The Synthesis of Polyol from Rice Bran Oil (RBO) through Epoxidation and Hydroxylation Reactions

Transcription

1 The Synthesis of Polyol from Rice Bran Oil (RBO) through Epoxidation and Hydroxylation Reactions by Edy Purwanto School of Chemical Engineering The University of Adelaide A thesis submitted for the degree of Master of Engineering Science July 00

2 Declaration This work contains no material which has been accepted for the award of any other degree or diploma in any university or other tertiary institution to Edy Purwanto and, to the best of my knowledge and belief, contains no material previously published or written by another person, except where due reference has been made in the text. I give consent to this copy of my thesis when deposited in the University Library, being made ailable for loan and photocopying, subject to the provisions of the Copyright Act 968. The author acknowledges that copyright of published works contained within this thesis (as listed below) resides with the copyright holders(s) of those works. Conference paper: E. Purwanto, Y. Ngothai, B. O Neill, and K. Bremmell, Optimization of epoxidation reaction of rice bran oil using response surface methodology, Proceedings: Chemeca th Australasian Chemical Engineering Conference, The Institution of Engineers, Perth, Australia, 7 30 September 009, ISBN: , CD-ROM. I also give permission for the digital version of my thesis to be made ailable on the web, via the University s digital research repository, the Library catalogue, the Australasian Digital Theses Program (ADTP) and also through web search engines, unless permission has been granted by the University to restrict access for a period of time. Mr. Edy Purwanto :. Date : 9 July 00 ii

3 Summary Polyurethanes are valuable polymers with a wide variety of applications. They are normally produced from polyol feedstocks derived from petroleum. As petroleum is a non-renewable resource, an alternative source of feedstock is sought. A potential source is rice bran oil. However, far too little attention has been paid to the utilization of rice bran oil as a potential raw material to produce polyol as it contains unsaturated fatty acids that can be converted to polyol and is the by product of rice milling process and ailable at very low cost. There are two sequential processes to produce polyol from rice bran oil, namely the epoxidation and hydroxylation reactions. In this work, the optimal conditions in the epoxidation reaction were investigated using acetic acid and formic acid as oxygen carriers in terms of reaction time and temperature. Furthermore, the reaction kinetics were also determined using formic acid as an oxygen carrier in the epoxidation step. Finally, the influence of reaction time and temperature in the hydroxylation step were also investigated in this study. In order to determine the optimal condition, the epoxidation reaction was performed in a three neck flask with the use of acetic and formic acid as oxygen carriers. Result shows that the conversion of iodine value increased with reaction time and temperature when acetic acid was used as an oxygen carrier (peroxyacetic acid). Interestingly, the oxirane content increased with reaction time and temperature then declined after hing achieved the optimal point. The optimal condition was achieved at a reaction time of 4.3 h and a temperature of 63.8 o C by performing response surface methodology. The conversion of iodine value also displayed similar behiour during the epoxidation reaction when formic acid was used as an oxygen carrier (peroxyformic acid), namely the conversion increased with reaction time and temperature. The measured rate constants were 0.7h - (40 o C), 0.304h - (50 o C), 0.374h - (60 o C), 0.45h - (70 o C) and 0.49h - (80 o C). The activation energy was.6 kj/mol and the epoxidation reaction was pseudo-first order with respect to the concentration of iii

4 double bonds in the oil. Interestingly, peroxyformic acid shows improved performance as indicated by higher content of maximal oxirane content 3.6% compared to peroxyacetic acid which is only.6%. The optimal condition with the use of formic acid as an oxygen carrier was obtained at reaction time of 4 h and temperature of 60 o C. In the hydroxylation step, results indicate that the hydroxyl value of polyol was a quadratic function of reaction time and temperature and the optimal condition was achieved at a reaction time of 5.5 min and temperature of 49 o C, with maximal hydroxyl value 6.5 mg KOH/g oil by performing response surface methodology. The viscosity of polyol increased with reaction time and temperature and resulted in polyol with viscosity in the range Temperature was found to he the most significant effect on the viscosity of polyol. The results of this study confirm the potential of rice bran oil as a feedstock for synthesis of polyol and show that the optimal conditions in the epoxidation and hydroxylation reactions are a key control variable to obtain a high quality of polyol. iv

5 Acknowledgement I would like to express my appreciation to numerous people who he greatly contributed and assisted me to complete this research study. In particular I would like to acknowledge: Dr Yung Ngothai, School of Chemical Engineering, University of Adelaide, as principal supervisor for the supervision, motivation, ideas, discussions, experience in the class and the opportunity to conduct research in the laboratory. A/Prof Brian O Neill, School of Chemical Engineering, University of Adelaide, as co-supervisor for the supervision, support, ideas, discussions and guidance for design of the experiments and the use of response surface methodology which is a new knowledge for me. Dr Kristen Bremmell, School of Pharmacy and Medical Sciences, University of South Australia, as co-supervisor for the supervision, ideas and discussions through this project. A/Prof Dzuy Nguyen, School of Chemical Engineering, University of Adelaide for permission to access Rheology Laboratory and viscometer device. Andrew Wright, Leanne Redding, Jason Peak and the workshop for assistance in the laboratory, construction and modification of apparatus; Thana Deawwanich for guidance to operate viscometer apparatus; Gideon Bani Kuncoro and Kan Li my colleagues, for support, motivation, and any discussions in the office. I would like to dedicate this thesis to my wife, Nanik Hasanah and my son, Nawfal Adiva Purwanto. I hope this thesis would provide a great contribution to the community and satisfy with the expectations of the related people. v

6 Table of Contents Declaration ii Summary iii Acknowledgment v Table of Contents vi List of Figures ix List of Tables x Nomenclature xv INTRODUCTION LITERATURE REVIEW 3. Rice Bran Oil (RBO) 3. Epoxidation Reaction 6.3 Hydroxylation Reaction 9.4 Synthesis of Polyol.5 Response Surface Methodology.6 Key Research Questions 3.7 Research Objectives 5.8 Significant/Contribution to the Discipline 5 3 MATERIALS and METHODS 8 3. Materials 8 3. Epoxidation Reaction Hydroxylation Reaction Epoxidation Test 3.4. Iodine Value Analysis 3.4. Oxirane Oxygen Content Analysis 3.5 Hydroxylation Test Hydroxyl Value Analysis Viscosity Analysis 4 4 EXPERIMENTAL RESULTS and DISCUSSION 5 4. Epoxidation of RBO - Acetic Acid as an Oxygen Carrier ( st Study) Experimental Design and Optimization of the Epoxidation Reaction 5 vi

7 4.. Statistical Analysis Effects of Reaction Time and Temperature on Reaction Conversion Effect of Reaction Time and Temperature on Oxirane Content 3 4. Epoxidation of RBO Formic Acid as an Alternate Oxygen Carrier ( nd Study) Effects of Reaction Time and Temperature on the Conversion Reaction Kinetics Effects of Reaction Time and Temperature on Oxirane Content Hydroxylation of Epoxidized RBO (3 rd Study) Experimental Design and Optimization of the Hydroxylation Reaction Statistical Analysis Effects of Reaction Time and Temperature on Hydroxyl Value Effects of Reaction Time and Temperature on Viscosity of Polyol 48 5 CONCLUSION 5 6 RECOMMENDATIONS FOR FUTURE RESEARCH 53 REFERENCES 54 Appendix A Calculation of Epoxidation Reaction 58 Appendix B Calculation of Hydroxylation Reaction 6 Appendix C Epoxidation Using Acetic Acid as an Oxygen Carrier ( st Study) 63 C. Experimental Design 63 C. Determination of Iodine Value and Conversion 65 C.3 Oxirane Content 66 C.4 Determination of the Optimal Condition 67 vii

8 Appendix D Epoxidation Using Formic Acid as an Oxygen Carrier ( nd Study) 69 D. Determination of Iodine Value and Conversion 69 D. Reaction Kinetics 73 D.3 Oxirane Oxygen Content 80 Appendix E Hydroxylation of Epoxidized Oil 84 E. Experimental Design 84 E. Determination of Hydroxyl Value 85 E.3 Determination of the Optimal Condition 88 E.4 Determination of the Viscosity of the Polyol Products 89 viii

9 List of Figures Figure - Structure of rice kernel 4 Figure - An epoxide 7 Figure -3 Epoxidation reaction 7 Figure -4 Ring opening mechanism using acid catalyst 0 Figure -5 Ring opening mechanism using base catalyst 0 Figure -6 Epoxidation reaction mechanism Figure -7 Hydroxylation reaction mechanism Figure 3- Experimental apparatus of epoxidation reaction 9 Figure 3- Experimental apparatus of hydroxylation reaction 0 Figure 4- Effects of reaction time (X ) and temperature(x ) on reaction conversion for acetic acid as an oxygen carrier 30 Figure 4- Effects of reaction time (X )and temperature (X ) on oxirane content for acetic acid as an oxygen carrier 3 Figure 4-3 Effects of reaction time (t) on reaction conversion (X) for formic acid as an oxygen carrier at different temperatures 34 Figure 4-4 Determination of activation energy for epoxidation using formic acid as an oxygen carrier 37 Figure 4-5 Effects of reaction time (t) and temperature on oxirane content (%) for formic acid as an oxygen carrier at different temperatures 39 Figure 4-6 Effects of reaction time (X ) and temperature (X ) on hydroxyl value of polyol 46 Figure 4-7 Effects of reaction time (X ) and temperature (X ) on viscosity of polyol 49 Figure 4-8 Sample of polyol produced 50 Figure D- Plot reaction time vs ln/(-x) at T = 40 o C 75 Figure D- Plot reaction time vs ln/(-x) at T = 50 o C 76 Figure D-3 Plot reaction time vs ln/(-x) at T = 60 o C 77 Figure D-4 Plot reaction time vs ln/(-x) at T = 70 o C 78 Figure D-5 Plot reaction time vs ln/(-x) at T = 80 o C 79 Figure D-6 Plot ln(k) vs /T 80 ix

