], Con-Sep [5], Swirl tube [6], DJ [7], ately, the downcomer, as the passage of liquid, gen- CTST [8], conical cap [9], cascade [10], JCPT [11], erally occupies a relatively small space versus the FGS-VT [12], Simul Flow [13], and many other trays. mass transfer element which the vapor passes through [14]. Why don’t we combine the downcomer and mass transfer element together? In doing so, the pas- Correspondence: J. Tao, School of chemical engineering, Hebei sage of both liquid and vapor are enlarged at the University of Technology, 300130, Tianjin, China. E-mail: [email protected] same time. In this paper, a new type tray named total Paper received: 29 October, 2016 spray tray (TST) is presented by setting an injection Paper revised: 17 January, 2017 tube throughout two adjacent plates. Using this tray, Paper accepted: 6 February, 2017 no flooding will occur because the vapor cannot https://doi.org/10.2298/CICEQ160829005R

523 W. RAN et al.: TOTAL SPRAY TRAY FOR DISTILLATION COLUMNS Chem. Ind. Chem. Eng. Q. 23 (4) 523−527 (2017) occupy the passages of liquid. More interestingly, as within the column, an organic glass column with 400 the process of liquid of upper plate flowing to the sub- mm inner diameter was used. In the experiment, the -plate does not consume the energy of the vapor, the injection tube (10) with 90 mm inner diameter was pressure drop of the tray is lower than other type placed in the setup with 300 mm tray spacing. On the trays. In this research, hydraulics experiments includ- side wall of the injection tube, 176 holes with 7 mm ing the flexibility and the pressure drop were done to diameter were opened. By contrast, a sieve tray evaluate the simple performance of TST, and its hyd- owned the same opening ratio (RP = 5%) to TST was rodynamics properties were discussed by comparison made in the experiment, where RP equals the total to traditional column trays. area of openings Ao divided by the cross-sectional

area of the column Ac. The constructions of sieve tray EXPERIMENTAL SETUPS and TST were illustrated by Figure 3, on which 231 holes with 6 mm diameter were opened. Structure of the TST The total spray tray consists of two parts: one is an injection tube with openings on its wall and the other is a liquid seal element. It is different from other trays the injection tube of TST is throughout two adjacent plates, as shown in Figure 1. When the device works, the overflowed liquid of the upper plate will drop down to the sub-plate directly under the action of gravity along the inner tube wall, and is then broken down into micro-droplets by the vapor coming from the sub-plate at the opening region, and finally splashed out of the injection tube with the vapor. When the vapor capacity changes, the clear liquid height of the upper plate will vary automatically due to the varied pressure drop. As a result, a new balance between liquid and vapor was achieved again. In Figure 2.Experimental setup for hydrodynamic experiment: other words, there no flooding occurs with TST. 1) blower, 2) frequency converter, 3) pitot tube, 4) weeping graduated cylinder, 5) weeping collection plate, 6) centrifugal pump, 7) sieve tray, 8) rotameter, 9) U-tube manometer, 10) TST, 11) entrainment collector, 12) entrainment graduated cylinder.

Figure 3. The construction of test trays: a) sieve and total spray trays; b) The injection tube of TST and its unfolder drawing.

When testing, water was pumped from the column bottom by a centrifugal pump (6) to the top of the column. The flow rate of water was monitored by a rotameter (8) next to the pump. The gas was supplied at the space between the sieve tray and the collection Figure 1. Operating principle of TST. plate for weeping by an air blower (1) controlled by a frequency converter (2). The flow rate of air was Hydrodynamics pilot plant quantified by a pitot tube (3) companied a digital anemometer. The pressure drop of the tested tray The hydrodynamic experiment setup is pre- was measured by two U-tube manometers (9) filled sented in Figure 2. To observe the phenomenon

524 W. RAN et al.: TOTAL SPRAY TRAY FOR DISTILLATION COLUMNS Chem. Ind. Chem. Eng. Q. 23 (4) 523−527 (2017) with water, which have an accuracy margin of 10 Pa. explanation for the liquid flow having little effect on The height of clear liquid can be read from the scales the pressure drop with TST. In other words, the dry labeled on the tower wall. pressure drop of TST (liquid flow = 0) is a significant constituent part of the total pressure drop. RESULTS AND DISCUSSION Comparison of pressure drop between TST and other Effect of liquid flow on pressure drop of TST trays Pressure drop represents the energy dissipated It is well known that sieve tray has traditionally caused by fluid flowing through the tray. Especially, in lower pressure drop. In the experiment, the pressure the days with the gradual scarcity of global energy drops of sieve tray were tested on the setup shown in resources, the lower pressure drop trays have attracted Figure 2, as well as the TST. Moreover, for learning lots attention. For TST, when the column works, the the properties of TST better, the pressure drop of resistance or the pressure loss is caused by two New VST, which is an earliest new vertical sieve tray parts. One is the holes, including the hole in the plate and its properties have been well studied [15] was and the holes on the injection tube, and the other is cited for comparison in Figure 5. From Figure 5, it can the liquid film. The pressure drop of caused by the be obtained that with increasing F0, the total pressure former can be denoted by the dry pressure and the drops of all trays increase although their trends are total pressure drop can be indicated by the wet pres- different: the one of sieve tray is the slowest, and the sure drop. one of the TST shows the sharpest rise. Moreover, The pressure drop under different liquid flow of under different ranges of F0, the pressure drops of three trays are different too. For example, when F0 is TST was shown in Figure 4, where F0 is the kinetic energy factor of gas based on the orifice gas velocity. less than 17, the TST has the smallest pressure drop, From Figure 4, it can be obtained that under a certain and when F0 is large than 17, the sieve tray has the liquid flow, the pressure drops increase rapidly with smallest pressure drop. the increasing of the F0. This may be understood well because pressure loss is directly proportional to the square of the vapor speed. However, with the increasing of the liquid flow, the pressure drop increase is low and about 100 Pa increase in pressure drop was found for every 1 m3 increase in liquid flow. This is because, for TST, the liquid of upper plate flowing to the sub-plate is done automatically under the action of gravity. When the liquid flow increases, the film thickness at the openings changes only slightly. For example, when the liquid flow changes from 1.2 to 2.4 m3/h, the film thickness increases about 1 mm. Thus, the energy loss for vapor is low. This gives a good

Figure 5. Pressure drop of different type trays.

Through the above analysis, we knew that the total pressure drop of a tray can be divided into two parts. One is the dry pressure produced by vapor passing the holes, and the other is the pressure loss caused by the liquid flowing. For sieve tray, the vapor passes through the hole only once. And for TST and New VST, it needs twice: One is the hole on the plate and the other is the holes of the injection. So, under

the higher F0, the pressure drop of TST and New VST must have the larger values than that of sieve tray.

However, when F0 is small, the energy loss caused liquid becomes the major components due to the dec- Figure 4. Pressure drop with different liquid flow. reased rate of vapor. For sieve tray, the vapor needs

525 W. RAN et al.: TOTAL SPRAY TRAY FOR DISTILLATION COLUMNS Chem. Ind. Chem. Eng. Q. 23 (4) 523−527 (2017) pass the clear liquid of upper plate, and for New VST, CONCLUSIONS the vapor need drawn the liquid to the holes on the side wall. So, TST has the lowest pressure drop In this work, the pressure drop of a new type tray (TST) was studied. Compared with sieve or New under the lower F0 as its liquid flow has little effect on VST trays, the TST can exhibit excellent perform- total pressure drop. For example, when F0 = 10, the pressure drop of TST is about 400 Pa and is two- ances in hydrodynamics. First, it can be operated -thirds that of the sieve tray. under higher F0 without flooding. Second, TST can provide much lower pressure drop than other trays at Effect of open-area ratio of injection tube on pressure small F0. Third, by the reasonable design of RA, the drop of TST pressure drop can be reduced significantly and

From the above analysis, we know that the reaches the level of sieve tray at higher F0. energy consumption of vapor passing the holes inc- For further research, the more detailed hydro- luding the holes on the plate and the holes on the side dynamics and the mass transfer properties of TST wall of injection tube is the main cause of pressure should be investigated. Besides no back mixing, we loss for TST. In order to further reduce the pressure have reason to believe that more advantages of TST drop, a larger open area may be efficient. In contrast are waiting to be found. Moreover, the research on to other traditional trays, the open area on the side- the TST provides us a novel direction on the develop- wall of the injection tube can be adjusted due to the ment of the column tray. higher space utilization of TST. In this paper, the Acknowledgement pressure drops of three injection tubes with different This work is support by national natural science RA were tested and are given in Figure 6, where RA equals the total open area on the side wall of injection foundation of china (NO. 11572357). tube divided by the cross-sectional area of the injection tube. As it can be seen from Figure 6, the pres- REFERENCES sure drop gradually decreased with the increasing of [1] H. Cong, X. Li, Z. Li, H. Li, X. Gao, Sep. Purif. Technol. R under the same F . For example, when F = 27, A 0 0 158 (2016) 293-301 the pressure drop of R = 1.7 is about 1050 Pa, which A [2] Z.G. Liang, R.B. Yuan, S.G. Lu, D.S. Duan, Chem. Ind. is closed to the value of sieve tray; however, that of Eng. Prog. 02 (2003) 180-182 RA = 1.3 and RA = 1 reaches 1300 and 1650 Pa, res- [3] J.D. Liu, S.Z. Zheng, H.H. Wang, Chem. Ind. Eng (China) pectively. That is, with the reasonable design of RA, 39 (2011) 47-50 the pressure drop of TST can be reduced significantly [4] R. Hauser, Pet. Ref. Eng. 29 (1999) 33-36 even at higher F0, which would benefit the energy [5] J.L. Bravo, K.A. Kusters, Chem. Eng. Prog. 96 (2000)33- decreasing in distillation operation greatly. –37 [6] P. Wilkinson, E. Vos, G. Konijn, H. Kooijman, G. Mosca, L. Tonon, Chem. Eng. Res. Des. 85 (2007) 130-135 [7] L.H. Wang, J.B. Ji, K.J. Yao, C.S. Xu, Chem. Ind. Eng (China) 28 (2000) 14-17 [8] J.D. Liu, J.H. Lv, J.P. Zhang, C.L. Li, J. Chem. Ind. Eng (China) 56 (2005) 1144-1149 [9] T. Zarei, R. Rahimi, A. Zarei, M. Zivdar, Chem. Eng. Proc. 64 (2013) 17-23 [10] J. Bausa, B. Pennemann, Chem. Eng. Res. Des. 99 (2015) 43-48 [11] J.M. Wang, Z.H. Shang, Q. Chen, Chem. Ind. Eng (China) 01 (1999) 17-20 [12] Q. Li, M. Zhang, X. Tang, L. Li, Z. Lei, Chem. Eng. Res. Des. 91(6) (2013) 970-976 [13] Q. Yang, G. Mosca, M. Roza, Chin. J. Chem. Eng. 18 (2010) 954-961 Figure 6. Pressure drop of TST with different RA. [14] W.L. Delnicki, J.L. Wagner, Chem. Eng. Prog. 66 (1970) 50-55 [15] W.H. Lai, Chem. Ind. Eng (China) 05 (1983)38-45.

526 W. RAN et al.: TOTAL SPRAY TRAY FOR DISTILLATION COLUMNS Chem. Ind. Chem. Eng. Q. 23 (4) 523−527 (2017)

WANG RAN TOTALNI SPREJ POD ZA DESTILACIONE TAO JINLIANG KOLONE: NOVA GENERACIJA PODA SA LING YANLI NIŽIM PADOM PRITISKA WEI FENG

LIU JIDONG Kao kritična komponenta kontakta para-tečnost, podovi su igrali važnu ulogu u perfor- School of chemical engineering, Hebei mansama destilacione kolone. U radu je istraživana hidrodinamika nove vrste poda, koji University of Technology, Tianjin, je nazvan totalni sprej pod (TSP). Pad pritiska sa različitim faktorom kinetičke energije

China (F0) je istraživan u sistemu vazduh-voda pod atmosferskim pritiskom. Rezultati ukazuju da nema plavljenja poda, pa TSP može obezbediti širu fleksibilnost u radu. Štaviše, pad

pritiska TSP je znatno niži od onog kod drugih podova. Opisana svojstva TSP će jako NAUČNI RAD pogodovati širenju proizvodnje i smanjenju energije u operaciji destilacije.

Ključne reči: destilacija, pod, pad pritiska, sprej.

527

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 529−536 (2017) CI&CEQ

YANJUN GUAN1 CFD INVESTIGATION OF SLUGGING XIUYING YAO2 BEHAVIOR IN A GAS-SOLIDS FLUIDIZED BED GUIYING WU2 1 KAI ZHANG Article Highlights 1 • A simple two-fluid model is applied to simulate the slugging behavior Beijing Key Laboratory of • Effect of viscous force on slugging dynamics is investigated Emission Surveillance and Control • Predicted results are in good agreement with the experimental findings for Thermal Power Generation, North China Electric Power Abstract University, Beijing, China The hydrodynamics of a slugging fluidized bed is investigated numerically by 2State Key Laboratory of Heavy Oil using a two-fluid model suggested by Brandani and Zhang [22]. Numerical Processing, China University of Petroleum, Beijing, China simulations are carried out in the platform of a commercial software package, CFX 4.4, by adding user-defined subroutines. The bed expansion ratio, the SCIENTIFIC PAPER pressure drop fluctuation and its power spectrum density at the equipment scale and the solids volume fraction distributions at the grid scale are predicted to UDC 66:004:519.876.5 explore the slugging fluidization characteristics for Geldart group D particles. By comparing with the experimental data in the literature, the numerical model is found to have better predictive ability when the viscous force term is considered in the gas phase momentum equation. Keywords: gas-solid fluidized bed, slugging, Geldart group D particles, CFD simulation.

Gas-solid fluidized beds have been extensively bed diameter [4]. The slugging behavior occurs easily used in petroleum refining and petrochemical pro- when the large particles, especially for Geldart group cesses because of their excellent heat and mass D particles, are employed, which should result in transfer properties, and effective utilization of active more unstable fluidization patterns [5]. For different catalytic surface. In parallel with industrial develop- types of fluidized beds, sometimes the slugging is ments, many academic efforts have been focused on encountered in the bed of high aspect ratio, often with the fluidization for providing a theoretical framework a small diameter compared to the particle size [6]. to strengthen this subject [1]. Grace et al. [2] pro- Generally, the slugging is categorized into three posed a flow regime map, including homogeneous, types: round-nosed slug, wall slug, and square-nosed bubbling, slugging, turbulent and fast fluidization. Up slug. Wu et al. [7] found that the fluidized beds with to now, the majority of studies on gas-solids fluidized Geldart group D particles could operate in the round- beds were centered on the bubbling or fast fluidization nosed slug, whilst the wall slug or square-nosed slug behaviors, which are common in industrial process. formed with the increase in superficial gas velocity or The slugging is characterized as the size of bub- static bed height. As it forms a complex and hetero- ble becomes almost the same as the diameter of geneous structure traversed by voids or bubbles, the fluidized bed [3]. It can be explained for bubbling beds slugging behavior is usually considered to be unfavor- by considering that the bubble growth reaches the able to the performance of fluidized bed reactor [8]. Therefore, it is necessary to understand the slugging characteristics for developing and scaling up the fluid- Correspondence: K. Zhang, Beijing Key Laboratory of Emission ized-bed reactors in the field of process industry. Surveillance and Control for Thermal Power Generation, North China Electric Power University, Beijing 102206, China. Computational fluid dynamics (CFD) modeling E-mail: [email protected] has been extended successfully to investigate the Paper received: 3 January, 2016 hydrodynamics of gas-solids systems. In general, Paper revised: 13 June, 2016 Paper accepted: 9 February, 2017 CFD models are divided into two categories: Eulerian- https://doi.org/10.2298/CICEQ160103007G –Lagrangian and Eulerian-Eulerian approaches. The

529 Y. GUAN et al.: CFD INVESTIGATION OF SLUGGING BEHAVIOR… Chem. Ind. Chem. Eng. Q. 23 (4) 529−536 (2017) former considers the solids phase at the particle level, CFD MODEL AND NUMERICAL PROCEDURE which is still too complicated to be applied in an engineering installation. While the latter considers both Mathematical model fluid phase and solids phase as two inter-penetrating The continuity and momentum balance equat- continua, and the two phases are described as separ- ions, describing gas and particle flows in the two-dim- ate conservation equations with appropriate interact- ensional cold model of fluidized bed, are given below: ion terms, which can handle the large-scale fluidized Continuity equations: bed since the point variables are averaged over local Gas phase: region. The Eulerian-Eulerian model, also called two- ∂ε  fluid model (TFM), has been suggested as a feasible g +∇()ε u =0 (1) ∂ gg approach for performing parametric investigations t and scale-up studies in the fluidized beds [9]. A set of Particle phase: physical or empirical models are employed to close ∂ε  the conservation equations in the TFM, especially for p +∇ε = ()ppu 0 (2) the solids phase stress term in the momentum equat- ∂t ions [10]. However, there is still no consensus on the Momentum equations: particle-phase viscosity in the solids momentum Gas phase: equation for CFD simulations based on TFM model  ∂()ερu   [11,12]. The solids-phase viscosity is generally def- ggg+∇ερ = ()gggguu ined by three approaches: i) as a constant of 1.0 ∂t (3)    based on experimental data of Grace [13], ii) as a =−βεερ − − ∇ + − ()uug p g p g gg F ad,g constant value of zero and iii) as a function of the granular temperature by kinetic theory of granular Particle phase: flow (KTFG) [14]. In recent twenty years, a few invest-  ∂()ερu   igations were carried out for studying the gas-solids ppp+∇ερ = ()ppppuu ∂t (4) slugging fluidization based on the TFM together with    =−βεερ − − ∇ + − KTFG [15-20], such as Pain et al. [15] investigated ()uup g p p p pg F ad,p the formation, elongation, coalescence and eruption where ε is the volume fraction ( εε+=1), t the time, of bubbles in the slugging fluidized bed. Zhang and  gp ρ the density, u the velocity vector, p the pressure, β Yu [16] obtained the effect of wall conditions on slug-  ging formation. Lettieri et al. [1,17] simulated the the inter-phase drag coefficient, g the acceleration behavior of fluidized beds under different flow reg- due to gravity, and Fad the additional force. The sub- imes, spanning from bubbling to slugging and turbul- scripts g and p indicate gas and particle phases, res- ent fluidization. pectively. Based on the inviscid two-fluid model of Gida- The inter-phase momentum transfer plays an spow [21], Brandani and Zhang [22] proposed a hyd- important role in modeling the particle-fluid interact- rodynamic model by introducing additional terms into ions. And the inter-phase drag term is a function of both gas- and solid-phases momentum balance the particle drag coefficient, CD, which can be derived equations. Series of simulations were implemented to as follows: explore the fluidization characteristics, such as homo-   3ερ uu− βε= pg p g −1.8 geneous fluidization of Geldart group A particles, CDg (5) bubbling and collapsing behavior for Geldart group B 4d p particles, and the pressure fluctuation characteristics C can be obtained from the empirical Dalla in the gas-solid fluidized bed [23-25]. The gas visco- D Valle relationship [26]: sity item was ignored in the gas phase momentum equation when an additional term was taken into 4.8 C =+(0.63 )2 (6) account and the relatively low operational gas velocity D Re was used in the homogeneous or bubbling fluidization regime. The effect of the gas viscosity term in the gas here:   phase momentum equation is explore to predict the ερ − gguudp g p gas-solid slugging fluidization behaviors, such as Re = (7) μ solids volume fraction, bed expansion, and pressure g drop fluctuation.

530 Y. GUAN et al.: CFD INVESTIGATION OF SLUGGING BEHAVIOR… Chem. Ind. Chem. Eng. Q. 23 (4) 529−536 (2017)

where dp is the particle diameter, μ is the dynamic The boundary conditions in the 2-D fluidized bed are: viscosity, and Re is the particle Reynolds number. i) in the bottom of bed, the gas inlet velocity is spe- Based on a linear stability analysis of Wallis [27] cified and the solids velocity is set to zero in the ver- and Gibilaro [28], the additional forces, Fad , are intro- tical direction. The gas and solids velocities are zero duced into the inviscid two-fluid model of Gidaspow in the horizontal direction; ii) in the top of bed, the [21] by Brandani and Zhang [22]. By compared constant static pressure boundary is adopted, i.e., experimental with predicted minimum bubbling points ambient atmosphere; iii) left and right walls are [29-31], the characteristic length for describing the treated as the no-slip velocity boundary conditions for discrete nature of the particles in this additional term both gas and solid phases; iv) front and back walls is of the order of particle diameter. The above method are regards as symmetry planes since the simulations can be used to predict the homogeneous, bubbling, are carried out in the 2-D bed. collapsing or jetting behaviors for Geldart group A or Numerical calculation is performed on the plat- B particles in the gas-solids fluidized bed [23-25] and form of a commercial CFD code, CFX 4.4. The conti- to explore the liquid-solid fluidization [32,33]. Due to nuity equations (Eqs. (1) and (2)) are solved directly the superficial gas velocity for the slugging pheno- by the software. In momentum equations (Eqs. (3) menon is much higher than that for the homogeneous and (4)), the pressure drop and gravity force terms or bubbling state in the gas-solids fluidized bed, the are defined in the command file, while the inter-phase effect of gas viscosity term in the gas momentum drag and additional force terms in Eqs. (5), (8) and (9) equation is introduced into the additional force term, are programmed by user-defined Fortran subroutines. Fad,g in the TFM suggested by Brandani and Zhang The numerical solutions of the modeling equations [22]. Accordingly, the additional forces are expressed are obtained by a finite volume method based on a as: collocated grid approach by used the Rhie-Chow   algorithm [34] to prevent checkerboard oscillations. Fd=−()12ε ρ + 2ε ρ gi ∇+∇ετ() (8) ad,g p g p gg g g SIMPLEC algorithm (semi-implicit method for pres-  sure linked equations-consistent algorithm) is used to Fd=+−∇2(12)ε ρ ε ρ giε (9) ad,p p p p p gg deal with the pressure-velocity coupling. Under-relax-  ation factors are 0.65 for gas (or solid) volume frac- where i is unit vector, which is the same direction as   tion and velocity vector, and 1.0 for pressure. Differ- acceleration g , τ is the viscosity tension. g ent differencing methods are used to treat the advect- Numerical procedure ion terms: central differencing scheme for gas or solids volume fraction, upwind differencing scheme In order to save computational resources, a for the shared pressure field, and hybrid differencing detailed CFD investigation is implemented in the 1.0 scheme for all velocity components. m (height)×0.1 m (width) 2-D fluidized bed. The struc- Initially, the bed keeps in the minimum fluidized ture grid of 160 (height)×16 (width) is employed state, and the initial bed height is set to H . The based on the grid independency analysis, and time 0 volume fraction of solids, ε , should be zero in the step is selected as 2×10-4 s. In order to validate the p dilute phase region above bed surface. However, ε is reliability of the present numerical model, the simul- p set to 10-10 during the practical computational proce- ated results are compared with the experimental data dure, so as to provide more realistic results for the reported by Yang et al. [18], which were obtained in particle velocity and ensure good convergence. In the the fluidized bed with an inner diameter of 0.1 m and dense phase region, the particles are uniformly sus- a height of 1.0 m. The physical properties of gas- pended at the minimum fluidization state. The initial solids phases are listed in Table 1. pressure profile is calculated from the hydrostatic bed Table 1. Physical properties of gas-solids system height based on the reference pressure:  PP=+ερ +−() ε ρ H − Hg Parameter Value 0mfgmfp01() (10) 3 Gas density, ρg, kg/m 1.225 3 Particle density, ρp, kg/m 1294 RESULTS AND DISCUSSION

Particle diameter, dp, μm 2200 Determination of the minimum slugging velocity Minimum fluidization velocity, umf, m/s 0.6

Minimum voidage, εmf 0.4 The formation and development of slugging are

Initial bed height, H0, m 0.3 depend upon the geometrical structure of fluidized bed, physical properties of gas-solid system, and

531 Y. GUAN et al.: CFD INVESTIGATION OF SLUGGING BEHAVIOR… Chem. Ind. Chem. Eng. Q. 23 (4) 529−536 (2017) operating condition. When the geometrical structure bed expansion. From the snapshots of solids volume and physical properties are determined, the super- fraction with small time interval, it is found that the ficial gas velocity plays an important role for slugging bubbles move upward and then merge into a dilute behaviors. Figure 1 shows the instantaneous distri- flow core in the upper part of the bed, while particles butions of solids volume fraction at different velocities. at the top of the slug fall back to the dense region in When the superficial gas velocity is less than 0.4 m/s, the lower part of the bed. The results from Yang et al. the bed height keeps constant either introducing gas [18] showed an increase in bubble size with inc- viscosity item (GVI) in Eq. (8) or not, since the bed reasing the superficial gas velocity, and the distinct maintains in fixed-bed state. The bed enters fluidiz- wall slug occur as the gas velocity higher than 0.8 ation state when the gas velocity is increased to 0.6 m/s. CFD simulation in this study indicates that a m/s. A bubble appears in the lower part of the bed stable and regular slugging is formed within the fluid- when the GVI is ignored, but a big bubble (slug) can ized bed, which is in good agreement with the find- be observed in the upper part of the bed when the ings of Yang et al. [18]. GVI is added. As the gas velocity is further increased Bed expansion to 0.8 m/s, distinct slugs are observed within the fluidized beds either considering GVI or ignoring GVI. Bed expansion process is mainly affected by the Based on experimental data, Stewart and movement of bubbles for Geldart group D particles, Davidson [3] proposed an experiential correlation for and it has an ability to describe the characteristics of predicting the minimum slugging velocity as below: bubbles in the fluidized bed. Accordingly, the time- averaged bed expansion ratio is one of the important =+ 0.5 uumsl mf 0.07( gD ) (11) parameters for scaling up the fluidized bed. Figure 3 compares experimental and computational bed exp- where u is the minimum slugging velocity, u the msl mf ansion ratios. Initially, the bed expansion ratio is set minimum fluidization velocity, g the gravitational as 1 in the fixed bed state; the computational results acceleration, and D the bed diameter. show a reasonably match with the experimental values For the gas-solid system and fluidized bed used owing to the superficial gas velocity is less than the in this study, the minimum slugging gas velocity is minimum fluidization velocity. The bed expansion ratio calculated as 0.67 m/s, which is close to the numer- increases with increasing the superficial gas velocity, ical value of 0.6 m/s with GVI or 0.8 m/s without GVI. when the superficial gas velocity is greater than the Solids volume fraction minimum fluidization velocity. The simulated bed exp- Figure 2 indicates the distribution of solids ansion ratios with GVI or without GVI are close to the volume fraction obtained at the gas velocity of 1.2 m/s experimental results [18] within a relative error of 12%. (twice of minimum fluidization velocity). The round- Pressure drop fluctuation and power spectrum density nosed slugs are obtained when GVI is not taken into Pressure drop fluctuation derives from bubble account (Figure 2a), however obvious wall slugs are movement in the gas-solids fluidized bed. Figure 4 formed when GVI is considered (Figure 2b). The exhibits the effect of GVI on the pressure drop at u = wave of bubbles appears in the lower part of bed and g 1.2 m/s. Bubble (or slug) phase and emulsion phase travels through the bed, which leads to a noticeable are distinguished, i.e., the peaks are considered as

Figure 1. Instantaneous distribution of solids volume fraction at different gas velocities.

532 Y. GUAN et al.: CFD INVESTIGATION OF SLUGGING BEHAVIOR… Chem. Ind. Chem. Eng. Q. 23 (4) 529−536 (2017)

Figure 2. Snapshots of solids volume fraction.

Figure 3. Comparison of simulated and experimental time-averaged bed expansion ratios.

533 Y. GUAN et al.: CFD INVESTIGATION OF SLUGGING BEHAVIOR… Chem. Ind. Chem. Eng. Q. 23 (4) 529−536 (2017)

Figure 4. Effect of gas viscosity term on the bed pressure drop fluctuation, ug = 1.20 m/s. the dense phase, whilst the troughs signify the bub- in Figure 2b. Qualitative and quantitative numerical bles or slug region. Furthermore, the time-averaged results with considering GVI show good agreement pressure drops predicted are about 2348.5 Pa by with the experimental data of Yang et al. [18]. ignored GVI and 2326.2 Pa by considered GVI, which are close to the experimental data of Yang et al. [18]. CONCLUSIONS Figure 5 presents the power spectral density obtained with a data acquisition frequency of 100 Hz The slugging behavior is investigated in a gas- at the superficial gas velocity of 1.2 m/s. The gas vel- –solid fluidized bed by using a hydrodynamic model ocity has a great influence on the power spectrum of proposed by Brandani and Zhang [22]. The main the bed pressure drop fluctuations. The energy distri- conclusions are: i) the gas viscosity item in the mom- bution is more uniform when GVI is not taken into entum equation has an effect on the slugging form- account, which is similar to that in the bubbling bed ation in the gas-solid fluidized bed; ii) simulated (Figure 5a). However, the maximum power spectrum minimum slugging gas velocity of 0.6 m/s (or 0.8 m/s) is obtained in the low-frequency region with higher with (or without) gas viscosity term is close to the pre- energy when GVI is added in Eq. (8) (Figure 5b). An diction of Stewart and Davidson [3]; iii) the comput- obvious peak frequency of 1.87 Hz is found close to ational bed expansion ratios are in agreement with the experimental result of 1.36 Hz [18], while the the experimental results of Yang et al. [18]; iv) the energy is relatively small in other high-frequency analysis of the power spectral density shows that a parts. Such a phenomenon represents the slug cycle higher energy maximum is found in the low-frequency of formation, development and breakup, which is cor- region with the gas viscosity term, which is in good responding to the snapshots of solids volume fraction agreement with the experimental findings.

Figure 5. Power spectrum density of the bed pressure drop fluctuation signals, ug = 1.20 m/s.

