cashew nut shell liquid as a green alternative to ... - pantheon ufrj

CASHEW NUT SHELL LIQUID AS A GREEN ALTERNATIVE TO FUNCTIONALIZE

POLYMERS: EXPERIMENTAL AND MODELING INVESTIGATION

Laura Pires da Mata Costa

Dissertação de Mestrado apresentada ao

Programa de Pós-graduação em Engenharia

Química, COPPE, da Universidade Federal

do Rio de Janeiro, como parte dos requisitos

necessários à obtenção do título de Mestre em

Engenharia Química.

Orientadores: José Carlos Costa da Silva Pinto

Amanda Lemette Teixeira

Brandão

Márcio Nele de Souza

Rio de Janeiro

Fevereiro de 2019




DISSERTAÇÃO SUBMETIDA AO CORPO DOCENTE DO INSTITUTO ALBERTO LUIZ

COIMBRA DE PÓS-GRADUAÇÃO E PESQUISA DE ENGENHARIA (COPPE) DA

UNIVERSIDADE FEDERAL DO RIO DE JANEIRO COMO PARTE DOS REQUISITOS

NECESSÁRIOS PARA A OBTENÇÃO DO GRAU DE MESTRE EM CIÊNCIAS EM

ENGENHARIA QUÍMICA.

Examinada por:

Prof. Fernando Gomes de Souza Jr, D.Sc.

Prof. Marina Damião Besteti, D.Sc.

Dr. Bruno Francisco Oeschler, D.Sc.

Prof. José Carlos Costa da Silva Pinto, D.Sc.

Prof. Amanda Lemette Teixeira Brandão, D.Sc.

RIO DE JANEIRO, RJ – BRASIL

FEVEREIRO DE 2019

Pires da Mata Costa, Laura

Cashew Nut Shell Liquid as a Green Alternative to

Functionalize Polymers: Experimental and Modeling

Investigation/Laura Pires da Mata Costa. – Rio de Janeiro:

UFRJ/COPPE, 2019.

XVIII, 159 p.: il.; 29,7cm.


Amanda Lemette Teixeira Brandão


Dissertação (mestrado) – UFRJ/COPPE/Programa de

Engenharia Química, 2019.

Referências Bibliográficas: p. 150 – 158.

1. LCC. 2. Cardanol. 3. Ácido anacárdico. 4.

Copolímero. I. Costa da Silva Pinto, José Carlos et al. II.

Universidade Federal do Rio de Janeiro, COPPE, Programa

de Engenharia Química. III. Título.

iii

I dedicate ...

iv

Acknowledgments

I am thankful ...

v

Resumo da Dissertação apresentada à COPPE/UFRJ como parte dos requisitos

necessários para a obtenção do grau de Mestre em Ciências (M.Sc.)

LÍQUIDO DA CASTANHA DE CAJU COMO UMA ALTERNATIVA VERDE PARA

FUNCIONALIZAR POLÍMEROS: INVESTIGAÇÃO EXPERIMENTAL E TEÓRICA


Fevereiro/2019




Programa: Engenharia Química

O líquido da castanha de caju (LCC) é um sub-produto da indústria de produção de

castanha de caju. Apesar disto, ele é riquissímo em compostos fenólicos que podem

ser utilizados como comonômeros na funcionalização de polímeros. No presente

trabalho, é estudada a polimerização radicalar desses materiais. Inicialmente, a

copolimerização em massa do cardanol ou LCC natural com estireno, metacrilato

de metila, acetato de vinila ou ácido acrílico foi estudada cineticamente e diversas

análises foram realizadas para caracterizar os copolímeros formados. Algumas

vantagem obtidas na copolimerização foi o aumento da estabilidade térmica e redução

da temperatura vítrea dos polímeros. Assim, um mecanismo cinético foi proposto e

um modelo matemático derivado para que fossem estimados parâmetros cinéticos

que permitiram a descrição apropriada de conversão dos monômeros, massas molares

médias e composição dos copolímeros. É detectado que os monômeros renováveis

exercem forte efeito inibitório nas polimerizações, fazendo com que a incorporação

deles nas cadeias copoliméricas seja difícil, mas não impossível. Com tais resultados,

polimerizações em suspensão foram realizadas com as condições que obtiveram

melhor desempenho em massa. Polímeros antes hidrofóbicos, se tornaram mais

hidrofílicos e a adição de cardanol nessas polimerizações se mostrou uma ótima

alternativa para controlar tamanho de partículas e manter suspensões estáveis por

longos períodos de tempo.

vi

Abstract of Dissertation presented to COPPE/UFRJ as a partial fulfillment of the

requirements for the degree of Master of Science (M.Sc.)




February/2019

Advisors: José Carlos Costa da Silva Pinto



Department: Chemical Engineering

Cashew nut shell liquid (CNSL) is a by-product of cashew nut industry. Despite

this, it is rich in phenolic compounds that can be used as comonomers in the

functionalization of polymers. In the present study, the radical polymerization of these

materials is investigated. Inittialy, the bulk copolymerization of cardanol or natural

CNSL with styrene, methyl methacrylate, vinyl acetate or acrylic acid was studied

kinetically and several analyzes were carried out to characterize the copolymers

formed. Some advantages obtained in the copolymerization were the increase in

thermal stability and reduction of the glass temperature of the polymers. Thus, a

kinetic mechanism was proposed and a mathematical model derived to estimate

kinetic paramers that allowed proper description of monomer conversion, average

molecular weights, and composition of the copolymers. It is detected that the

renewable monomers exert a strong inhibition effect on the polymerizations, making

the incorporation of them into the copolymer chains difficult, but not impossible. With

these results, suspension polymerizations were performed with the conditions that

obtained the best results in bulk polymerization. Polymers previously hydrophobics

became more hydrophilic; and the addition of cardanol in such polymerization

has proved to be an excellent alternative for particle size control and to stabilize

suspensions for long periods of time.

vii

Contents

Acknowledgments v

List of Figures xi

List of Tables xvii

1 Introduction 1

1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2 Theoretical Foundation About Polymers 3

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2.2 Types of Polymerization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2.3 Polymerization methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.3.1 Bulk Polymerization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.3.2 Solution Polymerization . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.3.3 Suspension Polymerization . . . . . . . . . . . . . . . . . . . . . . . . 13

2.3.4 Emulsion Polymerization . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.4 Molecular Weight Distribution (MWD) . . . . . . . . . . . . . . . . . . . . . 13