10 List of Tables Table - Typical fatty acid composition (wt%) of RBO 6 Table - Physical and chemical characteristics of RBO 6 Table 4- The range and levels of variables used in the RSM procedure to determine the optimum conditions for the epoxidation reaction of RBO 6 Table4- CCD and response in terms of conversion and oxirane content during the epoxidation of RBO 7 Table 4-3 Regression statistics for conversion and oxirane content for epoxidation of RBO 8 Table 4-4 Analysis of variance (ANOVA) for conversion and oxirane content for epoxidation of RBO 8 Table 4-5 Significance of regression coefficients for conversion of epoxidation of RBO 9 Table 4-6 Significance of regression coefficients for oxirane content of epoxidation reaction 9 Table 4-7 Rate constant value at different temperature for epoxidation using formic acid 36 Table 4-8 The range and levels of variables for hydroxylation reaction 4 Table 4-9 CCD and response in terms of hydroxyl value and viscosity of polyol 43 Table 4-0 Regression statistics for hydroxyl value and viscosity of polyol 44 Table 4- Analyses Of Variance (ANOVA) for hydroxyl value and viscosity of polyol 44 Table 4- Significance of regression coefficients for hydroxyl value of polyol 44 Table 4-3 Significance of regression coefficients for viscosity of polyol 45 Table A- Fatty acid composition of rice bran oil 58 Table C- Range and levels of variables 64 Table C- Experimental design of epoxidation using acetic acid as an oxygen carrier 64 Table C-3 Experimental results of iodine value & conversion 66 Table C-4 Experimental results of oxirane oxygen content 67 x

11 Table D- Experimental results of Iodine Value (IV) and Conversion (%X) at 40 o C 7 Table D- Experimental results of Iodine Value (IV) and Conversion (%X) at 50 o C 7 Table D-3 Experimental results of Iodine Value (IV) and Conversion (%X) at 60 o C 7 Table D-4 Experimental results of Iodine Value (IV) and Conversion (%X) at 70 o C 7 Table D-5 Experimental Results of Iodine Value (IV) and Conversion (%X) at 80 o C 7 Table D-6 Determination of k value at 40 o C 75 Table D-7 Determination of k value at 50 o C 75 Table D-8 Determination of k value at 60 o C 76 Table D-9 Determination of k value at 70 o C 77 Table D-0 Determination of k value at 80 o C 78 Table D- Calculation of activation energy 79 Table D- Experimental results of oxirane oxygen content at T = 40 o C 8 Table D-3 Experimental results of oxirane oxygen content at T = 50 o C 8 Table D-4 Experimental results of oxirane oxygen content at T = 60 o C 8 Table D-5 Experimental results of oxirane oxygen content at T = 70 o C 83 Table D-6 Experimental results of oxirane oxygen content at T = 80 o C 83 Table E- Range and levels of hydroxylation reaction 85 Table E- Experimental design of hydroxylation reaction 85 Table E-3 Saponification value prior to acetylation 86 Table E-4 Saponification value after acetylation 87 Table E-5 Hydroxyl value of polyol 88 Table E-6 Viscosity at shear rate s -, t = 0 min and T = 30 o C 90 Table E-7 Viscosity at shear rate s -, t = 0 min and T = 30 o C 90 Table E-8 Viscosity at shear rate 4.9 s -, t = 0 min and T = 30 o C 9 Table E-9 Viscosity at shear rate s -, t = 0 min and T = 30 o C 9 Table E-0 Viscosity at shear rate s -, t = 0min and T = 30 o C 9 Table E- Viscosity at shear rate s -, t = 0 min and T = 30 o C 9 Table E- Viscosity at shear rate 9.95 s -, t = 0 min and T = 30 o C 9 Table E-3 Viscosity at shear rate s -, t = 80 min and T = 40 o C 93 xi

12 Table E-4 Viscosity at shear rate s -, t = 80 min and T = 40 o C 93 Table E-5 Viscosity at shear rate 4.9 s -, t = 80 min and T = 40 o C 93 Table E-6 Viscosity at shear rate s -, t = 80 min and T = 40 o C 94 Table E-7 Viscosity at shear rate s -, t = 80 min and T = 40 o C 94 Table E-8 Viscosity at shear rate s -, t = 80 min and T = 40 o C 94 Table E-9 Viscosity at shear rate 9.95 s -, t = 80 min and T = 40 o C 95 Table E-0 Viscosity at shear rate s -, t = 60 min and T = 40 o C 96 Table E- Viscosity at shear rate s -, t = 60 min and T = 40 o C 96 Table E- Viscosity at shear rate 4.9 s -, t = 60 min and T = 40 o C 96 Table E-3 Viscosity at shear rate s -, t = 60 min and T = 40 o C 97 Table E-4 Viscosity at shear rate s -, t = 60 min and T = 40 o C 97 Table E-5 Viscosity at shear rate 55.00s -, t = 60 min and T = 40 o C 97 Table E-6 Viscosity at shear rate 9.95 s -, t = 60 min and T = 40 o C 98 Table E-7 Viscosity at shear rate s -, t = 40 min and T = 50 o C 99 Table E-8 Viscosity at shear rate s -, t = 40 min and T = 50 o C 99 Table E-9 Viscosity at shear rate 4.9 s -, t = 40 min and T = 50 o C 99 Table E-30 Viscosity at shear rate s -, t = 40 min and T = 50 o C 00 Table E-3 Viscosity at shear rate s -, t = 40 min and T = 50 o C 00 Table E-3 Viscosity at shear rate s -, t = 40 min and T = 50 o C 00 Table E-33 Viscosity at shear rate 9.95 s -, t = 40 min and T = 50 o C 0 Table E-34 Viscosity at shear rate s -, t = 0 min and T = 50 o C 0 Table E-35 Viscosity at shear rate s -, t = 0 min and T = 50 o C 0 Table E-36 Viscosity at shear rate 0.0 s -, t = 0 min and T = 50 o C 0 Table E-37 Viscosity at shear rate s -, t = 0 min and T = 50 o C 03 Table E-38 Viscosity at shear rate s -, t = 0 min and T = 50 o C 03 Table E-39 Viscosity at shear rate s -, t = 0 min and T = 50 o C 03 Table E-40 Viscosity at shear rate 5.05 s -, t = 0 min and T = 50 o C 04 Table E-4 Viscosity at shear rate s -, t = 00 min and T = 50 o C 05 Table E-4 Viscosity at shear rate s -, t = 00 min and T = 50 o C 05 Table E-43 Viscosity at shear rate 4.9 s -, t = 00 min and T = 50 o C 05 Table E-44 Viscosity at shear rate s -, t = 00 min and T = 50 o C 06 Table E-45 Viscosity at shear rate s -, t = 00 min and T = 50 o C 06 Table E-46 Viscosity at shear rate s -, t = 00 min and T = 50 o C 06 Table E-47 Viscosity at shear rate 9.95 s -, t = 00 min and T = 50 o C 07 xii

13 Table E-48 Viscosity at shear rate s -, t = 80 min and T = 60 o C 08 Table E-49 Viscosity at shear rate 4.9 s -, t = 80 min and T = 60 o C 08 Table E-50 Viscosity at shear rate s -, t = 80 min and T = 60 o C 08 Table E-5 Viscosity at shear rate 9.95 s -, t = 80 min and T = 60 o C 09 Table E-5 Viscosity at shear rate s -, t = 80 min and T = 60 o C 09 Table E-53 Viscosity at shear rate s -, t = 80 min and T = 60 o C 09 Table E-54 Viscosity at shear rate s -, t = 80 min and T = 60 o C 0 Table E-55 Viscosity at shear rate s -, t = 60 min and T = 60 o C Table E-56 Viscosity at shear rate s -, t = 60 min and T = 60 o C Table E-57 Viscosity at shear rate 4.9 s -, t = 60 min and T = 60 o C Table E-58 Viscosity at shear rate s -, t = 60 min and T = 60 o C Table E-59 Viscosity at shear rate s -, t = 60 min and T = 60 o C Table E-60 Viscosity at shear rate s -, t = 60 min and T = 60 o C Table E-6 Viscosity at shear rate 9.95 s -, t = 60 min and T = 60 o C 3 Table E-6 Viscosity at shear rate s -, t = 0 min and T = 70 o C 4 Table E-63 Viscosity at shear rate s -, t = 0 min and T = 70 o C 4 Table E-64 Viscosity at shear rate 4.9 s -, t = 0 min and T = 70 o C 4 Table E-65 Viscosity at shear rate s -, t = 0 min and T = 70 o C 5 Table E-66 Viscosity at shear rate s -, t = 0 min and T = 70 o C 5 Table E-67 Viscosity at shear rate s -, t = 0 min and T = 70 o C 5 Table E-68 Viscosity at shear rate 9.95 s -, t = 0 min and T = 70 o C 6 Table E-69 Viscosity at shear rate s -, t = 0 min and T = 50 o C 7 Table E-70 Viscosity at shear rate s -, t = 0 min and T = 50 o C 7 Table E-7 Viscosity at shear rate 4.9 s -, t = 0 min and T = 50 o C 7 Table E-7 Viscosity at shear rate s -, t = 0 min and T = 50 o C 8 Table E-73 Viscosity at shear rate s -, t = 0 min and T = 50 o C 8 Table E-74 Viscosity at shear rate s -, t = 0 min and T = 50 o C 8 Table E-75 Viscosity at shear rate 9.95 s -, t = 0 min and T = 50 o C 9 Table E-76 Viscosity at shear rate s -, t = 0 min and T = 50 o C 0 Table E-77 Viscosity at shear rate s -, t = 0 min and T = 50 o C 0 Table E-78 Viscosity at shear rate 4.9 s -, t = 0 min and T = 50 o C 0 Table E-79 Viscosity at shear rate s -, t = 0 min and T = 50 o C Table E-80 Viscosity at shear rate s -, t = 0 min and T = 50 o C Table E-8 Viscosity at shear rate s -, t = 0 min and T = 50 o C xiii

14 Table E-8 Viscosity at shear rate 9.95 s -, t = 0 min and T = 50 o C xiv

15 Nomenclature E activation energy (kj/mol) k reaction rate constant (h - ) m mass (g) n reaction orders n moles number (mol) N normality (N) R gas constant (J/mol.K) R SA SB t t T V x X X Y correlation coefficient saponification value after acetylation (mg KOH/g oil) saponification value before acetylation (mg KOH/g oil) reaction time (h, for epoxidation reaction) reaction time (min, for hydroxylation reaction) temperature (K) volume (ml) real value of independent variables coded value of independent variables reaction conversion response variables Greek x ρ τ step change value density (g/ml) shear stress (Pa) γ shear rate (s - ) µ viscosity () Abbreviations ANOVA analysis of variance DB double bonds EO epoxidized oil FA formic acid xv

16 FFA HV IV MMT MW PFA PL PPO RBO RSM US free fatty acids hydroxyl value iodine value million metric tons molecular weight peroxyformic acid phospholipids polypropylene oxide rice bran oil response surface methodology united states xvi