534 Y. GUAN et al.: CFD INVESTIGATION OF SLUGGING BEHAVIOR… Chem. Ind. Chem. Eng. Q. 23 (4) 529−536 (2017)

Nomenclature [7] G. Wu, J. Ouyang, B. Yang, Q. Li, F. Wang, Particuology 10 (2012) 72-78 CD particle drag force coefficient [8] X. Ku, T. Li, T. Lovas, Chem. Eng. Sci. 95 (2013) 94-106 dp particle diameter, m  [9] B. van Wachem, J. Schouten, C. van den Bleek, R. F additional force, N/m3 ad Krishna, J. Sinclair, AIChE J. 47 (2001) 1035-1051 g acceleration due to gravity, m/s2 [10] S. Hosseini, W. Zhong, M. Esfahany, L. Pourjafar, S. H bed height, m Azizi, J. Fluids Eng. 132 (2010) 041301 H0 static bed height, m 2 [11] J. Li, J. Ouyang, S. Gao, W. Ge, N. Yang, W. Song, Multi- P bed pressure, N/m scale simulation of particle-fluid complex-fluid complex 2 P0 atmospheric pressure, N/m systems (in Chinese), Science Press, Beijing, 2005 2 p pressure, N/m [12] D. Patil, M. van Sint Annaland, J.A.M. Kuipers, Chem. Re Reynolds number Eng. Sci. 60 (2005), 57-72 t time, s  [13] J.R. Grace, Can. J. Chem. Eng. 48 (1970) 30-33 u velocity vector, m/s [14] K. Zhang, P. Pei, S. Brandani, H. Chen, Y. Yang, Chem. umf minimum fluidization velocity, m/s Eng. Sci. 68 (2012) 108-119 Greek letters [15] C. Pain, S, Mansoorzadeh, C. de Oliveira, Int. J. Multiphas. Flow 27 (2001) 527-551 β interphase drag coefficient [16] S. Zhang, A. Yu, Powder Technol. 123 (2002) 147-165 ε volume fraction [17] P. Lettieri, L. Cammarata, G. Micale, J. Yates, Int. J. εmf minimum voidage Chem. React. Eng. 1 (2003) 1-19 μ viscosity, Pa·s 3 [18] F. Yang, J. Wang, X. Gu, Chin. J. Pro. Eng. 5 (2005) 597- ρ density, kg/m –600 Subscripts [19] N. Ratkovich, C. Chan, P. Berube, I. Nopens, Chem. Eng. g gas phase Sci. 64 (2009) 3576-3584 p particle phase [20] X. Wang, L. Yu, J. Wang, Int. J. Chem. React. Eng. 10 (2012) A77 x vector component in x direction [21] D. Gidaspow, Multiphase Flow and Fluidization, z vector component in z direction Academic Press, Boston, MA, 1994 Acknowledgements [22] S. Brandani, K. Zhang. Powder Technol. 163 (2006) 80- –87 Financial support from National Natural Science Foundation of China (U1610254), Key Scientific and [23] K. Zhang, S. Brandani, J. Bi, J. Jiang, Prog. Nat. Sci. 18 (2008) 729-733 Technological Project of Shanxi Province (MD2015- [24] P. Pei, K. Zhang, E. Lu, D. Wen, Petrol. Sci. 6 (2009) 69- -01), Fundamental Research Funds for the Central –75 Universities (2017MS020) and 111 Project (B12034) [25] Q. Wang, K. Zhang, H. Gu. Petrol. Sci. 8 (2011) 211-218 is gratefully acknowledged. [26] J. Dalla Valle, Micromeritics, Pitman, London, 1948 [27] G. Wallis, One-Dimensional Two-Phase Flow, McGraw- REFERENCES -Hill, New York, 1969 [1] P. Lettieri, G. Saccone, L. Cammarata, Chem. Eng. Res. [28] L. Gibilaro, Fluidization Dynamics, Butterworth Heine- Des. 82 (2004) 939-944 mann, London, 2001 [2] J. Grace, Powder Technol. 113 (2000) 242-248 [29] K. Rietema, H.W. Piepers, Chem. Eng. Sci. 45 (1990) 1627-1639 [3] P. Stewart, J. Davidson, Powder Technol. 1 (1967) 61-80 [30] A. Marzocchella, P. Salatino, AIChE J. 46 (2000) 901-910 [4] D. Gidaspow, P. Chaiwang, Chem. Eng. Sci. 97 (2013) 152-161 [31] K. Godard, J.F. Richardson, I. Chem. E. Symp. Ser. 30 (1968) 126-135 [5] Y. Jin, J Zhu, Z. Wang, Z. Yu, Fluidization Engineering Principles (in Chinese), Tsinghua University Press, [32] K. Zhang, G. Wu, S. Brandani, H. Chen, Y. Yang, Powder Beijing, 2001, p. 85 Technol. 227 (2012) 104-110 [6] P. Lettieri, R. di Felice, R. Pacciani, O. Owoyemi, Powder [33] K. Zhang, Y. Guan, X. Yao, Y. Li, X. Fan, S. Brandani, Technol. 167 (2006) 94-103 Powder Technol. 235 (2013) 180-191 [34] C. Rhie, W. Chow, AIAA J. 21 (1983) 1525-1532.

535 Y. GUAN et al.: CFD INVESTIGATION OF SLUGGING BEHAVIOR… Chem. Ind. Chem. Eng. Q. 23 (4) 529−536 (2017)

YANJUN GUAN1 CFD ISTRAŽIVANJE PONAŠANJA ČEPOVA U XIUYING YAO2 FLUIDIZOVANOM SLOJU GASNO-ČVRSTO GUIYING WU2 1 KAI ZHANG Hidrodinamika fluidizovanog sloja sa čepovima je istraživan korišćenjem modela sa dva 1Beijing Key Laboratory of Emission fluida koji su predložili Brandani i Zhang. Numeričke simulacije su izvedene na platformi Surveillance and Control for Thermal komercijalnog softverskog paketa CFKS 4.4, dodavanjem korisničkih potprograma. Power Generation, North China Electric Model proračunava odnos ekspanzije sloja, fluktuaciju pada pritiska i gustinu spektra Power University, Beijing, China snage na nivou opreme i raspodela sadržaja čvrste faze na nivou mreže u cilju odre- 2State Key Laboratory of Heavy Oil đivanja karakteristika fludizovanog sloja sa čepovima u slučaju za čestica Geldart group Processing, China University of D. Upoređivanjem sa eksperimentalnim podacima iz literature, utvrđeno je da numerički Petroleum, Beijing, China model ima bolju prediktivnu sposobnost kada se član viskozne sile uključi u jednačini količine kretanja za gasnu fazu. Ključne reči: fluidizovani sloj gas-čvrsta materija, formiranje čepova, Geldartove NAUČNI RAD čestice, CFD simulacija.

536 Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 537−545 (2017) CI&CEQ

SNEZANA SEVIC SIMULATION OF TEMPERATURE-PRESSURE BRANKO GRUBAC PROFILES AND WAX DEPOSITION IN PM Lucas Enterprises, Kać, Serbia GAS-LIFT WELLS

SCIENTIFIC PAPER Article Highlights • Temperature-pressure profiles in gas-lift wells can be simulated by Aspen HYSYS UDC 622:66:519.876.5 • Matching simulated with measured profiles depended on the number of pipes representing tubing • The more pipes included in the model, the better matching with measured data was achieved • The model can be used to determine wax deposit thickness distribution vs. well depth • Wax deposit profile can be used to plan wax cutting depth and frequency

Abstract Gas-lift is an artificial lift method in which gas is injected down the tubing- -casing annulus and enters the production tubing through the gas-lift valves to reduce the hydrostatic pressure of the formation fluid column. The gas changes pressure, temperature and fluid composition profiles throughout the production tubing string. Temperature and pressure drop along with the fluid composition changes throughout the tubing string can lead to wax, asphaltenes and inorganic salts deposition, increased emulsion stability and hydrate formation. This paper presents a new model that can sucesfully simulate temperature and pressure profiles and fluid composition changes in oil well that operates by means of gas-lift. This new model includes a pipe-in-pipe segment (production tubing inside production casing), countercurrent flow of gas-lift gas and producing fluid, heat exchange between gas-lift gas and the surrounding ambient – ground; and gas-lift gas with the fluid in the tubing. The model enables a better understanding of the multiphase fluid flow up the production tubing. Model was used to get insight into severity and locations of wax deposition. The obtained information on wax deposition can be used to plan the frequency and depth of wax removing operations. Model was developed using Aspen HYSYS software. Keywords: gas-lift; oil well; pressure-temperature profile; simulation; wax deposition.

Gas-lift is a method that uses an external source and lowers the flowing bottom hole pressure, creating of high-pressure gas for helping formation gas to lift a drawdown that allows the fluid to flow into the well- the well fluids. The compressed gas is injected down bore. During fluids flow through a tubing string in oil the production casing-tubing annulus, entering the producing well, pressure, temperature, phase’s ratio, tubing through the working gas-lift gas valves. As the composition of phases change. These changes are gas-lift gas enters the tubing, it forms bubbles, the result of different effects, such as the frictional lightens the formation fluids by reducing fluid density loss, fluid lifting, and heat transfer from the surround- ings. At the same time, the Joule-Thompson effect Correspondence: S. Sevic, PM Lucas Enterprises, Kać, Serbia. takes place. Typically, gas lift designs are based on E-mail: [email protected] natural gas as the injection gas. Early gas lift oper- Paper received: 14 October, 2016 Paper revised: 7 February, 2017 ations were conducted using air as the injection gas. Paper accepted: 10 February, 2017 Nitrogen and carbon dioxide offer good alternatives to https://doi.org/10.2298/CICEQ161014006S natural gas for gas lift [1,2].

537 S. SEVIC, B. GRUBAC: SIMULATION OF TEMPERATURE-PRESSURE… Chem. Ind. Chem. Eng. Q. 23 (0) 000−000 (2017)

One of the first theoretical models which des- amic multiphase flow simulator to simulate the tran- cribed single-phase fluid flow temperature as a func- sient dynamic gas-lift unloading process. tion of well depth and producing time was given by Flow assurance in oil and gas production is con- Ramey [3]. Sagar et al. [4] extended Ramey’s method sidered as the ability to produce fluids economically, for wellbore with multiphase flow. A unified model for from the reservoir to a production facility over the life of temperature distribution with approximate method to a field and in any environment. Deposition of any sort calculate the Joule-Thompson coefficient for two may lead to operational difficulties and production loss. phase mixture with a black oil in wellbore was pre- Gas-lift application can influence flow assurance sented by Alves et al. [5]. Hasan [6] developed ana- issues. Temperature drop can cause wax deposition. lytical model for the flowing fluid temperature in the Gas-lift gas can extract H2S and CO2 out of the oil drill pipe/tubing and in the annulus as a function of and water phase and by that effect can shift pH to well depth and circulation time. Hasan and Kabir car- greater values and support CaCO3 formation and sul- ried out research on the heat transfer in wellbore [7,8] fate reducing bacteria activity. Gas bubbling causes and fluid temperature profile in gas-lift wells [9]. They intensification of water and oil leading to increased also developed method for predicting two-phase gas/ emulsion stability. Introduction of light gas compo- /oil pressure-drop in vertical oil wells [10,11]. The nents can cause asphaltenes precipitation. Corrosion effect of tubing temperature and injection gas-lift gas in a tubing-casing annulus and increase of hydrate temperature on the gas-lift valve dome temperature formation temperature are often recorded in gas-lift were studied by Bertovic [12] and Faustinelli [13]. Xu operations. [14] made experimental study of three-phase flow in a Temperature reduction is the most common vertical pipe in order to investigate the influence of cause of wax deposition because wax solubility in gas injection and the average in situ phase fraction hydrocarbon fluids decreases as the temperature and pressure gradient. Cazarez et al. [15] developed drops. Pressure changes usually have a very small a model able to predict pressure, temperature and effect on wax precipitation temperatures and amounts. velocity profiles and volumetric fraction of the com- Wax deposition may cause operational problems in ponents. Bannwart et al. [16] done research on pres- production tubing string and surface pipelines during sure drop and pressure gradient in wells in three oil flow from the perforated/open hole interval to phase fluid flow. Temperature and pressure distri- surface and further down the oil pipeline. bution versus well depth in gas, gas-condensate and Numerous wax deposition models have been oil wells was extensively examined [17-19]. At the developed [26-29]. These models differ in the number beginning of the 21st century, two artificial neural net- of phases that were considered, wax formers, pro- work (ANN) models were developed to predict the perties of components. temperature of the flowing fluid at any depth in flowing There are a few tools designed specifically for oil wells [20]. modeling wax deposition such as pvtSimnova, PipeSim, The presence of multiple phases greatly com- OLGA, Aspen HYSYS, etc. Simulators can be divided plicates pressure drop calculations. This is due to the into static and dynamic. In the static models, the con- fact that the properties of each fluid present must be ditions of a system do not change with time, opposite taken into account. Also, the interactions between from the dynamic model. each phase must be considered. Mixture properties The authors of this paper could not find refer- must be used, and therefore the gas and liquid in-situ ences that deal with use of Aspen HYSYS in simul- volume fractions throughout the pipe need to be det- ation of temperature and pressure distribution in wells ermined. In general, multiphase correlations are that produce oil by means of gas-lift. This paper pre- essentially two-phase and not three-phase. Accord- sents a new model developed using Aspen HYSYS ingly, the oil and water phases are combined, and which can sucesfully simulate temperature-pressure treated as a pseudo single-liquid phase, while gas is profiles, fluid phase behavior, composition change considered a separate phase. Modeling and simul- and wax deposition as a function of well depth in the ation of multiphase system, even under steady-state gas-lift wells. This new model included pipe-in-pipe condition, is complex [21-24]. There are a few tools, segment (tubing-casing), countercurrent flow of gas- such as PipePhase, PipeSim, OLGA and Aspen lift gas and producing fluid, heat exchange between HYSYS, designed specifically for simulation and gas-lift gas and the surrounding ambient-ground; and analysis of complex multiphase systems. Li [25] dev- gas-lift gas with the fluid in the tubing. eloped dynamic model for sensitivity analysis of gas- lift wells using a commercially available OLGA dyn-

538 S. SEVIC, B. GRUBAC: SIMULATION OF TEMPERATURE-PRESSURE… Chem. Ind. Chem. Eng. Q. 23 (0) 000−000 (2017)

MODEL DEVELOPMENT The tubing was modeled as a circular pipe with a given diameter. Annular space was modeled as a Model Boundaries and Streams circular pipe with hydraulic diameter corresponding to To develop a model, it is necessary to define the an annular section. Hydraulic diameter of a circular inlet and outlet boundaries of the model, to establish tube with an inside circular tube was calculated as: the process flow along with the process parameters drr=−2() (1) and to define inlet and outlet streams. hoi

The model inlet boundaries were: where dh is a hydraulic diameter (mm), ro is an inside

– bottom of the well and radius of the outside tube (mm), ri is an outside radius – the inlet of the compressor for gas-lift gas. of the inside tube (mm). The model outlet boundary was the wellhead of 2. Countercurrent fluid flow: the production well. − Producing fluid from the well, along with the The inlet stream consisted of the following: accompanied gas-lift gas flows from the bottom of the – hydrocarbon fluid at the bottom of the well, well up, toward the wellhead, through the tubing – water stream at the bottom of the well and (inner space). – gas-lift gas. − Gas-lift gas flows from the top of the casing- The outlet stream was producing fluid from the tubing annulus to the bottom and enters the tubing well. through the working gas-lift valve. Figure 1 presents the scheme of gas-lift well 3. Heat exchange takes place as follows: operation. − Gas-lift gas exchanges heat with the surrounding ambient - ground and with the fluid in the tubing. − Fluid in the tubing exchanges heat with the gas-lift gas in the annular space. Heat transfer between gas-lift gas and ambient was modeled by the heat transfer option “Estimate heat transfer coefficient (HTC)” (pipe wall, inner HTC, and outer HTC were included), taking into account ground thermal gradient, as no other data were available [30]. Heat transfer between inner pipe and annular space was modeled only at nodes of a desired number of elements, and using estimated heat flow to obtain measured temperature at the end of a pipe segment. For this, the overall length of the pipe was broken down into a desired number of elements, and the end of each node was accounted for the heat flux using a mixer. Figure 1. Scheme of gas-lift wells operation. If the overall heat transfer coefficient and the Aspen HYSYS Software version 8.8 was used ambient temperature are specified, then the outlet for simulation. Peng-Robinson equation of state temperature is determined from the following equat- package was used. ions: =Δ Process flow modelling QUATLM (2) =− Process of gas-lift wells operation included the QQIN Q OUT (3) following elements relevant for the targeted simulation: where Q is an amount of heat transferred per unit of 1. Pipe-in-pipe system which includes: time (W), U is an overall heat transfer coefficient 2 2 − Production tubing – the inner pipe. (W/(m K)), A is an outer heat transfer area (m ), Δ − Production casing – the outer pipe. TLM is a log mean temperature difference (K), QIN is − Annulus “A” - space between the production a heat flow per unit of time of inlet stream (W) and tubing outer diameter and production casing inner QOUT is a heat flow per unit of time of (W) [30]. diameter, tubing-casing annular space.

539 S. SEVIC, B. GRUBAC: SIMULATION OF TEMPERATURE-PRESSURE… Chem. Ind. Chem. Eng. Q. 23 (0) 000−000 (2017)

RESULTS AND DISCUSSION compressed gas-lift gas was redirected towards producing oil wells at the gas manifold via separated Model validation pipelines. Calculations on gas-lift performance have Available production and process data were shown that gas pressure for the Well D should be tested to verify the reliability of developed model. The around 74 bar. Therefore, pressure was reduced to inlet parameters for the simulation were as follows: 74 bar through the choke. Gas was injected into the – Oil, gas and water production rates on the annular space of the well, and entered the tubing date when measurements by production logging tool through the working gas-lift valve. Gas-lift gas, along (PLT) were performed. with the fluid from the well, flowed up through the – Oil and gas composition at bottom hole con- tubing to the wellhead, and was transported by the ditions from the pressure, volume, and temperature pipeline to the main gathering station. The main pro- (PVT) analysis. cess parameters used in the simulation are shown in – Gas-lift gas injection rate, composition, inject- Table 1. ion temperature and pressure on the date when measurements by production logging tool (PLT) were per- Table 1. Process parameters used in the simulation formed. Parameter Unit Value – Ground thermal gradient for the region in the Well production data vicinity of the wells. Oil production rate t 14.85 – Existing well completion. Gas production rate S m3 12996 − Flow rates of producing fluids and gas-lift gas Water production rate S m3 3.3 on the date when measurements by production logging tool (PLT) were performed were accepted. Bottom hole conditions ° To develop and verify this approach, actual field Bottom hole temperature C 69.05 multiphase-flow data points were obtained from the Bottom hole pressure bar 87.49 existing measured data of temperature-pressure pro- Gas-lift gas files. Injection rate S m3 18077 Pressure at the top of the casing bar 73.5 Case study Temperature at the top of casing °C 4.45 Sales gas was used in the gas-lift operation. Gas was compressed to 125 bar, and transported Process flow diagram (PFD) is shown in Figure through the pipeline to the distribution manifold. The 2. Producing well – tubing-casing, gas-lift gas and pro-

Figure 2. Process flow diagram of gas-lift gas transportation to the casing-tubing annular space and from the wellhead to the oil treatment plant.

540 S. SEVIC, B. GRUBAC: SIMULATION OF TEMPERATURE-PRESSURE… Chem. Ind. Chem. Eng. Q. 23 (0) 000−000 (2017) ducing fluid streams, are presented by subflowsheet, In the next model, tubing was presented with 3 abbreviation T. (three) separate pipes, one from the bottom of the well to the working gas-lift valve depth, and 2 (two) Temperature-pressure profiles from the working gas-lift gas valve to the wellhead. In the first model, the tubing consisted of 2 (two) The annular space was presented with 2 (two) pipes pipes: the first pipe represented a part of the tubing (Figure 4). from the bottom of the well up to the working gas-lift The idea was to improve matching with the mea- valve. The second pipe represented the part of the tub- sured data. All available pressure drop correlations ing from the working gas-lift valve to the wellhead. for vertical flow were tested again. Simulation was The annular space was presented as a single pipe. performed in two ways: All pipes were divided into 3 segments, and each − The same pressure drop correlation for all segment was divided into 5 increments. All multi- pipes. phase pressure drop correlations for vertical flow of − Different correlations for pipes. gas-liquid systems available in HYSYS were checked. The result showed better matching of calculated Measured pressure and temperature profiles were temperature-pressure profiles with measured data used as a reference. Results showed poor matching compared to the first model. In the case when a single of calculated temperature-pressure profiles with mea- pressure drop correlation was accepted for all pipes, sured values no matter which of the correlations were the best matching was achieved using Hagedorn and used. Some pressure drop correlation overestimated Brown correlation. Results are shown in Figure 5. and some of them underestimated measured pres- It was noticed that different correlations applied sure and temperature distribution. The same results to each pipe provided better pressure loss estimates were obtained if pipe was divided in more segments than a single-correlation for the entire tubing. It could and increments. Figure 3 shows results of tempera- be explained by the fact that along the tubing, due to ture-pressure profiles obtained by 4 different multi- the temperature and pressure changes, liquid/vapor phase pressure drop correlations for vertical fluid flow. phase ratios changed, and by that phases superficial

Figure 3. Measured vs. calculated profiles for 2 pipes case: a. pressure; b. temperature.

541 S. SEVIC, B. GRUBAC: SIMULATION OF TEMPERATURE-PRESSURE… Chem. Ind. Chem. Eng. Q. 23 (0) 000−000 (2017)

Figure 4. Process flow diagram of fluids flow through the tubing and the annular space.

Figure 5. Measured vs. calculated profiles for 3 pipes case: a. pressure; b. temperature.

542 S. SEVIC, B. GRUBAC: SIMULATION OF TEMPERATURE-PRESSURE… Chem. Ind. Chem. Eng. Q. 23 (0) 000−000 (2017) velocities changed, which affected flow regime and AEA [30]. Results of wax deposit thickness profile pressure drop. versus well depth and wax deposit volume calculated Further improvement of matching between mea- by different models are shown in Figures 7 and 8. sured and calculated temperature-pressure profiles Software took into account deposited wax thick- using one multiphase pressure loss correlation was ness to calculate temperature - pressure profiles obtained by dividing tubing and annular space into when wax deposition calculation module was active more independent pipes. The best matching was (checked), leading to different temperature and pres- obtained when the tubing was divided into 10 (ten) sure at the outlet of the pipe compared to the case pipes. Temperature and pressure profiles are shown when calculation module was inactive (unchecked). in Figure 6. All other parameters, i.e., vapor fraction, liquid holdup, friction gradient, Reynolds numbers of gas Wax deposition simulation and liquid phases (and their velocities) differed, too. Wax deposition simulation was carried out using On the other hand, no changes of the overall fluid available models: Pederson, Conoco, Chung and

Figure 6. Measured vs. calculated profiles for 10 pipes a. temperature; b. pressure.

Figure 7. Wax deposit thickness profile calculated by different models.

543 S. SEVIC, B. GRUBAC: SIMULATION OF TEMPERATURE-PRESSURE… Chem. Ind. Chem. Eng. Q. 23 (0) 000−000 (2017)

Figure 8. Wax deposit volume calculated by different models. composition at the outlet of the pipe were observed gas and liquid velocities on temperature-pressure pro- although deposition took place inside the segments of files was not included in the developed model. the pipe. Thus, it was concluded that overall fluid compositional changes due to wax deposition along REFERENCES the pipe were not transferred between the pipes. Dif- ferences in ratios and compositions of present phases [1] H.W. Winkler, J.R. Blann, Production Operations Eng- ineering, SPE, Richardson, TX, 2006, pp. 521-623 (vapor, liquid, aqueous) at the inlet and the outlet of [2] G.Takacs, Gas Lift Manual, PennWell Corporation, Tulsa, the pipe depended on the temperature and pressure OK, 2005, pp.1-5 conditions. It is strongly recommended to check the [3] H.J. Ramey Jr., J. Pet. Technol. 14 (1962) 427-435 validity of PVT reports and composition analysis due [4] R. K. Sagar, D.R. Dotty, Z. Schmidt, 1989 SPE SPE to the fact that fluids composition data may be a Annu. Tech. Conf. Exhib., San Antonio, TX,1989, paper source of errors in multiphase flow calculations. SPE 19702 [5] I.N. Alves, F. J.S. Alhanti, O. Shoham, SPE Prod. Eng. 7 CONCLUSIONS (1992) 363-367 [6] A.R. Hasan, C.S. Kabir, SPE Prod. Facil. 11 (1996) 179- A model using Aspen HYSYS software to simul- –185 ate fluid phase behavior and temperature-pressure [7] A.R. Hasan, C.S. Kabir, 1991 SPE SPE Annu. Tech. profiles in wells that operate by means of gas-lift has Conf. Exhib., TX, 1991, paper SPE 22948 been developed. This model was tested against the [8] A.R. Hasan, C.S. Kabir, Western Regional Meeting, measured temperature-pressure data for wells depth Anchorage, AK, 1993, paper SPE 26098 up to 4600 m, pressures from 8 to 120 bar and [9] A.R. Hasan, C.S. Kabir, J. Pet. Sci. Eng. 86–87 (2012) temperatures from 5 to 110 °C. The developed model 127-136 can be used for wells without PLT data to obtain tem- [10] A.R. Hasan, C.S. Kabir, J. Pet. Sci. Eng. 4 (1990) 273- perature-pressure distribution and phase composit- –289 ions along the tubing. The results showed that match- [11] A.R. Hasan, C.S. Kabir, J. Pet. Sci. Eng. 72 (2010) 42-49 ing of calculated temperature-pressure profiles dep- [12] D. Bertovic, Production Operations Symposium, Okla- ended on the number of pipes representing tubing homa City, OK, (1997, paper SPE 37424 and annular space and the type of multiphase pres- [13] J.G. Faustinelli, D.R Doty, SPE Latin American and sure loss correlation. Different multiphase pressure Caribbean Petroleum Engineering Conference, Buenos loss correlations applied provided better pressure loss Aires, 2001, paper SPE 69406 estimates than a single correlation for the entire tub- [14] X.J. Z Jing-yu, L. Hai-fei, Int. J. Multiphase Flow 46 ing. The developed model can be used to get insight (2012) 1-8 into severity of the wax deposition and is helpful in [15] O. Cazarez, D. Montoya, A.G. Vital, A.C. Bannwart, Int. J. planning frequency of removing operations. Overall Multiphase Flow 36 (2010) 439-448 compositional changes along the pipe were not trans- [16] A.C. Bannwart, O.M.H. Rodrigez, F.E. Trevisan, C.H.M. ferred between the pipes, regardless of if wax deposit de Carvalho, J. Pet. Sci. Eng. 65 (2009) 1-13 was formed inside the pipe. Impact of the angle of [17] X. Zhao, J. Xu, World J. Modell. Simul. 4 (2008) 94-103 inclination, liquid viscosity, water cut, gas density, [18] I. Alves, F. Alhanati, O. Shoham, SPE Prod. Facil. (1992) 363-367 gas/liquid interfacial tension, the average superficial

544 S. SEVIC, B. GRUBAC: SIMULATION OF TEMPERATURE-PRESSURE… Chem. Ind. Chem. Eng. Q. 23 (0) 000−000 (2017)

[19] M.F. Zhou, X. Zheng, Asian J. Earth Sci. 26 (2015) 116- [25] L. Mengxia, R. Liao, Int. J. Heat Technol. 33 (2015) 237- –126 –245 [20] A. Zamani, P. Pourafshari, F. Rabiee, Gas Process. J. 2 [26] K.S. Pedersen, P.L. Christensen, J.A. Shaikh, Phase (2014) 81-85 Behavior of petroleum Reservoir Fluids, 2nd ed., CRC [21] F.F. Farshad, Engin. Comput. 17 (2000) 735-754 Press, Boca Raton, FL, 2015, pp. 229-257 [22] I.Y. Mohammed, Int. J. Curr. Engin. Technol. 4 (2014) [27] M. Stubsjøen, M.Sc. Thesis, Petroleum Norwegian 4057-4062 University of Science and Technology, Oslo, 2013, pp. [23] A.P. Szilas, Production and transport of oil and gas, 19-23 Development in Petroleum Science 3, Elsevier Scientific [28] J.A. Svendsen, AIChE J. 39 (1993) 1377-1388 Publishing Company, New York, 1975, pp. 169-192; 491- [29] E.D. Burger, T.K. Perkins, J.H. Striegler, J. Pet. Technol. –533 33 (1981) 1075-1086 [24] H. Hamedi, F. Rashidi, E. Khamehchi, Pet. Sci. Technol. [30] HYSYS 8.8 Operations Guide, AspenTech 2003 5.10- 29 (2011) 1305-1316 –5.60.

SNEZANA SEVIC SIMULACIJA PROFILA TEMPERATURE I PRITISKA BRANKO GRUBAC I TALOŽENJA PARAFINA U GAS-LIFT PM Lucas Enterprises, Kać, Srbija BUŠOTINAMA

NAUČNI RAD Gas-lift je mehanička metoda eksploatacije naftnih bušotina u kojoj gas utisnut u među- prostor između proizvodnog tubinga i proizvodnog kezinga, ulazi u proizvodni tubing kroz gas-lift ventile, čime se smanjuje hidrostatički pritisak stuba ležišnog fluida. Gas-lift gas utiče na promenu profila pritiska, temperature i sastava fluida u proizvodnom tubingu. Pad temperature i pritiska, uz promenu sastava fluida u tubingu, mogu da dovedu do taloženja parafina, asfaltena i neorganskih soli, povećanja stabilnosti emulzije i formiranje hidrata. Ovaj rad prikazuje novi model koji se može uspešno koristiti za simu- laciju promene profila temperature, pritiska i sastava fluida u stubu naftne bušotine koja radi u gas-liftu. Novi model uključuje protivstrujni protok gas-lift gasa i proizvedenog fluida kroz cev; razmenu toplote izmedju gas-lift gasa i okoline – zemlje, te gas-lift gasa i fluida u tubingu. Prikazani model omogućava bolje razumevanje multi-faznog protoka kroz proivodni tubing. Model je korišćen da se dobije uvid u nivo i mesto stvaranja taloga parafina. Dobijene informacije o taloženju parafina mogu se koristiti da se planira učestalost i dubina operacije uklanjanja parafina. Model je razvijen korišćenjem programa Aspen HYSYS.