3 Cashew Nut Shell Liquid – A Literature Review 16

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

3.2 Cashew Nut Shell Liquid (CNSL) . . . . . . . . . . . . . . . . . . . . . . . . . 18

3.3 Synthesis of polymers using CNSL . . . . . . . . . . . . . . . . . . . . . . . . 24

3.4 Polymer Functionalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

3.5 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

Kinetic Study of Bulk Copolymerization 39

4 Bulk Copolimerization: Methodology 40

4.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

4.2 Set up for bulk polymerizaion . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

4.3 Kinetic Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

4.4 Gravimetric Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

viii

4.5 Characterizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

4.5.1 Gel Permeation Chromatography (GPC) . . . . . . . . . . . . . . . . 43

4.5.2 Nuclear Magnetic Resonance (NMR) . . . . . . . . . . . . . . . . . . 43

4.5.3 Fourier-transform infrared spectroscopy (FTIR) . . . . . . . . . . . . 44

4.5.4 Thermogravimetric analysis (TGA) . . . . . . . . . . . . . . . . . . . 44

4.5.5 Differential Scanning Calorimetry (DSC) . . . . . . . . . . . . . . . . 44

5 Bulk Copolymerization: Experimental Results 46

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

5.2 Homopolymerization of cardanol and anacardic acid . . . . . . . . . . . . . 47

5.3 Copolymerization of cardanol and styrene . . . . . . . . . . . . . . . . . . . 55

5.4 Copolymerization of cardanol and methyl methacrylate . . . . . . . . . . . 67

5.5 Copolymerization of natural CNSL and methyl methacrylate . . . . . . . . 74

5.6 Copolymerization of cardanol and vinyl acetate . . . . . . . . . . . . . . . . 78

5.7 Copolymerization of cardanol and acrylic acid . . . . . . . . . . . . . . . . . 82


6 Bulk Copolymerization: Modeling 87

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

6.2 Kinetic Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

6.3 Material Balances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

6.4 Methodology and Numerical Solution . . . . . . . . . . . . . . . . . . . . . . 95

6.5 Parameters for homopolymerization . . . . . . . . . . . . . . . . . . . . . . 99

6.6 Parameter estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

6.7 Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

6.7.1 Styrene Homopolymerization . . . . . . . . . . . . . . . . . . . . . . 105

6.7.2 Styrene/Cardanol Copolymerizations at 110°C . . . . . . . . . . . . 107

6.7.3 Styrene/Cardanol Copolymerizations at 85°C . . . . . . . . . . . . . 110

6.7.4 Styrene/Cardanol Copolymerizations at 110°C for 10wt % . . . . . . 114

6.7.5 Methyl Methacrylate/Cardanol Copolymerizations at 85°C . . . . . 115

6.7.6 Methyl Methacrylate/Cardanol Copolymerizations at 110°C . . . . . 124


Kinetic Study of Suspension Copolymerization 128

7 Suspension Copolimerization: Methodology 129

7.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

7.2 Set Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

7.3 Kinetic Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

7.4 Gravimetric Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

7.5 Characterizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

ix

7.5.1 Contact Angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

7.5.2 Interfacial tension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

7.5.3 Particles Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

7.5.4 MEV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

7.5.5 Optical Microscopy (OM) . . . . . . . . . . . . . . . . . . . . . . . . . 133

8 Suspension Copolimerization: Results 134

8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

8.2 Styrene-Cardanol Copolymer Particles . . . . . . . . . . . . . . . . . . . . . 134

8.3 Methyl Methacrylate-Cardanol Copolymer Particles . . . . . . . . . . . . . 137


9 Conclusions 147

Bibliography 150

x

List of Figures

1.1 Logo from COPPETEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2.1 Some Ways to Classify a Polymer . . . . . . . . . . . . . . . . . . . . . . . . . 4

2.2 Free Radical Polymerization: Initiation . . . . . . . . . . . . . . . . . . . . . 5

2.3 Homolytic scission of BPO followed by the initiation of styrene polymer-

ization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.4 Thermal initiation of styrene . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.5 Free Radical Polymerization: Propagation . . . . . . . . . . . . . . . . . . . 6

2.6 Propagation of poly(styrene) . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2.7 Free Radical Polymerization: Termination (Disproportionation or Cou-

pling) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2.8 Termination of poly(styrene) . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.9 Synthesis of Nylon 66, a step-growth polymerization . . . . . . . . . . . . . 7

2.10 Decomposition of BPO (a) and AIBN (b) . . . . . . . . . . . . . . . . . . . . 8

2.11 Last model applied for copolymerization . . . . . . . . . . . . . . . . . . . . 10

2.12 Penultimate model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

3.1 Cashew . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3.2 An Example of Beneficiation process for Cashew Nut . . . . . . . . . . . . . 18

3.3 Cashew nut: the shell and the almond (adapted from Sistemas de Pro-

dução, EMBRAPA). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

3.4 Products obtained from cashew (adapted from AgroIndúsria do Caju no

Nordeste, ETENE/BNB, 1973). . . . . . . . . . . . . . . . . . . . . . . . . . . 19

3.5 Production of Cashew Nut between the years of 2006 and 2018

(IBGE/CEPAGRO) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.6 Components of the CNSL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.7 Advantages arising from functional groups of cardanol and anacardic acid 21

3.8 Decarboxylation of anacardic acid . . . . . . . . . . . . . . . . . . . . . . . . 21

3.9 Thermal Polymerization of cardanol. (OLIVEIRA, LINCOLN DAVI M. DE

, 2007) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

xi

3.10 Anacardic Acid: (i) 1-hydroxy-2-carboxy-3-pentadecyl benzene, (ii)

1-hydroxy-2-carboxy-3-(8-penta-decenyl)benzene, (iii) 1-hydroxy-

2-carboxy-3-(8,11-pentadecadienyl)benzene, and (iv) 1-hydroxy-2-

carboxy-3-(8,11,14 pentadecatrienyl)benzene. . . . . . . . . . . . . . . . . . 24

3.11 Chemical Transformations of Cardanol (BALGUDE and SABNIS, 2014) . . 25

3.12 Oxidative Polymerization of Cardanol using Fe-Salen as a catalyst (IKEDA

et al., 2000a) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

3.13 Structure of Fe-salen complex . . . . . . . . . . . . . . . . . . . . . . . . . . 26

3.14 Cationic Polymerization of Cardanol . . . . . . . . . . . . . . . . . . . . . . . 27

3.15 Synthesis of poly(cardanyl acrylate) and its film (MANJULA et al., 1992) . . 28

3.16 An epoxy novolac resin structure sold by Cardolite . . . . . . . . . . . . . . 32

3.17 Reaction scheme of anacardic acid polymerization catalyzed by peroxi-

dase (adapted from CHELIKANI et al. (2009)) . . . . . . . . . . . . . . . . . 36

3.18 Synthesis of anacardanyl acrylate (a) and anacardanyl methacrylate (b)

(adapted from PHILIP et al. (2007)) . . . . . . . . . . . . . . . . . . . . . . . 36

4.1 Set Up for Bulk Copolymerization . . . . . . . . . . . . . . . . . . . . . . . . 41

5.1 Reactive sites of cardanol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

5.2 GPC of cardanol oligomers at 85°C (a) and 110°C (b) . . . . . . . . . . . . . 48

5.3 GPC of natural CNSL oligomers at 85°C (a) and 110°C (b) . . . . . . . . . . 49

5.4 1H NMR of cardanol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

5.5 1H NMR of the reaction mixture for synthesize polycardanol by radical

polymerization using BPO after 4 hours . . . . . . . . . . . . . . . . . . . . . 50