17 INTRODUCTION Polyurethanes were first discovered by Professor Otto Bayer in 973 and are used in a wide variety of applications (Chian et al., 998). They he been exploited as coatings, thermoset and thermoplastic materials, adhesives and rigid or non-rigid foams (Guo et al., 006; Lligadas et al., 006; Wang et al., 009). Hence, the worldwide demand for polyols is projected to increase each year. Polyurethanes are generally produced from the reaction between polyols and diisocyanate (Pechar et al., 006). Polyols are normally derived from petroleum feedstocks and are known as petroleum-based polyols (Tu et al., 007). As the demand for polyols is increasing whilst the amount of petroleum is declining, alternative raw material for the production of polyols are needed. Worldwide rice production is roughly 500 million metric tons (MMT) per year (Jahani et al., 008). Rice bran oil (RBO) is a by product of the rice milling process; hence it cheap and readily ailable. RBO is a potential raw material that has not been explored as feedstock to produce polyol. RBO is attractive as it contains unsaturated bonds that can be converted to hydroxyl groups called polyols. Hence, RBO can be used as a raw material for polyol production. To date, rice bran has been used as a mixture of livestock food and as biomass fuel for boiler feed. The oil content of RBO is in the range of 5 3 wt% (Zullaikhah et al., 005). Hence, RBO has enormous potential as it is a renewable resource and research needs to be undertaken to develop and add value by converting the double bonds present in RBO to produce polyols. Moreover, the production cost of polyols could be reduced with the utilization of such by products and low priced feedstock. Polyol can be produced by two consecutive reactions, namely epoxidation and hydroxylation (Petrovic et al., 003). During the epoxidation stage, the double bonds in the vegetable oils are converted into epoxide (oxirane) groups whereas in the hydroxylation stage, the epoxide groups are converted to hydroxyl groups. The resulting product is a polyol as it contains more than one hydroxyl groups.

18 In the epoxidation of rubber seed oil using peroxyacetic acid, the ring opening of epoxide groups could be minimized when the epoxidation reaction was carried out at an intermediate temperature of o C. (Okieimen et al., 00). In the epoxidation of soybean oil using peroxyacetic acid followed by hydroxylation reaction using methanol, the hydroxyl value of soy-polyol of 69 mg KOH/g oil could be achieved whereas using olive oil as a raw material and m- chloroperoxybenzoic acid, the hydroxyl value of polyol product was 38 mg KOH/g oil (Petrovic et al., 003). However, far too little attention has been paid to the utilization of RBO as a potential raw material to produce polyol. Therefore, in the synthesis of polyol from RBO, the optimal operating conditions in the epoxidation and hydroxylation reactions he to be clearly defined to obtain a high quality of polyol since different raw materials will he different characteristics. High quality of polyol is indicated by high content of hydroxyl groups represented by hydroxyl value. The key goal of this project is to determine the optimal operating conditions for the epoxidation and hydroxylation reactions to produce a polyol synthesized from rice bran oil.

19 LITERATURE REVIEW This review will be divided into several parts. As mentioned in Chapter, the composition of potential feedstock is a crucial factor for the successful synthesis of polyol. Therefore, the first section will examine the potential of rice bran oil (RBO) as raw material for synthesis of polyol. The next section will consider the synthesis of polyol using epoxidation and hydroxylation reactions. As these reactions he strong influence on the quality of polyol product, a review on epoxidation and hydroxylation reactions will help to understand the process and contribution to polyol product. Following this, response surface methodology (RSM) will be discussed as a method to determine the optimal operating condition. Finally, research questions, objectives and significance of this research will be discussed as research direction and contribution of research topic to field area.. Rice Bran Oil (RBO) Nowadays, polyol plays an important aspect as a raw material to produce polyurethanes which he been widely used in polymer applications (Guo et al., 000, Hu et al., 00, Lligadas et al., 006, Wang et al., 007). The demand for polyol worldwide is predicted to increase each year. Generally, polyols are produced from petroleum feedstocks. In recent years, interest has increased in the utilization of vegetable oils as feedstock given environmental concerns and diminished ailability of raw materials for polyol production. Therefore, a number of researchers he studied vegetable oils as alternative feedstock to substitute for petroleum to provide sustainable development (Chen et al. 00; Okieimen et al., 00; Petrovic et al., 00 and 003; Goud et al., 006 and 007; Setyopratomo et al., 006; Dinda et al., 007; Benaniba et al., 007; Ikhuoria et al., 007; Lin et al., 008; Cai et al., 008; Mungroo et al., 008). One type of vegetable oils that can be utilized as an alternative feedstock is rice bran oil (RBO). Hence, it needs to be explored for various applications in particular polymer materials. The term rice bran oil refers to the oil that comes from rice bran (Oryza sativa) through an extraction process. Rice bran is a by product of rice milling process 3

20 produced from the outer layer of rice kernel or during conversion of brown to white rice. The rice kernel consists of endosperm, 70 7%; hull, 0%; bran, %; and embrio, - 3% (Ju et al. 005). The structure of rice kernel is illustrated in Figure - below: Figure - Structure of rice kernel (source Rice bran oil is widely used as cooking oil in numerous countries, in particular in Indonesia, China, Japan, India and Korea. These countries he attractive and productive agriculture especially for rice cultivation. Therefore, RBO is readily ailable and develops for edible oil in food processing application since it offers benefit for healthiness and rich on nutrition. RBO has different characteristic compared to other oils in terms of its unsaponifiable matter. Rice bran is a by product from the rice milling process. It is normally used as biomass fuel for boilers or as a food source for animals. Therefore, effort is needed to increase the added value of rice bran by extracting the oil content as it contains oil in the range 5 3% (Zullaikhah et al., 005). RBO can be utilized as a potential raw material for synthesis of polyol since it contains double bond, readily ailable and inexpensive raw material. The typical composition of RBO by weight percent (wt%) is (Ghosh, 007): Triglycerides : 8 84% Diglycerides : 3% Monoglycerides : % Free fatty acids (FFA): 6% Wax : 3 4% Glycolipids : 0.8% 4

21 Phospholipids (PL) : % Unsaponifiable matter : 4% The unsaturated fatty acid content in the oil is an important factor for synthesis of polyol as the double bonds are converted to polyol. The unsaturated fatty acid content is quantified by the iodine value as the double bonds adsorb iodine to form saturated compounds. The amount of iodine adsorbed indicates the number of double bonds hence unsaturated fatty acid concentration in the oil. In broad chemical terms, iodine value can be defined as the amount of iodine (grams) adsorbed by 00 grams of oil or fat (Ketaren, 005). Potential vegetable oils for use as raw material to produce polyol must posses an iodine value in the range of 80 to 40 gram I /00 gram oil to be economic. Vegetable oils with lower iodine values will produce polyol with a very low hydroxyl value. Therefore, the iodine value of feedstock is the key determinant of suitability for polyol synthesis (Petrovic et al., 003). RBO is a potentially attractive feedstock that can be used as a raw material for polyol production because it has high content (about 74%) of unsaturated fatty acids (oleic and linoleic acid) and an iodine value of 99.9 gram I / 00 gram oil. The composition, physical and chemical characteristics of RBO are primarily determined by soil conditions in the growing area. For example, the iodine value of RBO produced from a rice plant in Texas was 0.3 gram I /00 gram oil, whereas RBO extracted in North America States has an iodine value 99.9 gram I /00 gram oil. Table. provides a typical chemical composition of RBO measured by Jamieson (in Bailey, 95) whereas Table. provides the physical and chemical characteristic of RBO. Both tables are derived from rice cultivated in the Northern US (Bailey, 95). Saponification value indicates the amount of potassium hydroxide (milligrams) required to saponify gram oil or fat (Ketaren, 005). In addition, acid value shows the amount of potassium hydroxide (milligrams) required to neutralize free fatty acid in gram oil or fat (Ketaren, 005). 5

22 Table - Typical fatty acid composition (wt %) of RBO (Bailey, 95) Saturated fatty acid Composition Myrictic acid (C 4 H 8 O ) 0.5 % Palmitic acid (C 6 H 3 O ).7 % Stearic acid (C 8 H 36 O ).7 % Arachidic acid (C 0 H 40 O ) 0.5 % Lignoceric acid (C 4 H 48 O ) 0.4 % Unsaturated fatty acid Composition Oleic acid (C 8 H 34 O ) 39. % Linoleic acid (C 8 H 3 O ) 35. % Unsaponifiable matter 4.64% Table. Physical and chemical characteristics of RBO (Bailey, 95) Parameters Values Iodine value 99.9 g I /00 g oil Saponification value 85.3 mg KOH/g oil Acid value 73.7 mg KOH/ g oil. Epoxidation Reaction The use of modified plant oils as a renewable feedstock in the chemical industry has become more desirable. In particular, epoxidized fatty acid derivatives derived from vegetable oil sources may be utilized as stabilizers and plasticizers in polymers, as lubricant additives and as an intermediate product in the synthesis of polyol as constituents of urethane foam. The utilization of epoxidized oil has become more common over the past few years as such epoxidized oil derived from vegetable oils is environmentally friendly (Wu et al., 000). Moreover, plasticizers and additives for polymer PVC derived from vegetable oil-based he been shown to he improved performance in terms of high resistance to heat and light (Gan et al., 995). 6

23 The term epoxidized oil is generally understood to denote an oil that is derived from vegetable oils using the epoxidation reaction. Epoxidized oil contains epoxide groups or oxirane rings. The term epoxide may be defined as cyclic ethers which consist of three elements in the epoxide ring. The term oxirane is also usually used to refer to epoxide according to categorization by IUPAC (Solomons, 99). The chemical structure of epoxide according to Solomons (99) can be illustrated as follows: C C O Figure - An epoxide The general process for synthesis of epoxide groups is known as an epoxidation reaction wherein an alkene is reacted with an organic peroxy acid. The term peracid is frequently used and refers to a peroxy acid. The simplified epoxidation reaction is summarized as follows (Figure -3): O epoxidation R CH = CHR + R COOH R HC CHR + R COH O Alkene Peroxy acid Epoxide (or oxirane) Figure -3 Epoxidation reaction O The epoxidation reaction is an important step in the synthesis of polyol. It plays a key role in contributing to the final hydroxyl groups because epoxide groups will be converted to hydroxyl groups. A number of studies he shown that a variety of vegetable oils such as linseed oil, rubber seed oil, mahua oil, karanja oil, soybean oil, canola oil, cottonseed oil, sunflower oil, corn oil, jatropha oil and methyl esters of parkia biglobosa seed oil can be used in the epoxidation reaction to produce epoxidized oils (Chen et al., 00, Okieimen et al., 00; Petrovic et al., 00; Goud et al., 006 and 007; Ikhuoria et al., 007; Dinda et al., 007; Benaniba et al., 007; Cai et al., 008; Mungroo et al., 008). This is due to their high percentage of 7