Ključne reči: gas-lift, naftna bušotina, profil pritisak-temperatura, simulacija, talo- ženje parafina.

545

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017) CI&CEQ

1 AINHOA RUBIO-CLEMENTE KINETIC MODEL DESCRIBING THE UV/H2O2 EDWIN CHICA2 3 PHOTODEGRADATION OF PHENOL GUSTAVO A. PEÑUELA FROM WATER 1Departamento de Ciencia y Tecnología de los Alimentos. Article Highlights

Facultad de Ciencias de la Salud. • A kinetic model for organic pollutant degradation through the UV/H2O2 system is Universidad Católica San Antonio developed de Murcia UCAM, s/n. Guadalupe- • The model is validated using independent experimental data from literature for phenol Murcia, España abatement 2Departamento de Ingeniería • A high extent of agreement was found between the calculated and the experimental Mecánica, Facultad de Ingeniería, data • The kinetic model allows determining the optimal conditions for an efficient pollutant Universidad de Antioquia UdeA, removal Medellín, Colombia 3 Grupo GDCON, Facultad de Abstract Ingeniería, Sede de A kinetic model for phenol transformation through the UV/H O system was Investigaciones Universitarias 2 2 developed and validated. The model includes the pollutant decomposition by (SIU), Universidad de Antioquia • • •- • UdeA, Medellín, Colombia direct photolysis and HO , HO2 and O2 oxidation. HO scavenging effects of 2- - 2- - CO3 , HCO3 , SO4 and Cl were also considered, as well as the pH changes SCIENTIFIC PAPER as the process proceeds. Additionally, the detrimental action of the organic matter and reaction intermediates in shielding UV and quenching HO•was UDC 547.56:66:544.526.5:54 incorporated. It was observed that the model can accurately predict phenol

abatement using different H2O2/phenol mass ratios (495, 228 and 125), obtaining an optimal H2O2/phenol ratio of 125, leading to a phenol removal higher than 95% after 40 min of treatment, where the main oxidation species was HO•. The developed model could be relevant for calculating the optimal level of

H2O2 efficiently degrading the pollutant of interest, allowing saving in costs and time.

Keywords: H2O2 level, kinetic model, matrix background, phenol pollution, UV/H2O2.

The rapid economic growth of the last decade Phenol is a colorless to pink-colored solid or has led to an increase in the number of pollutants in dense liquid substance with a particular sweet odor. It the aquatic environment, positioning advanced oxid- is highly soluble in water but also in organic solvents. ation processes (AOPs) as alternatives to the con- It is a ubiquitous pollutant commonly found as a com- ventional techniques water treatment plants are oper- ponent of oil refinery wastes. It is also produced in ating with [1-6], which are unable to completely trans- pharmaceuticals and phenol manufacturing plants. form inhibitory organics such as phenol. Phenol can also be generated from the production of metallurgical coke from coal, and it is used in the manufacture of fertilizers, textiles and paints. Addit- ionally, phenol can be formed from natural sources, Correspondence: A. Rubio-Clemente, Departamento de Ciencia y such as the decomposition of organic matter [7]. Tecnología de los Alimentos. Facultad de Ciencias de la Salud. Universidad Católica San Antonio de Murcia UCAM, Avenida de Nevertheless, anthropogenic processes contribute in los Jerónimos, s/n. Guadalupe-Murcia. España. a larger extension in comparison with natural ones. E-mail: [email protected] Once phenol is in the environment, it can enter Paper received: 19 November, 2016 Paper revised: 24 January, 2017 aquatic environments due to its solubility in aqueous Paper accepted: 24 February, 2017 media, posing a risk for living beings. Furthermore, it https://doi.org/10.2298/CICEQ161119008R can react with other components, such as chlorine,

547 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017)

• •- increasing its hydrophobicity and, subsequently, its ten- UV/H2O2 system including the action of HO , O2 and • • 2- dency to be bioaccumulated through the food chain. HO2 . The HO scavenging effects of carbonate (CO3 ), - 2- - Phenol is a toxic compound. The toxicity of phe- bicarbonate (HCO3 ), sulfate (SO4 ) and chloride (Cl ) nol is related to the reactivity of the compound with ions were considered on the performance of the sys- biomolecules, being responsible for the death of org- tem as well as the pH changes in the bulk as the anisms when exposed to high amounts of this sub- oxidation process proceeds. Additionally, the detri- stance. It is also involved in reproductive problems, mental action of the organic matter and reaction inter- • reducing fertility and producing changes in appear- mediates in shielding UV and quenching HO was ance and behavior, since living being hormonal sys- incorporated. The model is based on the extensively tem function can be affected [7]. Furthermore, it exhi- accepted chemical and photochemical reactions, and bits mutagenic and carcinogenic potential [7,8]. As a the rate constants involved in the studied AOP. On matter of fact, researchers evidenced mutagenic acti- the other hand, MATLAB software was used to find vity of phenol from hamster fibroblasts [7]. In addition, the numerical solution to the set of ordinary differential it induces immunotoxic, hematological and physiolog- equations (ODE) characterizing the current model. In ical effects [8]. In fact, it is included by the US Envi- addition, the proposed kinetic model was validated by ronmental Protection Agency in the List of Priority Pol- comparing the experimental data from Alnaizy and lutants [7]. Therefore, it must be removed from water. Akgerman study [16] and the model predictions In recent decades, different AOPs have been obtained at different operating conditions in terms of used for destroying organic compounds. Among the phenol abatement. AOPs applied for water decontamination, the UV/

/H2O2 system implementation has risen drastically in EXPERIMENTAL recent years for water treatment purposes [6]. This process involves the photolysis of hydrogen peroxide A kinetic model describing the UV/H2O2 process was developed taking into account the impact of rad- (H2O2) using ultraviolet radiation (UV), resulting in the • • • • •- •- • production of hydroxyl radicals (HO ) [9-11]. The HO ical species such as HO , HO2 , O2 , CO3 , SO4 -, • • •- is very reactive and attack non-selectively a variety of H2ClO , Cl and Cl2 on the pollutant conversion. In organic and inorganic substances because their oxid- addition, the role of the dissolved organic matter ation potential (E = 2.8 V) [12] is higher than other (DOM), in terms of dissolved organic carbon (DOC), oxidizing agents such as ozone (E = 2.07 V) and and the degradation by-products in absorbing UV chlorine (E = 1.36 V), transforming those substances radiation and scavenging radicals were included. In the development of the proposed kinetic model, the into CO2, H2O and inorganic salts [10,13]. contribution of UV radiation in phenol transformation The efficiency of the UV/H2O2 system depends on various parameters, such as the oxidant dosage, was also taken into account. Furthermore, the model pollutant concentration and structure, UV-light inten- incorporates the change of pH in the solution as the sity and predominant wavelength, irradiation time, advanced oxidation process proceeds due to the solution initial pH, temperature and matrix constitu- formation of short chain organic acids. ents (including natural organic matter and inorganic • Mechanistic kinetic model ions, which may quench HO and/or shield UV radi- The chemical and photochemical reaction ation penetration into the bulk), among other para- scheme, reaction rate constants and values of the meters [3,10]. parameters involved in the UV/H O system are com- This homogeneous chemical oxidation process 2 2 piled in Tables 1 and 2. has several advantages in comparison to other water As it can be observed from the tables, 22 differ- treatment methods, such as no mass transfer limit- ent chemical species (including radicals, radical ions, ations occur and no sludge requiring a subsequent inorganic ions and molecules) and 51 reactions (dif- treatment and disposal is produced. Additionally, it ferentiated in photolysis reactions - first-order react- can be produced at room temperature and pressure ions, equilibrium reactions - acid-base reactions, and [6,10]. Nevertheless, this process has associated high second-order reactions, corresponding to the ele- electrical and oxidizing agent costs [10,14,15], and mentary reactions and those ones related to the mat- kinetic models seem to be a good option to optimize rix background) were considered. operating conditions without incurring in high oper- In addition, correlation factors ( δ ) were inc- ating costs. Ri luded in order to study the generation and consump- This work is aimed at proposing a kinetic model tion of some chemical species as the pH in the bulk is for the prediction of phenol decontamination by the δ = increased or reduced. In this sense, when Ri 1, the

548 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017)

Table 1. Elementary reactions and kinetic equations for species disappearance and formation in the UV/H2O2 system

No. Reaction Parameters and rate constants Ref. Kinetic equations −+ ⎯⎯k1 →+ −−21 R1 HO22 HO 2 H k =×3.7 10 s [17] dHO 1 22 =− k122HO =×10−− 1 1 dt k 2 2.6 10 M s − pK = 11.6 dHO2 a = k HO dt 122 dH+ = k HO dt 122 −+ +⎯⎯→k2 − R2 HO222 H H O dHO 2 −+ =−k HO H dt 22 dH+  −+ =−k HO H dt 22

dHO −+ 22 = k HO H dt 22 ν R HO⎯⎯→h 2HO• −−11 [9] 3 22 ε (254 nm)= 1800 M s dHO22 HO22 =−ϕ I δ HO22 a,HO 22R 3 −−11 dt ε − (254 nm)= 22800 M s HO2 • dHO −1 = 2ϕ I δ ϕ (254 nm)= 0.5 mol Ein HO22 a,HO 22R 3 HO22 dt −1 ϕ − = IIf=−−{1 exp[ 2.3 l (ε  H O + (254 nm) 0.5 mol Ein a,H22 O 0 H 22 O H 22 O 2 2 HO2 +++εεε− pH> 11.6 δδ== 0, 1 − HO2C C DOC  DOC  )]} RR24 HO2   

pH< 11.6 δδ== 1, 0 ε HO RR HO22 2 2 34 f = HO22 εεεε+++− HOHO 2 2− HO 2 C C DOC DOC 22 HO2  −•−ν +⎯⎯→+h − R4 HO22 H O 2HO OH dHO 2 =−ϕ δ −−I R dt HO22 a,HO 4 dHO• = ϕ δ 2 −−I R dt HO22 a,HO 4 dOH− = ϕ δ −−I R dt HO22 a,HO 4 =−−ε + IIf−−0HO22{1 exp[ 2.3 l ( H O a,HO22 HO 22

+++εεε− − HO2C C DOC  DOC  )} HO2    ε − − HO2 = HO2  f − − HO2 εεεε+++ HOHO 2 2− HO 2 C C DOC DOC 22 HO2 

••−+k3 51− • R5 HO⎯⎯→ O + H k =×1.58 10 s [18  22 3 dHO2 • 10−− 1 1 =−  k =×1.0 10 M s k 32HO 4 dt  p4.8K = a •− dO2 = k HO• dt 32 + dH • = k HO dt 32 •− + • k4 •− R6 OH+⎯⎯→ HO  22 dO2 =−k OH•− + dt 42 dH+ =−k OH•− + dt 42 • dHO2 •− + = k OH dt 42

549 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017)

Table 1. Continued

No. Reaction Parameters and rate constants Ref. Kinetic equations

••−+k5 711−− R7 HO+⎯⎯→++ HO O H HO k =×2.7 10 M s [19] dHO 22 2 2 5 22 =−k HO HO• δ 522R7 pH> 4.8 δδ== 1, 0 dt RR78 dHO• pH< 4.8 δδ== 0, 1  • RR78 =−k HO HO δ dt 522R7 •− dO2 = k HO HO• δ dt 522R7 dH+  • = k HO HO δ dt 522R7

••k5 R8 HO+⎯⎯→+ HO HO HO dHO 22 2 2 22 =−k HO HOº δ dt 522R8 dHO• =−k HO HO• δ dt 522R8 • dHO2 • = k HO HO δ dt 522R8 •− • − −− +⎯⎯→+k6 =×911 • R9 HO HO22 HO OH k 6 7.5 10 M s [20] dHO =−k HO•− HO dt 62 − dHO2 =−k HO•− HO dt 62 • dHO2 = k HO•− HO dt 62 − dOH •− = k HO HO dt 62 •• −− +⎯⎯→k7 =×911 • R10 HO HO H22 O k 7 5.5 10 M s [19] dHO =−k HO•• HO dt 7 

dHO •• 22 = k HO HO dt 7  •• −− +⎯⎯→+k8 =×911 • R11 HO HO222 H O O k 8 6.6 10 M s [21] dHO =−k HO•• HO dt 82 • dHO2 •• =−k HO HO dt 82

••−k9 − 911−− •− R12 HO+⎯⎯→+ O O OH k =×7.0 10 M s [9]  22 9 dO2 =−k HO••− O dt 92 dHO• =−k HO••− O dt 92 − dOH ••− = k HO O dt 92

••k10 511−− • R13 HO+⎯⎯⎯ HO→+ H O O k =×8.3 10 M s [18]  22 222 10 dHO2 =−k HO•• HO dt 10 2 2

dHO •• 22 = k HO HO dt 10 2 2

550 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017)

Table 1. Continued

No. Reaction Parameters and rate constants Ref. Kinetic equations •• k11 −−11 R14 H O+⎯⎯⎯ HO→++ HO H O O k = 3 M s [22] dHO 22 2 2 2 11 22 =−k HO HO• dt 11 2 2 2 • dHO2 =−k HO HO• dt 11 2 2 2 • dHO • = k HO HO dt 11 2 2 2 •− • − k12 −−11 R15 HO+⎯⎯⎯ O→++ HO O OH k = 0.13 M s [22] dHO 22 2 2 12 22 =−k HO O•− dt 12 2 2 2 •− dO2 =−k HO O•− dt 12 2 2 2 dHO• = k HO O•− dt 12 2 2 2 dOH−  •− = k HO O dt 12 2 2 2

••−k13 − 711−− • R16 HO+⎯⎯⎯ O→+ HO O k =×9.7 10 M s [18]  22 22 13 dHO2 =−k HO••− O dt 13 2 2 •− dO2 =−k HO••− O dt 13 2 2 - dHO2 ••− = k HO O dt 13 2 2 hν R17 CDOC?⎯⎯→ + ε , ϕ – dC CC =−ϕ I dt Ca,C dDOC = ϕ I dt Ca,C IIf=−−{1 exp[ 2.3 l (ε  HO + a,C 0 CHO22 2 2

+++εεε− − HO2 CDOC C  DOC )]} HO2    • k14 R18 CHO+⎯⎯⎯→+ DOC? kk• = – dC C, HO 14 =−k CHO• dt 14  dHO• =−k CHO• dt 14 

dDOC • = k CHO dt 14  • +⎯⎯⎯k15 →+ = R19 CHO2 DOC? kk• 15 – dC • C, HO2 =−k CHO dt 15 2 • dHO2 =−k CHO• dt 15 2

dDOC • = k CHO dt 15 2 •− +⎯⎯⎯k16 →+ = R20 CO2 DOC? kk•− 16 – dC •− C,O2 =−k CO dt 16 2 •− dO2 =−k CO•− dt 16 2

dDOC •− = k CO dt 16 2

551 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017)

Table 2. Effect of the matrix background in the UV/H2O2 system

No. Reaction Parameters and rate constants Ref. Kinetic equations • k17 * R21 DOC+⎯⎯⎯ HO→+ H CO H O kk• == [23] dDOC 23 2 DOC, HO 17 =−k DOC HO• 17  =×2.0 108-1-1 M s dt dHO• =−k DOC HO• dt 17  * dHCO23 • = k DOC HO dt 17  hν R22 DOC⎯⎯→ ? εϕ, – dDOC DOC DOC =−ϕ I dt DOC a,DOC =−−ε + IIfa,DOC 0 DOC{1 exp[ 2.3 l ( H O H 2 O 2 22 +++εεε− − HO2C C DOC  DOC  )]} HO2    ε DOC DOC f = DOC εεεε+++− HOHO 2 2− HO 2 C C DOC DOC 22 HO2  −• •−− 2 k18 811−− 2− R23 CO+⎯⎯⎯ HO→+ CO OH =× [19]  33k18 3.9 10 M s dCO3 =−k CO2−• HO dt 18 3 dHO• =−k CO2−• HO dt 18 3 dOH− = k CO2−• HO dt 18 3 dCO•− 3 2−• = k CO HO dt 18 3 −− k18 611−− − R24 HCO+⎯⎯⎯ HO→+ CO H O =× [19]  332k19 8.5 10 M s dHCO3 =−k HCO− HO dt 19 3 dHO =−k HCO− HO dt 19 3 dCO− 3 −  = k HCO HO dt 19 3

•−k − • 511−− R25 HO+⎯⎯⎯ CO20 →+ HCO HO k =×4.3 10 M s [24] dHO22 •− 22 3 3 2 20 =−k HO CO dt 20 2 2 3 •− dCO3 =−k HO CO•− dt 20 2 2 3 − d HCO3 = k HO CO•− dt 20 2 2 3 • dHO2 •− = k HO CO dt 20 2 2 3 −•− −• k21 2 711−− − R26 HO+⎯⎯⎯ CO→+ CO HO =× [24]  23 3 2 k 21 3.0 10 M s dHO2 =−k HO−•− CO dt 21 2 3 •− dCO3 =−k HO−•− CO dt 21 2 3 2− dCO3 = k HO−•− CO dt 21 2 3 • dHO2 −•− = k HO CO dt 21 2 3

552 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017)

Table 2. Continued

No. Reaction Parameters and rate constants Ref. Kinetic equations •− •− − k22 2 811−− •− R27 OCO+⎯⎯⎯→+ COO =× [25]  23 32 k 22 6.5 10 M s dO2 =−k OCO•− •− dt 22 2 3 •− dCO3 =−k OCO•− •− dt 22 2 3 2− dCO3 = k OCO•− •− dt 22 2 3 −• •− k23 511−− − R28 HSO+⎯⎯⎯ HO→+ SO H O =× [26]  442k 23 3.5 10 M s dHSO4 =−k HSO−• HO dt 23 4 dHO• =−k HSO−• HO dt 23 4 •− dSO4 −• = k HSO HO dt 23 4 •− − + • k24 2 711−− •− R29 SO+⎯⎯⎯ H O→++ SO H HO =× [26]  422 4 2 k 24 1.2 10 M s dSO4 =−k SO•−  H O dt 24 4 2 2 dHO 22 =−k SO•−  H O dt 24 4 2 2 2− dSO4 = k SO•−  H O dt 24 4 2 2 dH+ = k SO•−  H O dt 24 4 2 2 dHO• 2 •− = k SO H O dt 24 4 2 2 •− • − + k25 2 911−− •− R30 SO+⎯⎯⎯ HO→++ SO H O =× [26]  42 4 2k 25 3.5 10 M s dSO4 =−k SO•− HO • dt 25 4 2 • dHO2 =−k SO•− HO • dt 25 4 2 2− dSO4 = k SO•− HO • dt 25 4 2 + dH •− • = k SO HO dt 25 4 2 −• •− R Cl+⎯⎯⎯ HOk26 → HClO =×911−− [26] − 31 k 26 4.3 10 M s dCl =−k Cl−• HO dt 26  dHO• =−k Cl−• HO dt 26  •− d HClO −• = k Cl HO dt 26  •− + • R HClO+⎯⎯⎯ Hk27 → H ClO =×10−− 1 1 [26] •− 32 2 k 27 3.0 10 M s dHClO =−k HClO•− H + dt 27  dH+ =−k HClO•− H + dt 27  • dHClO2 = k HClO•− H + dt 27 

553 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017)

Table 2. Continued

No. Reaction Parameters and rate constants Ref. Kinetic equations

••k28 41− • R33 HClO⎯⎯⎯→+ HO Cl =× [26]  22 k 28 5.0 10 s dHClO2 =−k HClO• dt 28 2 dCl• = k HClO• dt 28 2 •− •− R Cl+⎯⎯⎯ Clk29 → Cl =×911−− [26] • 34 2 k 29 8.5 10 M s dCl =−k Cl•− Cl dt 29  dCl− =−k Cl•− Cl dt 29  •− dCl2 = k Cl•− Cl dt 29  •− • • − k30 911−− •− R35 Cl+⎯⎯⎯ HO→+ HClO Cl =× [27]  2 k 30 1.0 10 M s dCl2 =−k Cl•− HO • dt 30 2 dHO• =−k Cl•− HO • dt 30 2 d HClO• = k Cl•− HO • dt 30 2 − dCl •− • = k Cl HO dt 30 2 •− • − + k31 411−− •- R36 Cl+⎯⎯⎯ H O→++ HO 2Cl H =× [26]  222 2 k 31 4.1 10 M s dCl2 =−k Cl•−  H O dt 31 2 2 2 dHO 22 =−k Cl•−  H O dt 31 2 2 2 • dHO2 = k Cl•−  H O dt 31 2 2 2 dCl− = 2ClHOk •−  dt 31 2 2 2 dH+ = k Cl•−  H O dt 31 2 2 2 ••−+ R Cl+⎯⎯⎯ H Ok32 →++ HO Cl H =×911−− [26] • 37 22 2 k 32 1.1 10 M s dCl =−k Cl•  H O dt 32 2 2 dHO 22 =−k Cl•  H O dt 32 2 2 • dHO2 = k Cl•  H O dt 32 2 2 dCl− = k Cl•  H O dt 32 2 2 + dH • = k Cl H O dt 32 2 2 •− • − + k33 911−− •− R38 Cl+⎯⎯⎯ HO→+ O 2Cl + H =× [26]  22 2 k 33 3.0 10 M s dCl2 =−k Cl•− HO • dt 33 2 2 • dHO2 =−k Cl•− HO • dt 33 2 2 dCl− = 2ClHOk •− • dt 33 2 2

554 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017)

Table 2. Continued

No. Reaction Parameters and rate constants Ref. Kinetic equations •− • − + k33 911−− + R38 Cl+⎯⎯⎯ HO→+ O 2Cl + H k =×3.0 10 M s [26]  22 2 33 dH •− • = k Cl HO dt 33 2 2 •− •− − k34 911−− •− R39 Cl+⎯⎯⎯ O→+ O 2Cl =× [26]  22 2 k 34 2.0 10 M s dCl2 =−k Cl•− O •− dt 34 2 2 •− dO2 =−k Cl•− O •− dt 34 2 2 − dCl •− •− = 2ClOk  dt 34 2 2

•− k35 R40 CCO+⎯⎯⎯→+ DOC? k35 – dC 3 =−k CCO•− dt 35 3 •− dCO3 =−k CCO•− dt 35 3

dDOC •− = k CCO dt 35 3

•− k36 R41 CSO+⎯⎯⎯→+ DOC? k36 – dC 4 =−k CSO•− dt 36 4 •− dSO4 =−k CSO•− dt 36 4

dDOC •− = k CSO dt 36 4

• k37 R42 CHClO+⎯⎯⎯→+ DOC? k37 – dC 2 =−k CHClO• dt 37 2 • dHClO2 =−k CHClO• dt 37 2

dDOC • = k CHClO dt 37 2

• k38 R43 C+⎯⎯⎯ HClO→+ DOC ? k38 – dC =−k CHClO• dt 38  d HClO• =−k CHClO• dt 38 

dDOC • = k C HClO dt 38 

• k39 R44 CCl+⎯⎯⎯→+ DOC? k39 – dC =−k CCl• dt 39  • dCl • =−k CCl dt 39 

dDOC • = k CCl dt 39  •− +⎯⎯⎯k40 →+ R45 CCl DOC? k40 – dC •− 2 =−k CCl dt 40 2 •− dCl2 •− =−k CCl dt 40 2

dDOC •− = k CCl dt 40 2

555 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017)

Table 2. Continued

No. Reaction Parameters and rate constants Ref. Kinetic equations −+ * k41 10 -1 * R46 HCO⎯⎯⎯→+ HCO H k =×1.0 10 s [28]  23 3 41 dHCO23 =−k HCO* =×3-1-1 41 2 3 k 42 4.5 10 M s dt = − p6.3K a d HCO3 = k HCO* dt 41 2 3 dH+ = k HCO* dt 41 2 3 −+ k42 * − R47 HCO+⎯⎯⎯ H→ H CO  323 dHCO3 =−k HCO−+ H dt 42 3 dH+ =−k HCO−+ H dt 42 3 * dHCO23 −+ = k HCO H dt 42 3

−−+k43 2 10 -1 − R48 HCO⎯⎯⎯→+ CO H k =×1.0 10 s [28]  33 43 dHCO3 =−k HCO− =×−1-1-1 43 3 k 44 4.5 10 M s dt 2− pK = 10.3 dCO3 a = k HCO− dt 43 3 dH+ = k HCO− dt 43 3 −+ − 2 k44 2− R49 CO+⎯⎯⎯ H→ HCO  33 dCO3 =−k CO2−+ H dt 44 3 dH+ =−k CO2−+ H dt 44 3 − d HCO3 = k CO2−+ H dt 44 3

−−+k45 2 10 -1 − R50 HSO⎯⎯⎯→+ SO H k =×1.0 10 s [26]  44 45 dHSO4 =−k HSO− =×11 -1 -1 45 4 k 46 4.510 M s dt 2− p1.9K = dSO4 a = k HSO− dt 45 4 dH+ = k HSO− dt 45 4 −+ − 2 k46 2− R51 SO+⎯⎯⎯ H→ HSO  44 dSO4 =−k SO2−+ H dt 46 4 dH+ =−k SO2−+ H dt 46 4 − dHSO4 = k SO2−+ H dt 46 4

species considered in the particular reaction Ri were Considering the reactions, constants and para- considered; while such as species were obviated if meters summarized in Tables 1 and 2, the degrad- δ = 0 . A more detailed explanation about how the ation of the pollutant can be estimated as indicated in R pHi changes in the solution were addressed has been Eq. (1): reported in Rubio-Clemente et al. [29]. Moreover, the dC •• ODE set of the chemical species involved in the sys- −=ϕ Ik +HO C + k  HO C + dt C a,C 14  15  2 tem can also be consulted in Rubio-Clemente et al. 40 (1) •− [29]. ++kOCCA k  R  16 2  ii   i =35

556 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017) where: batch cylindrical photoreactor made on Pyrex glass. •• The considered UV-light intensity (I ) and illuminated ϕ Ik,  HO C , k  HO  C , 0 C a,C 14 15  2  path length (l) were 1.516×10-6 Ein L-1 s-1 (radiation of 40 •− 254 nm > 90% and power = 15 W) and 63.5 mm, res- kkROC andCA   16 2  ii   pectively [16]. The adopted quantum yield and molar i =35 extinction coefficient of phenol (φPhenol and εPhenol ) correspondingly refer to the specific contributions of were 0.07 mol Ein-1 [30] and 51600 M-1 m-1 [16], res- • • •- UV radiation, and the oxidation of HO , HO2 , O2 and • • • pectively. In turn, the assumed values of the kinetic - - • • the formed anions radicals (AR) (CO3 , SO4 , H2ClO , - • • •- rate constants of phenol with HO (k14), O2 (k16) and HClO , Cl and Cl2 ) to the overall pollutant deg- CO •- (k ) were respectively 6.6×109, 5.8×103 and radation. 3 35 2.2×107 M-1 s-1 [31-33]. Additionally, a phenol working For examining the role of the terms mentioned solution of 2.23×10-3 M prepared in deionized water above, four parameters (d, f, g and h) were included was selected and the effect of different H2O2/phenol in Eq. (1), resulting in Eq. (2): ratios (495, 228 and 125), obtained by varying the ini-

dC •• tial level of H2O2 and keeping the concentration of −=ϕ Ik +HO C dk +  HO  C f + dt Ca,C 14 15  2  phenol constant [16], were studied. 40 (2) •− Model hypotheses ++kgkRhOC  CA  16 2  ii   i =35 Since the reaction water used in the UV/H2O2 experiments was deionized water [16], the con- Consequently, for evaluating the pollutant rem- tribution of SO •-, H ClO•, HClO•, Cl• and Cl •- were oval by the exclusive action of UV radiation, d = f = 4 2 2 not included in the study of phenol conversion. There- = g = h = 0 (i.e., the initial level of oxidant radical spe- • fore, among the anions considered in the model, only cies is equal to zero). In turn, HO action was con- - 2- HCO3 and CO3 were studied. In this sense, to sidered when d = 1 and f = g = h = 0. For studying the •- • • examine the contribution of CO (due to the reaction contribution of HO , O - and other anion radicals 3 2 2 between HO• and HCO - or CO2-), along with the included in Table 2, f, g and h were equaled to 1, 3 3 photolysis and the oxidation of the pollutant through respectively. HO•, d = h = 1 and f = g = 0. On the other hand, to Since the mathematical model considers the analyze the role of HO • in the pollutant degradation, contribution of different anions, such as CO 2-, HCO -, 2 3 3 as well as HO• and CO •-, d = f = h = 1 and g = 0. In SO 2- and Cl-, in the pollutant conversion, it might be 3 4 turn, when d = f = g = h = 1, and no inorganic anions used for the treatment of water with salt content. This different from HCO - and CO 2- are present in the stu- enables the applicability of the developed kinetic 3 3 died water, in addition to the species mentioned model to be expanded depending on the character- above, the kinetic model considers the action of O •-. istics of the water of interest to be treated. 2 On the other hand, and taking into account the Numerical analysis high oxidation potential and kinetic reaction rate • In order to obtain the numerical solution of the constant, HO is generally the radical developing the proposed kinetic model for the degradation of organic main action in pollutant conversion [6,14,17,23,34- –37]. Therefore, a special attention to the evolution of pollutants using the UV/H2O2 system, MATLAB soft- its concentration with time must be paid. In this vein, ware and ODE15S function were used. Additionally, • the differential rate equations representing the con- the evolution of HO level during the reaction time can centration evolution of the pollutant throughout the be expressed as indicated in Eq. (3): reaction time were plotted. • dHO =+−22ϕδϕδII−− HO22 a,HO 22RR 3HO a,HO 4 RESULTS AND DISCUSSION dt 22 −−kkHO HO••δδ HO  HO − 522RR 7 522   8 Taking into account the reactions compiled in −−−kkHO•− HO  HO •• HO Table 1 and 2, as well as the parameters and the rate 627  (3) constants involved in each chemical and photochem- −−+kkHO•• HO  HO ••− O 8292  ical reaction, a kinetic model was developed. The kin- ++kkHO HO••− HO  O − etic model was validated using the experimental 11 2 2 2 12  2 2  2  values from Alnaizy and Akgerman [16], where phe- −−••  − kk14CHO 17  DOCHO  nol was used as the probe compound and the UV/  