5.6 1H NMR of cardanol after 8h for reaction with BPO . . . . . . . . . . . . . . 51

5.7 1H NMR of CNSL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

5.8 1H NMR of CNSL after 8h of reaction . . . . . . . . . . . . . . . . . . . . . . 53

5.9 FTIR for cardanol after 4 and 8 hours of reaction . . . . . . . . . . . . . . . . 54

5.10 Inhibition Mechanism for Cardanol . . . . . . . . . . . . . . . . . . . . . . . 54

5.11 TGA and DTG analysis for cardanol and natural cashew nut shell liquid . . 55

5.12 Chemical structure of polystyrene . . . . . . . . . . . . . . . . . . . . . . . . 56

5.13 Photo for synthesized polystyrene or poly(styrene-co-cardanol) at 110°C . 56

5.14 Conversion and molecular weights for poly(Sty-co-C), 85°C . . . . . . . . . 57

5.15 Conversion and molecular weights for poly(Sty-co-C), 110°C . . . . . . . . 58

5.16 TGA and DTG curves for poly(Sty-co-C), 85°C . . . . . . . . . . . . . . . . . 59

5.17 TGA and DTG curves for poly(Sty-co-C), 110°C . . . . . . . . . . . . . . . . 59

5.18 1H-NMR spectrum of poly(styrene-co-cardanol) prepared with 5 wt% of

cardanol at 110°C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

5.19 FTIR spectrum of poly(styrene-co-cardanol) prepared with at 85°C . . . . . 63

5.20 FTIR spectrum of poly(styrene-co-cardanol) at 110°C . . . . . . . . . . . . . 64

xii

5.21 FTIR spectrum of poly(styrene-co-cardanol) at condition 2.5wt % and 85°C . 65

5.22 FTIR spectrum of poly(styrene-co-cardanol) at condition 2.5wt % and

110°C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

5.23 FTIR spectrum of poly(styrene-co-cardanol) at condition 10wt % and

110°C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

5.24 Methyl methacrylate, MMA, structure . . . . . . . . . . . . . . . . . . . . . . 67

5.25 Conversion and molecular weights for P(MMA-co-C), 85°C . . . . . . . . . 68

5.26 Conversion and molecular weights for P(MMA-co-C), 110°C . . . . . . . . . 68

5.27 TGA and DTG curves for P(MMA-co-C), 85°C . . . . . . . . . . . . . . . . . . 70

5.28 TGA and DTG curves for P(MMA-co-C), 110°C . . . . . . . . . . . . . . . . . 70

5.29 FTIR analysis for P(MMA-co-C), 85°C . . . . . . . . . . . . . . . . . . . . . . 71

5.30 FTIR analysis for P(MMA-co-C), 110°C . . . . . . . . . . . . . . . . . . . . . 72

5.31 NMR spectrum for P(MMA-co-C), 5% w/w, 110°C . . . . . . . . . . . . . . . 74

5.32 Conversion for P(MMA-co-CNSL), 85 and 110°C . . . . . . . . . . . . . . . . 75

5.33 Conversion for P(MMA-co-CNSL) and P(MMA-co-C), 85°C . . . . . . . . . 75

5.34 Conversion for P(MMA-co-CNSL) and P(MMA-co-C), 110°C . . . . . . . . . 76

5.35 TGA and DTG curves for P(MMA-co-CNSL) and P(MMA-co-C), 85°C . . . . 77

5.36 TGA and DTG curves for P(MMA-co-CNSL) and P(MMA-co-C), 110°C . . . 77

5.37 Structure of poly(vinyl acetate), PVAc . . . . . . . . . . . . . . . . . . . . . . 78

5.38 Conversion and molecular weight for P(VAc-co-C), 85°C . . . . . . . . . . . 79

5.39 1H NMR for P(VAc-co-C), 5% w/w, 85°C . . . . . . . . . . . . . . . . . . . . . 80

5.40 TGA and DTG curves for P(VAc-co-C), 85°C . . . . . . . . . . . . . . . . . . . 81

5.41 FTIR for P(VAc-co-C), 85°C . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

5.42 Polyacrylic acid, PAA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

5.43 Conversion for P(AA-co-C) at 85°C . . . . . . . . . . . . . . . . . . . . . . . . 84

5.44 TGA curves for for P(AA-co-C) at 85°C . . . . . . . . . . . . . . . . . . . . . . 85

6.1 Employed parameter estimation procedure (BRANDÃO (2017)) . . . . . . . 102

6.2 Predicted and experimental results for styrene homopolymerization at

90°C (condition A5). (MCI = model confidence interval; MP = model pre-

dictions and ED = experimental data). . . . . . . . . . . . . . . . . . . . . . . 106

6.3 Predicted and experimental results for styrene homopolymerization at

110°C (condition A6). (MCI = model confidence interval; MP = model

predictions and ED = experimental data). . . . . . . . . . . . . . . . . . . . . 106

6.4 Joint confidence region for parameters r1,2 and r2,1 when parameter fi nh

was fixed at 0.731. The red center dot is the estimated values of r1,2 and

r2,1 (1.628, 0.026). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

xiii

6.5 Predicted and experimental results for styrene/cardanol copolymeriza-

tion at condition A1. (MCI = model confidence interval; MP = model pre-





6.7 Predicted and experimental results for cardanol compositions at condi-

tions A1 and A2. (MP = model predictions and ED = experimental data). . 110

6.8 Joint confidence region for parameters r1,2 and r2,1 when parameter fi nh

was fixed at 0.628. The red center dot is the estimated values of r1,2 and

r2,1 (3.172, 0.015). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111







6.11 Predicted and experimental results for cardanol compositions at condi-

tions A3 and A4 (MP = model predictions). . . . . . . . . . . . . . . . . . . . 113

6.12 Predicted and experimental results for cardanol copolymerization at

110°C for 10wt % (MP = model predictions). . . . . . . . . . . . . . . . . . . 114

6.13 Predicted and experimental results for cardanol compositions at 110°C

for 10wt % (MP = model predictions). . . . . . . . . . . . . . . . . . . . . . . 115

6.14 Predicted and experimental results for MMA/cardanol copolymer con-

version and molecular weights at conditions 2.5 wt % (MP = model pre-

dictions). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

6.15 Predicted and experimental results for cardanol composition in a

MMA/cardanol copolymer at 85°C and 2.5wt % (MP = model predictions). 117


version and molecular weights at conditions 5 wt % (MP = model predic-

tions). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117


MMA/cardanol copolymer at 85°C and 5wt % (MP = model predictions). . 118


version and molecular weights at conditions 10wt % (MP = model pre-

dictions). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118


MMA/cardanol copolymer at 85°C and 10wt % (MP = model predictions). 118

xiv


version and molecular weights at conditions 2.5 wt % considering a 4th

parameter, t1,2 (MP = model predictions). . . . . . . . . . . . . . . . . . . . . 121