24 unsaturated bonds (indicated by high iodine values). For example, soybean oil has the iodine value of 5 g I /00 g oil while mahua oil with the iodine value 88 g I /00 g oil. In these studies, various factors influencing the kinetics of the epoxidation step were investigated and are discussed below. In Okieimen et al s study (00), epoxidation reaction by utilizing rubber seed oil as a raw material to produce epoxidized rubber seed oil was investigated. The influence of mole ratio of acetic acid and hydrogen peroxide to the oil and the temperature versus the percentage oxirane oxygen content of epoxidized oil during the epoxidation reaction were discussed. They concluded that the concentrations of acetic acid as an oxygen carrier and hydrogen peroxide as an oxidizing agent he a positive effect in term of increasing the formation of epoxide groups. It was also reported that the ring opening of oxirane groups could be minimized when the reaction was carried out at an intermediate temperature range (50 60 o C). One of the limitations of this paper is that it did not discuss the mole ratio of unsaturated bonds in the oil to hydrogen peroxide or acetic acid. A study of the epoxidation of soybean oil was conducted by Petrovic et al. (00). Their results indicated that at higher temperature (80 o C), acetic acid is more effective as an oxygen carrier than formic acid to attain high oxirane oxygen content, whereas formic acid was more effective than acetic acid if the epoxidation reaction was performed at 40 and 60 o C. In this study, Petrovic et al. (00) also investigated the reaction kinetics for epoxidation by determining the rate constant and activation energy. An investigation of the epoxidation reaction of mahua oil using hydrogen peroxide was performed by Goud et al. (006). They studied the influence of various factors including catalyst type, temperature, molar ratio of reactant and mixing speed on the epoxidation reaction. They concluded sulfuric acid is the best catalyst for the epoxidation reaction producing a high conversion of double bonds to oxirane groups when the epoxidation reaction was performed at an intermediate temperature of o C to reduce the hydrolysis reaction. In this work, they reported that the economic value of mahua oil could be increased by converting the oil to an epoxidized mahua oil. 8

25 In 006, Goud et al. investigated the possible use of karanja oil as a feedstock to produce epoxidized oil. They found that at intermediate temperature in the range of 55 to 65 o C the conversion of double bonds to oxirane groups was optimal and the reaction time was minimized. It was further also reported that a molar ratio of acetic acid to karanja oil is 0.5 mole and a mole ratio of.5 for hydrogen peroxide to oil was the optimal concentration for epoxidation of karanja oil. However, this study would he been much more interesting had the authors included the steps in determining the reaction kinetics. A previous study by Purwanto et al. (006) reported that at temperatures of 60 o C and higher, the hydroxyl value of the polyol product decreased with reaction time. Hence, a temperature of 60 o C and 4 h reaction time is the optimal condition for epoxidation of soybean oil as it has produced polyol with highest content of hydroxyl value of 5. mg KOH/g oil. In addition, this research results in polyol viscosities in the range of 59.3 to 78.0 centipoises. Dinda et al. (008) recently synthesized epoxidized oil from cottonseed oil using hydrogen peroxide and inorganic acids as a catalyst. These workers reported that the optimal operating conditions were achieved if the epoxidation reaction was performed using H SO 4 (best catalyst) at a concentration approximately % by weight, using a mole ratio of hydrogen peroxide to oil in the range.5.0, a mole ratio of acetic acid as an oxygen carrier to oil of 0.5 at temperatures in the range of o C. Using these conditions, a high content of oxirane groups was achieved and degradation of oxirane ring was minimized..3 Hydroxylation Reaction Throughout this research report, the term hydroxylation reaction is used to refer to the process of introducing hydroxyl groups into unsaturated bonds to the oil. There are various sources of hydroxyl groups that can be used in the hydroxylation process such as alcohols and water. Before the hydroxylation reaction occurs, the oxirane ring must be opened. There are two ways to open an oxirane ring to permit the 9

26 hydroxylation reaction to occur. Firstly, ring opening using acid catalyst is performed (Solomon, 99). The mechanism of this process can be described as: +H + C C C C -H + + O O H H-O-H HO C + C O H -H + HO C C OH H Figure -4 Ring opening mechanism using acid catalyst Another pathway to open oxirane ring uses a base catalyst (Solomon, 99). The mechanism of this process is as follows: ROH RO:- + C C RO C C O- RO C C OH + RO:- O Figure -5 Ring opening mechanism using base catalyst A number of studies he found that the hydroxylation reaction is a crucial step in the synthesis of polyol since it also contributes to the hydroxyl value of polyol by introducing hydroxyl groups into the oxirane groups (Petrovic et al., 003; Setyopratomo et al., 006; Guo et al., 006; Lin et al., 008). They investigated various epoxidized vegetable oils as raw materials to produce polyols such as epoxidized soybean oil, safflower oil, sunflower oil, canola oil, corn oil, olive oil and peanut oil. A variety of factors that may influence in the hydroxylation were also studied for process optimization. Extensive work on the hydroxylation of epoxidized oil was undertaken by Petrovic et al. (003). In this study, they found that the ratio of alcohol and water to vegetable oils in the mixture of alcohol and water is important for success in the hydroxylation reaction. Groups of alcohol that can be used for hydroxylation reaction include methanol, ethanol, n-propanol, isopropanol and n-butanol. The study would he 0

27 been more convincing if they had included the influence of temperature and reaction time on the quality of polyol product. Kinetic studies of oxirane degradation using methanol he been investigated by Lin et al. (008). They were able to produce polyol by reacting epoxidized soybean oil and methanol without any catalyst in the mixture. The result indicated that hydroxylation of epoxidized soybean oil was first and second order to the concentration of epoxidized oil and methanol concentration, respectively. It is envisaged that the hydroxylation reaction would he been faster had they used catalyst to open the oxirane ring by using either an acid or base catalyst. The comparison of performance of acid and base catalysts that influence the hydroxyl value of polyol products has been investigated for the hydroxylation of palm oil (Setyopratomo et al., 006). An acid catalyst represented by H SO 4 shows better performance than a base catalyst represented by NaOCH 3 indicated by a higher content of hydroxyl groups. The optimal operating condition using acid catalyst was achieved at temperature of 50 o C and h reaction time..4 Synthesis of Polyol From the previous section, it is clear that polyol can be synthesized from vegetable oils that contain unsaturated bonds through two consecutive steps involving epoxidation and then hydroxylation reactions. In the epoxidation step, the reaction mechanism can be described as follows (Okieimen et al., 00):. Formation of peroxyacid: RCOOH + H O RCOOOH + H O. Epoxidation reaction: R CH=CHR + RCOOOH R CH CHR + RCOOH O Figure -6 Epoxidation reaction mechanism

28 In the hydroxylation step, the reaction mechanism can be depicted below (Purwanto et al., 006):. Reaction of epoxidized oil with water: R CHOCHR + H O R -CH-CHR OH OH. Reaction of epoxidized oil with alcohol: R CHOCHR + R 3 OH R -CH-CHR R 3 O OH Figure -7 Hydroxylation reaction mechanism A variety of research studies he been undertaken using various raw materials to synthesize polyol. So far, however, there has been little discussion about the synthesis of polyol from rice bran oil (RBO). In fact, RBO is a cheap and readily ailable by product and waste in the milling rice process which has high content of unsaturated bonds. Thus, it is our belief that the synthesis of polyol from RBO needs to be explored..5 Response Surface Methodology Response surface methodology (RSM) is method for performing process optimization using mathematics and statistics to construct a model of the process to determine the optimal process conditions (Myers, et al., 995; Montgomery, 005). The relationship between independent variables (x) and dependent (response) variables (y) is usually unknown and complex and RSM is applied to determine an estimation equation that represents closely the relationship between x and response variables (y). For ease of application, a low-order polynomial model is normally employed to represent estimation function. The first order model is the simplest model employed (Equation -): y = β + β x + β x β x 0 k k + ε (-) or second order model (Equation -): y = k β 0 + βi xi + βii xi + β ij xi x j + ε (-) i= k i=

29 The next step is to perform an experimental design and then this step is followed by data analysis. The actual values of variables are normally expressed in coded values for analysis. The coded value is given by equation (-3): X i xi x0 = x i (-3) Where: X i = the coded value of an independent variable x i = the real value of an independent variable x 0 = the real value of an independent variable at the centre point x i = the value of step change Following this, the range and level of variables is determined to indicate scope of variables includes the experimental design to determine the model parameters. The fitted surface is then employed to analyze the response surface. If the resulting response surface is statistically a good fit then the response surface will be a close approximation of the real system. Application of RSM for process optimization is a commonly used strategy. The developed model can be easily maximized or minimized to determine the best response variables. Thus in process optimization of a particular chemical process, RSM can determine the optimal operating condition and offer benefits such as reduced experimental time, a less complexity, and highly efficient process..6 Key Research Question Over the last a few decades, polyol derived from petroleum oil as feedstock has been used to produce polyurethane to fulfill the world s needs for this polymer. Recently, a number of studies he been conducted to investigate the use of vegetable oils as a renewable and sustainable feedstock to replace the use of petroleum oil. Many researchers he concluded that vegetable oil can be exploited as an alternative raw material to substitute for petroleum oil to produce polyol. Examples include soybean oil, safflower oil, olive oil, canola oil, cottonseed oil, palm oil and rapeseed oil (Petrovic et al., 003). Vegetable oil based polyol is a sustainable material which 3