/H2O2 process was carried out in a completely mixed

557 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017)

2−• −• -1 -1 −−kkCO  HO HCO  HO − ficient (ԑDOC(280 nm)) of 35967 M m [38] since this 18 3  19 3  parameter was not considered in Alnaizy and Akger- −−−kkHSO−•  HO Cl −• HO 23 4  26 man study [16]. −k Cl•− HO • Finally, the pH variation during the reaction time 30 2 was included in the proposed kinetic model; as Nevertheless, since HO• is a highly reactive spe- according to Alnaizy and Akgerman [16] the pH dec- cies, it can be assumed that HO• level change during reases as the oxidation process proceeds. In this time is negligible. Therefore, HO• concentration can vein, those kinetic reactions significantly dependent be expressed as Eq. (4): on the pH of the bulk, such as R3, R4, R7 and R8, were

• promoted and discriminated as explained in Rubio- HO=++ (2ϕδϕδII 2 −− HO22 a,HO 22RR 3HO a,HO 4 Clemente et al. [29].  22 ++kkHO HO••− HO  O )/ 11 2 2 2 12  2 2  2 Degradation of phenol by direct photolysis and UV/H O /(kkk HOδδ+++ HO HO− 2 2 522752286RR   2 (4) Considering the previously mentioned hypo- ++kkkkHO••− O +++ C DOC 8 2 9 2 14 17   theses and in order to study the effect of UV light in −−− phenol conversion, parameters d, f, g and h from Eq. ++kkCO2  HCO  + kHSO+ 18 3 19  3 23 4 (2) were equaled to 0. Thus, Eq. (2) is reduced to Eq. +ClCl)kk−•−+   (6): 26 30  2  dC Considering that the oxidant is in a high level, −=ϕ I (6) dt Ca,C due to the high H2O2/phenol ratios studied (i.e., δδ+>> • kk52275228HORRii  HO  kX , where: Additionally, for examining the effect of HO and •- −••− CO , d and h were equaled to 1, and f and g, to 0. kX=+++ kHO k HO k O 3  ii 628292 Therefore, Eq. (2) is simplified as Eq. (7): 2−− ++kkCDOCCO   + k + k HCO  + 40 14 17   18 3 19  3  dC • −=ϕ Ik +HO C + k C A R (7) −−•− Ca,C 14  ii +++ kkkHSO ClCl dt i =35 23 4 26  30 2 and In Figure 1, the experimental data and those ones calculated by the proposed kinetic model con- 22ϕδϕδII+>>−− HO22 a,HO 22RR 3HO a,HO 4 cerning the transformation of phenol by direct photo- 22 lysis and the contribution of HO• and CO •- for >>••− +  3 ()kk11HO 2 2 HO 2 12  HO 2 2 O 2 -3    2.23×10 M phenol and a H2O2/phenol ratio of 495 for a reaction time of 220 min are illustrated. Eq. (4) can be simplified to Eq. (5). k and X are i i From the figure, it can be observed that model • the rate constants between HO and species i, and predictions are in a good agreement with experimen- the concentration of species i, respectively: tal data both related to phenol direct photolysis alone

22ϕ IIδ + ϕ −−δ and indirect oxidation with a correlation factor of HO22 a,HO 22RR 3HO a,HO 4 HO• = 22 (5) 98.25 and 99.34%, respectively. kkHOδδ+ HO 52275228RR  Additionally, it can be found that phenol initial level was reduced about 40% after 220 min of treat- Additionally, the initial concentration of DOC dif- ment considering the exclusive effect of UV radiation. ferent from the DOC provided by phenol was assumed This can be explained by the photolytic activity of to be equal to 0. phenol under the irradiation wavelength used in the In agreement with several authors [9,14,37], the studied AOP. However, higher removal of phenol is decrease in DOC by means of direct photolysis (R 22 obtained when H O is added to the system, resulting from Table 2) was obviated. 2 2 in more than 90% of removal, primarily due to the As the process proceeds, DOC from the photo- action of HO•. Between HO• and CO •-, the main role oxidation of phenol and the intermediates can impair 3 in removing phenol was developed by HO•, due to the phenol transformation since it can absorb UV-light rate constant between phenol and CO •- is compar- and react with reactive oxygen species (ROS), such 3 • atively lower than the rate constant between phenol as HO . and HO•; and because the CO •- level was found to With regard to the amount of light quenched by 3 be in a concentration in the range of 10-15 M. This fact DOC, it was assumed a DOC molar extinction coef-

558 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017)

Figure 1. Phenol degradation predicted data (lines) versus experimental data (symbols). Operating conditions: -3 [phenol]0 = 2.23×10 M; H2O2/phenol = 495; t = 220 min.

•- was determinant in obviating the action of CO3 in 1,3-benzenediol and 1,4-benzenediol, corresponding transforming phenol for the studied working con- to catechol, resorcinol and hydroquinone, respect- ditions. ively, were reported [16]. Additionally, p-benzoqui- • •- On the other hand, the effect of HO2 and O2 none was found among phenol degradation by-pro- was investigated. It was observed that both species ducts [16]. Once these intermediates are generated, were in the system in the range of 10-7 M; a very the oxidation action of the oxidant species is con- larger concentration in comparison with HO• (∼10-14 tinued, especially that of HO•, as explained above. Con- M). Nevertheless, as the pH of the reaction medium sequently, organic acids, like maleic, formic and oxa- •- decreases, the amount of O2 is reduced. Addition- lic acid, among others, are yielded, resulting in a dec- ally, the reaction rate constant between phenol and rease of the pH in the bulk, as indicated previously. this radical is 5.8×103 M-1 s-1 [32]. Therefore, the con- •- Influence of initial concentration of H2O2 tribution of O2 in phenol degradation can be neglected. Keeping constant the initial level of phenol • (2.23×10-3 M), the effect of H O initial concentration, With regard to the action of HO2 in converting 2 2 • considering H O /phenol mass ratios of 125, 228 and phenol, phenol-HO2 kinetic reaction rate constant 2 2 must be known. For this purpose, the kinetic reaction 495, is represented in Figure 2, where the expe- rate constant calculated in Rubio-Clemente et al. [29] rimental data and the predictions from the proposed was used, which was equal to 1.6×103 M-1 s-1, con- kinetic model are compared. It can be observed that 3 -1 -1 sistent with literature data ((2.7±1.2)×10 M s ) [39]. the optimal H2O2/phenol ratio causing a transform- Including this value in the ODE set of the proposed ation of phenol higher than 95% after 40 min of treat- kinetic model, the degree of agreement between the ment is 125, which corresponds to Alnaizy and Akger- experimental and the proposed data was calculated man findings [16]. A further increase of H2O2/phenol to be 99.57%. In this vein, it was found that the cor- ratio from 125 to 228 results in a reduction of phenol relation factor of the model slightly increases in degradation, which is magnified when H2O2/phenol • ratio is equal to 495. 0.23%. Therefore, although the amount of HO2 is far • greater than the amount of HO , in this case its action It is important to note that in the UV/H2O2 sys- in the UV/H O system for transforming phenol can be tem, when H2O2 is irradiated with a wavelength of 254 2 2 • obviated; probably due to the low reactivity of this nm, HO is produced according to R3 from Table 1, radical with the pollutant. Therefore, the main role in which exhibits a high oxidation potential [12] and phenol conversion for the considered experimental attacks the pollutant, leading to a higher phenol rem- conditions was developed by HO•. oval in a shorter reaction time. Therefore, it could be During phenol oxidation, different reaction inter- thought than an increase in H2O2 level would lead to a mediates are formed; among them 1,2-benzenediol, rise in phenol conversion. However, two opposing

559 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017)

Figure 2. Phenol degradation predicted data (lines) versus experimental data (symbols). Operating conditions: -3 [phenol]0 = 2.23×10 M; H2O2/phenol = 495 (—), 228 (- • -), 125 (∙∙∙∙); t = 220 min.

effects occur. A rise in the initial level of H2O2 results The detrimental effect of an excess of H2O2 is in a higher production of HO•, which is available to supported by Figure 3, representing the time-evol- • attack the organic pollutant. Nevertheless, when H2O2 ution profiles of the HO level for the H2O2/phenol is in excess, it acts scavenging HO•, as indicated in ratios studied. From the figure, it can be observed that reaction R7 from Table 1. Furthermore, even though a the ratio providing the lowest and the highest concen- • high H2O2 concentration can generate a great number tration of HO throughout the reaction time is 495 and • of HO2 , they have a lower oxidation potential in com- 125, respectively. These results are coincident with • • • parison with HO . Moreover, HO2 can react with HO those findings illustrated in Figure 2, where the H2O2/ • (R11 from Table 1). In addition, when HO is in excess, /phenol ratio providing the lowest and the highest it can suffer recombination (R10 from Table 1), red- phenol conversion extent was, correspondingly, 495 ucing the number of HO• available in the bulk. Hence, and 125. the overall efficiency of the applied AOP in removing Consequently, and in order to save in operating phenol, and eventually leading to its mineralization is costs, H2O2 optimal level achieving an efficient pol- decreased. lutant removal must be found. Nonetheless, this can

• -3 Figure 3. Estimated HO time-profile. Operating conditions: [phenol]0 = 2.23×10 M;

H2O2/phenol = 495 (—), 228 (- • -), 125 (∙∙∙∙); t = 220 min.

560 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017) be an arduous and expensive task. In this sense, the l optical path length of the photoreactor (mm) proposed kinetic model allows calculating the optimal ODE ordinary differential equations ϕ concentration of the oxidizing agent to be used when x quantum yield of the photolysis of a -1 implementing the UV/H2O2 system for an efficient species (mol Ein ) phenol removal without incurring high costs related to Ri (photo)chemical reaction the use of an unnecessary amount of oxidant. ROS reactive oxygen species t irradiation time (min) CONCLUSION UV ultraviolet [X] concentration of a species X A kinetic model for pollutant degradation by the Acknowledgements UV/H2O2 process was developed. The model predictions were checked with experimental results from The authors thank the Colombian Administrative Alnaizy and Akgerman work [16] for phenol direct Department of Science, Technology and Innovation photolysis alone and in combination with indirect oxid- (COLCIENCIAS) for its financial support. ation at different H2O2/phenol ratios. It was found that the proposed kinetic model fits well with the experi- REFERENCES mental data, which is a probe of the accuracy of the model. [1] R. Zhang, Y. Yang, C.H. Huang, L. Zhao, P. Sun, Water Res. 103 (2016) 283-292 Additionally, the model provides an understand- [2] M.D. Murcia, N.O. Vershinin, N. Briantceva, M. Gomez, ing of the impact of the initial level of H2O2 being able E. Gomez, E. Cascales, A.M. Hidalgo, Chem. Eng. J. 266 to optimize this parameter. In fact, it was observed (2015) 356-367 that the optimal H O concentration enabling a rem- 2 2 [3] A. Rubio-Clemente, R.A. Torres-Palma, G.A. Peñuela, oval of the pollutant higher than 95% after 40 min of Sci. Total Environ. 408 (2014) 201-225. treatment corresponded to a H2O2/phenol ratio of 125. • [4] A. Rubio-Clemente, E. Chica, G.A. Peñuela, Ambiente The action of HO2 in transforming phenol was also Agua 8 (2013) 93-103 studied due to the high concentration of these species [5] M. Dopar, H. Kušić, N. Koprivanac, Chem. Eng. J. 173 • -7 formed during the process compared to HO (∼10 M (2011) 267-279 -14 >> ∼10 M). It was found that, although the degree of [6] N. Daneshvar, M.A. Behnajady, M. Mohammadi, M.S. agreement of the correlation factor was increased Seyed, Desalination 230 (2008) 16-26 from 99.34 to 99.57%, the radical developing the [7] J. Michalowicz, W. Duda, Pol. J. Environ. Stud. 16 (2007) main action in phenol abatement under the operating 347-362 • conditions was HO . [8] N.S. Gad, M.S. Saad, Glob. Veter. 2 (2008) 312-319 The developed kinetic model could be of rel- [9] J.C. Crittenden, S. Hu, D.W. Hand, S.A. Green, Water evance in the optimization of the operating conditions Res. 33 (1999) 2315-2328 when implementing the UV/H2O2 process, allowing [10] M. Litter, N. Quici, Recent Pat. Eng. 4 (2010) 217-241 saving in costs and time. [11] B.A. Wols, C.H.M. Hofman-Caris, Water Res. 46 (2012) 2815-2827 Nomenclature [12] J. Saien, H. Nejati, J. Hazard. Mater. 148 (2007) 491-495 AOP advanced oxidation process [13] D.B. Hasan, A.R.A. Aziz, W.M.A.W. Daud, Chem. Eng. AR anion radicals Res. Design 90 (2012) 298-307 C target compound [14] W. Song, V. Ravindran, M. Pirbazari, Chem. Eng. Sci. 63 E oxidation potential (V) (2008) 3249-3270 ε X molar extinction coefficient of a species X [15] C. Comninellis, A. Kapalka, S. Malato, S.A. Parsons, I. (M-1m-1) Poulios, D. Mantzavinos, J. Chem. Technol. Biotechnol. DOC dissolved organic carbon 83 (2008) 769-776 DOM dissolved organic matter [16] R. Alnaizy, A. Akgerman, Adv. Environ. Res. 4 (2000) δ 233-244 Ri correction factor [17] P. Kralik, H. Kušić, N. Koprivanac, A.L. Božić, Chem. f x fraction of the UV radiation absorbed by a species Eng. J. 158 (2010) 154-166 I intensity of the UV radiation absorbed by a [18] H.J. Bielski, H.J. Benon, D.E. Cabelli, L.A. Ravindra, A.B. a,x Alberta, J. Phys. Chem. Ref. Data 14 (1985) 1041-1100 species (Ein L-1 s-1) -1 -1 [19] G.V. Buxton, C.L. Greenstock, W.P. Helman, A.B. Ross, I0 incident UV radiation intensity (Ein L s ) -1 -1 -1 J. Phys. Chem. Ref. Data 17 (1988) 513-886 ki second order (M s ) or first order (s ) reaction rate constant

561 A. RUBIO-CLEMENTE et al.: KINETIC MODEL DESCRIBING THE UV/H2O2… Chem. Ind. Chem. Eng. Q. 23 (4) 547−562 (2017)

[20] H.S. Christensen, K. Sehested, H. Corftizan, J. Phys. [30] T. Alapi, A. Dombi, J. Photochem. Photobiol., A 188 Chem. 86 (1982) 15-88 (2007) 409-418 [21] K. Sehested, O.L. Rasmussen, H. Fricke, J. Phys. Chem. [31] L.M. Dorfman, G.E. Adams, Nati. Bur. Stand. Technol. 46 72 (1968) 626-631 (1973) [22] J. Weinstein, H.J. Benon, H.J. Bielski, J. Am. Chem. Soc. [32] T. Yasuhisa, H. Hideki, Y. Muneyoshi, Int. J. Biochem. 25 101 (1979) 58-62 (1993) 491-494 [23] B.A. Wols, D.J.H. Harmsen, E.F. Beerendonk, C.H.M. [33] S.N. Chen, M.Z. Hoffman, Radiat. Res. 56 (1973) 40-47 Hofman-Caris, Chem. Eng. J. 255 (2014) 334-343 [34] H. Yao, P. Sun, D. Minakata, J.C. Crittenden, C.H. Huang, [24] Z.D. Draganic, A. Negron-Mendoza, K. Sehested, S.I. Environ. Sci. Technol. 47 (2013) 4581-4589 Vujosevic, R. Navarro-Gonzales, M.G. Albarran-Sanchez, [35] H. Kušić, N. Koprivanac, A.L. Božić, S. Papić, I. Peternel, I.G. Draganic, Radiat. Phys. Chem. 38 (1991) 317-321 D. Vujević, Chem. Biochem. Eng. Q. 20 (2006) 293-300 [25] P. Neta, R.E. Huie, A.B. Ross, J. Phys. Chem. Ref. Data [36] M. Edalatmanesh, R. Dhib, M. Mehrvar, Int. J. Chem. 17 (1988) 1027-1284 Kinet. 40 (2007) 34-43 [26] J. De Laat, G.T. Le, B. Legube, Chemosphere 55 (2004) [37] W.T.M. Audenaert, Y. Vermeersch, S.W.H. Van Hulle, P. 715-723 Dejans, A. Dumoulin, I. Nopens, Chem. Eng. J. 171 [27] G.D. Fang, D.D. Dionysiou, Y. Wang, S.R. Al-Abed, D.M. (2011) 113-126 Zhou, J. Hazard. Mater. 227-228 (2012) 394-401 [38] J. Peuravuori, K. Pihlaja, Anal. Chim. Acta 337 (1997) [28] P. Mazellier, É. Leroy, J. De Laat, B. Legube, New J. 133-149 Chem. 26 (2002) 1784-1790 [39] Z. Kozmér, E. Arany, T. Alapi, E. Takács, L. Wojnárovits, [29] A. Rubio-Clemente, E. Chica, G.A. Peñuela, in Waste- A. Dombi, Radiat. Phys. Chem. 102 (2014) 135-138. water Treatment and Resource Recovery, R. Farooq, Ed., InTech Open, Rijeka, In press

1 AINHOA RUBIO-CLEMENTE KINETIČKI MODEL UV/H2O2 FOTODEGRADACIJE EDWIN CHICA2 FENOLA IZ VODE GUSTAVO A. PEÑUELA3

1 Departamento de Ciencia y Razvijen je i proveren kinetički model za konverziju fenola u UV/H2O2 sistemu. Model • Tecnología de los Alimentos. Facultad uključuje razgradnju zagađujućih materija direktnom fotolizom i oksidacijom sa HO , • • • 2- - 2- - de Ciencias de la Salud. Universidad HO2 i O2 , hvatanje HO pomoću, CO3 , HCO3 , SO4 i Cl , kao i promenu pH za vreme Católica San Antonio de Murcia UCAM, procesa. Takođe, razmotreno je štetno delovanje organskih materija i intermedijera s/n. Guadalupe-Murcia, España reakcije u UV zaštiti i smanjenju HO• fluorescencije. Uočeno je da model može precizno 2Departamento de Ingeniería predvideti smanjenje fenola korišćenjem različitih masenih odnosa H2O2/fenol (495, 228 Mecánica, Facultad de Ingeniería, i 125). Model definiše optimalni odnos H2O2/fenola od 125 sa uklanjanjem fenola više od Universidad de Antioquia UdeA, • 95% nakon 40 min tretmana, ukoliko HO je glavna oksidaciona vrsta. Razvijeni model Medellín, Colombia može biti relevantan za izračunavanje optimalnog nivoa H O za efikasnu degradaciju 3Grupo GDCON, Facultad de 2 2 zagađivača od interesa, čime se znatno štedi na vremenu i troškovima. Ingeniería, Sede de Investigaciones

Universitarias (SIU), Universidad de Ključne reči:. nivo H2O2, kinetički model, zagađenje fenolom, UV/H2O2. Antioquia UdeA, Medellín, Colombia

NAUČNI RAD

562 Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 563−571 (2017) CI&CEQ

NINA TRUPEJ INVESTIGATION OF THE THERMODYNAMIC MAŠA KNEZ HRNČIČ PROPERTIES OF THE BINARY SYSTEM MOJCA ŠKERGET VITAMIN K3/CARBON DIOXIDE ŽELJKO KNEZ

University of Maribor, Faculty of Article Highlights Chemistry and Chemical • Solubility of vitamin K3 in CO2 as a function of T and P was measured Engineering, Maribor, Slovenia • Data at temperatures of 40, 60 and 80 °C and pressures up to 40 MPa are given • Partial molar volumes were determined by a vibrating tube densitometer • SCIENTIFIC PAPER Partial molar volumes of vitamin K3 in CO2 are large and negative • Solubility data correlation is performed by a density-based model UDC 615.356:546.264-31:544.3:66 Abstract

The binary system of vitamin K3 and CO2 was investigated at temperatures of 40, 60 and 80 °C up to a pressure of 40 MPa. Solubility was measured by a static-analytic method. Partial molar volumes were determined by a method

involving a vibrating tube densimeter. The solubility of vitamin K3 in CO2 is found as a function of pressure and temperature. The highest solubility (31.16×10-4 mol∙mol-1) was attained at a pressure of 25.40 MPa at temperature of 40 °C. However, at temperature of 60 °C and a pressure of 24.02 MPa, the solubility was 18.79×10-4 mol∙mol-1. Solubility was lower at a temperature of 80 °C and a pressure of 22.06 MPa (6.48×10-4 mol∙mol-1). The partial molar

volumes are negative and the dissolved vitamin K3 has a minor impact on the density of the solution of K3 in CO2 compared to the density of the pure CO2. Keywords: solubility, density, partial molar volume, food industry, food processing.

Processes with sub- and supercritical fluids with defined physical shape [3]. Chemical compounds have high potential and have been researched for can be processed by the use of supercritical fluids; for applications in different industry branches, like the example, obtaining fine powders of crystalline mat- food industry [1]. One of the major advantages of erial with micronization processes like PGSS (par- applying supercritical fluids in food industry is the pos- ticles from gas saturated solutions). Phase equilib- sibility of processing fine products without traces of rium data is crucial for planning processes involving harmful organic solvents which need no further pur- supercritical fluids, like PGSS and also extraction, ification. Moreover, supercritical fluids are considered chromatography, chemical reactions and chemical as environmentally safe and are not harmful to food engineering [4], therefore it is important to investigate components since they allow processing at lower these data (for example solubility in the supercritical temperatures [2]. fluid, mixture density, partial molar volumes, compres- 60% of products that are sold by chemical com- sibility, dielectric constant etc.). These properties are panies are crystalline, polymeric or amorphous solids defined as strong functions of temperature and pressure, especially near the critical conditions. The critical parameter that controls the feasibility Correspondence: Ž. Knez, University of Maribor, Faculty of of the process is the solubility of the substance in the Chemistry and Chemical Engineering, Smetanova 17, SI-2000 supercritical fluid. Due to some hindrances consider- Maribor, Slovenia. E-mail: [email protected] ing the models predicting the solubilities of solids in Paper received: 4 August, 2016 supercritical fluids, there is an essential need to dev- Paper revised: 23 January, 2017 elop new methods to determine these data experi- Paper accepted: 1 March, 2017 mentally [5]. In the literature, reviews with thermodyn- https://doi.org/10.2298/CICEQ160804009T

563 N. TRUPEJ et al.: INVESTIGATION OF THE THERMODYNAMIC PROPERTIES… Chem. Ind. Chem. Eng. Q. 23 (4) 563−571 (2017) amic data of compound solubilities in supercritical Determination of the vibration period and sampling fluids (SCF) are available [3,6], as well as the solu- procedure bilities of vitamins in supercritical fluids. Before each measurement the entire apparatus The knowledge of thermodynamic properties of was properly cleaned with ethanol and pressurized an investigated system is fundamental from the eco- with CO2 and N2. The autoclave was filled with a suf- nomic point of view, since the solubility is the critical ficient amount of vitamin K3, approximately 5 g, con- parameter that controls the feasibility of the process. nected to the densimeter (Figure 1a). Two different authors report solubility of vitamin After the components were connected, the ent- K3 in CO2 [7,8]. In the present work, in addition to the ire system was heated to the desired temperature by solubility of the vitamin K3 in carbon dioxide, the par- the thermostatic baths and electric wires. Mixing was tial molar volumes and the density of the vitamin dis- performed using a magnetic stirrer. After opening the solved in the supercritical fluid are presented, which valves, the hand pump was set to the minimum vol- to our knowledge has not yet been investigated. ume and the whole system was depressurized with a Despite the importance of these data, there is vacuum pump. Then, CO2 was introduced into the still a lack of knowledge on the interactions between system to access the desired pressure. After a one- the solute and solvent. Interactions between mole- hour wait for equilibrium, the stirring was stopped and cules are described by partial molar volumes and are an additional hour was required for the sedimentation important for applications of supercritical fluid tech- of vitamin K3 as it was experimentally determined. nology in industry [9]. Valve V2 was then closed and the hand pump (Nova Swiss piston pump) was turned to the direction of its EXPERIMENTAL maximal volume, which enabled the transport of the saturated carbon dioxide with vitamin K to the den- Materials 3 simeter. A pressure drop occurred. Nitrogen with purity 99.999 wt.% and carbon The equipment and the method are described in dioxide with purity 99.995 wt.% were obtained from literature [9], but due to the pressure drop and in

Messer (Slovenia). Vitamin K3 (Menadione) with cat- order to ensure saturation of CO2 with vitamin K3, a alogue number CAS 58-27-5 and absolute ethanol modification of the apparatus and measuring proce- (≥99.8 wt.%) were obtained from Sigma-Aldrich. Milli-Q dure was made. water was used for the calibration of the equipment. When valve V1 was closed, valve V2 was The substances were used without further purific- opened and the hand pump was turned back to the ation. Vitamin K3 and Milli-Q water were degassed beginning position to its minimal volume and the fluid before application. with saturated vitamin K3 was returned back into the Methods autoclave to the starting conditions (Figure 1a). Then, the stirring was restarted, the equilibrium was estab- Vibrating tube densimeter lished and the circulation of the mixture was repeated. A vibrating tube densimeter was used for den- After the third equilibrium establishment and fluid sity determination. The main part of the densimeter is circulation, the vibration period of the fluid saturated the Anton Paar DMA 512 unit, which is connected to with vitamin K3 was measured with the DMA 60 unit. the DMA 60 unit. The interior of the DMA 512 unit was With further equilibrium establishment the mole frac- thermostated with a thermostatic bath. Inside the tion of vitamin K3 in the mixture was constant. Fluid DMA 512 unit the U-tube was thermostated with a cir- density was calculated from the vibration period. After -3 culating water bath with an accuracy of ±5×10 K and determining the vibration period, the valves V6 and the temperature inside the U-tube was measured with V7 were closed (the pressure increased for appro- a GTH 1150 digital thermometer NiCr-Ni with accur- ximately 0.2 MPa), the pipeline between the valves acy 1%. The autoclave and the U-tube were equipped V2 and V7 was removed and a sampling system was with a manometer Nuova Fima EN837-1 with an connected (Figure 1b). About 0.1 mL sample was accuracy of 0.25% for pressures lower than 60 MPa. taken through the sampling valve V7 into a trap with The apparatus was validated by measuring sol- ethanol where the vitamin was solubilised. The exp- ubility of systems of substances for which data is anding valve was purged with 5-7 mL of ethanol. The available in literature. Experimental data was com- amount of CO2 was measured with a disposal of water pared to the literature and the deviation was lower in a graduated cylinder. The content of vitamin K3 in the than 3%. Some results were already published by the sample was determined with a Varian 50 probe UV/ thermodynamic part of our research group [9]. /Vis spectrophotometer at a wavelength of 251 nm.

564 N. TRUPEJ et al.: INVESTIGATION OF THE THERMODYNAMIC PROPERTIES… Chem. Ind. Chem. Eng. Q. 23 (4) 563−571 (2017)

Figure 1. Apparatus for determination of solubilities and densities of supercritical fluid/solid mixtures; a) equilibrium establishment, b) sampling: TI, temperature indicator, PI, pressure indicator, V, valve, V3, saftety valve, V6, and V7, sampling valves, A ,autoclave, CP, chromatographic pump, SP, syrine pump, H, heating, DMA 512, apparatus with the vibrating tube, DMA 60, measuring unit, T, thermostat.