MMA/cardanol copolymer at 85°C and 2.5wt % considering a 4th param-

eter, t1,2 (MP = model predictions). . . . . . . . . . . . . . . . . . . . . . . . . 121


version and molecular weights at conditions 5 wt % considering a 4th pa-

rameter, t1,2 (MP = model predictions). . . . . . . . . . . . . . . . . . . . . . 122


MMA/cardanol copolymer at 85°C and 5wt % considering a 4th parame-

ter, t1,2 (MP = model predictions). . . . . . . . . . . . . . . . . . . . . . . . . 122


version and molecular weights at conditions 10wt % considering a 4th



MMA/cardanol copolymer at 85°C and 10wt % considering a 4th param-



version and molecular weights at 110°C and 2.5 wt % considering a 4th



MMA/cardanol copolymer at 110°C and 2.5wt % considering a 4th pa-

rameter, t1,2 (MP = model predictions). . . . . . . . . . . . . . . . . . . . . . 125


version and molecular weights at 110°C and 10wt % considering a 4th



MMA/cardanol copolymer at 110°C and 10wt % considering a 4th param-


7.1 Set Up for Suspension Polymerizations . . . . . . . . . . . . . . . . . . . . . 130

8.1 Photos of pendant upward drop tensiometry used for some non-reported

interfacial tension calculations. (a) was the test for cardanol-water and

(b) for styrene + cardanol in water. . . . . . . . . . . . . . . . . . . . . . . . . 136

8.2 Conversion values for styrene-cardanol suspension polymerizations . . . 136

8.3 Conversion values for methyl methacrylate-cardanol suspension poly-

merizations with 1wt% of BPO . . . . . . . . . . . . . . . . . . . . . . . . . . 137

xv

8.4 SEM photo for methyl methacrylate-cardanol particles produced with

10wt% of cardanol and 1wt% of BPO after 6 hours of polymerization . . . 138

8.5 SEM photo for methyl methacrylate-cardanol particles produced with

5wt% of cardanol and 1wt% of BPO after 4 hours of polymerization . . . . 139

8.6 Particle size distribution for copolymer particles produced with 1wt% of

BPO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

8.7 Photo of a becker contaning the reaction mixture of methyl methacrylate

and 5wt% of cardanol one day after the reaction . . . . . . . . . . . . . . . . 141

8.8 Optical microscopy images for the center (a) and bottom (b) for methyl

methacrylate-cardanol copolymer particles . . . . . . . . . . . . . . . . . . 141

8.9 Conversion values for methyl methacrylate-cardanol suspension poly-

merizations with 5wt% of cardanol and 1 or 3wt% of BPO . . . . . . . . . . 142

8.10 Particle size distribution for PMMA or copolymer particles made with 1

or 3wt% of BPO respectively . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

8.11 Conversion data for methyl methacrylate-cardanol suspension polymer-

izations with 5wt% (a) or 10wt% (b) of cardanol and 1wt% of AIBN . . . . . 143

8.12 Particle size distribution for 5wt% cardanol copolymer particles made

with 1wt% of BPO or AIBN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

8.13 Experimental set up for PMMA-water contact angle measure . . . . . . . . 144

8.14 Photos taken while a contact angle analysis was being performed for a

copolymer of MMA and cardanol . . . . . . . . . . . . . . . . . . . . . . . . . 145

xvi

List of Tables

3.1 Difference between natural and technical CNSL (adapted from

MAZZETTO et al. (2009)) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

3.2 Composition of natural CNSL measured by gas chromatography and

mass spectroscopy (GC-MS) (adapted from MAZZETTO et al. (2009)) . . . 23

3.3 Properties of cardanol and anacardic acid. . . . . . . . . . . . . . . . . . . . 23

3.4 Patents using cardanol as a reactive molecule . . . . . . . . . . . . . . . . . 29

3.5 Patents using cardanol to produce polyurethanes . . . . . . . . . . . . . . . 33

4.1 Kinetic Experiments Done . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

5.1 Conversion for homopolymerization of cardanol with 1% w/w of BPO . . . 47

5.2 Conversion for homopolymerization of natural cashew nut shell liquid

with 1% w/w of BPO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

5.3 Glass transition temperatures for P(STy-co-C) . . . . . . . . . . . . . . . . . 60

5.4 Cardanol contents of copolymer samples (wt%) for P(STy-co-C) . . . . . . 62

5.5 DSC analysis for P(MMA-co-C) . . . . . . . . . . . . . . . . . . . . . . . . . . 69

5.6 Copolymer composition for P(MMA-co-C), 85°C . . . . . . . . . . . . . . . 73

5.7 Copolymer composition for P(MMA-co-C), 110°C . . . . . . . . . . . . . . . 73

5.8 DSC analysis for P(MMA-co-CNSL) . . . . . . . . . . . . . . . . . . . . . . . 78

5.9 Copolymer composition for P(VAc-co-C), 85°C . . . . . . . . . . . . . . . . . 79

6.1 Parameter used for benzoyl peroxide, BPO . . . . . . . . . . . . . . . . . . . 99

6.2 Parameters for styrene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

6.3 Parameters for methyl methacrylate, MMA . . . . . . . . . . . . . . . . . . . 101

6.4 Equations for gel and glass effects for methyl methacrylate, MMA . . . . . 101

6.5 Experimental conditions used for parameter estimation . . . . . . . . . . . 103

6.6 Parameters used for cardanol . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

6.7 Estimated model parameters for styrene homopolymerization reactions. . 105

6.8 Estimated model parameters for cardanol/styrene copolymerization re-

actions performed at 110°C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107


actions performed at 90°. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

xvii


actions performed with 10wt % at 110°. . . . . . . . . . . . . . . . . . . . . . 114

6.11 Estimated model parameters for cardanol/MMA copolymerization reac-

tions, 85°C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

6.12 Boundaries for estimated model parameters considering ktm12. . . . . . . 119


tions considering ktm12, 85°C . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

6.14 Estimated model parameters (without the exponentials) for car-

danol/MMA copolymerization reactions considering ktm12, 85°C . . . . . . 120


tions considering ktm12, 110°C . . . . . . . . . . . . . . . . . . . . . . . . . . . 124

6.16 Estimated model parameters (without exponentials) for cardanol/MMA

copolymerization reactions considering ktm12, 110°C . . . . . . . . . . . . . 125

7.1 Conditions for Suspension Polymerizations . . . . . . . . . . . . . . . . . . 131

8.1 Interfacial tension between styrene or styrene-cardanol with distilled water135

8.2 Interfacial tension between styrene or styrene-cardanol with distilled water137

xviii

Chapter 1

Introduction

Due to an increase on environmental concerns, green chemistry and waste manage-

ment are topics that have been discussed and studied in the 21th century. Thus, many

renewable sources are being used, mainly those that are derivatives by-products or

wastes are gaining a new presentation. In the polymerization industry, most of the

polymers are still petroleum-derived and it has been suffering harsh criticism due to

environment impact. Consequently, alternative sources of monomers have been in-

vestigated. One of them is cashew nut shell liquid (CNSL) which is a mixture of pheno-

lic compounds obtained from the fruit of cashew tree.