30 can be efficiently produced through two-consecutive steps, epoxidation and hydroxylation reactions. Rice bran oil is a potential feedstock source. The unsaturated fatty acid content in RBO is indicated by its iodine value. RBO with an iodine value of 99.9 g I /00 g oil has 74.3% of total unsaturated fatty acid (oleic and linoleic acid) in the oil (Bailey, 95). A key criterion to economically produce polyol is to use vegetable oil with an iodine value in the range of 80 to 40 to accomplish high hydroxyl value and consequently allow production high viscosity of polyol product (Petrovic et al., 003). To date, the normal route to produce polyol is by converting unsaturated bonds in the oil in a two step process using epoxidation followed by hydroxylation reactions. Research conducted by Purwanto et al. (006) and Lin et al. (008) has shown that the epoxidation and hydroxylation reactions contribute to hydroxyl value of the final product of polyol. In the epoxidation step, unsaturated bonds in RBO are converted to produce epoxy groups indicated by the percentage of oxygen content. A high percentage of oxygen content in the epoxidized oil has more epoxy groups. Therefore, this research investigated the influence of the reaction time, reaction temperature and type of peroxy acid (peroxyacetic acid and peroxyformic acid) as oxygen carrier in order to determine the optimal operation conditions in the epoxidation reaction. In the hydroxylation step, the epoxy groups in the epoxidized oil are converted to form hydroxyl groups called polyol. A good quality of polyol is indicated by a high hydroxyl value. Therefore, intensive research needs to be conducted to investigate key process variable in the hydroxylation step that influence the final quality of the product of polyol. In other words, the optimum operating conditions in the hydroxylation step also need to be quantified, such as reaction time and temperature to attain high quality of polyol products synthesized from rice bran oil. 4

31 .7 Research Objectives This study is divided into two steps. The first step deals with the determination of optimal operating conditions in the epoxidation step in terms of reaction time and temperature. The aims of the first step in this study are: to investigate the optimal operating condition in the epoxidation step using acetic acid and formic acid as oxygen carriers (peroxyacetic and peroxyformic acid respectively) to attain epoxidized oil with high oxygen content. to determine the reaction kinetics in the epoxidation step using formic acid as an oxygen carrier (peroxyformic acid). to compare the performance of acetic acid and formic acid as oxygen carriers in the epoxidation step in converting double bonds to epoxy groups. The optimum operating conditions obtained in the epoxidation step will be used to generate epoxidized oil samples for the next step. The second stage of this research considers the determination of the optimal operating conditions in the hydroxylation step for the synthesis of polyol. The aim of the second step in this study is to investigate the optimal reaction time and temperature in the hydroxylation step to attain polyol of high quality indicated by a high content of hydroxyl value with a reasonable viscosity. At present there is no fixed specification for polyol viscosity..8 Significant/Contribution to the Discipline Vegetable oil particularly rice bran oil (RBO) is an alternative raw material to produce polyol for the next generation. This is important since it can lessen the world s reliance on petroleum oil and the limited amount of non-renewable resources can be sed as a heritage for future generations. Polyol derived from vegetable oil could overcome the problem in the application of polyol derived from petroleum since polypropylene oxide (PPO) triols derived from petroleum oil tend to undergo oxidation processes, thus these materials are rather unstable. By contrast, polyol from vegetable oils are resistant to oxidative degradation. Other benefits of 5

32 polyurethanes produced from vegetable oil based-polyol are stability under thermal stress and improved dielectric properties (Guo et al., 000). Over the past few decades, the world s demand for polyurethanes has increased and these polyurethane compounds can also be produced from vegetable oils (John et al., 00). Therefore, the use of RBO in particular, could increase the economic value of RBO. Furthermore, the ailability of RBO as a raw material can be guaranteed since RBO-oil is a renewable resource. A high content of unsaturated bonds indicated by high iodine value of RBO allows high yield of hydroxyl groups, a key requirement for economic production of RBO based polyols. Epoxidation and hydroxylation reactions are crucial steps to produce a high quality of polyol as indicated by a high hydroxyl value with reasonable viscosities. The synthesis of polyol with the utilization of soybean oil, corn oil, safflower oil, sunflower oil, canola oil, olive oil and peanut oil has been studied extensively by Petrovic, Guo, and Jni (003). However, there is significant lack of information on the use of RBO as a starting material to produce polyol. As the epoxidation and hydroxylation steps also contribute to the final product of polyol, the optimum operating conditions for reaction time and temperature in the epoxidation and hydroxylation reactions must be determined. The best form of peroxy acid as oxygen carrier (peroxyacetic acid or peroxyformic acid) in the epoxidation step must also be investigated to determine the key parameters hydroxyl value and the viscosity of polyol product. High grade polyols will he a high hydroxyl value with reasonable viscosity. Determination of the optimum operating conditions for polyol production in the epoxidation and hydroxylation stages will result in a cheap, more efficient process. Both reaction time and temperature he a close relationship with energy consumption in the process hence, optimum reaction time and temperature could prevent energy waste so that an efficient process could be attained. This could produce many benefits such as reducing production cost and maximizing selection process. 6

33 In addition, if we can use RBO as a raw material to produce polyol, it can se petroleum as a non-renewable resource. This research can also support the establishment of vegetable oil based polyol industries. Hence, it will automatically create more work opportunities and also contribute largely to the national economy. Furthermore, this research can encourage and support other research to use renewable natural resources. Therefore, this research generally can encourage the creation of renewable natural resources-based industry. 7

34 3 MATERIALS and METHODS The purpose of this research is as an initial study of the synthesis of polyol from rice bran oil to increase its added value and to determine the optimal reaction conditions. The two consecutive steps for the synthesis of polyol are epoxidation and hydroxylation reactions. These reactions were investigated to determine the optimal operating conditions in the epoxidation reaction using acetic acid and formic acid where the oxygen carriers and the reaction kinetics using formic acid in the epoxidation step and hydroxylation reaction are subjects of evaluation. Formic acid was selected to study the reaction kinetics in the epoxidation step as the formation rate of peroxyformic acid is faster than peroxyacetic acid (Kirk-Othmer Encyclopedia of Chemical Technology, 965). This chapter outlines the research methods applied in this investigation. 3. Materials The raw materials for this study of the epoxidation reaction were rice bran oil (Old Fashioned Foods Ltd) which was purchased from a supermarket; glacial acetic acid (Chem Supply); formic acid (Chem Supply); hydrogen peroxide 30 wt% (Ajax Finechem); sulfuric acid wt% (Ajax Finechem). Furthermore, methanol (Chem Supply); isopropanol (Chem Supply); sulfuric acid wt% (Ajax Finechem); and water were used in the hydroxylation reaction. 3. Epoxidation Reaction The epoxidation reaction was carried out in a 500 ml three neck flask batch reactor, equipped with agitator, reflux condenser and thermocouple. The three neck flask was immersed in a heating mantle whose temperature could be controlled to less than ± K. A schematic of the experimental apparatus for the epoxidation reaction is illustrated in Figure 3-. 8

35 Notes: A. Three neck flask B. Mechanical stirrer C. Condenser D. Heating mantle E. Retort and stand F. Thermocouple Figure 3- Experimental apparatus of epoxidation reaction In order to meet the research objectives for the epoxidation step, two initial studies were performed. The first study determined the optimal operating condition for the epoxidation reaction as a function of reaction time and temperature using acetic acid as the oxygen carrier (peroxyacetic acid). A response surface methodology (RSM) was then performed to obtain the optimal operating condition for this epoxidation reaction using peroxyacetic acid. The second study focused on determining the optimal operating condition using formic acid as the oxygen carrier (peroxyformic acid) and investigated the reaction kinetics. The optimal condition obtained from epoxidation step in terms of oxygen carrier, reaction time and temperature was used for the next step in the synthesis of polyols from RBO through epoxidation and hydroxylation reactions. Prior to the epoxidation reaction, RBO was analyzed to determine its initial iodine value. The RBO was then used as the feedstock for the epoxidation stage. The experimental method for the epoxidation step (adapted from Goud et al., 007) is as follows: rice bran oil (RBO) in the amount of 00 ml was placed in the 500 ml three neck flask equipped with reflux condenser. Formic acid or acetic acid at a molar ratio of 0.5: to the oil and sulfuric acid catalyst 3% weight of hydrogen peroxide and oxygen carrier was added into RBO. A hydrogen peroxide molar ratio 9

36 of.5: to the oil was then added drop-wise into the mixture. This feeding strategy was required to oid overheating the system as the epoxidation reaction is highly exothermic. The reaction was well mixed and was performed at a stirring speed of 600 rpm under isothermal conditions at several temperatures and reaction times. The product of the reaction was next cooled and decanted to effect a separation of the organic-soluble compounds (epoxidized oil) from water-soluble phase. The epoxidized oil was then washed with warm water (in small aliquots) to remove residual contaminants. The product was then analyzed to determine its iodine value and oxirane content. 3.3 Hydroxylation Reaction The hydroxylation reaction was performed in a 000 ml glass reactor, equipped with stirrer, reflux condenser and thermocouple. The reactor was placed on a heating plate with temperature control. A photograph of the experimental apparatus for hydroxylation reaction is presented as Figure 3-. Figure 3- Experimental apparatus of hydroxylation reaction 0

37 As noted earlier in order to determine the optimal operating condition for the hydroxylation step, the optimal condition obtained from epoxidation reaction was applied to produce the epoxidized oil feedstock. Therefore, epoxidation reaction at the optimal operating condition was followed by hydroxylation reaction. To meet the research objectives of determining the overall optimal process conditions, the optimal operating parameters for the hydroxylation reaction were investigated in terms of reaction time and temperature. Then RSM was performed to deduce the optimal operating condition in the hydroxylation reaction. The procedure for hydroxylation reaction (adapted from Petrovic et al., 003) is as follows: 50 ml epoxidized oil was hydroxilated using a mixture of alcohols (methanol and isopropanol), water and sulfuric acid as a catalyst. Mixture of alcohol (methanol and isopropanol) with molar ratio of 4: to the oil each and water at molar ratio of : were mixed with the epoxidized RBO and sulfuric acid catalyst in the reactor. The reaction was performed at several fixed temperatures and reaction times. The reaction product (polyol) was then washed with warm water (in small aliquots) to remove contaminants and then decanted to effect a separation of the organic-soluble compounds (polyol) from water-soluble ones. The resulting polyol produced from the hydroxylation process was then analyzed using two key parameters: the hydroxyl value and the viscosity of the product polyol. Again RSM was performed to determine the optimal operating conditions for the hydroxylation process. 3.4 Epoxidation Test This section summaries the analysis methods used to evaluate yield from the epoxidation reaction evaluation in terms of the two key variables, namely iodine value and oxirane oxygen content Iodine Value Analysis Iodine value was determined by applying the Wijs method (Ketaren, 005; Siggia, 963; Sudarmaji et al., 997). The sequence of the procedure is as follows: 0. to 0.5