Density determination with a vibrating tube ρρ− K = 12 (2) ττ22− densimeter 12 The most recent method for the determination of The main disadvantage of this calibration proce- partial molar volume is the densimeter [10,11]. A vib- dure is that it is time consuming. Moreover, it is rec- rating tube densimeter is basically a spring-mass sys- ommended that the difference between the density tem in which the frequency of vibration of the tubing is fluids is lower than 500 kg∙m-3 and in cases when the measured (vibration period) and related to the fluid densities of the reference fluids are not “well known” density. Eq. (1) was applied for calculating the density this is not the appropriate calibration method [12]. (ρ/kg∙m-3): Therefore, new calibration processes with important ρ =−+ττ22ρ K ()11 (1) limitations were developed to simplify the procedure, as reported by the authors [13]. where K (kg∙s2∙m-1) is the characteristic parameter of In the present research, the reference fluids were the U-tube, τ (s-1) the vibration period of the U-tube Milli-Q water and nitrogen. The density data at differ- when filled with a fluid of an unknown density, ρ1 ent temperatures and pressures for calibration fluids -3 -1 (kg∙m ) is the density of reference fluid 1 and τ1 (s ) are well known and were obtained from literature [14]. the vibration period of the U-tube is filled with refer- Solubility maximum ence fluid 1. Previously, the characteristic parameter K has to The existence of a maximum concentration of be determined by Eq. (2) at each temperature and solids in supercritical fluids was observed before pressure by measuring the vibration period of two ref- [6,15–17], and is of great importance, especially when erence fluids of known densities: employing supercritical fluids for extraction of solid

565 N. TRUPEJ et al.: INVESTIGATION OF THE THERMODYNAMIC PROPERTIES… Chem. Ind. Chem. Eng. Q. 23 (4) 563−571 (2017)

compounds [16]. The concentration extrema of solid- were determined for vitamin K3 before and after exp-

–fluid phase equilibria can be derived as: osure to CO2 with a Mettler Toledo DSC 20 apparatus

p with a temperature step of 5 K/min as: gs*g=−1 ln y 2m,22(VVp )d (3) ΔH RT  cr ()%= 100m2 (6) p * Δ 2 Hm1 where y g is the mole fraction of the low-volatility 2 The crystallinity degree of vitamin K3 before and component 2 in the fluid phase [16]. The solubility Δ after exposure to CO2 was compared. Hm1 is the extremum (minimum and maximum) is a function of enthalpy of fusion of vitamin K3 before exposure and pressure can be related to the partial molar volume ΔH the enthalpy of fusion after exposure to CO2. g m2 (V2 ) of the solute (component 2) in the supercritical phase (component 1). When the molar volume of the RESULTS AND DISCUSSION s* pure solid component Vm,2 equals the partial molar volume of the dissolved solid in the fluid phase Solubility comparison (V s*−V g ) the concentration extrema occurs [16,17]. m,2 2 The solubility data are presented in Figure 2 Partial molar volumes (and Tables under Supplementary data available from corresponding author upon request). Compared to lit- The partial molar volumes V g (m3∙mol-1) were 2 erature values [7,8], the measured solubilities of vit- calculated after determining the densities of the sys- amin K in CO obtained in the present research are tem of vitamin K in CO by the vibrating tube densi- 3 2 3 2 on the same order of magnitude, however they are meter according to the following equation [9]: somewhat lower, probably as a consequence of the ρ g 1dn different methods used for the solubility determination V =−()MV 2 (4) 22mρ ()+ 2gdy [7]. K/CO32 nn12 2 The solubility difference (at the solubility max- where ρ (g∙cm-3) is the density of the binary imum) between the data obtained in the present work K/CO32 3 -1 system of CO2 and vitamin K3 and Vm (cm ∙mol ) is and Knez-Škerget and Johannsen-Brunner is: at 40 -1 the volume of the mixture. M2 (g∙mol ) is the molar °C the solubility is for 23% lower compared to Knez mass of vitamin K3, n1 and n2 the number of moles of and Škerget and 52% lower compared to Johannsen g ° CO2 and vitamin K3 and y 2 the mole fraction of vit- and Brunner. At a temperature of 60 C the solubility amin K3. measured in the frame of our research is for 44% Partial molar volumes were calculated from the lower compared to the one of Knez and Škerget slope of the linear function of densities versus the (Figure 3), and at a temperature of 80 °C and at pres- mole fraction of vitamin K3 that was plotted for the sure of 23 MPa for 64% lower compared to the one of data up to the maximum solubility of vitamin K3 in Knez and Škerget (Figures 4 and 5). The discrepancy

CO2. Therefore the results of partial molar volumes is increasing with temperature, probably due to the are presented to the pressure of the maximum sol- solubility decrease (due to the lower solvent power of ubility (approximately 25 MPa). CO2) and therefore a higher discrepancy in solubility calculations. Solubility data correlation The non-polar CO2 requires high pressures to The solubility data have been correlated with a dissolve even small amounts of high-molecular-mass simple density model (log y2 against log ρ). Solubility compound [19] like vitamin K3 and therefore sampling of a crystalline solute in a supercritical fluid as a of low concentrated sample is difficult [7]. One of the function of solvent density is expressed by [18]: reasons for lower solubility compared to literature =−θ ρ could be the pressure decrease while circulating the log yC20 log (5) saturated fluid with vitamin through the circuit into the .-1 where y2 (mol∙mol ) is the molar fraction of K3 in vibrating U-tube. During expansion of the fluid with -3 supercritical fluid and ρ (mol∙m ) [14] is the density of vitamin K3, the equilibrium was probably disturbed the pure supercritical fluid, C0 is a constant and θ is and some amount of the vitamin precipitated. After the slope of the curve of log y2 as a function of log ρ. the expansion probably compression took place, and it is assumed that the entire amount of precipitated Thermal analysis vitamin was not completely dissolved in CO2. The sol- The melting points and the degree of crystallinity ubility has a clear trend (Figure 2) with some devi- based on melting points and the change in enthalpy ations, probably due to the fact that the depressoris-

566 N. TRUPEJ et al.: INVESTIGATION OF THE THERMODYNAMIC PROPERTIES… Chem. Ind. Chem. Eng. Q. 23 (4) 563−571 (2017)

g ° Figure 2. Molar fraction ( y 2 ) of vitamin K3 in vitamin K3/CO2 solution at 40 C and comparison to the literature of Johannsen and Brunner, 1997 [7] and Knez and Škerget, 2001 [8].

g ° Figure 3. Molar fraction ( y 2 ) of vitamin K3 in vitamin K3/CO2 solution at 60 C and comparison to the literature of Knez and Škerget, 2001 [8].

g ° Figure 4. Molar fraction ( y 2 ) of vitamin K3 in vitamin K3/CO2 solution at 80 C and comparison to the literature of Knez and Škerget, 2001 [8].

567 N. TRUPEJ et al.: INVESTIGATION OF THE THERMODYNAMIC PROPERTIES… Chem. Ind. Chem. Eng. Q. 23 (4) 563−571 (2017)

g ° Figure 5. Molar fraction ( y 2 ) of vitamin K3 in vitamin K3/CO2 solution at temperatures 40, 60 and 80 C and its approximation. ation rate was not exactly the same before each of vitamin K3 were measured by a gas pycnometer sampling and a consequence could be that the amount (Micromeritics, AccuPyc II 1340) at 294 K. The mea- of the precipitated vitamin is different resulting in a sured data are available under supplementary data). deviation of solubility. Solubility data correlation The solubilities of vitamin K3 in supercritical CO2 in the presented research reach a maximum at tem- The density-based model (log y2 against log ρ) peratures of 40 and 60 °C. The isotherm at tempera- is normally used for solubility data that increase with ture of 40 °C is higher than at 60 °C and at 60 °C the pressure. The determined constants of the model for isotherm is higher than at a temperature of 80 °C (Fig- the vitamin K3/CO2 solutions and the AARD (%) of the ure 2). The solubility decreases with increasing tem- model from the data are given in supplementary data). perature. Generally, the solubility behaviour is influ- Method comparison enced by two effects: the decrease of solvent power In the literature [7,8], vitamin K with the same of the fluid with increasing temperature due to the 3 purity as in this study (98%) was used. Furthermore, decreasing density and the increase of solubility due CO with 99.995% purity was used in the present to the increasing vapour pressure of the substance with 2 work and in the research of Knez and Škerget, while increasing temperature [7]. In the present research, Johannsen and Brunner used CO with purity of the decrease of solvent power dominates over the 2 99.95% and a static analytical method with direct increasing vapour pressure of vitamin K3. coupling of an equilibrium cell to a supercritical fluid Another explanation for a solubility difference chromatographic system as described in literature [20]. compared to literature could be the structure of the In addition, Knez and Škerget also used a static ana- vitamin K3. Vitamin K3 was added into the autoclave in lytic method but with a different sampling method with a crystalline and a powderous form. After equilibrium a glass trap, and the samples taken from the auto- was established, the supercritical CO2 saturated with clave were analysed on UV/Vis spectrophotometry [8, vitamin K3 was circulated through the pipelines to the 21]. The results of the present research were also U-tube. During this transport through the pipelines, by obtained by a static analytic method and a sampling means of a hand pump, first a pressure drop occurred method similar to the one of Knez and Škerget. The and small amounts of the vitamin probably precipit- difference between all three methods is the sampling, ated, leading to change in the morphology of the solid consequently the pressure drop due to the sampling, (e.g., degree of crystallinity), resulting in the change and the analysis method. Also other authors reported of the amount of vitamin dissolved in CO2. The crys- differences in solubilities between different researches tallinity degree, determined in air atmosphere by for the same system [22]. A small difference in the DSC, changed for 16%. purity or in the protocol can provide high deviations – The melting points of the two previously men- even more than one order of magnitude, as reported tioned forms of vitamin K3 (before and after it was in literature [23]. A detailed description of expe- exposed to CO2 in the densimeter) were measured by rimental methods for phase equilibria at high pres- an HP DSC 1 (STAR System, Mettler Toledo) at atmospheric conditions. The densities of both forms

568 N. TRUPEJ et al.: INVESTIGATION OF THE THERMODYNAMIC PROPERTIES… Chem. Ind. Chem. Eng. Q. 23 (4) 563−571 (2017) sures with their specific advantages and weaknesses 21 MPa at temperature 60 °C and from pressure of can be found in the literature [24,25]. approximately 10 MPa to pressure of approximately 23 MPa at temperature 80 °C. Density of solution of vitamin K3 in CO2 The interesting aspect of supercritical fluids is Partial molar volumes and solubility maximum their microscopic inhomogeneity, clustering and the The molar volume of the pure solid compound most well-known are the long-range density fluc- (vitamin K3) was calculated from: tuations near the critical point, which can be studied M by the partial molar volume. The partial molar volume V s* = 2 (7) m,2 ρ s* measurements can explain the unusual behaviour of 2 solutes in supercritical fluids [26]. Much research of ρ s* -3 where 2 (g∙cm ) is the density of the solid vitamin partial molar volumes has already been done and K3 determined at room temperature of 294 K and there are many reports that indicate a large negative atmospheric pressure with the helium pycnometer partial molar volume, which are dependent on particle -1 (Supplementary data), M2 (g∙mol ) is the molar mass size, shape, interaction strength, polarity and on inter- -1 of vitamin K3 and is 172.18 g∙mol . The molar vol- actions between solute-solvent and solvent-solvent umes of the pure solid vitamin K3 (determined with molecules [26]. Eq. (7)) were converted to the unit 10-3 m3∙mol-1 and The densities of vitamin K3/CO2 solutions are s* × -3 are: Vm,2 (before exposure to CO2) = 0.1288 10 presented in Figure 6 (and in Tables under Supple- 3 -1 s* m ∙mol and Vm,2 (after exposure to CO2) = mentary data). By comparing the densities of CO2/ 0.1317×10-3 m3∙mol-1. The difference between the vit- /vitamin K3 solutions to the pure CO2 densities [14] amin before and after exposure to CO2 is approx- (Figure 6), it is clear that the dissolved amount of vit- s* imately 2 % and the average of Vm,2 before and after amin in the CO2 has a minor impact on the density of s* exposure to CO2 was considered as Vm,2 and is the fluid. 0.13×10-3 m3∙mol-1. With increasing temperature at constant pres- Figure 7 (and Tables under Supplementary data) sure, the density of the fluid with the dissolved solid shows that all partial molar volumes at all investigated compound decreases and it increases with increasing temperatures and pressures are negative. Below and pressure at constant temperature. In the present study, near the critical point, the partial molar volumes are the density steps where the densities increased the large and negative, which indicates that the clustering fastest took place at following conditions: from pres- of the solvent molecules (CO2) is a consequence of sure of approximately 7 MPa to pressure of approx- the strong attractive forces between solvent and sol- ° imately 13 MPa at temperature 40 C, from pressure ute particles [11]. The compressibility of supercritical of approximately 9 MPa to pressure of approximately solutions is high and they have large free volumes,

Figure 6. Density of vitamin K3/CO2 solution and comparison to density of pure CO2 (Hunter, Lias) at temperatures 40, 60 and 80 °C.

569 N. TRUPEJ et al.: INVESTIGATION OF THE THERMODYNAMIC PROPERTIES… Chem. Ind. Chem. Eng. Q. 23 (4) 563−571 (2017)

Figure 7. Partial molar volumes of vitamin K3 in vitamin K3/CO2 solution at temperatures 40, 60 and 80 °C.

therefore solvent molecules move into energetically than 1%, amounts of moles of CO2 and vitamin K3 (n1 favourable locations and form clusters around the sol- and n2) were determined gravimetrically with an inac- ute molecule [9]. In the supercritical region, the local curacy of approx. 1% (the values are considered by a density increases (due to the solvent molecules around square function, therefore the value is multiplied by the solutes) and a divergence of density fluctuations two). Slope of the linear function of densities versus appears [11]. In literature, similar conclusions were the mole fraction of vitamin K3 that was plotted for the g made for other systems with a similar or a different data up to the maximum solubility (dρ/ dy 2 ) con- partial molar volume determination methods [4,9,27– tributes highest value to the total inaccuracy which –29]. ranges up to 10%. If the total inaccuracy is evaluated

Partial molar volume of vitamin K3 in CO2 dec- as a summary of uncertainties of each variable of the reases with increasing temperature at isobaric con- equation, it is estimated to be around 14.5%. ditions (Figure 4). The partial molar volume at pressure of 0.34 MPa and at temperature 80 °C has an CONCLUSIONS extremely negative value and is not represented in the figure. Solubility is a function of temperature – with inc- s* reasing temperature the solubility of vitamin K3 in CO2 After comparing the calculated Vm,2 , which is × -3 3 -1 s* decreases at isobaric conditions due to the lower den- 0.13 10 m ∙mol , with the absolute value of Vm,2 at maximum concentration it is seen that at 40 °C the sity of the supercritical fluid and lower solubility s* × -3 3 -1 ° s* power. The calculated partial molar volumes of vit- Vm,2 is 2.45 10 m∙mol , at 60 C the Vm,2 is 19.40×10-3 m3∙mol-1 and at 80 °C and pressure of amin K3 in CO2 are large and negative, especially s* × -3 3 -1 before and near the critical point, which indicates 22.06 MPa the Vm,2 is 59.54 10 m ∙mol . The calculated partial molar volume depends on the solubility major interactions between the supercritical fluid and vitamin molecules. The calculated partial molar vol- of vitamin K3 in CO2. The differences could be a con- ume of the solid in the supercritical fluid at the max- sequence of the precipitation of vitamin K3 during the fluid circulation period, which is a result of pressure imum solubility point is lower than the molar volume s* of the pure solid compound. However, the results cor- drop. Furthermore, for calculations of the Vm,2 the density of the pure vitamin should be measured at the respond quite well, namely that at maximum solubility experimental temperatures 40, 60 and 80 °C by a the partial molar volume of the pure solid compound pycnometer equipped with a heater. The inaccuracy is equal to the partial molar volume of the solid compound in dissolved in the solvent, at temperature of for partial molar volume of vitamin K3 in CO2 is con- ° siderable. This was well expected due to the fact that 40 C. Comparing the solubility data to literature partial molar volumes were calculated after determin- shows that different authors report different solubility ing densities of the system according to Eq. (4). Each data for the same binary system differ due to different segment of the equation (except constant values such experimental equipment and analytical methods. It is appropriate to apply the density of the mixture vitamin as M2 - molar mass of vitamin K3) contributes to the total inaccuracy. It is assumed that inaccuracy of den- K3/CO2 in density based models or equations of state, sity measurements is lower than 0.5% (for all iso- due to the small difference between density of pure therms), the one of the volume of the mixture is lower

570 N. TRUPEJ et al.: INVESTIGATION OF THE THERMODYNAMIC PROPERTIES… Chem. Ind. Chem. Eng. Q. 23 (4) 563−571 (2017) fluid and of the mixture with a dissolved solid com- [12] A.T. Sousa, C. Nieto de Castro, R. Tufeu, B. Le Neindre, pound. High Temp. - High Pressures 24 (1992) 185–194 [13] M.J. Comuñas, J.-P. Bazile, A. Baylaucq, C. Boned, J. Acknowledgements Chem. Eng. Data 53 (2008) 986–994 Special thanks to Janko Trupej for the gram- [14] E. Hunter, S. Lias, Chemistry WebBookNIST Standard matical corrections of this article, to Marko Krainer Reference Database, http://webbook. nist. gov/chemistry and Igor Krmelj for the tehnical support, to Aljana (accessed 15 March 2014) Petek and Darja Pečar for helping with the theoretical [15] U. Haarhaus, P. Swidersky, G. M. Schneider, J. Chem. part, to the Slovenian Research Agency (ARRS) and Eng. Data 8 (1995) 100–106 to the research programme group P2 – 0046: Separ- [16] R. Kurnik, R. Reid, AIChE J. 27 (1981) 861–863 ation processes and production design for the finan- [17] D. Tuma, G.M. Schneider, Fluid Phase Equilib. 158 (1999) 743–757 cial support. [18] Ž. Knez, A. Rižner-Hraš, K. Kokot, D. Bauman, Fluid Phase Equilib. 152 (1998) 95–108 REFERENCES [19] Z. Yang, H.-J. Yang, J. Tian, C.-Y. Guo, H. Kim, J. Chem. [1] G. Brunner, J. Food Eng. 67 (2005) 21–33 Eng. Data 56 (2011) 1191–1196 [2] C. Turner, J.W. King, L. Mathiasson, J. Chromatogr., A [20] M. Johannsen, G. Brunner, Fluid Phase Equilib. 95 936 (2001) 215–237 (1994) 215–226 [3] M. Škerget, M. Knez Hrnčič, Ž. Knez., High Temp. - High [21] Ž. Knez, R. Steiner, J. Supercrit. Fluids 5 (1992) 251–255 Pressures 56 (2011) 694–719 [22] M.D. Saldaña, L. Sun, S.E. Guigard, F. Temelli, J. [4] T. Funazukuri, C.Y. Kong, S. Kagei, Ind. Eng. Chem. Supercrit. Fluids 37 (2006) 342–349 Res. 41 (2002) 2812–2818 [23] A. Tabernero, E.M. Martín del Valle, M.A. Galán, Ind. [5] M. Knez Hrnčič, E. Markočič, N. Trupej, M. Škerget, Ž. Eng. Chem. Res. 52 (2013) 18447–18457 Knez, J. Supercrit. Fluids 87 (2014) 50–58 [24] R. Dohrn, J.M.S. Fonseca, S. Peper, Annu. Rev. Chem. [6] R.B. Gupta, J.-J. Shim, Solubility in supercritical carbon Biomol. Eng. 3 (2012) 343-367 dioxide, CRC press, Boca Raton, FL, 2006 [25] S. Peper, R. Dohrn, J. Supercrit. Fluids 66 (2014) 2–15 [7] M. Johannsen, G. Brunner, J. Chem. Eng. Data 42 (1997) [26] X. Zhang, L. Gao, Z. Liu, J. He, J. Zhang, B. Han, J. 106–111 Supercrit. Fluids 23 (2002) 233–241 [8] Ž. Knez, M. Škerget, J. Supercrit. Fluids 20 (2001) 131– [27] C.Y. Kong, T. Funazukuri, S. Kagei, G. Wang, F. Lu, T. –144 Sako, J. Supercrit. Fluids, A 1250 (2012) 141–156 [9] D. Pečar, V. Doleček, J. Supercrit. Fluids 40 (2007) 200– [28] C.Y. Kong, N.R. Withanage, T. Funazukuri, S. Kagei, J. –207 Supercrit. Fluids 37 (2006) 63–71 [10] J. Zhang, X. Zhang, B. Han, J. He, Z. Liu, G. Yang, J. [29] C.Y. Kong, N.R. Withanage, T. Funazukuri, S. Kagei, J. Chem. Eng. Data 22 (2002) 15–19 Chem. Eng. Data 50 (2005) 1635–1640. [11] C.Y. Kong, T. Siratori, T. Funazukuri, G. Wang, J. Chro- matogr., A 1362 (2014) 294–300

NINA TRUPEJ TERMODINAMIČKA SVOJSTVA BINARNOG MAŠA KNEZ HRNČIČ SISTEMA VITAMIN K3/UGLJEN-DIOKSID MOJCA ŠKERGET

ŽELJKO KNEZ Binarni sistem vitamina K3 i CO2 je analiziran na temperaturama 40, 60 i 80 °C i pritisku University of Maribor, Faculty of do 40 MPa. Rastvorljivost je merena statičko-analitičkom metodom. Delimične molarne Chemistry and Chemical zapremine određene su metodom koja uključuje denzitometar sa i vibrirajućom cevi. Engineering, Maribor, Slovenia Rastvorljivost vitamina K3 u CO2 je određena kao funkcija pritiska i temperature. Naj- veća rastvorljivost (31,16×10-4 mol∙mol-1) je postignuta pri pritisku 25,40 MPa na NAUČNI RAD temperaturi 40 °C, dok je na temperaturi 60 °C i pritisku 24,02 MPa, rastvorljivost iznosila 18,79×10-4 mol∙mol-1. Rastvorljivost (6,48×10-4 mol∙mol-1) je bila niža na temperaturi od 80 °C i pritisku od 22,06 MPa. Parcijalni molske zapremine su negativne, a

rastvoreni vitamin K3 ima mali uticaj na gustinu rastvora K3 u CO2 u poređenju sa gus-

tinom čistog CO2. Ključne reči: rastvorljivost, gustina, parcijalna molska zapremina, prehrambena industrija, prerada hrane.

571

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 573−580 (2017) CI&CEQ

SEMA AKYALCIN KINETIC STUDY OF THE HYDRATION OF Anadolu University, Department of PROPYLENE OXIDE IN THE PRESENCE OF Chemical Engineering,Eskisehir, HETEROGENEOUS CATALYST Turkey

SCIENTIFIC PAPER Article Highlights • The hydration of propylene oxide was investigated detailed • - UDC 544.4:66.097.3:66.093.4:54 Kinetic model was expressed in the presence of Lewatit MonoPlus M500/HCO3 • The reactions follow a series-parallel irreversible homogeneous mechanism • Temperature dependencies of kinetic parameters were determined

Abstract The kinetics of the hydration of propylene oxide was studied using a pressurized batch reactor for both uncatalyzed and heterogeneously catalyzed - reactions. Lewatit MonoPlus M500/HCO3 was used as heterogeneous cat- - alyst, which showed better performance than Dowex Marathon A/HCO3 . The effects of the parameters, namely internal and external diffusion resistances, temperature, catalyst loading and mole ratios of reactants, on the reaction rate were studied. The uncatalyzed and heterogeneously catalyzed reactions were proven to follow a series-parallel irreversible homogeneous mechanism. The temperature dependencies of the rate constants appearing in the rate expressions were determined. Keywords: hydration reaction, propylene oxide, propylene glycol, dipro- - pylene glycol, Lewatit MonoPlus M500/HCO3 , reaction kinetics.

Propylene glycol (PG) is an important inter- The hydration of PO can occur either non-cat- mediate for numerous chemicals in the production of alytically or catalytically. Non-catalytic hydration of alkyd resins for paints and varnishes, in hydraulic PO is carried out at high temperature with a large fluids, and in heat transfer fluids. PG has very low amount of water (12 to 20 mol water (W)/mol PO) [4]. toxicity, therefore it is directly used in foods, pharma- This process causes high energy consumption and ceuticals and cosmetics. Industrially, propylene glycol also increases the cost contributed by the purification is produced by the hydration of propylene oxide (PO) of the product. Acid or base catalysts can be used to which proceeds via a series-parallel reaction. During enhance reaction rates or PG selectivity. Both homo- the reaction, dipropylene glycol (DPG) and tripropyl- genous and heterogeneous catalysts can be used for ene glycol (TPG) are produced as by-products [1,2]. the hydration of PO. While sulfuric acid [5] and salts of DPG is a chemical intermediate in the manufacture of some acids [6,7] are used as homogeneous catalysts, high-performance, unsaturated polyesters resins, ion-exchange resins [8,9], solid base [10] and solid polyurethanes, and plasticizers. DPG is used in hyd- acid [11] catalysts can serve as heterogeneous cat- raulic brake-fluid formulations, cutting oils, industrial alysts. Compared to homogeneous catalysts, hetero- soaps, and solvents [1,3]. TPG is also used as a geneous catalysts are noncorrosive and can be easily solvent, as textile soaps, and lubricants [1]. removed from the reaction mixture by decantation and filtration. Additionally, heterogeneous catalyst

suppresses side reactions, leading to higher purity of Correspondence: Anadolu University, Department of Chemical Engineering, 26555, Eskisehir, Turkey. product [12]. E-mail: [email protected] Liu et al. synthesized PG and DPG by the Paper received: 19 November, 2016 hydration of PO over solid base catalysts. They rep- Paper revised: 24 January, 2017 Paper accepted: 24 February, 2017 orted that sol-gel derived Na2O-ZrO2 showed the https://doi.org/10.2298/CICEQ170203011A excellent performance among the others and under

573 S. AKYALCIN: KINETIC STUDY OF THE HYDRATION OF PROPYLENE OXIDE… Chem. Ind. Chem. Eng. Q. 23 (4) 573−580 (2017) the optimal reaction conditions, PO conversion, PG Marathon A in chloride forms were converted in the selectivity and DPG selectivity reached 99.9, 46.2 and bicarbonate form and used as catalysts [9,15]. The 41.3%, respectively [10]. Shaikhutdinov et al. reported properties of the ion-exchange resins reported by the kinetics of propylene oxide hydration in the pre- manufacturer are given in Table 1. sence of bis(ethane-1,2-diol)molybdate. They sug- Apparatus gested a mathematical description of PO disappearance and PG formation [13]. Kozlovsky et al. studied The experiments were performed in a 450 mL the kinetics of hydration of ethylene and propylene stainless steel, high-pressurized batch reactor (Parr oxides in concentrated aqueous solutions and pro- 4562), equipped with a heating jacket, a cooling coil, posed the first-order kinetics of ethylene or propylene a sampling line, two oblong windows, and a catalyst oxide consumption. They reported that 96% PG yield basket with uniflow stirrer. A controller (Parr 4843) was obtained in the presence of 0.25 mol/L sodium was used to control temperature and stirrer speeds in bicarbonate at 94 °C when the W/PO mole ratio was the reactor and to monitor the reactor pressure. 33.7/1 [7]. The same reaction was also studied in a Procedure batch reactor by Reman and Van Kruchten using - A known amount of the fresh catalyst in a cat- Amberlite 400/HCO3 , immobilized anions of salts on alyst basket and a predetermined amount of water, heterogeneous carrier. They reported that the select- providing the desired W/PO mole ratio, were loaded ivity of PG and DPG was 95.5 and 3.9%, respectively, into the reactor. A nitrogen line to the stainless steel when the conversion of PO was 99% at 120 °C. They feeding vessel and the reactor were connected to the compare catalytic and non-catalytic PO hydration system. The reactor content was heated to the react- reaction data. Analysis of the product which is ion temperature, and it was kept constant during the obtained from non-catalytic reaction at 160 °C reaction. Meanwhile, the amount of the PO, providing showed that the selectivity of PG and DPG was found the desired W/PO mole ratio, was introduced into the as 81.5 and 17.8%, respectively when more than 99% reactor after heating to the reaction temperature in a of PO had been converted [9]. Horbatenko et al. stainless steel vessel. The total volume was kept investigated stepwise and concerted mechanisms for constant at 150 mL in all experiments. Then, the the formation of PG and DPG over ZSM-5 zeolite. sample was collected in a cold trap and analyzed to They report that although the formation of the PG is determine the initial composition of the reaction mix- faster, the formation of DPG as a byproduct cannot be ture. Thereafter, a 1 mL liquid sample was periodic- avoided using ZSM-5 zeolite [14]. ally withdrawn from the reactor for analysis. In the present work, the kinetic model was established for representing the hydration of PO in the Analysis presence of Lewatit MonoPlus M500/HCO - which 3 An HP 7890 gas chromatograph equipped with showed better performance than Dowex Marathon flame ionization detector (GC-FID) was used to det- A/HCO -. The kinetic model was rearranged for the 3 ermine PG, DPG and TPG concentrations in all formation of PG and DPG. samples and a high polarity HP-INNOwax capillary column was used to separate the compounds in the MATERIALS AND METHODS samples. The FID and injector temperatures were set ° Materials to 250 C. The column temperature was programmed to be increased from 100 to 220 °C at a rate of 8 Propylene oxide (99.0%) was purchased from °C·min-1. 1,3-butanediol (≥99.0%) was used as an int- Acros. Deionized water (18.2 MΩ) was used in all ernal standard. Mass balance calculations based on experiments. Lewatit MonoPlus M500 and Dowex the stoichiometric equations given in Eqs. (1a-c) were

Table 1. Properties of ion-exchange resins

Resin Lewatit MonoPlus M500 Dowex Marathon A Manufacturer LANXESS Deutschland GmbH. Dow Chemical Co. + + Functional group -[Ph-CH2-N(CH3)3] -[Ph-CH2-N(CH3)3] % Moisture 48-55 50-60 Total exchange capacity(equiv./l) >1.3 1.2 Particle size(mm) 0.62 0.525-0.625 Polymer type Gelular Gelular

574 S. AKYALCIN: KINETIC STUDY OF THE HYDRATION OF PROPYLENE OXIDE… Chem. Ind. Chem. Eng. Q. 23 (4) 573−580 (2017)

used to determine the concentration of PO and W. obtained from the experimental data (rexp) as shown The initial concentration of PO was also determined below: by the concentration of the products obtained when minϕ =−()rr2 (3) the reaction was completed.  calc exp All data samples To determine the equivalent concentration of bicarbonate in the catalysts, the catalysts were Using the experimental data for each tempe- treated with 10% NaCl solution [16], and then the rature, reaction rate constants were calculated. solution was titrated with 0.1 N HCl solution [17]. Uncatalyzed reaction RESULTS AND DISCUSSION The uncatalyzed reactions were carried out at 358, 373, and 388 K with W/PO mole ratio of 7.5/1.