Although many people think that the fruit is the cashew apple, the actual fruit is

the nut with a kidney-shaped seed hidden within, the cashew kernel. Hence, the CNSL

is found inside of the cashew nut shells. It can be extracted by thermal, solvent or by

mechanical means. And depending on the method, it can be classified as natural and

technical. Natural CNSL is rich in anacardic acid, while technical CNSL is rich in car-

danol. Cardanol is a salicylic acid with a long side chain, while anacardic acid also has

a carboxylic acid group. The side chain varies in the number of unsaturation rang-

ing from 0 to 3. Both compounds are amphiphilic and, truly, are a mixture of several

closely related organic compounds. (LUBI and THACHIL (2000), LOCHAB et al. (2014),

OTSUKA et al. (2017)).

Thus, although the funcionalities of CNSL are large, it is still considered a by-

product of the cashew almond industry, being normally released to the environment

and discarded as waste. Besides biological applications (as anticancer, antioxidant and

antibacterial agent), it has been applied on many industrial applications like brake lin-

ings, surface coating, paints and varnishes. But it can undergo many polymerization

reactions because all compounds from CNSL contain at least a phenolic group (poly-

condensation) and a meta-substituted unsaturated long aliphatic chain (C15) (polyad-

dition). Mainly with cardanol, due to the hydroxyl group, polycondensation reactions

with formaldehyde (or a less toxic alternative) are well studied. However, very few is

1

known about how they react on radical addition polymerization (BHUNIA et al. (1999),

BALGUDE and SABNIS (2014), MUBOFU and MGAYA (2018)).

Finally, the worldwide annual production of raw cashew nuts stands at approxi-

mately 2.1 million tons. Knowing that 25wt%. of the total cashew fruit is the nut, and

30-35 wt%. of the shell is CNSL, depending on the method of extraction, million tons of

CNSL is produced and obtained annually. However, it does not have this demand, and

is even used as furnace fuel to consume it. Therefore, the aim of this study is the use

of CNSL, or its derivatives (e.g. cardanol) as a co-monomer to functionalize polymers

via free radical polymerization, proposing to be an alternative waste disposal of CNSL.

Then, consequently, the functionalization may happen due to the presence of carboxyl

and hydroxyl groups presented in the cashew nut compounds (TIAN et al. (2012)).

Firstly, bulk copolymerization of cardanol/natural CNSL with styrene, methyl

methacrylate or vinyl acetate is studied – kinetics experiments were done for conver-

sion data; GPC, NMR, FTIR, DSC and TG analysis are interpreted; and, after proposing

a model, some kinetic parameters are estimated. Because of the long side chain, steric

hindrance is expected. Also, inhibitory effects can happen due the phenolic moiety

and appearance of stable radicals.

Then, as suspension polymerization kinetics is similar to the bulk polymerization

in kinetic aspects, new experiments were done is suspension based on those already

obtained. In this case, functionalized particles are made, and the porosity, hydropho-

bicity and size are studied. These particles can be used for bioconjugation, immobi-

lization of enzymes, as a drug carrier and other uses.

Figure 1.1: Logo from COPPETEX

1.1 Motivation

2

Chapter 2

Theoretical Foundation About Polymers

2.1 Introduction

In this chapter, a brief theoretical introduction about polymers is discussed. Poly-

mers are macromolecules formed by joining together repeating units (monomers),

building a long molecule with high molecular mass. However, there are several ways

of building them. First, the kinetic scheme can be chain or step-growth polymeriza-

tion. You can use more than one monomer, enabling the formation of a copolymer,

terpolymer and so on. Also, a single monomer can form more than one bond, cre-

ating a network; or a catalyst or a chain transfer agent can be added to control the

polymerization. And you can do all of this in bulk, solution, suspension or emulsion

polymerization. Every single aspect changes the final characteristics of the polymer

made. Therefore, it is important to understand each process. Figure ?? shows ways to

classify a polymer.

2.2 Types of Polymerization

Polymerization is the process that small units known as monomers are joined by

covalent bonds, building high-molecular-weight polymers. This can be done by two

processes: chain-growth polymerization and step-growth polymerization. If the reac-

tion is started by a reactive site (a cation, an anion, or a radical), then a chain-growth

polymerization occurs. However, if more than one monomer is used and they have

a complementary chemical function, enabling the occurrence of a chemical reaction,

then a step-growth polymerization happens.

3

Figure 2.1: Some Ways to Classify a Polymer

Chain-Growth Polymerization

Chain polymerization (which is also called by vinyl, olefin or addition polymeriza-

tion) starts when a reactive site (a cation, an anion, or a radical) is formed. It repre-

sents 40% of worldwide resin production (ODIAN, 2004). Basically, three stages occur,

and high molecular mass polymers are already formed during the first instants (RO-

DRIGUEZ et al., 2015).

• Initiation: the creation of an active center - a free radical, a carbanion or a car-

bonium ion. This active specie reacts with a monomer, adding to the double

bond, regenerating another radical and initiating the polymerization. In the case

of radical polymerization, radicals can be formed by the thermal dissociation

(usually 50 – 150°C ) of benzoyl peroxide (BPO) or azobisisobutyronitrile (AIBN)

(BILLMEYER, 1984). Monomers that polymerizes through free radical polymer-

ization have at least one unsaturation. Taking the initiation of styrene with BPO

as an example, the reaction happens as Figure 2.3 shows.

In addition, thermal initiation of a specie, like the thermal initiation of styrene,

can also take place as shown in Figure 2.4. Other ways to stimulate the creation

of radicals is the use of light (UV or LED), which is a great way to control the

polymerization rates.

• Propagation: the addition of more monomer to the reactive site (or active cen-

4

Figure 2.2: Free Radical Polymerization: Initiation

Figure 2.3: Homolytic scission of BPO followed by the initiation of styrene polymeriza-tion

Figure 2.4: Thermal initiation of styrene

ter) at the growing chain end. A common hypothesis is that the propagation

rate constant (kp ) is the same for every living chain and does not depend on the

length (long-chain hypothesis). This is based on the idea that the reactivity of

a growing polymer chain depends only on the last unit added and not the pre-

vious added units. Again, taking the polymerization of styrene as an example,

propagation happens as shown in Figure 2.6.