38 grams of sample were placed into the flask. 0 ml of chloroform was then added to the sample. Following this, 5 ml Wijs iodine solution was added. Using the same procedure, blank solution was also prepared. The mixture was then stored in a dark place for at least 30 minutes at temperature of 5±5 o C and after that, 0 ml of 5 wt% potassium iodide (KI) solution and 50 ml of water were added into the mixture. The iodine content in the mixture was then titrated using 0. N sodium thiosulfate solution until the yellow colour of the solution almost disappeared. A few drops of starch indicator solution were then added and titration was continued until the blue color completely disappeared. The iodine value was calculated using the following equation: ( B A) x N x (. 69) Iodine value = (3-) C Where: A = Volume of Na S O 3 solution required for titration of the sample (ml). B = Volume of Na S O 3 solution required for titration of the blank solution (ml). C = weight of sample (gram). N = normality of Na S O 3 solution Oxirane Oxygen Content Analysis The oxirane content of the epoxidized oil must be quantified to determine the conversion of unsaturated bonds in RBO to oxirane groups. The procedure (adapted from Siggia, 963) is as follows: a calculated amount of epoxidized oil was added into a flask. 5 ml of ethyl ether was used to wash the flask side and then 0 ml of the hydrochlorination reagent (0. N HCl in ethyl ether) was added into the flask. Simultaneously, a blank solution was prepared using an identical procedure. The mixture was then allowed to stand for 3 hours at room temperature. The mixture was then titrated with standard 0. N sodium hydroxide solution. Prior to this, a few drops of phenolphthalein indicator solution and 5 0mL of ethanol solution were added. The percentage of oxirane content was calculated using the following equation:

39 % oxirane oxygen content ( V V ) b s x N x 6 x 00 = (3-) W x 000 Where: V b = volume of NaOH used for blank (ml). V s = volume of NaOH used for sample (ml). N = normality of NaOH W = weight of the sample (gram) 3.5 Hydroxylation Test This section summaries the analysis techniques and methodology used for evaluation of the hydroxylation step evaluation in terms of hydroxyl value and viscosity Hydroxyl Value Analysis The hydroxyl content of hydroxilated oil (polyol) must be determined as the hydroxyl value of RBO based-polyol is a key measure of the quality of the resultant polyol product. The procedure was adapted from the work of Ketaren (005). Prior to this, the saponification value of the sample must be determined before and after the acetylation process. In the acetylation process, the hydroxyl groups in the oil are converted to ester then analyzed in terms of their saponification value before and after the acetylation process to determine the amount of hydroxyl groups. In the acetylation process, 0 ml of oil sample was mixed with 0 ml of acetic anhydride in a flask. The mixture was then boiled for hours. Following this, 50 ml of water was added into the mixture and the mixture was boiled for 5 minutes. It was then cooled and washes water then used to effect a separation from the mixture. After that, 50 ml of water was added and mixture was boiled for 5 minutes. This procedure was repeated several times until the washing water was neutral. After the washing water was separated, acetylated oil was dried over sodium sulfate anhydrous and then filtered. 3

40 In the saponification process, a calculated amount of oil was placed into a flask. 0 ml of alcoholic potassium hydroxide (0.5 N) was then added into the flask and the mixture was boiled to form saponificable oil. The mixture was then cooled and the flask side was washed with little water. After that, a few drops of phenolphthalein indicator were added and the mixture was then titrated with 0.5 N hydrochloric acid solution until the red color entirely disappeared. Using the same procedure, a blank solution was prepared and then titrated with hydrochloric acid solution. The hydroxyl value of polyol was calculated using the following equation: Where: Where: Saponification value ( B) A x = (3-3) G A = milliliters of 0.5N HCl required for blank solution. B = milliliters of 0.5N HCl required for sample. G = weight of the sample (g) = a half value of molecular weight of KOH Hydroxyl value SA SB = (3-4) ( SA) SA = saponification value after acetylation process SB = saponification value before acetylation process 3.5. Viscosity Analysis The viscosity of RBO based-polyol were measured at 5 o C using Brookfield viscometer VT 550 connected to computer software. The procedure is as follows: the computer connected to viscometer apparatus was turned on. Prior to turning on the viscometer apparatus, the job manager in the computer program was opened. After that, the polyol samples were filled in into the cup. Following this, the cup was inserted into temperature control vessel and secured by screwing at the bottom. In order to control temperature at 5 o C, a water bath equipped with controller with a fixed set point temperature in the water bath. Once the file job was opened on the computer, the speed level was selected and then rotor started to rotate. The collected data was checked in data manager and then changed into shear stress (τ ) and shear rate ( γ ). The viscosity was measured as the ratio of shear stress and shear rate. 4

41 4 EXPERIMENTAL RESULTS and DISCUSSION This section summarizes the results obtained from the experiments performed in the laboratory. Specifically, it will discuss critically the synthesis of polyol from RBO through epoxidation and hydroxylation steps. Hence, it is divided into three sections consisting of epoxidation using acetic acid and formic acid as oxygen carriers followed by hydroxylation of epoxidized oil to determine the optimal operating conditions for the two stage involved in the synthesis of polyol from RBO. 4. Epoxidation of RBO - Acetic Acid as an Oxygen Carrier ( st Study) During the epoxidation reaction, double bonds in the oil are converted to oxirane rings. Hence, the objective function for determining the optimal condition in the epoxidation reaction using acetic acid as an oxygen carrier is to maximize the amount of oxirane groups as indicated by a high amount of oxirane oxygen content. Response surface methodology (RSM) provides an efficient experimental strategy to study the influence of imposed variables to discover a final optimum condition (Montgomery 005, p. 405). Furthermore, RSM has additional benefits such as it allows determination of interaction effects between variables (Wang et al., 008) and ses time as a reduced number of experiments are required (Doddapaneni et al., 007). 4.. Experimental Design and Optimization of the Epoxidation Reaction Full factorial central composite design (CCD) was employed and the total number of k treatment combinations can be represented as + k + n0 (Doddapaneni et al., 007). Where: k = factorial design k = star point k = the number of independent variables 5

42 n 0 = the number of replications at the centre point This first study evaluated the effect of two independent variables (reaction time and temperature) of epoxidation reaction on the response variables (% conversion and % oxirane content). The central values of the independent variables in the epoxidation of RBO using acetic acid as an oxygen carrier were a batch reaction time of 4 hours at a temperature of 60 o C. The two independent variables to be optimized were coded X and X at five levels (-, -, 0,, ) using the equation below: X i xi x0 = x i (4-) Where: X i = the coded value of an independent variable x i = the real value of an independent variable x 0 = the real value of an independent variable at the centre point x i = the value of step change The distribution of coded X and X at five levels is indicated in Table 4-. At this stage, 4 points factorial design, 4 star points and 3 replicates at the central points (all factors at level 0) were performed to fit with the second order polynomial model. Hence, experiments were conducted for the epoxidation of RBO using acetic acid as an oxygen carrier. Table 4- The range and levels of variables used in the RSM procedure to determine the optimum conditions for the epoxidation reaction of RBO Variables Symbol coded Range and levels Reaction times (h) Temperatures ( o C) X X

43 4.. Statistical Analysis A second-order polynomial model (Equation 4-) was fitted to represent the experimental data presented in Table 4-. Y = β + (4-) 0 + β X + β X + β X + β X β X X Where Y i ( i = ) is the response ( Y = % conversion and Y = % oxirane content); β 0 is a constant; β and β represent the linear coefficients; whilst β and β are the quadratic coefficients; β is the interaction coefficient. Table 4- CCD and response in terms of conversion and oxirane content during the Exp. No. epoxidation of RBO Reaction time (h) Temperature ( o C) Conversion (%) Oxirane (%) X X Y Y Microsoft Excel was used to analyze the results in the form of an analysis of variance (ANOVA). For each reaction time and corresponding temperature, the conversion ( Y ) and oxirane content ( Y ) were determined for the epoxidation process. The correlation of the responses Y and Y to the coded values of variables was estimated using 7

44 multiple linear regression. The obtained regression coefficients are presented in Table 4-3 and the analysis of variance (ANOVA) is provided as Table 4-4. Analysis of regression statistics from the analysis of variance indicated a good fit and reasonable significance F < 0.0. Table 4-3 Regression statistics for conversion and oxirane content for epoxidation of RBO Regression statistics Conversion ( Y ) Oxirane content ( Y ) Multiple R R Square Adjusted R Square Standard Error Observations Table 4-4 Analysis of variance (ANOVA) for conversion and oxirane content for epoxidation of RBO Source DF SS MS F Significance F Conversion Regression Residual Total Oxirane Regression content Residual Total

45 Table 4-5 Significance of regression coefficients for conversion of epoxidation of RBO Coefficients Standard Error t Statistic P-value Intercept E-06 a X b X b X X X X a Significant at 0.% (p<0.00) b Significant at.0% (p<0.0) Table 4-6 Significance of regression coefficients for oxirane content of epoxidation reaction Coefficients Standard Error t Statistic P-value Intercept E-07 a X c X b X b X b X X a Significant at 0.% (p<0.00) b Significant at.0% (p<0.0) c Significant at 5.0% (p<0.05) The model is a good fit of the experimental data as indicated by high values of the correlation coefficients (R ) for the responses. The data analysis for regression was undertaken using Microsoft Excel whereas Matlab plotting software was employed to visualize the results. 9

46 4..3 Effects of Reaction Time and Temperature on Reaction Conversion The experimental data from the central composite design was fitted with a secondorder polynomial model by performing multiple linear regressions. The correlation between conversions of iodine value Y and the two independent variables (reaction time and temperature) in coded units after applying of response surface methodology can be represented by equation: Y X + 3.9X.49X.69X = 68 X X (4-3) Where Y (conversion of iodine value in %) is the response, and X and X are the coded values of the independent variables (reaction time and temperature). Figure 4- Effects of reaction time (X ) and temperature (X ) on reaction conversion for acetic acid as an oxygen carrier The result indicates that the conversion of iodine value in the RBO increases linearly with the rise of reaction time and temperature, and at a faster rate with temperature than with reaction time (Figure 4-). These effects are positive and significant at p < 0.0 confidence level for both reaction time and temperature (Table 4-5). 30