Reaction mechanism and kinetic modeling The calculated values of k1, k2, and k3 with 95% The experiments were performed in a high-pres- confidential interval are shown in Table 2. surized batch reactor to observe the effects of the –1 –1 parameters on the reaction rate. For each experi- Table 2. Calculated values of k1, k2 and k3 (L mol ∙min ) for the uncatalyzed reaction with 95% confidential interval ment, the parameter under investigation was varied 4 4 4 while keeping the others constant. Temperature, K k1×10 k2×10 k3×10 The experimental data show that the formation 358 0.439±0.051 0.768±0.075 0.356±0.048 of TPG is negligible in the presence of the catalyst 373 1.08±0.11 2.411±0.172 9.855±1.891 while it is significant in the uncatalyzed reaction. 388 2.319±0.066 5.865±0.738 21.265±2.32 Therefore, the hydration of PO can be considered to proceed in a three step series-parallel reactions for Applying the Arrhenius equation using the uncatalyzed reactions and in a two-step series- values given in Table 2, the reaction rate constants as parallel reactions for the catalyzed reactions [3,4]: a function of temperature were determined as follows:

k1 kT=−exp(11.52 7713 / ) (4) W + PO ⎯⎯→ PG (1a) 1 =− kT2 exp (16.87 9421/ ) (5) k2 PG + PO ⎯⎯→DPG (1b) =− kT3 exp (43.45 19086/ ) (6) k3 DPG + PO ⎯⎯→ TPG (1c) Catalyzed reaction

If each step given in Eqs. (1a-c) is assumed as The effects of the parameters, namely catalyst an irreversible, bimolecular and constant density type, external and internal diffusion resistances, tem- reaction, the reaction rate expressions can be written perature, catalyst loading, and mole ratio of reactants as follows: on the reaction rate were observed for catalyzed hydration reaction. dC −=PO kC C + k C C + kC C (2a) dt 1WPO 2PGPO 3DPGPO Effect of catalyst type The reaction was conducted at 358 K with W/PO dC PG =− kC1WPO C k 2PGPO C C (2b) mole ratio of 7.5/1 in the presence of two types of dt - catalyst, namely Lewatit MonoPlus M500/HCO3 and dC Dowex Marathon A/HCO -. Each catalyst containing DPG =− 3 kC2PGPO C kC 3DPGPO C (2c) - dt the same amount of HCO3 , which are corresponding to 0.15 mol HCO -/L in equivalent, was used in the dC 3 TPG = kC3 DPG C PO (2d) experiment. For the prescribed conditions, the aver- dt age reaction rates can be calculated using Eq. (7): where k1, k2, and k3 are the reaction rate constants. C Eqs. (2a-d) were tested with the experimental data by − rCAd using the nonlinear regression analysis with Polymath  −=r C0 (7) 6.10 software. The aim of the data fitting procedure is A,ave − CC0 to minimize the mean square differences between the −=− calculated values of the reaction rate (rcalc) with values where rCtA d/dA was obtained from the slope of concentration-time curve [18]. The average reaction

575 S. AKYALCIN: KINETIC STUDY OF THE HYDRATION OF PROPYLENE OXIDE… Chem. Ind. Chem. Eng. Q. 23 (4) 573−580 (2017) rate calculated using the conversion range from 0 to presence of different particle sizes of the catalyst. 90% and PG selectivity based on the percentage of However, Lewatit MonoPlus M500 has a uniform size PO converted into PG for the conversion of 90% are (monodisperse) of 0.62 mm. Therefore, the tempera- given in Table 3. ture criteria was applied to investigate the effect of the Table 3 shows that, as expected, the average internal diffusion, since internal diffusion is less tem- reaction rate and PG selectivity for the reactions with perature-dependent than reaction step [20,21]. In this the catalyst are much higher than those of the respect, Figure 2 shows that the average reaction rate reactions without catalyst. Additionally, Lewatit calculated using the conversion range from 0% to - MonoPlus M500/HCO3 showed better performance 90% doubles for every 10 K increase in temperature. - than Dowex Marathon A/HCO3 . Therefore, Lewatit Thus, it is reasonable to neglect the effect of internal - MonoPlus M500/HCO3 was used as a catalyst for the pore diffusion, considered as a physical step. hydration of PO and the kinetic model was developed These results confirm the literature findings claim- in the presence of this catalyst. ing that the influence of external and internal diffusion can be neglected for most of the hydration reactions External and internal diffusion significance catalyzed by ion exchange resins [18,22,23]. Prior to the kinetic study, it is important to study the influence of external and internal mass transfer Effect of catalyst loading resistances. The effect of the external mass transfer The experiments were performed at different - resistance on the reaction rate is directly related to Lewatit MonoPlus M 500/HCO3 loadings of 0-0.15 - the stirring speed [19]. To observe the effect of ext- mol HCO3 /L in equivalent in the mixture with W/PO ernal diffusion on the reaction rate, the experiments mole ratio 7.5/1 at 358 K. The effect of catalyst load- were performed at 358 K, with the catalyst loading of ing on the conversion of PO is given in Figure 3a. The - 0.15 mol HCO3 /L in equivalent (Lewatit MonoPlus average reaction rates can be calculated from the - M500/HCO3 ) and the W/PO mole ratio of 7.5/1, under experimental data given in Figure 3a by using Eq. (7). 350 and 600 rpm. The results are shown in Figure 1. A plot of the average reaction rate versus catalyst Figure 1 shows that the reaction rate does not loading is given in Figure 3b. change with the stirring speed, which indicates that As seen from Figure 3b, the average reaction − the external diffusion is not the rate-controlling step. rates, rPO,ave , increase linearly with the catalyst load- Therefore, all experiments were carried out at ing under expressed conditions. This is the expected constant stirring speed of 350 rpm. result, since the active surface area is proportional to To observe the effect of internal diffusion on the the amount of catalyst. It also confirms that the effects reaction rate, the reaction can be performed in the of the internal and external diffusions are negligible

Table 3. Average reaction rate and PG selectivity at 358 K with W/PO mole ratio of 7.5/1

− –1 –1 Catalyst type Conversion range rPO,ave / mol L ∙min PG Selectivity, % Uncatalyzed 0-0.90 0.00443 82.1 - Dowex Marathon A/ HCO3 0-0.90 0.0157 92.7 - Lewatit MonoPlus M500/HCO3 0-0.90 0.0182 95.0

- Figure 1. Effect of stirring speed on the conversion of PO at 358 K, W/PO = 7.5/1 and catalyst loading of 0.15 mol HCO3 /L in equivalent.

576 S. AKYALCIN: KINETIC STUDY OF THE HYDRATION OF PROPYLENE OXIDE… Chem. Ind. Chem. Eng. Q. 23 (4) 573−580 (2017)

Figure 2. Effect of temperature on the average reaction rate at 358 K with W/PO mole ratio of 7.5/1 and a - catalyst loading of 0.15 mol of HCO3 /L in equivalent.

1.0PO 0.020 0.018 0.8 0.016 0.014 R² = 0.9979 0.6 0.012 0.010 (mol/L.min) uncatalyzed 0.008

0.4 ave 0.053 mol HCO3-/L )

PO 0.006 0.082 mol HCO3-/L 0.2 (-r 0.004 0.13 mol HCO3-/L 0.002 0.15 mol HCO3-/L 0.000

Conversion of Propylene Oxide, X Oxide, of Propylene Conversion 0.0 0 500 1000 1500 0 0.05 0.1 0.15 Time, min Catalyst loading - (mol HCO3 /L in equivalent) (a) (b)

Figure 3. a) Effect of the catalyst loading on conversion of PO; b) effect of the catalyst loading on the average reaction rate (358 K with W/PO mole ratio of 7.5/1).

[24]. From Figure 3b, the mathematical expression version range from 0 to 90% for the prescribed con- between the average reaction rate and the catalyst ditions are given in Figure 5. loading can be given by: Figure 5 shows that PG selectivity increases by increasing W/PO mole ratio. This is an expected molHCO − −=3 + rCPO,ave0.0907 cat ( ) 0.0043 (8) result, since the reaction given in Eq. (1b) will be sup- L pressed due to the low concentration of PO and PG in where 0.0043 corresponds to the average uncat- dilute solution. Similarly, the average reaction rate of alyzed reaction rate. This result clearly confirmed the PO will not increase with high W/PO mole ratio since experimental result given in Table 3. the concentration of PO decreases as the reaction Effect of reactants mole ratio proceeds. The intercept of the two curves given in Figure 5 will give the optimum W/PO mole ratio of 6/1. Hydration reactions were realized at W/PO mole ratios of 1.9/1, 4.7/1, 7.5/1 and 11.8/1, while main- Effect of temperature - taining a catalyst loading of 0.13 mol of HCO3 /L in The experiments were realized at different tem- equivalent and a temperature of 358 K. The results peratures of 338, 348, 358 and 368 K under the are shown in Figure 4. constant reaction conditions, with W/PO mole ratio of - As seen from Figure 4, the conversion of PO 7.5/1 and catalyst loading of 0.15 mol HCO3 /L in increases with an increase in the mole ratio of W/PO. equivalent. The calculated reaction rate constants The PG selectivity at 90% conversion and the values with the confidence limits for each temperature are of average reaction rates calculated using the con- given in Table 4. In the prescribed conditions, the formation of tripropylene glycol is negligible.

577 S. AKYALCIN: KINETIC STUDY OF THE HYDRATION OF PROPYLENE OXIDE… Chem. Ind. Chem. Eng. Q. 23 (4) 573−580 (2017)

- Figure 4. Effect of W/PO mole ratio on conversion of PO at 358 K and a catalyst loading of 0.13 mol of HCO3 /L in equivalent.

Figure 5. Effect of the W/PO mole ratio on the average reaction rate and PG selectivity at 358 K and a catalyst loading of - 0.13 mol of HCO3 /L in equivalent.

Table 4. Calculated reaction rate constants (L mol–1 min–1) with =− kT2 exp(20.53 10782 / ) (12) 95% confidence interval

4 4 When Eqs. (9) and (10) were compared with Temperature, K k1×10 k2×10 Eqs. (11) and (12), the activation energies for k1 and 338 0.449±0.096 0.109±0.037 k2 decrease by 4.76 and 0.04 kJ/mol, respectively, 348 0.886±0.200 0.303±0.195 - with a catalyst loading of 0.15 mol HCO3 /L in equi- 358 1.905±0.312 0.786±0.347 valent. Consequently, the reaction rate and PG sel- 368 3.584±0.249 1.655±0.581 ectivity increase in the presence of Lewatit MonoPlus - M500/HCO3 . It also indicates that PG selectivity dec- Applying the Arrhenius equation using the reases with temperature in the presence of the same values given in Table 4, the reaction rate constants as amount of catalyst. a function of temperature were determined to be: Comparison of experimental data with model results kT=−exp(15.72 8705 / ) (9) 1 When the reaction rate expression given in Eq. =− (2a) is rewritten in terms of PO conversion, the fol- kT2 exp(20.53 10777 / ) (10) lowing equation can be obtained: To compare the activation energies for uncat- dx alyzed and catalyzed reactions, uncatalyzed reaction −=rCPO = C2(1 − x ) × PO PO00 PO PO rate constants given in Eqs. (4) and (5) were rear- dt (13) × ()()() −++++ ranged by equalizing the corresponding frequency kS1PO2PO3PO x k P x k D x  factors given in Eqs. (9) and (10): The predicted PO conversion based on the =− kT1 exp(15.72 9277 / ) (11) model was calculated by numerical integration of Eq.

578 S. AKYALCIN: KINETIC STUDY OF THE HYDRATION OF PROPYLENE OXIDE… Chem. Ind. Chem. Eng. Q. 23 (4) 573−580 (2017)

(13). The results obtained at temperatures of 338-368 -1,2-diol) molybdate. It is reported that the selectivity K under the condition of: of PG in the presence of molybdenum glycolates is SC==/7.5 C , about 97-99% at W/PO mole ratio of 34/1 and 128/1 WPO00 PC==/0 C , [13]. Although this result seems to be better than that PG00 PO DC==/0 C , of present study, that is an expected result since the DPG00 PO C = 4.89 mol/L, PG selectivity increases by increasing W/PO mole PO0 - - in the presence of 0.15 mol HCO3 /L in equivalent are ratio. Therefore, Lewatit MonoPlus M500/HCO3 can given in Figure 6. As seen in Figure 6, there is a good be recommended for the hydration of propylene oxide. agreement between the calculated curves and the experimental points.

Figure 6. Experimental points and calculated curves from kinetic model at different temperatures.

CONCLUSION Acknowledgement This work is supported by Scientific Research Two types of catalysts, Dowex Marathon A/ Project Commission of Anadolu University (Project /HCO - and Lewatit MonoPlus M500/HCO -, have 3 3 No. 1605F453). been tested in the hydration of propylene oxide. Pre- liminary investigation showed that Lewatit MonoPlus Nomenclature - M500/HCO3 has better performance than the other. DC= / C DPG00 PO Consequently, the kinetic model was developed in the DPG=dipropylene glycol - presence of Lewatit MonoPlus M500/HCO3 . The hyd- k = rate constant ration of propylene oxide was found to follow the PC= / C PG00 PO series-parallel irreversible reaction mechanism in PG=propylene glycol which the reaction rate expression, based on the con- PO=propylene oxide sumption of propylene oxide, can be given by Eq. SC= / C WPO00 (2a). The conversion of propylene oxide increases T = temperature (K) with the increase of temperature and catalyst loading. TPG = tripropylene glycol It is also observed that the average reaction rate xi = mole fraction of component i increases linearly with catalyst loading in accordance - Subscripts with Eq. (8). Lewatit MonoPlus M500/HCO3 reduces the activation energy of PG production by 4.76 kJ/mol ave = average - with catalyst loading of 0.15 mol HCO3 /L in equi- 0 = at initial condition valent. The selectivity of PG obtained at W/PO mole ratios of 1.9/1 and 11.8/1 is about 78 and 96%, res- REFERENCES pectively, while keeping the catalyst loading of 0.13 [1] R.E. Kirk, D.F. Othmer, In Encyclopedia of Chemical mol of HCO -/L in equivalent at temperature of 358 K. 3 Technology, 4th ed., Vol. 12, John Wiley &Sons, New The kinetics of hydration of propylene oxide was also York, 2001, p. 365 studied in the literature in the presence of bis(ethane-

579 S. AKYALCIN: KINETIC STUDY OF THE HYDRATION OF PROPYLENE OXIDE… Chem. Ind. Chem. Eng. Q. 23 (4) 573−580 (2017)

[2] A. Chauvel, G. Lefebvre, Petrochemical Processes, Edit- [14] Y. Horbatenko, J.P. Perez, P. Hernandez, M. Swart, M. ions Technip, Paris, 1989 Sola, J. Phys. Chem., C 118 (2014) 21952-21962 [3] Dow Chemical Company, Product Safety Assessment: [15] E.M.G.A. Van Kruchten, US 5874653 (1999) Dipropylene Glycol, 2013 [16] V.F. Shvets, R.A. Kozlovskiy, I.A. Kozlovskiy, M.G. Maka- [4] R.E. Kirk, D.F. Othmer, In Encyclopedia of Chemical rov, J.P. Suchkov, A.V. Koustov, Org. Process Res. Dev. Technology, 4th ed., Vol. 20, John Wiley &Sons, New 9 (2005) 768-773 York, 2001, p. 134 [17] D.A. Skoog, D.M. West, F.J. Holler, Fundamental of [5] A.L. Benham, F. Kurata, AIChE J. 1 (1955) 118-124 Analytical Chemistry, Saunders College Pub, Philadel- [6] T. Masuda, K. Asona, N. Hori, S. Ando, US 4937393 phia, PA, 1996 (1990) [18] M.R. Altiokka, S. Akyalçın, Ind. Eng. Chem. Res. 48 [7] I.A. Kozlovsky, R.A. Kovzlovsky, A.V. Koustov, M.G. (2009) 10840-10844 Makarov, J.P. Suchkov, V.F. Shvets, Org. Process Res. [19] S.H. Ali, A. Tarakmah, S.Q. Merchant, T. Al-Sahhaf, Dev. 6 (2002) 660-664 Chem. Eng. Sci. 62 (2007) 3197-3217 [8] R. Jaganathan, R.V. Chaudhari, P.A. Ramachandran, [20] H.S. Foggler, Elements of Chemical Reaction Engineer- AIChE J. 1 (1984) 1-7 ing, Prentice Hall, Englewood Cliffs, NJ, 1999 [9] W.G. Reman, E.M.G.A. Van Kruchten, US 5488184 [21] C.H. Bartholomew, R.J. Farrauto, Fundamentals of Ind- (1996) ustrial Catalytic Processes, John Wiley & Sons, New [10] Z. Liu, W. Zhao, F. Xiao, W. Wei, Y. Sun, Catal. York, 2006 Commun. 11 (2010) 675-678 [22] Y. Liu, Z. Zhou, G. Yang, Y. Wu, Z. Zhang, Ind. Eng. [11] J. Liu, J. Yang, C. Li, Q. Yang, J. Porous Mater. 16 (2009) Chem. Res. 49 (2010) 3170-3175 273-281 [23] G. Yang, P. Wu, Z. Zhou, X. He, W. Meng, Z. Zhang, Ind. [12] M.R. Altiokka, A. Citak, Appl. Catal., A 239 (2003) 141- Eng. Chem. Res. 51 (2012) 15864-15871 –148 [24] R.J. Madon, M. Boudart, Ind. Eng. Chem. Fundam. 21 [13] R.Z. Shaikhutdinov, L.A. Petukhov, V.N. Sapunov, K.E. (1982) 438-447. Kharlampidi, A.A. Petukhov, Kinet. Catal. 51 (2010) 50- –55

SEMA AKYALCIN KINETIKA HIDRATACIJE PROPILEN-OKSIDA U Anadolu University, Department of PRISUSTVU HETEROGENOG KATALIZATORA Chemical Engineering,Eskisehir, Turkey Kinetika hidratacije propilen oksida u odsustvu i prisustvu heterogenog katalizatora je proučavana u šaržnom reaktoru pod pritiskom. Heterogeni katalizator LevatitMonoPlus NAUČNI RAD - M500/HCO3 je pokazao bolje performanse nego katalizator Dovek Marathon A/HCO3. Analizirani su efekti nekih parametara, kao što su otpornost na unutrašnju i spoljašnju difuziju, temperatura, količina katalizatora i molski odnos reaktanata, na brzinu reakcije. Nekatalizovane i heterogeno katalizovane reakcije prate paralelni ireverzibilni homogeni mehanizam. Takođe, određena je temperaturna zavisnost konstanti brzina koje se pojavljuju u izrazima za brzinu.

Ključne reči: reakcija hidratacije, propilen-oksid, propilen-glikol, dipropilen-glikol, - LevatitMonoPlus M500/HCO3 , kinetika reakcije.

580 Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 23 (4) 581−588 (2017) CI&CEQ

ILAYDA ACAROGLU DEGITZ1 RECYCLE OF GOLD MINE TAILINGS SLURRY MUGE SARI YILMAZ2 2 INTO MCM-41 MESOPOROUS SILICA WITH SABRIYE PISKIN HIGH SPECIFIC SURFACE AREA 1Chemical Engineering Department, Yeditepe University, Article Highlights Istanbul, Turkey • MCM-41 was synthesized from tailings slurry from gold mine treatment plant for the first 2 Faculty of Chemical and time Metallurgical Engineering, • Reaction time, temperature, and pH adjustments played a significant role in the syn- Chemical Engineering Department, thesis o Yıldız Technical University, • The optimum condition was found to be at 30 C for 60 minutes with a pH of 11 • Istanbul, Turkey The sample synthesized at optimum condition had a high surface area • The structural properties were found to be close to those obtained from a pure silica source SCIENTIFIC PAPER

Abstract UDC 622.342’17:66:54 In this study, MCM-41 was prepared from the gold mine tailings slurry as an alternative silica source. The effects of the reaction time, temperature, and pH adjustments on the synthesis of MCM-41 were investigated. The optimum condition for the synthesis process of pure and highly ordered MCM-41 from tailings slurry was found to be at 30 °C for 60 min with a pH of 11. In addition, MCM-41 was synthesized from a pure silica source, in order to compare the structural properties of the obtained sample from that of the slurry produced at optimum conditions. The samples synthesized from tailings slurry had a high surface area of 983 m2/g and a pore volume of 0.74 cm3/g. The structural properties of the samples produced from the slurry were closely related with those that were synthesized from a pure silica source. Results indicate that the tailings slurry has the potential to be used as a cheap, alternative silica source which can be used for the synthesis of a pure and ordered mesoporous silica MCM-41. Keywords: MCM-41; synthesis; waste treatment; gold mine.

Since the new family of mesoporous molecular M41S family members have pore sizes that can sieves, known as M41S, was discovered in 1992 by vary within a range of 15-100 Å, and wide surface researchers at the Mobil Oil Corporation, mesoporous areas. The M41S family has three main members molecular sieves have attracted much attention by classified by their different structural features: 3-D scientific research projects and for practical applic- cubic pore structured, MCM-48; 1-D hexagonal pore ations [1]. The main driving force for researchers’ structured, MCM-41; and the unstable pore layer attention to M-41 has been the numerous potential structured, MCM-50 [3]. Among all the M41S mem- applications for use in different fields, such as catal- bers, MCM-41 has been receiving much attention due ysis and adsorption [2]. to its peculiar characteristics of being unidirectional and having a hexagonal, honeycomb-like structure.

Owing to this well-formed structure, which provides a Correspondence: S. Piskin, Faculty of Chemical and Metal- lurgical Engineering, Chemical Engineering Department, Yıldız high specific surface area and pore volume, uniformity Technical University, Istanbul, Turkey. and controllability of pore size, MCM-41 has become a E-mail: [email protected] remarkable member of the M41S family [2,4-7]. Paper received: 19 April, 2016 Paper revised: 6 September, 2016 MCM-41 is produced in aqueous, alkaline con- Paper accepted: 7 April, 2017 ditions using three main components: structure-direct- https://doi.org/10.2298/CICEQ160419012A ing surfactants, a source of silica, and an acidic or

581 I.A. DEGITZ et al.: RECYCLE OF GOLD MINE TAILINGS SLURRY… Chem. Ind. Chem. Eng. Q. 23 (4) 581−588 (2017) base catalyst [7]. Typical silica sources include meter using CuKα (λ = 0.15406 nm) radiation. The sodium silicate, tetraethyl orthosilicate (TEOS) [8], tet- diffraction data were recorded in the scanning range ramethylammonium silicate (TMA-silicate) [9], alkosil- of 2θ = 0.58–8° for small angle XRD (SAXRD) and anes, alkylamines, sodium water glass, and fumed 2θ = 5–80° for wide angle XRD (WAXRD) with a scan- silica [10]. The high costs of synthesizing MCM-41 ning speed of 1°/min. The concentrations of Si and Al materials have mainly been attributed to the cost- of the synthesized sample from slurry at optimum liness of its templates which promote the need for conditions were analyzed using a Perkin Elmer using a cheap silica sources in lieu of more expensive Optima 2100 DV inductively coupled plasma optical ones for synthesis [10,11]. emission spectrometry (ICP-OES) device. Nitrogen Scientists have considered waste products such physisorption measurements of synthesized MCM-41 as rice husk, bottom ash, fly ash, and iron ore tailings samples from slurry and pure silica sources were car- for the synthesis of MCM-41 [12-14]. Unbeknownst to ried out at 77 K using a Micromeritics ASAP 2020 this project group, no effort had previously been dir- surface area and porosimetry analyzer. The pore size ected to introduce a synthesis of MCM-41 from tail- distributions and pore volumes were calculated by the ings slurry. A tailings slurry with a high content of SiO2 Barrett–Joyner–Halenda (BJH) method using the can be obtained from the treatment plant of a gold desorption isotherm. Prior to the experiments, samples mine [15]. Every year, approximately 277.882 t of tail- were outgassed under vacuum at 300 °C. The mor- ings slurry is produced during the extraction pro- phology of products was observed using a CamScan cesses of gold at gold mine treatment facilities [16]. Apollo 300 scanning electron microscope (SEM). Thus, the conversion of the slurry into usable pro- High resolution transmission electron microscopy ducts is rather important for helping solve environ- (HRTEM) observations were carried out by a JEOL- mental and economic issues. -2100 HRTEM. In this paper, the novel synthesis of MCM-41 The extraction of silicon (Si) from Bergama Ovacık Gold Mine’s tailings slurry and the influences of synthesis parameters are reported Extraction of silica from tailings slurry. Tailings for the first time. The reaction time, temperature, and slurry used in this study was obtained from the Ber- pH were varied. The influences of these parameters gama Ovacık Gold Mine Treatment Plant in Turkey. on the synthesis of MCM-41 were investigated by The slurry was prepared according to the previous XRD analysis. In addition, textural and morphological study [17]. Extraction of silica (Si) from tailings slurry properties of synthesized MCM-41 at optimum react- was studied in fusion method [18]. The fusion process ion conditions were carried out. For the comparison of was performed by mixing slurry and sodium hydroxide the structural properties of the obtained sample, MCM- at a ratio of 1:1. The extracted solution was examined -41 was also synthesized from a pure silica source. with Perkin Elmer Optima 2100 DV ICP-OES. The amounts of Al, Si and Na in the solution were 1145, -1 EXPERIMENTAL 28030 and 72800 mg L , respectively. The Si/Al ratio of 24.48 in the extracted solution was close to the Materials best quality MCM-41 obtained from the extracted sol- For usage in this study, the tailings slurry was ution of fly ash, which has the highest Si/Al ratio [19]. supplied from the Bergama Ovacık Gold Mine Treat- Synthesis of MCM-41 from tailings slurry and pure ment Plant in Turkey. The tailings slurry was com- silica posed of SiO (77.70%), Al O (12.10%) and small 2 2 3 HDTMA-Br surfactant was added into distilled amounts of other metal oxides [17]. It was used in this water under ultrasonic irradiation at different tempera- study without any reduction or elimination of impur- tures (30, 40, 50, 60 and 70 °C) for determining the ities. Hexadecyltrimethylammonium-bromide (HDTMA- influence of temperature on the synthesis process of -Br, Merck) was used as a surfactant. Sulphuric acid MCM-41. After the solution became clear, Si solutions and sodium hydroxide were purchased from Merck obtained from a fusion process were added into the chemicals. All materials were used as received. Tet- mixture. Stirring under ultrasonic irradiation was con- raethyl orthosilicate (TEOS, Merck, 99% pure) was tinued for different reaction times (15, 30, 60, 90 and used as a pure silica source. 180 min) for determining the influence of time on the Characterization synthesis process. After reaction times were finished, Powder X-ray diffractograms of samples were the pH adjustments were carried out with sulphuric recorded on a Philips Panalytical X’Pert-Pro diffracto- acid for different acidic and base pH values for determining the influence of pH on the synthesis process.

582 I.A. DEGITZ et al.: RECYCLE OF GOLD MINE TAILINGS SLURRY… Chem. Ind. Chem. Eng. Q. 23 (4) 581−588 (2017)

Precipitation was observed during the pH adjustment. RESULTS AND DISCUSSION Three different synthesis parameters were examined using a one-factor-at-a-time approach [20]. After the Effects of reaction parameters on the synthesis of pH adjustments, the samples were stored for 24 h at MCM-41 room temperature and then the precipitated products In order to investigate the effects of the reaction were filtered. The filtered products were washed thor- parameters on the formation of the MCM-41 mat- oughly with distilled water, dried in oven for a day, erials, different syntheses were conducted by indi- and calcined at 550 °C for 6 h for removal of residual vidually changing and testing the parameters. The surfactants. values of the reaction parameters were chosen After preparing MCM-41 from slurry at optimum according to which were related to the synthesis of conditions, the synthesis of MCM-41 was performed mesoporous materials in the literature [12,21-23]. using a pure silica source for comparison. Firstly, Effects of reaction time. In order to study the HDTMA-Br was dispersed in distilled water under effects of the reaction time on the synthesis of MCM- ultrasonic irradiation at 30 °C. The designated amount -41, the synthesis experiments with different reaction of 1 M of sodium hydroxide solution was added. Next, times (15, 30, 60, 90 and 180 min) were performed, TEOS was added into the mixture drop by drop. After whereas the other variables were held constant (tem- 60 min, the pH value of the sample was adjusted to perature 50 °C and pH 11). Figure 1 reveals the XRD 11. Finally, the solution was kept at room temperature patterns of the samples that were produced at differ- for 24 h. The precipitated solid was separated by fil- ent reaction times. A strong intensity peak (100) dir- tration and washed with distilled water and then dried ectly indicates the presence of the MCM-41 structure, at room temperature. The remaining surfactant was and the two weak intensity peaks (110) and (200) can removed through calcination at 550 °C for 6 h. The be indexed on a hexagonal structure with 2-D hexa- synthesized samples obtained from slurry at optimum gonal p6mm symmetry [24]. As shown in Figure 1, conditions and pure silica accordingly were hereafter MCM-41 mesoporous material was obtained from labeled as M-41-S and M-41-PS, respectively. samples synthesized at 30, 60, 90 and 180 min, whereas it was not obtained from sample synthesized at 15 min. Compared to other samples peaks, the diffraction peaks of the sample synthesized at 60 min

Figure 1. XRD patterns of the synthesized samples at different reaction times: a) 15, b) 30, c) 60, d) 90 and e) 180 min.