5

Figure 2.5: Free Radical Polymerization: Propagation

Figure 2.6: Propagation of poly(styrene)

• Termination: the disappearance of a living chain (active center) when this one

reacts with another. Because these chains lose their radicals and functional activ-

ity, they cannot grow any longer and are called dead chains. In general, two types

of termination can occur. If it is through combination (coupling), this means that

two molecules are getting together forming a single molecule. However, if they

lose their radicals, forming two dead chains, it is a disproportionation. This hap-

pens because one polymeric chain transfers a hydrogen atom to another living

chain, forming a saturated and an unsaturated dead chain. The final molecular

weight of a dead polymer formed by combination is twice the one obtained by

disproportionation (RODRIGUEZ et al., 2015). Continuing the styrene example,

termination by disproportionation or coupling is shown in Figure 2.8.

Figure 2.7: Free Radical Polymerization: Termination (Disproportionation or Cou-pling)

6

Figure 2.8: Termination of poly(styrene)

Step-Growth Polymerization

If two monomers have a complementary chemical function, and each of them has

functionality equal or greater than two, a step-growth polymerization can take place.

It is important that they have the same functionality, otherwise the polymerization will

cease as the functional groups will not be able to react with the other. Also, it is impor-

tant that the molecular amount added corresponds to the stoichiometric ratio. Figure

2.9 shows the synthesis of Nylon 66 by step-growth polymerization in which carboxy-

lates of the acid reacts with the amine group.

Figure 2.9: Synthesis of Nylon 66, a step-growth polymerization

The use of an initiator is not required. Instead, a catalyst is used to accelerate the

polymerization. But one of the biggest differences between step-growth polymeriza-

tion and chain-growth is that, in this process, a polymer with high molecular weight

is formed only when all monomer is consumed. Moreover, throughout the process, a

small molecule is being eliminated. For example, during esterification, which is a reac-

tion between an alcohol and an acid, molecules of water are being formed. Therefore,

conversion can be controlled by removing this molecule to force the equilibrium to the

direction of products production due to the Le Chatelier’s principle.

7

Initiator and Its Half-Life

Initiators are used to initialize the polymerization, as show in Figure 2.2. Benzoyl

peroxide (BPO) and azobisisobutyronitrile (AIBN) are the most common. Their chem-

ical structure is shown in Figure 2.10. They cleave in the presence of an external stim-

ulus (e.g.: temperature or light), generating two radicals. The decomposition is a first-

order process (RODRIGUEZ et al., 2015).

Figure 2.10: Decomposition of BPO (a) and AIBN (b)

−d [I ]

d t= ki [I ] (2.1)

l n[I ]

[I0]=−ki t (2.2)

If half of the initiator concentration is left, the time took is known as half-life. Thus, we

have:

t1/2 = 0.693/ki (2.3)

And ki can be correlated by an Arrhenius expression, where Ei is the energy of acti-

vation for the decomposition process, A is the collision frequency factor, R is the gas

constant and T is the absolute temperature (K ).

ki = Aexp

(−Ei

RT

)(2.4)

8

However, not every molecule of initiator dissociates generating two radicals and suc-

cessfully reacting with monomer. Therefore, it is said that only a certain fraction f of

initiator is successful. Equation 2.5 shows how to obtain f , being n the number of mols

formed by the decomposition of 1 mol of initiator and, thus, n = 2 for BPO or AIBN. In

general, it is used a value in the range of 0.8±0.2 (RODRIGUEZ et al., 2015). However,

f decays as the viscosity increases and monomer concentration decreases (ODIAN,

2004).

f = i ni t i at i on r ate

n ∗ i ni t i at i or decomposi t i on r ate(2.5)

Chain Transfer

Sometimes, the molecular weight is smaller than excepted. A reason for this may be

chain transfer. Some species present during polymerization can act as chain transfer

agents (CTAs): solvent, initiator, monomer, impurities, or polymer. They capture, in an

irreversible manner, the polymeric radical by some weakly bound atom (normally, a

hydrogen atom), resulting in a dead chain, but initiating a new one (RODRIGUEZ et al.,

2015). Transfer to monomer with subsequent polymerization of the double bond lead

to the formation of branched molecules, changing the molecular weight distribution

(BILLMEYER, 1984).

Inhibition and Retardation

A retarder is a substance (e.g. an impurity) that can react with a radical to form

products incapable of adding monomer. If the retarder is very effective, it becomes

an inhibitor. Therefore, inhibitors are species that reacts with radicals to give stable

species and, therefore, reducing the polymerization rate or ceasing it. Thus, the length

of the polymer chains gets smaller too. Hydroquinone and diphenylamine are fre-

quently used to prevent undesired polymerization or to stop it. Oxygen also acts as

an inhibitor, and total removal of it can be a key to allow some polymerizations sys-

tems (e.g.: living radical polymerization). "Induction period" is the time that takes to

normalize the reaction rate (without retarders).

Gel and Glass Effects

A controlled or theoretical polymerization (with initiation, propagation and termi-

nation steps) proceeds by first-order kinetics if the initiator concentration does not

vary largely and its efficiency is independent of monomer concentration. This means

that the polymerization rate is proportional to monomer concentration so that, in gen-

eral, it is detected a first-order kinetics. However, at high concentrations, the overall re-

action rate decreases with time as monomer and initiator are being consumed. How-

9

ever, mainly in bulk polymerizations, a deviation from first-order kinetics happens,

increasing the reaction rate and the molecular weight. This behavior is known as gel

effect (or autoaccelaration or Trommsdorff effect).

This effect happens because, as the conversion increases, viscosity also increases,

thereby decreasing the rate at which polymer molecules diffuse through the medium

as termination is diffusion controlled for most liquid-phase polymerizations. There-

fore, molecules cannot find each other and terminate, increasing their lifetime

(BILLMEYER, 1984).

Also, as living chains cannot move around freely to react with monomers due to

their size, steric hindrance and a viscous medium; the propagation rate is affected.

This is known as glass effect.

Homo- and co-polymerization

Figure 2.11: Last model applied for copolymerization

For free-radical polymerization, if more than one monomer is used, polymers with

more than one type of mere in the same backbone can be formed. And, depending on

the technique used and the reactivity of the monomers, different types of copolymers

can be formed. A good way to represent the polymerizations is the terminal model,

which considers that the reactivity of the polymeric radical only depends on the nature

of the last mere added and not the rest of the chain. This method is very popular for

co-polymerizations as it is not very challenge to solve and is a good approximation.

Figure 2.11 illustrates the last model.

However, sometimes it is used the penultimate model (Figure 2.12) that considers

the reactivity of the last to mere of the chain. But now, instead of four propagation re-

actions, eight reactions need to be solved or eight kinetic parameters need to be known

and they are not easy to determine experimentally (BILLMEYER, 1984).