47 Unsaturated double bonds present in the oil are converted to oxirane rings through the epoxidation reaction as indicated by the decrease in the iodine value. This value in the RBO represents the concentration of double bonds and it decreases with reaction time (Petrovic et al., 00). Therefore, the reaction conversion increases with reaction time and temperature. These findings are also consistent for epoxidation of cottonseed oil with those Dinda et al. (008) who also reported a similar behiour Effect of Reaction Time and Temperature on Oxirane Content The empirical relationship between oxirane content and the two independent variables (reaction time and batch temperature) in coded units is represented as: Y X X 0.7X 0.65X 0. 6 = X X (4-4) The regression result (Table 4-6) shows that oxirane content in the RBO is a quadratic function of reaction time and temperature. All the observed effects were significant at p < 0.05 (Table 4-6). As can be seen from Figure 4-, the amount of oxirane content increases with reaction time and temperature and then that value reaches the maximal level. Following this, the oxirane content decreases with reaction time and temperature. 3

48 Figure 4- Effects of reaction time (X ) and temperature (X ) on oxirane content for acetic acid as an oxygen carrier There are two major reactions involved in the epoxidation reaction (Petrovic et al, 00; Milchert et al., 009). During the first stage, peroxyacetic acid is formed from the reaction of acetic acid and hydrogen peroxide as summarized by: CH 3 COOH + H O CH 3 COOOH + H O (4-5) During the second stage, epoxidized oil is produced from the reaction between peroxyacetic acid and double bonds in the oil, illustrated as: CH 3 COOOH + CH CH CH CH + CH 3 COOH (4-6) O The maximum oxirane content was achieved at a reaction time (coded) 0.8 and at a temperature (coded) 0.38 (Appendix C.4). These coded values were attained using calculation to determine the function s stationary points by taking partial derivatives of the polynomial equation of oxirane content. The real values of variables at the 3

49 optimal condition was deduced by converting these coded values to the original scale using equation (4-) and the optimal condition for the epoxidation reaction occurred with reaction time of 4.3 h at a batch temperature of 63.8 o C. At higher reaction time and temperature, the reaction results in lower oxirane content. This is a result of a high temperature of epoxidation reaction fouring a high rate of oxirane ring opening thereby producing a product with reduced oxirane content (Purwanto et al., 006). Therefore, higher side reaction products may be formed above the optimal temperature such as reaction between oxirane rings and acetic acid or water and a dimerization reaction may occur (Petrovic et al., 00; Milchert 009). These findings are also consistent with those of Dinda et al. (008) who found at higher temperature (60 o C or higher) the relative conversion to oxirane increased to an optimal operating point and then declined gradually in the epoxidation of cottonseed oil. 4. Epoxidation of RBO Formic Acid as an Alternate Oxygen Carrier ( nd Study) This section presents results and critical discussions for epoxidation using formic acid as an oxygen carrier (peroxyformic acid generated in situ). Experiments were performed to investigate reaction conversion and oxirane content of epoxidized oil. The results will be utilized to study the influence of reaction time and temperature on the reaction conversion, kinetics and oxirane content. The purpose is to determine the optimal operating condition utilizing peroxyformic acid generated in situ and also to determine if formic acid is better than acetic acid as an oxygen carrier. 4.. Effects of Reaction Time and Temperature on the Conversion The results of this study indicate the reaction conversion increases with increasing reaction time and temperature (Figure 4-3). It means that the concentration of double bonds decreases with increasing reaction time and temperature since double bonds are represented by the iodine value (Petrovic et al., 00). This finding supports previous research with the utilization of cottonseed oil as a raw material to produce epoxidized oil (Dinda et al., 008). This result can be explained by the fact that 33

50 through the epoxidation reaction double bonds in the oil were converted to epoxidized oil. A model that can illustrate for rate constant increases with temperature is Arrhenius law (Levenspiel, 97; Fogler, 006): k = k e (4-7). E / RT 0 Where: k 0 = frequency factor E = activation energy R = gas constant T = absolute temperature (in Kelvin) From Equation (4-7), it is clear that reaction constant ( k ) is a function of reaction temperature ( T ). If the reaction temperature goes up, the reaction constant value will increase. Hence, it will increase the reaction rate of epoxidation and in this case a higher final conversion resulted X (%) t (h) 40 C 50 C 60C 70 C 80 C Figure 4-3 Effects of reaction time (t) on reaction conversion (X) for formic acid as an oxygen carrier at different temperatures 34

51 4.. Reaction Kinetics Epoxidation reaction using peroxyformic acid generated in situ was performed with a high stirring speed at 600 rpm to oid mass transfer resistance and to achieve an homogenized two phase system. Research that has been conducted by Goud et al. (006) confirms that epoxidation of mahua oil is not controlled by mass transfer resistance if stirring speed exceed 500 rpm. Therefore, in this research stirring speed at 600 rpm was used to ensure that the epoxidation reaction was controlled by chemical reaction. The experimental results for conversion of iodine value were employed to determine the reaction kinetics in the epoxidation using peroxyformic acid in terms of reaction order, rate constant and activation energy. There are two major reactions involved in the epoxidation reaction (Petrovic et al. 00). During the first stage, peroxyformic acid is formed from the reaction of formic acid and hydrogen peroxide as summarized by: HCOOH + H O HCOOOH + H O (4-8) During the second stage, epoxidized oil is produced from the reaction between peroxyformic acid and double bonds in the oil illustrated as: HCOOOH + CH CH CH CH + HCOOH (4-9) O General form of rate equation for conversion of double bonds by peroxyformic acid in Equation (4-0) may be written as: Where: [ ] d DB n n = k3 [ DB] [ PFA] (4-0) dt [ DB ] = Molar concentration of double bonds [ PFA ] = Molar concentration of peroxyformic acid k 3 = Reaction rate constant n = Reaction orders with respect to the double bonds concentration 35

52 n = Reaction orders with respect to the peroxyformic acid concentration If it is assumed that epoxidation is pseudo-first order with respect to the double bonds, the rate equation for pseudo-first order can be expressed as: k = k PFA Where: [ ] 3 n [ ] d DB = k [ DB] (4-) dt Then rate equation data for the epoxidation reaction using peroxyformic acid is fitted with Equation (4-). If X is expressed as the conversion of double bonds in the oil, after integration the equation above can be defined as [ DB] = [ DB] ( X ) and the equation (4-) can be integrated to give: ln = k. t X 0 (4-) The rate constant ( k ) value for each temperature can be determined as the slope of the plot of 7. ln X versus reaction time (t) and the results summarized in Table 4- Table 4-7 Rate constant value at different temperatures for epoxidation using formic acid Temperature ( o C) k value (h - ) R The result shows that k value increases with reaction temperatures. The present findings are consistent with the Arrhenius equation which states that the rate constant (k) value is dependent on temperature (T). The most interesting finding was that the rate equation model indicated a good fit as a high significance is indicated by high value of the correlation coefficient. Hence, it can be concluded that epoxidation of RBO using formic acid as an oxygen carrier is pseudo-first order 36

53 with respect to the concentration of double bonds. Another important finding was that at higher temperature, epoxidation reaction deviates from pseudo-first order model since the plot of X ln versus reaction time ( ) t become less linear as indicated by reducing correlation coefficient values as temperature increases (Table 4-7). The experimental data was taken at five different temperatures of 40, 50, 60, 70 and 80 o C. Samples were taken every hour reaction time and maximal 6 hour reaction time. The activation energy of epoxidation reaction with the use of peroxyformic acid was calculated from k values at different temperatures by performing Arrhenius equation: k = k e (4-3). E / RT 0 E ln k = ln k0 (4-4) RT The activation energy was calculated from the slope plot of 4-4) ln k versus / T (Figure ln(k) ln(k)= -76.(/T) R = /T Experimental values Figure 4-4 Determination of activation energy for epoxidation using formic acid as an oxygen carrier 37

54 The result indicates the activation energy of epoxidation reaction using peroxyformic acid is.6 kj/mol. This value of activation energy is lower than the results of Petrovic et al. (00) for epoxidation of soybean oil using peroxyformic acid (E = kj/mol). The lower value for the activation energy of rice bran oil is probably related to the characteristic of the feed source. Soybean oil has a higher concentration of double bonds in the oil with an iodine value of 5 gram iodine per 00 gram oil whereas in this research the iodine value of rice bran oil is 98. gram iodine per 00 gram oil Effects of Reaction Time and Temperature on Oxirane Content In the synthesis of polyol, epoxidation reaction is a key aspect to obtain polyol products with high hydroxyl value. Hence, the optimal operating condition in the epoxidation step must be clearly defined to achieve high content of epoxy groups since it will be used to generate raw material to produce polyol. In the epoxidation of RBO using formic acid as an oxygen carrier as indicated in Figure 4-5, the oxirane content of epoxidized oil increases with reaction time at lower temperatures (40 and 50 o C). From Figure 4-5, it is apparent that at higher temperatures (60, 70 and 80 o C) the oxirane content increases with reaction time then achieves optimal level before declining. These findings are also consistent with those of Dinda et al. (008) who found for epoxidation of cottonseed oil that at higher temperature (60 o C or higher) the relative conversion to epoxy groups increased to an optimal operating point and then declined gradually. Maximum oxirane content of 3.6% was achieved at 4 hours reaction time and temperature of 60 o C. Hence, the results of this study indicate that the optimal operating condition for epoxidation reaction was achieved at a reaction time of 4 hours and a temperature of 60 o C. 38

55 3.5 3 % oxirane C 50 C 60 C 70 C 80 C t (h) Figure 4-5 Effects of reaction time (t) on oxirane content (%) for formic acid as an oxygen carrier at different temperatures Given the question of which acid (acetic or formic acid) is the most suitable oxygen carrier for epoxidation reaction, this study found that formic acid shows better performance than acetic acid as indicated by higher content of epoxy groups with a 3.6% of oxirane content in the epoxidized oil for formic acid. In contrast to early findings using acetic acid in Section 4., however only about.6% of maximum value of oxirane content could be achieved. It seems possible that these results are due to a different reactivity of acetic acid and formic acid involved in the epoxidation reaction to produce peroxyacetic acid and peroxyformic acid. At the first stage before the epoxidation reaction, peroxyacetic acid or peroxyformic acid were formed from the reaction between hydrogen peroxide and acetic acid or formic acid as follows: CH 3 COOH + H O CH 3 COOOH + H O (peroxyacetic acid) (4-5) HCOOH + H O HCOOOH + H O (peroxyformic acid) (4-6) Formic acid and hydrogen peroxide in the reaction mixture will achieve equilibrium to form peroxyformic acid at a faster rate than acetic acid (Kirk-Othmer, 965). Therefore a possibility to react with double bonds in the oil in the next step of the 39