583 I.A. DEGITZ et al.: RECYCLE OF GOLD MINE TAILINGS SLURRY… Chem. Ind. Chem. Eng. Q. 23 (4) 581−588 (2017)

were sharp and well-resolved, which was interpreted the best quality product (with maximum a0 values) as an indication of a higher ordering of the mesopore was found to be the sample synthesized at 30 °C. channels [22,25]. Effects of pH value. Figure 3 shows the XRD In order to verify such improvement in the deg- patterns of the samples that were produced at low ree of ordering, hexagonal unit cell parameters (a0) and high pH values whereas the other variables rem- values are calculated using the following equation [26]: ained constant (time 60 min and temperature 30 °C). Only the samples at pH 2, 3 and 3.5 exhibited a dif- a0 = 2dhkl /√3 (1) fraction peak relative to the (100) plane while diffract- In this equation, dhkl was obtained from the peak ions corresponding to other peaks found at higher in the XRD pattern using Bragg’s equation (2λdsinθ, angles were not observed (Figure 3a). A poor crys- where λ = 0.15406 nm was used for the CuKα line. tallinity of the material and the distortion of the long The value of a0 was equal to internal pore diameter range ordering of the mesoporous structure which plus the wall thickness of one pore [6,21,26]. The may or may not include poorly structured hexagonal sample with the highest a0 value resulted in the best arrays are typically indicated by a broader and less quality product. It should be noted, however, that intense main peak [25]. According to this, it was according to Hui and Chao that there are no rules for understood that ordered, hexagonally structured MCM- the assertion of the incorporation of the metal ions -41 mesoporous material was not obtained under an within the MCM-41 structures, due to its amorphous acidic medium. On the other hand, it was successfully structure in which the bond lengths and the angles of obtained from the samples that were synthesized at the metal ions on the silica surface is fluctuating. Fur- pH 10 and 11 which contained the three characteristic thermore, they also explain that the size of the cations peaks with a better definition and higher intensity. and their coordination states play important roles for Compared to other samples peaks, the diffraction the possibility of their incorporation into the skeleton peaks of sample synthesized at pH 11 were sharp [11]. The a0 values of the synthesized samples were and well-resolved, which can be interpreted as an given in Table 1. It was seen that the a0 value of the indication of higher ordering of the mesopore chan- synthesized sample started to decrease by time after nels [22]. The a0 values of the synthesized sample at 60 min, that is, longer reaction time made the pore pH 10 and 11 were calculated as 39.94 and 41.16, size decrease [27]. The best quality product (with respectively. Consequently, the best quality product maximum a values) was found as the sample syn- 0 (with maximum a0 values) was found to be the sample thesized at 60 min. synthesized at pH 11. Effects of reaction temperature. In order to study the effects of the reaction temperature on the syn- Characterization of M-41-S and M-41-PS thesis of MCM-41, synthesis experiments were car- The SAXRD and WAXRD patterns of the syn- ried out at different temperatures (30, 40, 50, 60 and thesized samples are given in Figure 4. It was seen 70 °C), whereas the other variables were held cons- that all samples demonstrated three well-resolved tant (time 60 min and pH 11). Figure 2 shows the characteristic diffraction peaks, which suggested that XRD patterns of the samples that were produced at there are ordered hexagonal mesoporous structures different reaction temperatures. As shown in the fig- in the SAXRD patterns. According to the WAXRD pat- ure, MCM-41 mesoporous material was obtained from terns of samples, only a broadened diffraction peak all samples. Compared to the other samples peaks, can be observed, which confirms the existence of the diffraction peaks of the sample synthesized at 30 amorphous silica, the samples are composed of only °C were sharp and well-resolved, which can be inter- one phase. preted as an indication of a higher ordering of the The amounts of Si and Al of the M-41-S were -1 mesopore channels [22]. The a0 values of the synthe- found as 210380 and 15740 mg kg , respectively. In sized sample at 30, 50 and 60 °C were calculated as our previous study, it was found that Si/Al ratio and Si 41.16, 40.12 and 37.59, respectively. Consequently, content, thereby chemical composition of waste play

Table 1. Values of d100 peaks and a0 of samples synthesized at different reaction times, temperatures, and pH

Parameter Reaction time, min Temperature, °C pH 30 60 90 180 30 50 60 2 10 11 2θ / ° 2.6010 2.5428 2.5446 2.6558 2.4787 2.5428 2.7143 2.8982 1.9815 2.4787 d100 33.9678 34.7449 34.7208 33.2667 35.6466 34.7449 32.5504 30.4604 34.5855 35.6466 a0 39.2226 40.1199 40.0921 38.4131 41.1611 40.1199 37.5860 35.1726 39.9359 41.1611

584 I.A. DEGITZ et al.: RECYCLE OF GOLD MINE TAILINGS SLURRY… Chem. Ind. Chem. Eng. Q. 23 (4) 581−588 (2017)

Figure 2. XRD patterns of the synthesized samples at different reaction temperatures: a) 30, b) 40, c) 50, d) 60 and e) 70 °C.

(a) (b)

Figure 3. XRD patterns of the synthesized samples at a) low and b) high pH values. a significant role in the synthesis of ordered MCM-41. study, it can be seen that highly ordered MCM-41 with The minimum Si/Al ratio was found as 10.30 to syn- an Si/Al ratio 13.37 was successfully synthesized thesize of highly ordered MCM-41 [19]. Chang et al. from slurry. demonstrated that MCM-41 from fly ash can be syn- The nitrogen adsorption-desorption isotherms thesized with an Si/Al of 13.4 [28]. In the present and the corresponding pore size distribution curves of

585 I.A. DEGITZ et al.: RECYCLE OF GOLD MINE TAILINGS SLURRY… Chem. Ind. Chem. Eng. Q. 23 (4) 581−588 (2017)

(a) (b)

Figure 4. XRD patterns of the samples: a) small and b) wide angle.

M-41-S and M-41-PS are shown in Figure 5. The cific surface area and pore volume of samples were nitrogen sorption isotherm of M-41-S had the type IV found as being 983 and 0.74 cm3/g for the M-41-S, as shape, typical of a uniform mesoporous material. The 1223 and 0.83 cm3/g for M-41-PS. Although, the tex- sharp inflection between the relative pressures of tural properties of M-41-S were inferior to those of M- 0.25-0.45 is attributed to the uniformity of pore size -41-PS, to the best of the authors’ knowledge, the sur- distribution. The adsorption isotherm for M-41-PS, face area of M-41-S had the highest BET surface with its S-shape, was similar to the previous studies area obtained for MCM-41 samples synthesized from [29,30]. It also displays a larger adsorbed volume different wastes [12-14]. than the M-41-S. The BJH pore diameter distributions SEM and HRTEM images of the samples are of M-41-S and M-41-PS are relatively narrow with a shown in Figure 6a and b. It can be seen that the maximum at 2.89 and 2.93 nm, respectively. The spe- morphological features of the M-41-S were quite simi-

Figure 1. XRD patterns of the synthesized samples at different reaction times: a) 15, b) 30, c) 60, d) 90 and e) 180 min.

586 I.A. DEGITZ et al.: RECYCLE OF GOLD MINE TAILINGS SLURRY… Chem. Ind. Chem. Eng. Q. 23 (4) 581−588 (2017)

(a)

(b)

Figure 6. Images of the samples: a) SEM and b) HRTEM. lar to M-41-PS. SEM images indicate that the samples of 11, which were confirmed by XRD analysis. The displayed spherical shaped particles with a size of prepared MCM-41 at optimum reaction conditions has <0.4 μm and non-spherical, irregularly agglomerated high specific surface area (983 m2/g) and a total pore particles which were similar to those reported in pre- volume (0.74 cm3/g). Structural properties of the vious studies [31]. HRTEM images show the well- samples obtained from slurry were found to be close -ordered hexagonal arrays of uniform channels. The to those obtained by pure silica source. In view of pore size of the samples was determined as approx- environmental and economic aspects, the synthesis imately 2.60 nm for the M-41-S and 2.75 nm for the of MCM-41 from tailings slurry may provide solutions M-41-PS, consistent with the result of nitrogen ads- to the disposal problem of slurry and recycle the orption analysis. waste into a useful material.

CONCLUSIONS REFERENCES

Ordered MCM-41 molecular sieves were suc- [1] A. Vinu, T. Mori, K. Ariga, Sci.Tech. Adv. Mater. 7 (2006) cessfully synthesized from tailings slurry, as a source 753-771 of silica. The results demonstrate that the reaction [2] S. Wang, T. Dou, Y. Li, Y. Zhang, X. Li, Z. Yan, J. Solid parameters (reaction time, temperature and pH value) State Chem. 177 (2004) 4800-4805 played a significant role in the synthesis of the [3] G. Øye, J. Sjöblom, M. Stöcker, Adv. Colloid Interface ordered mesoporous silica MCM-41. The best quality Sci. 89-90 (2001) 439-466 MCM-41 was obtained under the following conditions: [4] B.A. Ali, Master Thesis, Middle East Technical University, Ankara, 2010 a reaction time of 60 min at 30 °C and a synthesis pH

587 I.A. DEGITZ et al.: RECYCLE OF GOLD MINE TAILINGS SLURRY… Chem. Ind. Chem. Eng. Q. 23 (4) 581−588 (2017)

[5] U. Ciesla, F. Schüth, Micropor. Mesopor. Mat. 27 (1999) [20] J. Waryzwoda, J. Porous. Mater. 13 (2006) 37-47 131-149 [21] C.-F. Cheng, D.H. Park, J. Klinowski, M. Hargreaves, L. [6] Y. Güçbilmez, J. Eng. Architect. Fac. Eskişehir F. Gladden, J. Chem. Soc., Faraday Trans. 93 (1997) Osmangazi Univ. 23 (2010) 63-81 359-363 [7] C. Kresge, M. Leonowicz, W. Roth, J. Vartuli, J. Beck, [22] I. Majchrzak-Kucęba, W. Nowak, Int. J. Miner. Process. Nature 359 (1992) 710-712 101 (2011) 100-111 [8] M.P. Mokhonoana, N.J. Coville, J. Sol-Gel Sci. Technol. [23] X. Zhang, Z. Zhang, J. Suo, S. Li, Stud. Surf. Sci. Catal. 54 (2010) 83-92 129 (2000) 23-30 [9] K.M. Reddy, C. Song, Catal. Lett. 36 (1996) 103-109 [24] R. Xu, W. Pang, J. Yu, Q. Huo, J. Chen, Chemistry of [10] H. Misran, R. Singh, S. Begum, M.A. Yarmo, J. Mater. zeolites and related porous materials: synthesis and Process. Technol. 186 (2007) 8-13 structure, John Wiley & Sons, New York, 2009 [11] K.S. Hui, C.Y.H.Chao, J. Hazard. Mater., B 137 (2006) [25] Y. Cesteros, G.L. Haller, Micropor.and Mesopor. Mater. 1135-1148 43 (2001) 171-179 [12] R.M. Braga, J.M. Barros, D.M. Melo, M.A. Melo, F.d.M. [26] T. Gaydhankar, V. Samuel, P. Joshi, Mater. Lett. 60 Aquino, J.C.d.O. Freitas, R.C. Santiago, J. Therm. Anal. (2006) 957-961 Calorim. 111 (2013) 1013-1018 [27] H.I. Meléndez-Ortizn, L.A. Garcıá-Cerda, Y. Olivares-Mal- [13] G. Chandrasekar, K.-S. You, J.-W. Ahn, W.S. Ahn, Micro- donado, G. Castruita, J.A. Mercado-Silva, Y.A. Perera- por. Mesopor. Mat. 111 (2008) 455-462 -Mercado, Ceram. Int. 38 (2012) 6353–6358 [14] H. Yu, X. Xue, D. Huang, Mater. Res. Bull. 44 (2009) [28] H.L. Chang, C.M. Chun, I.A. Aksay, W.H. Shih, Ind. Eng. 2112 Chem. Res. 38 (1999) 973-977 [15] M.S. Yilmaz, S. Piskin, J. Taiwan Inst.Chem. Eng. 46 [29] P. Carraro, V. Elías, A.G. Blanco, K. Sapag, S. Moreno, (2015) 176-182 M. Oliva, G. Eimer, Micropor. Mesopor. Mat. 191 (2014) [16] İ.M. Türkkal, Master Thesis, Yildiz Technical University, 103-111 Istanbul, 2006 [30] T. Zhao, L. Cuignet, M.M. Dîrtu, M. Wolff, V. Spasojevic, [17] M.S. Yılmaz, S. Pişkin, IJCEBS 1 (2013) 211-213 I. Boldog, A. Rotaru, Y. Garcia, C. Janiak, J. Mater. Chem., C 3 (2015) 7802-7812 [18] N. Shigemoto, H. Hayashi, K. Miyaura, J. Mater. Sci. 28 [31] D. Kumar, K. Schumacher, C.d.F. von Hohenesche, M. (1993) 4781-4786 Grün, K. Unger, Colloids Surfaces, A 187 (2001) 109- [19] M. Sari Yilmaz, N.Karamahmut Mermer, J. Sol-Gel Sci. –116. Technol. 78 (2016) 239–247

ILAYDA ACAROGLU DEGITZ1 RECIKLAŽA JALOVINE IZ RUDNIKA ZLATA U MUGE SARI YILMAZ2 OBLIKU MEZOPOROZNE SILIKE MCM-41 SA 2 SABRIYE PISKIN VELIKOM SPECIFIČNOM POVRŠINOM 1Chemical Engineering Department, Yeditepe University, Istanbul, Turkey U ovom radu je pripremljena silika MCM-41 iz suspenzije jalovine rudnika zlata kao 2 Faculty of Chemical and Metallurgical alternativnog izvora silicijum oksida. Proučavani su efekti vremena reakcije, temperature Engineering, Chemical Engineering i pH sintezu silike MCM-41. Utvrđeni su sledeći optimalni uslovi sinteze čiste i visoko Department, Yıldız Technical organizovane MCM-41 iz jalovine: vreme reakcije od 60 min, temperatura 30 °C i pH 11. University, Istanbul, Turkey Pored toga, silika MCM-41 je sintetisana iz čistog izvora silicijum oksida, kako bi se uporedile strukturna svojstva dobijenog uzorka sa uzorkom koji je sintetisan iz jalovine pod optimalnim uslovima. Uzorci sintetizovani iz jalovine imaju veliku specifičnu povr- NAUČNI RAD 2 3 šinu od 983 m /g i zapreminu pora od 0,74 cm /g. Strukturna svojstva uzoraka proiz- vedenih iz jalovine su bila vrlo slični svojstvima uzoraka sintetisanih iz čistog silicijum- -oksida. Rezultati pokazuju da jalovina može da se koristi kao jeftini, alternativni izvor silicijum oksida za sintezu čiste i organizovane mezoporozne silike MCM-41.

Ključne reči: MCM-41, sinteza, tretman otpada, rudnik zlata.

588 Journal of the Association of Chemical Engineers of Serbia, Belgrade, Serbia

CI&CEQ Vol. 23 Contents: Issues 1–4 YEAR 2017

No. 1 Dušan Lj. Petković, Miloš J. Madić, Goran M. Radenković, The effects of passivation parameters on pitting potential of biomedical stainless steel ...... 121 Marija Mihajlović, Marija Stanojević, Mirjana Stojanović, Je- lena Petrović, Jelena Milojković, Marija Petrović, Zorica Marija Kodric, Sandra Stojanovic, Branka Markovic, Dragan Lopičić, To what extent do soft mechanical activation Djordjevic, Modelling of polyester fabric dyeing in the and process parameters increase the efficiency of presence of ultrasonic waves ...... 131 different zeolite/phosphate rock fertilizer mixtures? ...... 1 Aishi Zhu, Shanshan Liu, Kanfeng Wu, Chuan Ren, Mao- Fatih Ilhan, Kaan Yetilmezsoy, Harun Akif Kabuk, Kubra qian Xu, Comparing of hot water and acid extraction of Ulucan, Tamer Coskun, Busra Akoglu, Evaluation of polysaccharides from proso millet ...... 141 operational parameters and its relation on the stoi- chiometry of Fenton’s oxidation to textile wastewater ...... 11 No. 2 Javad Ahmadishoar, S. Hajir Bahrami, Barahman Movas- sagh, Seyed Hosein Amirshahi, Mokhtar Arami, Rem- oval of Disperse Blue 56 and Disperse Red 135 dyes Bohuš Kysela, Jiří Konfršt, Zdeněk Chára, Ivan Fořt, Struc- fromaqueous dispersions by modified montmorillonite ture of turbulent velocity field in the discharge stream nanoclay ...... 21 from a standard Rushton turbine impeller ...... 151 Alireza Ebrahiminezhad, Yahya Barzegar, Younes Ghas- Lina Lv, Rui Gao, Jianbin Yang, Zhigang Shen, Yanbo Zhou, emi, Aydin Berenjian, Green synthesis and charac- Jun Lu, Investigation of organic desulfurization additives terization of silver nanoparticles using Alcea rosea affecting the calcium sulfate crystals formation ...... 161 flower extract as a new generation of antimicrobials ...... 31 Besheir Ahmed A. Abd-El-Nabey, Sherif El-Housseiny, Essam Aleksandra Mišan, Bojana Šarić, Ivan Milovanović, Pavle Khamis, Ashraf Moustafa Abdel-Gaber, Corrosion pro- Jovanov, Ivana Sedej, Vanja Tadić, Anamarija Mandić, tection and antifouling properties of varnish-coated steel Marijana Sakač, Phenolic profile and antioxidant pro- containing natural additive ...... 169 perties of dried buckwheat leaf and flower extracts ...... 39 Gamze Dalgic, F. Ilter Turkdogan, Kaan Yetilmezsoy, Emel Yajing Zhang, Yu Zhang, Fu Ding, Kangjun Wang, Xiaolei Kocak, Treatment of real paracetamol wastewater by

Wang, Baojin Ren, Jing Wu, Synthesis of DME by CO2 Fenton process ...... 177 hydrogenation over La2O3-modified CuO–ZnO–ZrO2/ K.D. Radosavljević, A.V. Golubović, M.M. Radišić, A.R. /HZSM-5 catalysts ...... 49 Mladenović, D.Ž. Mijin, S.D. Petrović, Amoxicillin photo- Tatjana Kaluđerović Radoičić, Nevenka Bošković-Vragolo- degradation by nanocrystalline TiO2 ...... 187 vić, Radmila Garić-Grulović, Mihal Đuriš, Željko Grbav- Ivana S. Lončarević, Aleksandar Z. Fišteš, Dušan Z. Rakić, čić, Friction factor for water flow through packed beds of Biljana S. Pajin, Jovana S. Petrović, Aleksandra M. Tor- spherical and non-spherical particles ...... 57 bica, Danica B. Zarić, Optimization of the ball mill pro- Hai-Peng Gou, Guo-Hua Zhang, Kuo-Chih Chou, Prepar- cessing parameters in the fat filling production ...... 197 ation of titanium carbide powder from ilmenite con- Nilay Gizli, Merve Arabacı, Enhanced sorption of Cu(II) ions centrate ...... 67 from aqueous solution by ionic liquid impregnated nano- Milana M. Zarić, Mirko Stijepovic, Patrick Linke, Jasna Sta- silica and nano-alumina particles ...... 207 jić-Trošić, Branko Bugarski, Mirjana Kijevčanin, Target- Waheed Ur Rehman, Waheed Zeb, Amir Muhammad, Wajid ing heat recovery and reuse in industrial zone ...... 73 Ali, Mohammad Younas, Osmotic distillation and quality Sandra Raquel Kunst, Lilian Vanessa Rossa Beltrami, Mari- evaluation of sucrose, apple and orange juices in hollow elen Longhy, Henrique Ribeiro Piaggio Cardoso, Tiago fiber membrane contactor ...... 217 Lemos Menezes, Célia de Fraga Malfatti, Effect of diiso- Milenko Košutić, Jelena Filipović, Zvonko Nježić, Vladimir decyl adipate concentration in hybrid films applied to S. Filipović, Vladimir M. Filipović, Bojana Blagojević, tinplate ...... 83 Flakes product supplemented with sunflower and dry Milovan Janković, Snežana Sinadinović-Fišer, Olga Gove- residues of wild oregano ...... 229 darica, Jelena Pavličević, Jaroslava Budinski-Simendić, Zahra Beagom Mokhtari-Hosseini, Razieh Hosseinabadi, Kinetics of soybean oil epoxidation with peracetic acid Ashrafalsadat Hatamian Zarmi, The reduction of oil pol- formed in situ in the presence of an ion exchange resin: lutants of petroleum products storage-tanks sludge Pseudo-homogeneous model ...... 97 using low-cost adsorbents ...... 237

Salah H. Aljbour, Sultan A. Tarawneh, Adnan M. Al-Harah- Li Chen, Guoyan Luan, Morphology control of MnO2 nano- sheh, Evaluation of the use of steelmaking slag as an particles: effect of P123 polymer in ethanol-water aggregate in concrete mix: A factorial design approach .... 113 system ...... 245

589

CI&CEQ Vol. 23 Contents: issues 1–4 YEAR 2017

Mohsen Beigi, Mehdi Torki-Harchegani, Mahmood Mah- bengalense for removal of methyl violet from aqueous moodi-Eshkaftaki, Prediction of paddy drying kinetics: A solutions ...... 399 comparative study between mathematical and artificial Petar Stojanovic, Marija Savic, Dusan Trajkovic, Jovan Ste- neural network modeling ...... 251 panovic, Miodrag Stamenkovic, Mirjana Kostic, The Seyed Mohammad Sadegh Hosseini Davarani, Hassan effect of false-twist texturing parameters on the Hashemipour, Alireza Talebizadeh, Oleylamine-mod- structure and crimp properties of polyester yarn ...... 411 ified impregnation method for the preparation of a highly Meng Peng, Xiaojun Yang, Jiaxiong Wu, Hua Yuan, Chao- efficient Ni/SiO2 nanocatalyst active in the partial oxid- fan Yin, Yuanxin Wu, Application of vanadium-incor- ation of methane to synthesis gas ...... 259 porated phosphomolybdate supported on the modified Arpad Kiralj, Tatjana Vulić, Dunja Sokolović, Radmila Šeće- kaolin in synthesis of diphenyl carbonate by oxidative rov Sokolović, Pero Dugić, Separation of oil drops from carbonylation with phenol ...... 421 water using stainless steel fiber bed ...... 269 Mohsen Beigi, Numerical simulation of potato slices drying Damjan Konovšek, Zdravko Praunseis, Jurij Avsec, Gorazd using a two-dimensional finite element model ...... 431 Berčič, Andrej Pohar, Simon Zavšek, Milan Medved, Underground coal gasification – The Velenje coal mine energy and economic calculations ...... 279 No. 4 Erratum Notice ...... 291 Majid Aliabadi, Effect of electrospinning parameters on the air filtration performance using electrospun polyamide-6 No. 3 nanofibers ...... 441 Rahim Shojaat, Afzal Karimi, Naghi Saadatjoo, Soheil Aber, Lina Lv, Jianbin Yang, Zhigang Shen, Yanbo Zhou, Jun Lu, Dye removal from artificial wastewater using hetero- Optimizing the characteristics of calcium sulfate geneous bio-Fenton system ...... 447 dihydrate in the flue gas desulfurization process: Ana Belščak-Cvitanović, Viktor Nedović, Ana Salević, Saša Investigation of the impurities in slurry – Cl-, Fe3+ and 2+ Despotović, Draženka Komes, Miomir Nikšić, Branko Mn ...... 293 Bugarski, Ida Leskošek Ćukalović, Modification of func- Wenyuan Fan, Xiaohong Yin, Bubble formation in shear- tional quality of beer by using microencapsulated green thinning fluids: Laser image measurement and a novel tea (Camellia sinensis L.) and Ganoderma mushroom correlation for detached volume ...... 301 (Ganoderma lucidum L.) bioactive compounds ...... 457 R.A.F. Oliveira, G.H. Justi, G.C. Lopes, Grid convergence Danica Savanović, Radoslav Grujić, Sladjana Rakita, Alek- study of a cyclone separator using different mesh sandra Torbica, Ranko Bozičković, Melting and crys- structures ...... 311 tallization DSC profiles of different types of meat ...... 473 Shaobai Li, Jungeng Fan, Shuang Xu, Rundong Li, Jingde Ali Sakin, Irfan Karagoz, Numerical prediction of short-cut Luan, The influence of pH on gas-liquid mass transfer in flows in gas-solid reverse flow cyclone separators ...... 483 non-Newtonian fluids ...... 321 Larissa Nayhara Soares Santana Falleiros, Bruna Vieira Bojana Ž. Bajić, Damjan G. Vučurović, Siniša N. Dodić, Cabral, Janaína Fischer, Carla Zanella Guidini, Vicelma Zorana Z. Rončević, Jovana A. Grahovac, Jelena M. Luiz Cardoso, Miriam Maria de Resende Eloízio Júlio Dodić, The biotechnological production of xanthan on Ribeiro, Improvement of recovered activity and stability vegetable oil industry wastewaters. Part I: Modelling of the Aspergillus oryzae β-galactosidase immobilized and optimization ...... 329 on duolite A568 by combination of immobilization Nilgün Ertaş, A comparison of industrial and homemade methods ...... 495 bulgur in turkey in terms of physical, chemical and Dung Van Nguyen, Harifara Rabemanolontsoa, Shiro Saka, nutritional properties ...... 341 Fed-batch fermentation of nipa sap to acetic acid by Jelena Filipović, Shahin Ahmetxhekaj, Vladimir Filipović, Moorella thermoacetica (f. Clostridium thermoace- Milenko Košutić, Spelt pasta with increased content of ticum) ...... 507 functional components ...... 349 Soraya Seankham, Senad Novalin, Suwattana Pruksasri, A. Sudha, V. Sivakumar, V.Sangeetha, K.S. Priyenka Devi, Kinetics and adsorption isotherm of lactic acid from Improving enzymatic saccharification of Cassava stem fermentation broth onto activated charcoal ...... 515 using peroxide and microwave assisted pre-treatment Wang Ran, Tao Jinliang, Ling Yanli, Wei Feng, Liu Jidong, techniques ...... 357 Total spray tray (TST) for distillation columns: A new Didem Özçimen, M. Ömer Gülyurt, Benan İnan, Optimiz- generation tray with lower pressure drop ...... 523 ation of biodiesel production from Chlorella protothe- Yanjun Guan, Xiuying Yao, Guiying Wu, Kai Zhang, CFD coides oil via ultrasound assisted transesterification ...... 367 investigation of slugging behavior in a gas-solids Liangchao Li, Bin Xu, CFD simulation of gas-liquid floating fluidized bed ...... 529 particles mixing in an agitated vessel ...... 377 Snezana Sevic, Branko Grubac, Simulation of temperature- Zixia Li, Yin Ran, Wenqi Zhang, Wei Sun, Tinghai Wang, –pressure profiles and wax deposition in gas-lift wells ..... 537 Industrial application of gasoline aromatization and Ainhoa Rubio-Clemente, Edwin Chica, Gustavo A. Peñuela,

desulfurization technology in Hohhot refinery ...... 391 Kinetic model describing the UV/H2O2 photodegradation Muhammad Imran Din, Kiran Ijaz, Khalida Naseem, Bio- of phenol from water ...... 547 sorption potentials of acid modified Saccharum

590 CI&CEQ Vol. 23 Content: issues 1–4 YEAR 2017

Nina Trupej, Maša Knez Hrnčič, Mojca Škerget, Željko Ilayda Acaroglu Degitz, Muge Sari Yilmaz, Sabriye Piskin, Knez, Investigation of the thermodynamic properties of Recycle of gold mine tailings slurry into MCM-41

the binary system vitamin K3/carbon dioxide ...... 563 mesoporous silica with high specific surface area ...... 581 Sema Akyalcin, Kinetic study of the hydration of propylene Contents: Vol. 23, Issues 1–4, 2017 ...... 589 oxide in the presence of heterogeneous catalyst ...... 573 Author Index, Vol. 23, 2017 ...... 593

591

Journal of the Association of Chemical Engineers, Belgrade, Serbia

CI&CEQ Vol. 23 Author Index YEAR 2017

A C

A. Sudha (3) 357 Carla Zanella Guidini (4) 495 Adnan M. Al-Harahsheh (1) 113 Célia de Fraga Malfatti (1) 83 Afzal Karimi (4) 447 Chaofan Yin (3) 421 Ahmetxhekaj Shahin (3) 349 Chuan Ren (1) 141 Ainhoa Rubio-Clemente (4) 547 Aishi Zhu (1) 141 D Ali Sakin (4) 483 Alireza Ebrahiminezhad (1) 31 Despotović Saša (4) 457 Alireza Talebizadeh (2) 259 Didem Özçimen (3) 367 Amir Muhammad (2) 217 Dodić M. Jelena (3) 329 Ashraf Moustafa Abdel-Gaber (2) 169 Dodić N. Siniša (3) 329 Ashrafalsadat Hatamian Zarmi (2) 237 Dugić Pero (2) 269 Avsec Jurij (2) 279 Dung Van Nguyen (4) 507 Aydin Berenjian (1) 31 Đ B Đorđević Dragan (1) 131 Đuriš Mihal (1) 57 Bajić Ž. Bojana (3) 329 Baojin Ren (1) 49 E Barahman Movassagh (1) 21 Belščak-Cvitanović Ana (4) 457 Edwin Chica (4) 547 Benan Inan (3) 367 Eloízio Júlio Ribeiro (4) 495 Berčič Gorazd (2) 279 Emel Kocak (2) 177 Besheir Ahmed A. Abd-El-Nabey (2) 169 Essam Khamis (2) 169 Bin Xu (3) 377 Blagojević Bojana (2) 229 F Bohuš Kysela (2) 151 Bošković-Vragolović Nevenka (1) 57 F. Ilter Turkdogan (2) 177 Bozičković Ranko (4) 473 Fatih Ilhan (1) 11 Bruna Vieira Cabral (4) 495 Filipović Jelena (2) 229; (3) 349 Budinski-Simendić Jaroslava (1) 97 Filipović M. Vladimir (2) 229 Bugarski Branko (1) 73; (4) 457 Filipović S. Vladimir (2) 229; (3) 349 Busra Akoglu (1) 11 Fišteš Z. Aleksandar (2) 197 Fu Ding (1) 49