After doing the mass balance and assuming the quasi-steady state to determine the

composition of a copolymer of a copolymerization, it is used the copolymer equation

10

Figure 2.12: Penultimate model

(Equation 2.6), also known as Mayo-Lewis equation (MAYO and LEWIS, 1944). This

equation shows that the copolymer composition does not depend on the initiation or

termination rates nor the presence of inhibitors, CTAs or solvents. Therefore, for het-

erogeneous systems (like suspension), those same reactivity ratios obtained for bulk

polymerization can be used. Moreover, this equation introduced the reactivity ratio

to determine if the monomer will react more frequently to its own specie or the other

monomer (Equations 2.7 and 2.8) (BILLMEYER, 1984).

d [M1]

d [M2]= [M1]

[M2]

r1[M1]+ [M2]

[M1]+ r2[M2](2.6)

r1 = k11

k12(2.7)

r2 = k22

k21(2.8)

The reactivity of monomers and radicals in copolymerization is determined by the

nature of the substituents on the double bond of the monomer:

• - The substituents may activate the double bond, making the monomer more

11

reactive;

• - They may stabilize the resulting radical by resonance;

• - They may provide steric hindrance at the reaction site.

Besides the Mayo-Lewis equation, other methods used to calculate the reactivity

ratios are: KELEN and TÜDOS (1974), TIDWELL and MORTIMER (1965), and FINE-

MAN and ROSS (1950). These methods can give discrepancies between the values

TIDWELL and MORTIMER (1970). Also, experimental design and the technique used

to analyze the data can influence in the precision of experimentally determined reac-

tivity ratios (ZALDÍVAR et al., 1998).

2.3 Polymerization methods

Depending on the final application of the polymer, they can be obtained using sev-

eral polymerization methods. If they occur in only one phase, they are called homo-

geneous – like the bulk and solution polymerization. If not, it is a heterogeneous poly-

merization – suspension and emulsion polymerizations. In this dissertation, only the

bulk and suspension polymerization are used so they will be explored deeply.

2.3.1 Bulk Polymerization

Using for free radical polymerization, the advantage of a bulk polymerization is

that anything besides the monomer and the initiator is required. Therefore, the final

polymer has a high purity. However, as the polymerization is an exothermic reaction

and viscosity increases, control this system is difficult. Heat removal is impeded by

high viscosity and low thermal conductivity. Also, the removal of traces of unreacted

monomer from the final product gets difficult because of low diffusion rates and low

surface-to-volume ratio, compromising the purification of the final product. Conver-

sion of all monomer is complicated for the same reason (RODRIGUEZ et al., 2015).

2.3.2 Solution Polymerization

Using a solvent can be very beneficial because it can help with the temperature

and viscosity control of the system. The reaction happens with full miscibility of the

monomer and initiator in solvent. The kinetics is like the one of bulk polymerization.

However, sometimes, depending on the solvent being used, undesired chemical inter-

actions can happen between the species and the solvent. Besides it, purification and

recirculation of the solvent increases the cost of the polymer production, and the pro-

ductivity decays (CANEVAROLO JR., 2006).

12

2.3.3 Suspension Polymerization

If the monomer is insoluble in water, bulk polymerization can occur inside of sus-

pended droplets of polymer in water, which acts like the heat transfer medium. How-

ever, although the viscosity within the beads rises, increasing reaction rate, the vis-

cosity of the continuous phase (water) almost does not change, improving the heat

transfer. To avoid coalescence, an suspension agent is used, commonly poly(vinyl al-

cohol) (PVA). Moreover, polymer particles produced through suspension can get from

10 to 1000 µm diameter. A way to control them is by changing the concentration of

suspending agent and the stirring rate.

A typical suspension procedure is the following: a reactor is filled with water con-

taining PVA (0.25% w/w in relation to H2O), this solution is heated up to the reac-

tion temperature and stirring happens throughout the whole polymerization. Then

the mixture of monomer and initiator is added to the reactor. After a certain amount

of time, when the conversion is close to 100%, the heat bath and the stirring can be

turned off. Finally, the suspension can be filtrated to obtain pearl-like beads of poly-

mer (RODRIGUEZ et al., 2015).

2.3.4 Emulsion Polymerization

As it happens on suspension polymerization, the monomer is dispersed on a con-

tinuous phase. Frequently, water is used as the continuous phase, but emulsion inver-

sion occurs using an oil. It also uses an emulsifying (surfactant) agent, which is present

in form of micelles – hydrophobic tails oriented inward and hydrophilic heads out-

ward. These micelles appear when the concentration is higher than the critical micelle

concentration (CMC). However, the initiator is hydrophilic and, because the micelles

are smaller than droplets, the surface area is much greater, enabling to capture free

radicals generated in the aqueous phase. Therefore, in emulsion polymerization, the

system consists of monomer-swollen polymer particles and droplets of monomer that

keeps getting into the old micelles. Particles between 50nm and 5µm can be obtained

by this method (RODRIGUEZ et al., 2015).

2.4 Molecular Weight Distribution (MWD)

Even for a homo-polymerization, there are infinite reactions happening at the same

time. You can’t know for sure if a living chain is going to propagate or terminate, but

there is a probability to each effect. Therefore, some of them propagate more while

others are losing their radicals “prematurely”. Thus, each reaction produces a mix-

ture of polymers with the same composition, but different degrees of polymerization.

13

Chain length distribution is a property that relates each chain with size i with the fre-

quency of the chain with the same i present, Pi . With mass basis, it becomes:

Wi = Pi i M M (2.9)

Experimentally, the distribution is determined by use of techniques such as high-

pressure liquid chromatography (HPLC) or size-exclusion chromatography (SEC). The

frequency of each size/molar is frequently normalized, enabling comparison between

different distributions. Therefore, it is possible to have a probability density function

(PDF) that treats the size of polymeric chain as a random variable with a certain prob-

ability to exist or not (LEMOS, 2014).

P nor mi = Pi∑∞

i=1 Pi(2.10)

W nor mi = Wi∑∞

i=1 Wi(2.11)

There is a chain size distribution known as Most Probable Distribution, or Schulz-

Flory Distribution, that is frequently found on polymeric systems (Equation 2.12). q is

the probability of propagation and Ptot is the total amount of chains (?).

Pi = Ptot q i−1(1−q) (2.12)

To simplify the distribution, although they might me ambiguous, it is highly com-

mon to represent the distribution by distribution moments. The medium average

length in a numeric base in and the medium size in mass base iw are:

in =∑∞

i=1 i Pi∑∞i=1 Pi

(2.13)

iw =∑∞

i=1 iWi∑∞i=1 Wi

(2.14)

And the number average molar mass Mn and mass average molar mass Mw are calcu-

lated by:

Mn =∑∞

i=1 Mi Ni∑∞i=1 Ni

= in M M (2.15)

14

Mw =∑∞

i=1 M 2i Ni∑∞

i=1 Mi Ni= iw M M (2.16)

Also, polydispersity index (PI) is a measure of the variance of the chain size distribu-

tion, defined by Equation 2.17, being σ2 the variance of the distribution.