56 epoxidation reaction will be higher. For this reason, peroxyformic acid was selected as the best peroxy acid for the epoxidation step at the optimal operating condition at a reaction time of 4 hour and a batch temperature of 60 o C prior to determining the optimal condition for the hydroxylation step in the synthesis of polyol. Another important finding was that at higher reaction times and temperatures than the optimal conditions a lower oxirane content will result. A possible explanation for this observation might be that higher reaction times and temperatures four a high rate of oxirane ring opening thereby producing epoxidized oil with a lower oxirane content (Purwanto et al., 006). Therefore, side reaction products may be formed as the oxirane ring may be decomposed due to reaction mixture contains materials that are likely to react with the oxirane rings such as sulfuric acid, formic acid, and water (Milchert et al., 009). The unwanted reaction in the epoxidation reaction is illustrated as follows: o Hydrolysis reaction: H + CH CH + H O CH CH (4-7) O OH OH o Acylation reaction: H + CH CH + HCOOH CH CH (4-8) O OH OCOOH Reaction temperatures higher than 60 o C result in lower oxirane content indicated by a reduced amount of oxirane content with reaction time (Figure 4-5). This result may be explained by the fact that epoxidation reaction using peroxy acid in this case peroxyformic acid is highly exothermic (Milchert et al., 009). Hence, high temperatures during the epoxidation reaction may cause the decomposition rate of epoxy groups to be higher than the formation rate. As a result, lower epoxy groups will be produced. 40

57 4.3 Hydroxylation of Epoxidized RBO (3 rd Study) During the hydroxylation reaction, epoxy groups in the epoxidized oil are converted to hydroxyl groups. Hence, the objective function for determining the optimal condition in the hydroxylation reaction is to maximize the concentration of hydroxyl groups as indicated by a high hydroxyl value Experimental Design and Optimization of the Hydroxylation Reaction A full factorial central composite design (CCD) was employed and the total number k of treatment combinations can be represented as + k + n0 (Doddapaneni et al., 007). Where: k = factorial design k = star point k = the number of independent variables n 0 = the number of replications at the centre point This third study evaluated the effect of two independent variables (reaction time and temperature) of the hydroxylation reaction on the response variables (hydroxyl value and viscosity). The central values of the independent variables in the hydroxylation of epoxidized oil were a batch reaction time of 0 minutes at a temperature of 50 o C. The two independent variables to be optimized were coded X and X at five levels (-, -, 0,, ) using the equation (4-9). X i = x x i x i 0 (4-9) Where: X i = the coded value of an independent variable x i = the real value of an independent variable x 0 = the real value of an independent variable at the centre point x i = the value of step change The distribution of coded X and X at five levels is indicated in Table 4-8. At this stage, 4 points factorial design, 4 star points and 3 replicates at the central points (all 4

58 factors at level 0) were performed to fit with the second order polynomial model. Hence, experiments were conducted for the hydroxylation of epoxidized oil. Table 4-8 The range and levels of variables for hydroxylation reaction Variables Symbol coded Range and levels Reaction times (min) Temperatures ( o C) X X Statistical Analysis A second-order polynomial model (Equation 4-0) was fitted to represent the experimental data presented in Table 4-9. Y = β + (4-0) 0 + β X + β X + β X + β X β X X Where Y i ( i = ) is the response ( Y = hydroxyl value and Y = viscosity); β 0 is a constant; β and β represent the linear coefficients; whilst β and β are the quadratic coefficients; β is the interaction coefficients. 4

59 Table 4-9 CCD and response in terms of hydroxyl value and viscosity of polyol Exp. No. Reaction time Temperature Hydroxyl value Viscosity (minute) ( o C) (mg KOH/g oil) () X X Y Y The software Microsoft Excel was performed to analyze the results in the form analysis of variance (ANOVA). For each reaction time and corresponding temperature, the hydroxyl value ( Y ) and viscosity ( Y ) were measured following the hydroxylation reaction. The correlation of the responses Y and Y to coded values of variables was estimated by multiple linear regression. The obtained regression statistics are presented in Table 4-0 and the analysis of variance (ANOVA) is provided as Table 4-. Analysis of regression statistics and analysis of variance indicated a good fit and significance F < 0.05). 43

60 Table 4-0 Regression statistics for hydroxyl value and viscosity of polyol Regression statistics Multiple R R Square Adjusted R Square Standard Error Observations Hydroxyl Value ( Y ) Viscosity ( Y ) Table 4- Analyses Of Variance (ANOVA) for hydroxyl value and viscosity of polyol Source DF SS MS F Significance F Hydroxyl value Regression (mg KOH/g oil) Residual Total Viscosity Regression () Residual Total Table 4- Significance of regression coefficients for hydroxyl value of polyol Coefficients Standard Error t Statistic P-value Intercept E-06 a) X X X a) X b) X X b) a Significant at 0.% (p<0.00) b Significant at 5.0% (p<0.05) 44

61 Table 4-3 Significance of regression coefficients for viscosity of polyol Coefficients Standard Error t Statistic P-value Intercept a) X X b) X X X X a Significant at 0.% (p<0.00) b Significant at.0% (p<0.0) c Significant at 5.0% (p<0.05) The model is a good fit of the experimental data as indicated by the high values of the correlation coefficients (R ) for the responses. The data analysis for regression was undertaken using Microsoft Excel whereas Matlab plotting software was employed to visualize the results Effects of Reaction Time and Temperature on Hydroxyl Value The empirical relationship between hydroxyl value and the two independent variables (reaction time and temperature) in coded units is illustrated as: Y X X 3.66X 8.77X = 63 X X (4-) Based on Table 4-, all the observed data was significant at p < The results of this study indicate that hydroxyl value is a quadratic function of reaction time and temperature. 45

62 Figure 4-6 Effects of reaction time (X ) and temperature (X ) on hydroxyl value of polyol The hydroxylation reaction using acid catalyst follows S N mechanism of reaction with the formation of carbocation (Solomons, 99). There are three main steps in the opening ring of oxirane of epoxidized oil using acid catalyst. The detail of each step is summarized as follows: Step : C C H + C C (carbocation) (4-) + O O H Step :.. + C C + H O H HO C C O H (4-3) + O H H 46

63 .. + C C + CH 3 OH HO C C OCH 3 (4-4) + O H H C C + CH 3 CHCH 3 OH C C O CH (4-5) + O.. OH H CH 3 H + CH 3 Step 3: + -H + HO C C O H OH C C OH (4-6) H + -H + HO C C OCH 3 OH C C OCH 3 (4-7) H CH 3 CH 3 + -H + OH C C O CH OH C C O CH (4-8) H CH 3 CH 3 The maximal hydroxyl value was achieved at a coded reaction time 0.37 and at coded temperature These coded values were determined by finding stationary points by calculation of partial derivatives of the polynomial equation for the hydroxyl value. The real values of variables at the optimal operating condition in the hydroxylation reaction were deduced by converting the coded values to their original values using equation (4-9). The optimal condition for hydroxylation reaction occurred with reaction time of 6 min at temperature of 49 o C and produced a polyol with a maximum hydroxyl value of 6.5 mg KOH/g oil. The result suggests that RBO is suitable as a potential feedstock for polyol production compared to previous research by Petrovic et al. (003) for epoxidation of corn oil resulted a polyol with hydroxyl value of 40 mg KOH/g oil. 47

64 As illustrated in Figure 4-6, the hydroxyl value of hydroxilated oil (polyol) increases with reaction time and temperature and then that value reaches an optimal level. After that, the hydroxyl value decreases with reaction time and temperature. Setyopratomo et al., (006) he also reported a similar behiour for polyol synthesis from palm oil. A possible explanation is that after attaining the optimal condition, the hydroxyl groups were substituted by a strong nucleophile CH 3 O - from methanol excess in the mixture, as nucleophile CH 3 O - is stronger than nucleophile OH. Below is the level of reactivity for some nucleophiles according to Solomon (99): RO - > HO - - >> RCO > ROH > H O Another possible explanation for this might be that the obtained hydroxyl groups react with epoxy groups (Kiatsimkul et al., 008). Explanation regarding this reason will be discussed more in Section Effects of Reaction Time and Temperature on Viscosity of Polyol The experimental data of central composite design was fitted with a second-order polynomial model by performing multiple linear regressions. The correlation between viscosity of polyol products ( Y ) and the two independent variables (reaction time and temperature) in coded units after applying of response surface methodology can be correlated by the following equation: Y = 4 + X X (4-9) X X X X. 380 Where Y (viscosity of polyol in centipoise ()) is the response and X, X are the coded values of independent variables (reaction time and temperature, respectively). 48

65 Figure 4-7 Effects of reaction time (X ) and temperature (X ) on viscosity of polyol The result shows that the viscosity of polyol increases linearly with the rise of temperature and this effect are positive and significant at p < 0.0 for temperature effect only (Table 4-3). As can be seen from Figure 4-7 the viscosity of polyols increases with reaction time and temperature and at a faster rate with temperature than with reaction time. Through the hydroxylation reaction, viscosity of polyols increases with reaction time and temperature until the optimal condition is attained, as hydroxyl groups were introduced into epoxy groups to produce polyol with higher molecular weight. In this research, results indicate that after achieving the optimal hydroxyl value, the viscosity of polyols increases with reaction time and temperature. A possible explanation for this might be that unexpected reaction take place in the mixture such as polymerization or cross linking. Molecular weight or chain length is the key factor influences the viscosity of polymer materials. The viscosity of polymers increases with molecular weight or chain length (Billmeyer, 984). In the mixture, hydroxyl groups may react with oxirane groups through the reaction depicted below (Kiatsimkul et al., 008): 49

66 OR OR RC CR RC CR + R C CR O (4-30) OH O R C CR OH This reaction may result in a product with higher molecular weight and will produce polyols with higher viscosity. The viscosity of polyols in this research is in the range between and is considered as a reasonable viscosity. This study confirms that the viscosity of polyol resulted from this research is close to the typical characteristic of methyl esters polyol with viscosity of 0. Pa.s (00 ) (Petrovic, 008). The viscosity of the polyol product in this research is lower than another type of polyol 73 with viscosity of 5 Pa.s (5000 ) undertaken by Petrovic (008). A possible explanation for this might be that as methanol was present in the reaction, it may cause a transesterification reaction as methanol is the most reactive alcohol (Petrovic et al., 003). Thus, methyl ester polyol with lower viscosity product will result. Thus, the amount or ratio of alcohol is a crucial factor in the hydroxylation step to achieve polyol with desirable viscosity. Figure 4-8 Sample of polyol produced 50