593

CI&CEQ Vol. 23 Author Index YEAR 2017

G K

G.C. Lopes (3) 311 K.S. Priyenka Devi (3) 357 G.H. Justi (3) 311 Kaan Yetilmezsoy (1) 11; (2) 177 Gamze Dalgic (2) 177 Kai Zhang (4) 529 Garić-Grulović Radmila (1) 57 Kaluđerović Radoičić Tatjana (1) 57 Golubović V. A. (2) 187 Kanfeng Wu (1) 141 Govedarica Olga (1) 97 Kangjun Wang (1) 49 Grahovac A. Jovana (3) 329 Khalida Naseem (3) 399 Grbavčić Željko (1) 57 Kijevčanin Mirjana (1) 73 Grubac Branko (4) 537 Kiralj Arpad (2) 269 Grujić Radoslav (4) 473 Kiran Ijaz (3) 399 Guiying Wu (4) 529 Knez Hrnčič Maša (4) 563 Guo-Hua Zhang (1) 67 Knez Željko (4) 563 Guoyan Luan (2) 245 Kodrić Marija (1) 131 Gustavo A. Peñuela (4) 547 Komes Draženka (4) 457 Konovšek Damjan (2) 279 H Kostić Mirjana (3) 411 Hai-Peng Gou (1) 67 Košutić Milenko (2) 229; (3) 349 Harifara Rabemanolontsoa (4) 507 Kubra Ulucan (1) 11 Harun Akif Kabuk (1) 11 Kuo-Chih Chou (1) 67 Hassan Hashemipour (2) 259 Henrique Ribeiro Piaggio Cardoso (1) 83 L Hua Yuan (3) 421 Larissa Nayhara Soares Santana Falleiros I (4) 495 Leskošek Ćukalović Ida (4) 457 Ilayda Acaroglu Degitz (4) 581 Li Chen (2) 245 Irfan Karagoz (4) 483 Liangchao Li (3) 377 Ivan Fořt (2) 151 Lilian Vanessa Rossa Beltrami (1) 83 Lina Lv (2) 161; (3) 293 J Ling Yanli (4) 523 Linke Patrick (1) 73 Janaína Fischer (4) 495 Liu Jidong (4) 523 Janković Milovan (1) 97 Lončarević S. Ivana (2) 197 Javad Ahmadishoar (1) 21 Lopičić Zorica (1) 1 Jianbin Yang (2) 161; (3) 293 Jiaxiong Wu (3) 421 M Jing Wu (1) 49 Jingde Luan (3) 321 M. Ömer Gülyurt (3) 367 Jiří Konfršt (2) 151 Madić J. Miloš (1) 121 Jovanov Pavle (1) 39 Mahmood Mahmoodi-Eshkaftaki (2) 251 Jun Lu (2) 161; (3) 293 Majid Aliabadi (4) 441 Jungeng Fan (3) 321 Mandić Anamarija (1) 39 Maoqian Xu (1) 141

594 CI&CEQ Vol. 23 Author Index YEAR 2017

Marielen Longhy (1) 83 Radenković M. Goran (1) 121 Marković Branka (1) 131 Radišić M. M. (2) 187 Medved Milan (2) 279 Radosavljević D. K. (2) 187 Mehdi Torki-Harchegani (2) 251 Rahim Shojaat (4) 447 Meng Peng (3) 421 Rakić Z. Dušan (2) 197 Merve Arabaci (2) 207 Rakita Sladjana (4) 473 Mihajlović Marija (1) 1 Razieh Hosseinabadi (2) 237 Mijin Ž. D. (2) 187 Rončević Z. Zorana (3) 329 Milojković Jelena (1) 1 Rui Gao (2) 161 Milovanović Ivan (1) 39 Rundong Li (3) 321 Miriam Maria de Resende (4) 495 Mišan Aleksandra (1) 39 S Mladenović R. A. (2) 187 Mohammad Younas (2) 217 S. Hajir Bahrami (1) 21 Mohsen Beigi (2) 251; (3) 431 Sabriye Piskin (4) 581 Mokhtar Arami (1) 21 Sakač Marijana (1) 39 Muge Sari Yilmaz (4) 581 Salah H. Aljbour (1) 113 Muhammad Imran Din (3) 399 Salević Ana (4) 457 Sandra Raquel Kunst (1) 83 N Santana Falleiros (4) 495 Savanović Danica (4) 473 Naghi Saadatjoo (4) 447 Savić Marija (3) 411 Nedović Viktor (4) 457 Sedej Ivana (1) 39 Nikšić Miomir (4) 457 Sema Akyalcin (4) 573 Nilay Gizli (2) 207 Senad Novalin (4) 515 Nilgün Ertas (3) 341 Sevic Snezana (4) 537 Seyed Hosein Amirshahi (1) 21 Nj Seyed Mohammad Sadegh Hosseini Davarani (2) 259 Nježić Zvonko (2) 229 Shanshan Liu (1) 141 Shaobai Li (3) 321 P Sherif El-Housseiny (2) 169 Shiro Saka (4) 507 Pajin S. Biljana (2) 197 Shuang Xu (3) 321 Pavličević Jelena (1) 97 Sinadinović-Fišer Snežana (1) 97 Petković Lj. Dušan (1) 121 Soheil Aber (4) 447 Petrović D. S. (2) 187 Sokolović Dunja (2) 269 Petrović Jelena (1) 1 Soraya Seankham (4) 515 Petrović Marija (1) 1 Stajić-Trošić Jasna (1) 73 Petrović S. Jovana (2) 197 Stamenković Miodrag (3) 411 Pohar Andrej (2) 279 Stanojević Marija (1) 1 Praunseis Zdravko (2) 279 Stepanović Jovan (3) 411 Stijepović Mirko (1) 73 R Stojanović Mirjana (1) 1 Stojanović Petar (3) 411 R.A.F. Oliveira (3) 311 Stojanović Sandra (1) 131

595 CI&CEQ Vol. 23 Author Index YEAR 2017

Sultan A. Tarawneh (1) 113 Wang Ran (4) 523 Suwattana Pruksasri (4) 515 Wei Feng (4) 523 Wei Sun (3) 391 Š Wenqi Zhang (3) 391 Wenyuan Fan (3) 301 Šarić Bojana (1) 39 Šećerov Sokolović Radmila (2) 269 X Škerget Mojca (4) 563 Xiaohong Yin (3) 301 T Xiaojun Yang (3) 421 Xiaolei Wang (1) 49 Tadić Vanja (1) 39 Xiuying Yao (4) 529 Tamer Coskun (1) 11 Tao Jinliang (4) 523 Y Tiago Lemos Menezes (1) 83 Tinghai Wang (3) 391 Yahya Barzegar (1) 31 Torbica M. Aleksandra (2) 197; (4) 473 Yajing Zhang (1) 49 Trajković Dušan (3) 411 Yanbo Zhou (2) 161; (3) 293 Trupej Nina (4) 563 Yanjun Guan (4) 529 Yin Ran (3) 391 V Younes Ghasemi (1) 31 Yu Zhang (1) 49 V. Sangeetha (3) 357 Yuanxin Wu (3) 421 V. Sivakumar (3) 357 Vicelma Luiz Cardoso (4) 495 Z Vučurović G. Damjan (3) 329 Vulić Tatjana (2) 269 Zahra Beagom Mokhtari-Hosseini (2) 237 Zarić B. Danica (2) 197 W Zarić M. Milana (1) 73 Zavšek Simon (2) 279 Waheed Ur Rehman (2) 217 Zdenĕk Chára (2) 151 Waheed Zeb (2) 217 Zhigang Shen (2) 161; (3) 293 Wajid Ali (2) 217 Zixia Li (3) 391

596 Journal of the Association of Chemical Engineers of Serbia, Belgrade, Serbia

Vol. 23 List of Referees in 2017 CI&CEQ

Chemical Industry and Chemical Engineering Quarterly (CI&CEQ) thanks the following people who have gene- rously contributed their time to the peer review process. The efficient continuing run of CI&CEQ could not be possible without their contribution.

Abou El Kacem Qaiss Moroccan Foundation for Advanced Science, Innovation and Research (MAScIR), Institute of Nanomaterials and Nanotechnology (NANOTECH), Laboratory of Polymer Processing, Rabat, Morocco; email: [email protected] Agnieszka Krzyzaniak Royal HaskoningDHV, PO Box 8520, Netherlands; email: [email protected] Ahmed. M. Saeed Department of Chemistry, College of Science, University of Wasit, Wasit, Iraq; email: [email protected] Aldrin Calderon Industrial Technology Development Institute, Department of Science and Technology, Taguig, Philippines; email: [email protected] Alicja Kucharska Department of Fruit, Vegetable and Cereals Technology, Wrocław University of Environmental and Life Sciences, Chełmońskiego 37, Wrocław, Poland; email: [email protected] Amit Kr. Singh Electrical and Electronics Engineering, Amity School of Engineering and Technology, Amity University, India; email: [email protected], [email protected] Ana Karla Abud Food Technology Department, Federal University of Sergipe, São Cristóvão, Sergipe, Brazil; email: [email protected] Ana María Chaux-Gutiérrez Bioscience, Language and Physical Sciences (IBILCE), UNESP-São Paulo State University, Rua Cristovão Colombo 2265, São José do Rio Preto, SP, Brazil; email: [email protected] Andrea Hinkova Institute of Chemical Technology in Prague, Department of Carbohydrates and Cereals, Technicka 5, Czech Republic; email: [email protected] Andreas Möller Quantachrome GmbH, Rudolf-Diesel-Straße 12, Odelzhausen, Germany; email: [email protected] Andrzej Krasinski Warsaw University of Technology, Faculty of Chemical and Process Engineering, Warynskiego 1, Warsaw, Poland; email: [email protected], [email protected] Angel Martin University of Valladolid, Industrial Engineering School, Department of Chemical Engineering and Environmental Technology, High Pressure Processes Group, C/Dr. Mergelina, s/n, Valladolid, Spain; email: [email protected] Anita Radovnikovic European Commission, Joint Research Centre (JRC), Directorate for Health, Consumers and Reference Materials, Via E. Fermi 2749, Ispra, VA, Italy; email: [email protected] Anjali Agrawal Department of Fabric and Apparel Science, Lady Irwin College, Delhi University, New Delhi 110001, India; email: [email protected] Arumugam Shanmuga Sundaram Department of Mechanical Engineering, Sri Chandrasekharendra Saraswathi Viswa Maha Vidyalaya, Enathur, Kanchipuram, Tamil Nadu, India; email: [email protected] Can Ertekin Department of Farm Machinery, Faculty of Agricultural Engineering, Akdeniz University, 07070, Antalya, Turkey; email: [email protected] Candida Milone Department of Electronic Engineering, Chemistry and Industrial Engineering, Contrada di Dio, Italy; email: [email protected] Carlos A. García-González Departamento de Farmacología, Farmacia y Tecnología Farmacéutica, Facultad de Farmacia, R+D Pharma group (GI-1645), and Health Research Institute of Santiago de Compostela (IDIS), Universidade de Santiago de Compostela, Santiago de Compostela, Spain; email: [email protected]

Vol. 23 List of Referees in 2017 CI&CEQ

Carmen Cuadrado Departamento de Tecnología de Alimentos, SGIT-INIA, Ctra. de La Coruña km 7.5, Spain; email: [email protected] Carmen Falagán Rodriguez Bangor University, Brambell Building, Deiniol Road, Bangor, United Kingdom; email: [email protected] Chau-Chyun Chen Department of Chemical Engineering, Texas Tech University, Lubbock, TX, United States; email: [email protected] Cheng Huaide Key Laboratory of Salt Lake Resources and Chemistry, Qinghai Institute of Salt Lakes, Chinese Academy of Sciences, No 18 Xinning Road, Xi'ning, Qinghai, China; email: [email protected] Chinmaya Prasad Lenka Improvement Engineer, Sadara Chemical Company, Saudi Arabia; email: [email protected], [email protected] Christian John Engelsen SINTEF Building and Infrastructure Institute, Richard Birkelands vei 3, Trondheim, Norway; email: [email protected] Christos Kordulis Department of Chemistry, University of Patras, Patras, Greece; email: [email protected] Chuanfang Yang Key Laboratory of Green Process and Engineering, Institute of Process Engineering, Chinese Academy of Sciences, Chine; email: [email protected] Conglu Zhang School of Pharmaceutical Engineering, Shenyang Pharmaceutical University, Shenyang, China; email: [email protected] Danielli Alessandra Reino Olegário da Silva Universidade Federal do Paraná, Pós-Graduação em Engenharia de Alimentos, Curitiba, PR, Brazil; email: [email protected] Dejan Skala Faculty of Technology and Metallurgy, University of Belgrade, Karnegijeva 4, Serbia; email: [email protected], [email protected] Deniz Akin Şahbaz Afyon Kocatepe University, Faculty of Engineering, Department of Chemical Engineering, Afyonkarahisar, Turkey; email: [email protected] Dhia H. Hussain Department of Chemistry, College of Science, University of Al-Mustansiriyah, Baghdad, Iraq; email: [email protected] Dhoni Hartanto Department of Chemical Engineering, Faculty of Engineering, Universitas Negeri Semarang, Kampus Sekaran Gunungpati, Semarang, Indonesia; email: [email protected] Di Wu Vice Chief Engineer, Production Chemistry, Daqing Oilfield Engineering Co., Ltd., Tel. +86-459-5903858 Mobile: +86-18645997136 email: wud_dod@petrochina.com.cn Dragan D. Govedarica Faculty of Technology, Department of Oil and Petrochemical Engineering, University of Novi Sad, Bulevar cara Lazara 1, Novi Sad, Serbia; email: [email protected] Dunja Sokolović Faculty of Technical Sciences, University of Novi Sad, Trg Dositeja Obradovića 6, 21000 Novi Sad, Serbia; email: [email protected], [email protected] E. Abdul Jaleel Chemical Engineering Department, National Institute of Technology, Calicut, India; email: [email protected] Ebru K. Akpinar Mechanical Engineering Department, Firat University, Elazig, Turkey; email: [email protected] Emilio Benfenati IRCCS - Istituto di Ricerche Farmacologiche Mario Negri, Laboratory of Environmental Chemistry and Toxicology, Via G. La Masa 19, Italy; email: [email protected] Eric Fossum Department of Chemistry, Wright State University, 202 Oelman Hall, 3640 Colonel Glenn Highway, United States; email: [email protected] Eugeniusz Milchert Instytut Technologii Chemicznej Organicznej, Wydział Technologii i Inzynierii Chemicznej, Zachodniopomorski Uniwersytet Technologiczny w Szczecinie, ul. Pułaskiego 10, Szczecin, Poland; [email protected], [email protected] Evanly Vo National Institute for Occupational Safety and Health, National Personal Protective Technology Laboratory, 626 Cochrans Mill Road, Pittsburgh, PA, United States; email: [email protected] Evgenij V. Pylypchuk Nanomaterials Department, Chuiko Institute of Surface Chemistry of the National Academy of Sciences of Ukraine, 17 General Naumov Str., Kyiv, Ukraine; email: [email protected] Fawaz Husaini Master of Public Health Program Study, Medicine Faculty, Lambung Mangkurat University, South Kalimantan, Indonesia; email: [email protected]

2 Vol. 23 List of Referees in 2017 CI&CEQ

Feng Ling Yang School of Mechanical EngineeringShandong University Jinan China; email:, [email protected] G.S. Kapur Indian Oil Corporation Limited, R and D Division, India; email: [email protected] Giorgos Markou Department of Natural Resources Management and Agricultural Engineering, Agricultural University of Athens, Iera Odos 75, Athens, Greece; email: [email protected] Gökhan Gürlek Ege University, Engineering Faculty, Department of Mechanical Engineering 35100 Bornova Kampüs İzmir, Turkey; email: [email protected] Gracielle Johann Bioprocesss Engineering and Biotechnology Course, Federal Technological University of Paraná, Estrada para Boa Esperança, Km 04, Dois Vizinhos, PR CEP, Brazil; email: [email protected] Hee Taik Kim Department of Chemical Engineering, Hanyang University, 1271 Sa 3-dong, Sangnok- gu, South Korea; email: [email protected] Helena Passos CICECO - Aveiro Institute of Materials, University of Aveiro, Campus Universitário de Santiago, 3810-193 Aveiro, Portugal; email: [email protected] Hengbo Yin Faculty of Chemistry and Chemical Engineering, Jiangsu University, Zhenjiang, China; email: [email protected] Heuy Dong Kim Andong National University, Andong, South Korea; email: [email protected] Hicham Zaitan Laboratory LCMC, Faculty of Science and Technology, Sidi Mohamed Ben Abdellah University, B.P. 2202, Fez, Morocco; email:[email protected] Huan Zhou Tianjin Key Laboratory of Marine Resources and Chemistry, College of Chemical Engineering and Materials Science, Tianjin University of Science and Technology, China; email:, [email protected] Huang Si School of Mechanical and Automotive Engineering, South China University of Technology, China; email: [email protected] Hwei Voon Lee Nanotechnology & Catalysis Research Centre (NanoCat), Institute Postgraduate Studies, University Malaya, Kuala Lumpur, Malaysia; email: [email protected] Ibrahim Doymaz Department of Chemical Engineering, Yildiz Technical University, Esenler, Istanbul, Turkey; email: [email protected] Ioan Iordache National Research and Development, Institute for Cryogenics and Isotopic Technologies, I.C.I.T. Ramnicu Valcea, 4 Uzinei Street, Romania; [email protected] Ivan Fort Department of Process Engineering, Faculty of Mechanical Engineering, Czech Technical University in Prague, Technická 6, 166 07 Prague 6, Czech Republic; email: [email protected] Jarmila Trpcevska Faculty of Metallurgy, Technical University of Košice, Letná 9, 040 01 Košice, Slovak Republic; email: [email protected] Jeff Gostick McGill University, Department of Chemical Engineering, Montreal, Quebec H3A 0C5, Canada; email: [email protected], [email protected] Jéssica López Pastén Food Engineering Department, La Serena University, Raúl Bitrán Avenue, La Serena, Region of Coquimbo, Chile; email: [email protected] Jilin Cao Hebei Provincial Key Lab of Green Chemical Technology and High Efficient Energy Saving, College of Chemical Engineering and Technology, Hebei University of Technology, Tianjin, China; email: [email protected] Joanna Karcz Department of Chemical Engineering, West Pomeranian University of Technology, Szczecin, Poland; email: [email protected] Johan Jacquemin Université Francois Rabelais, Laboratoire PCM2E, Parc de Grandmont, Tours, France; email: [email protected] Jonjaua Ranogajec University of Novi Sad, Faculty of Technology Novi Sad, Bulevar cara Lazara 1, 21 000 Novi Sad, Serbia; email: [email protected] Jorge Cuéllar Antequera Department of Chemical Engineering, University of Salamanca, Plaza de los Caídos 1- 5, Salamanca, Spain; email: [email protected] Juan A. Reyes-Labarta Department of Chemical Engineering, University of Alicante, Ap. Correos 99, Alicante 03080, Spain; email: [email protected] Juan Carlos Rendón-Angeles Research Institute of Advanced Studies, National Polytechnic Institute, Ceramic Engineering, Saltillo Campus, Ramos Arizpe, Coahuila, Mexico; email: [email protected] Juan Manuel Peralta Hernández Exactas, Universidad de Guanajuato, Cerro de la Venada s/n, Pueblito de Rocha, Guanajuato, Mexico; email: [email protected]

3 Vol. 23 List of Referees in 2017 CI&CEQ

Khaled M. Elattar Department of Chemistry, Faculty of Science, Mansoura University, Mansoura 35516, Egypt; email: [email protected] L. Rong School of Biological Science and Medical Engineering, Beihang University, China; email: [email protected], [email protected] Lazar A. Morgenstern Israel; email: [email protected] Mallika Boonmee Department of Biotechnology, Faculty of Technology, Khon Kaen University, Khon Kaen, Thailand; email: [email protected] Mallika Boonmee Department of Biotechnology, Faculty of Technology, Khon Kaen University, Khon Kaen, Thailand; email: [email protected] Mandy Klauck Hochschule für Technik und Wirtschaft - Univ. of Applied Sciences, Department of Chemical Engineering, Friedrich-List-Platz 1, Germany; email: [email protected] Manuel Moliner Instituto de Tecnología Química, Universidad Politécnica de Valencia (UPV), Consejo Superior de Investigaciones Científicas (CSIC), Valencia, Spain; email: [email protected] Marcio A. Mazutti Department of Chemical Engineering, Federal University of Santa Maria, UFSM, Av. Roraima, 1000, Brazil; email: [email protected] María Salomé Álvarez Álvarez Department of Chemical Engineering, University of Vigo, Vigo, Spain; email: [email protected] Martin Bertau TU Bergakademie Freiberg, Institut für Technische Chemie, Leipziger Straße 29, Freiberg, Germany; [email protected] Mateja Kert Department of Textiles, University of Ljubljana, Slovenia; email: [email protected] Mei Hong College of Chemical Engineering, Nanjing Forestry University, 159 Longpan Road, Nanjing, China; email: [email protected]; [email protected] Menwer Attarakih Thermal Process Engineering, The University of Kaiserslautern, Kaiserslautern, Germany; email: [email protected] Meysam Torab Mostaedi Nuclear Science and Technology Research Institute, Tehran, Iran; email: [email protected] Michael Berer Polymer Competence Center Leoben GmbH, Roseggerstrasse 12, Leoben, Austria; email: [email protected] Michael Steiger Fachbereich Chemie, Institut für Anorganische und Angewandte Chemie, Universität Hamburg, Martin-Luther-King-Platz 6, Germany; email: [email protected] Miguel A. Pedroza-Toscano Departamento de Investigación, Centro Universitario UTEG, Guadalajara, JAL, Mexico; email: [email protected] Min Pan School of Environmental Science and Engineering, Xiamen University of Technology, Xiamen, P. R. China; email: [email protected], [email protected] Mirjana Kijevčanin Faculty of Technology and Metallurgy, University of belgrade, Belgrade, Serbia; email: [email protected] Moacir Cardoso Elias Departamento de Ciência e Tecnologia Agroindustrial, Universidade Federal de Pelotas, Pelotas, RS, Brazil; email: [email protected] Mohammad Zahid Hossain REC Silicon email: [email protected] Morten Hammer SINTEF Energy Research, P.O. Box 4761 Sluppen, Trondheim, Norway; emil: [email protected] Moye Ajao Research Unit on Energy Efficiency and Sustainable Development of the Forest Biorefinery, Chemical Engineering Department, Polytechnique Montreal, C.P. 6079 Succ. Centre-Ville, Montréal, QC, Canada; email: [email protected] Myung Shin Ryu Department of Materials Science and Engineering, KAIST, 291 Daehak-ro, Yuseong-gu, Republic of Korea; e-mail: [email protected] Navdeep Singh Department of Civil Engineering, Dr. B R Ambedkar National Institute of Technology, Jalandhar, India; email: [email protected] Nicolas Gañan IDTQ - Grupo Vinculado PLAPIQUI, CONICET, Universidad Nacional de Córdoba, Av. Velez Sarsfield 1611, Córdoba, Argentina; email: [email protected] Nor Azwadi Department of Thermofluids, Faculty of Mechanical Engineering, Universiti Teknologi Malaysia, 81310 UTM Skudai, Johor Bahru, Malaysia; email: [email protected] Norma Alias CSNano, Ibnu Sina Institute for Scientific and Industrial Research, Universiti Teknologi Malaysia, UTM, Johor Bahru, Johor, Malaysia; email: [email protected]

4 Vol. 23 List of Referees in 2017 CI&CEQ

Nurettin Eltugral Department of Metallurgical and Materials Engineering, Karabuk University, Karabuk, Turkey; email: [email protected] Oldřich Živný Institute of Plasma Physics of the CAS, v.v.i, Za Slovankou 1782/3, Prague 8, Czech Republic; email: [email protected] Owen J. Curnow Department of Chemistry, University of Canterbury, Private Bag 4800, Christchurch, New Zealand; email: [email protected] Parimal Pal Environment & Membrane Technology Laboratory, Department of Chemical Engineering, National Institute of Technology, Durgapur, India; email: [email protected] Pascale Chalier Unité Mixte de Recherche, Ingénierie des Agropolymères et Technologies Emergentes 1208 (UMR IATE), INRA, SupAgro, Cirad, Université de Montpellier II, Place Viala, Montpellier cedex 01, France; email: [email protected] Paulina Sosa Wetsus and Wageningen University, NL, email: [email protected] Paweł Parzuchowski Warsaw University of Technology, Faculty of Chemistry, Noakowskiego 3, Poland; email: [email protected] Qing Zhao Key Laboratory for Ecological Metallurgy of Multimetallic Ores (Ministry of Education), Northeastern University, Shenyang, China; email: [email protected] Quan Zhu College of Chemical Engineering, Sichuan University, Chengdu, China; email: [email protected] Radwan Dweiri Department of Materials Engineering, Faculty of Engineering, Al-Balqa' Applied University-BAU, Jordan; email: [email protected] Razana Alwee Soft Computing Research Group, Faculty of Computing, Universiti Teknologi Malaysia, 81310 Skudai, Johor, Malaysia; email: [email protected] Regiane Victória de Barros Fernandes Institute of Agricultural Sciences, Campus de Rio Paranaíba, Federal University of Viçosa, Brazil; email: [email protected] René Ruby-Figueroa Programa Institucional de Fomento a la Investigación, Desarrollo e Innovación, Universidad Tecnológica Metropolitana, Ignacio Valdivieso 2409 P.O. Box, San Joaquín, Santiago, Chile; email: [email protected] Ricardo Arbach Fernandes de Oliveira Federal University os São Carlos, Department of Chemical Engineering, São Carlos - São Paulo – Brazil; email: [email protected] Roman Kozlovskiy Departments of Petrochemistry and Biotechnology, D. Mendeleev University of Chemical Technology of Russia, Moscow, Russia; email: [email protected] Rome-Ming Wu Department of Chemical and Materials Engineering, Tamkang University, Tamsui, Taiwan; email: [email protected] Rosiah Rohani Department of Chemical and Process Engineering, Faculty of Engineering and Built Environment, Universiti Kebangsaan Malaysia, UKM, Bangi, Selangor, Malaysia; email: [email protected] S.G. Da Cruz Metallurgical and Materials Engineering Department, PEMM, COPPE, Federal University of Rio de Janeiro, P.O. Box 68505, Rio de Janeiro, RJ, Brazil; email: [email protected] Sayeda Saleh Mohamed Chemistry of Natural and Microbial Products Department, National Research Center, Egypt; email: [email protected] Seyfi Şevik Hitit University, Vocational School of Technical Sciences, Electrical and Energy, Çorum, Turkey; [email protected] Seyhun A. Kipçak Department of Chemical Engineering, Faculty of Chemical and Metallurgical Engineering, Yildiz Technical University, Davutpasa Street No. 127, Esenler, Istanbul, Turkey; email: [email protected] Seyhun Kipcak Yıldız Technical University, Davutpasa Campus, Faculty of Chemical and Metallurgical Engineering, Department of Chemical Engineering, Chemical Technologies, Istanbul, Turkey; email: e-mail: [email protected] Sharmila Govindasamy Department of Industrial Biotechnology, Government College of Technology, Coimbatore, India; email: [email protected] Sibel Yağcı Food Engineering Department, Karamanoğlu Mehmetbey University, Karaman, Turkey; email: [email protected] Sitaraman Krishnan Department of Chemical and Biomolecular Engineering, Clarkson University, Potsdam, NY, United States; email: [email protected] Sohail Yasin University of Lille 1, Sciences et Technologies, France; email: [email protected]

5 Vol. 23 List of Referees in 2017 CI&CEQ

Stoyan Gaydardzhiev University of Liege, GeMMe – Minerals Processing and Recycling, Sart-Tilman, Allée de la Découverte, Liège, Belgium; email: [email protected] Susana Casal REQUIMTE, Faculty of Pharmacy, University of Porto, Rua de Jorge Viterbo Ferreira, 228, Portugal; email: [email protected] Svetozar Musić Ruder Bošković Institute, P.O. Box 180, Croatia; email: [email protected], [email protected] Tamer Mohamed Ismail Mechanical Engineering Department, Suez Canal University, Ismailia, Egypt; email: [email protected] Tatjana Trtić-Petrović Laboratory of Physics, Vinča Institute of Nuclear Sciences, University of Belgrade, P.O. Box 522, Belgrade, Serbia; email: [email protected] Teofil Jesionowski Institute of Chemical Technology and Engineering, Faculty of Chemical Technology, Poznan University of Technology, Berdychowo 4, Poznan, Poland; email: [email protected] Thanakorn Rojanakorn Department of Food Technology, Faculty of Technology, Khon Kaen University, Khon Kaen, Thailand; email: [email protected] Thanit Swasdisevi School of Energy, Environment and Materials, King Mongkut's University of Technology Thonburi, 126 Pracha U-tid Road, Thailand; email: [email protected] Thongthai Witoon National Center of Excellence for Petroleum, Petrochemicals and Advance Material, Department of Chemical Engineering, Faculty of Engineering, Kasetsart University, Bangkok, Thailand; email: [email protected] Torrez R.M. Irigoyen Centro de Investigación y Desarrollo en Criotecnología de Alimentos (CIDCA), Universidad Nacional de La Plata-CONICET-CIC, Calle 47 y 116, La Plata, Argentina; email: [email protected] Tsvetina Dobrovolska Institute of Physical Chemistry Bulgarian Academy of Sciences, Institute of Physical Chemistry, Sofia, Bulgaria; email: [email protected] Tuncay Yilmaz Engineering Faculty, Food Engineering Department, Manisa Celal Bayar University, Manisa, Turkey; email: [email protected] Valentin N. Sapunov Mendeleev University of Chemical Technology of Russia, Moscow 125047, Russian Federation; email: [email protected] Valter Bruno Silva Quinta de Prados, Apartado 1013, Vila Real, Portugal; email: [email protected] Velisa Vesovic Department of Earth Science & Engineering, Imperial College London, United Kingdom; emil: [email protected] Venu Borugadda Catalysis and Chemical Reaction Engineering Laboratories, Department of Chemical and Biological Engineering, University of Saskatchewan, Saskatoon, SK, Canada; email: [email protected] Vilma Kaškonienė Department of Biochemistry and Biotechnologies, Vytautas Magnus University, Vileikos Str. 8, Lithuania; email: [email protected] Wihdrawn by authors Recenzenti pristali ali nisu na vreme poslali recenzije iako su cise puta opominjani I autori povukli rad. Zato su oni I skinuti sa liste Yang Liu Key Laboratory of Energy Thermal Conversion and Control of Ministry of Education, School of Energy and Environment, Southeast University, Nanjing, China; email: [email protected] Yi Zhengji Key Laboratory of Functional Organometallic Materials of Hengyang Normal University, College of Hunan Province, Hengyang, Hunan 421008, China: email: [email protected] Yong Jia School of Chemical Engineering, Nanjing University of Science and Technology, China; email: [email protected] Yu Hui School of Chemical Engineering, Sichuan University, Chengdu, China; email: [email protected] Yuting Tan School of Chemical Science and Engineering, Royal Institute of Technology, Stockholm, Sweden; email: [email protected]