PD I = Mw

Mn

= iw

in

= 1+ σ2

in2 (2.17)

15

Chapter 3

Cashew Nut Shell Liquid – A Literature

Review

3.1 Introduction

The cashew tree is a plant native to the Brazilian northeast, which was taken in

the 16th and 17th centuries to India, Africa and other tropical countries. It belongs

to the family Anacardiaceae, in which the specie Anacardium occidentale is the most

explored variation. From it, one obtains the pseudo fruit (popularly known as cashew

apple) and the cashew (the almond), two products of high economic interest. For il-

lustration purposes, Figure 3.1 shows the combination of the cashew nut (true fruit)

and the peduncle (pseudo fruit), in which is consumed as fruit in nature or to produce

juices (e.g. cajuína), sweets and others.

Therefore, to obtain the cashew nut, it needs to undergo a process of beneficiation

in order to remove the shell (pericarp) that surrounds it. In Brazil, this process is about

90% mechanized. An example of this process is shown in Figure 3.2. However, mainly

in the northeast interior, the artisan process is still very used.

Figure 3.3 is a picture of the cashew nut - formed by the shell and the almond.

The shell is constituted by the epicarp (outer layer), endocarp (inner layer) and the

mesocarp (middle layer). It is in the mesocarp that the cashew nut shell liquid (CNSL)

is found. Truly, this is an oil composed of several phenolic lipids, among them the

anacardic acid and the cardanol.

Figure 3.4 revels the richness and diversity of products that can be obtained from

cashew. Looking at the cashew nut ramification, followed by the shell, many CNSL

applications are show. However, while CNSL has the potential to be a valuable raw

16

Figure 3.1: Cashew

material, it is still considered a low value-added waste in the country. Therefore, re-

search is this field is highly required so that innovative and sustainable products can

be developed by the Brazilian industry.

About the cashew production, since 1990, there were already about 582 thousand

hectares of plantations. Today, there are more than 607 thousand hectares such that,

in relation to the fruit species in Brazil, the area with cashew trees only loses to the

planted with orange trees (IBGE, 2016). 75% of the cashew trees are in the states of

Ceará, Rio Grande do Norte and Piauí (IBGE, 2016).

According to the IBGE census of 2016, Ceará represent approximately 61% of the to-

tal cashew nut produced in Brazil, followed by the states of Rio Grande do Norte (19%),

Piauí (13%), Bahia (3%), Maranhão (2%) and Pernambuco (1%). For 2018, 141,388

tons were produced in Brazil. However, due mainly to climatic conditions, produc-

tion varies annually (Figure 3.5). In 2006 and 2008, the maximum production were

obtained – 243 thousand tons.

In relation to the world market, according to the latest data from Food and Agricul-

ture Organization of the United Nations (FAOSTAT), Brazil is the 4th largest exporter

of cashew nuts, behind Vietnam, India and the Gulf of Guinea (MUBOFU, 2016). The

largest importers of the Brazilian cashew nut are the United States, Canada, the Nether-

lands, Argentina, the United Kingdom and Mexico (Secretariat of Foreign Trade, SE-

CEX, Brazilian Ministry of Development, Industry and Foreign Trade).

17

Figure 3.2: An Example of Beneficiation process for Cashew Nut

3.2 Cashew Nut Shell Liquid (CNSL)

CNSL accounts for approximately 25% of the nut weight and is a by-product of

cashew nut production, reaching one million tons annually MAZZETTO et al. (2009).

Currently, it has a low added value and is even used as an energy source (furnace fuel).

However, the liquid is rich in non-isoprenoid phenolic lipids and is a good a material in

the synthesis of organic products. Since 1970s, CNSL has been used as a raw material

in the manufacture of insecticides, germicides, antioxidants, thermal insulators, plas-

ticizers, paints, vanishes and others applications (GEDAM and SAMPATHKUMARAN

18

Figure 3.3: Cashew nut: the shell and the almond (adapted from Sistemas deProdução, EMBRAPA).

Figure 3.4: Products obtained from cashew (adapted from AgroIndúsria do Caju noNordeste, ETENE/BNB, 1973).

(1986), LUBI and THACHIL (2000), MUBOFU (2016)).

Figure 3.6 shows the phenolic lipids found in the CNSL. All of them have a side

chain with fifteen carbon side chain located at the meta-position with respect to the

benzene ring. This chain may be saturated (C15H31) or unsaturated: mono- (C15H29),

di- (C15H27) or tri- (C15H25), whose occurrences happens at carbon 8, 11 and 14. Be-

cause of this side chain, polymers formed partially with these lipids are flexible since

the chain acts as a “plasticizer”.

Cardanol (3-n-pentadecylphenol) has a hydroxyl in the aromatic ring as do the

other compounds. However, anacardic acid (3-n-pentadecylsalicylic acid) has a car-

boxyl group in the ortho position; cardol (5-npentadecylresorcinol) has a second

hydroxyl (diphenol) and 2-methyl-cardol (2-methyl-5-n-pentadecylresorcinol) has a

19

Figure 3.5: Production of Cashew Nut between the years of 2006 and 2018(IBGE/CEPAGRO)

Figure 3.6: Components of the CNSL

methyl group between two hydroxyls. These compounds are very interesting as due to

their multiple reactive sites that can generate derivatives or impart various functional-

ities to the molecule target product. Figure 3.7 shows characteristics of cardanol and

anacardic acid due to their functional groups and how this can impart on applications.

Aiming to obtain the CNSL out of the nut, if a cold extraction method is used, uti-

lizing a mechanical press or solvents (cold extrusion method), the CNSL is classified as

natural. However, it can also be extracted using a thermal method: hot oil bath, roast-

ing or solar cooker. In the former, the hot CNSL itself is used to heat the nuts and it

20

Figure 3.7: Advantages arising from functional groups of cardanol and anacardic acid

is the most common method applied industrially (KUMAR et al., 2018). Using these

methods, the cashews are submitted to high temperatures (higher than 180°C ). Thus,

the anacardic acid is decarboxylated and converted into cardanol, releasing carbon

dioxide and being called “technical CNSL” (Figure 3.8).

Figure 3.8: Decarboxylation of anacardic acid

In addition, during heating, temperatures above 150°C and the presence of oxygen

lead to the appearance of polymeric substances. This means that thermal and oxida-

tive polymerization occurs (MAHANWAR and KALE, 1996). Figure 3.9 shows a mech-

anism for thermal polymerization of cardanol. Supercritical CO2 extraction was also

reported and, despite had obtained the best yield among other methods, practically

did not contain anacardic acid (PATEL et al., 2006) and it is not very common (KUMAR

et al., 2018).

Table 3.1 shows the difference between the composition of natural and technical

CNSL. In Brazil, the thermal-mechanic process is commonly used, obtaining the tech-

nical CNSL in which is rich in cardanol. The table also shows that anacardic acid con-

stitutes in average 77% of natural CNSL. However, in the literature are relates of natural

CNSL containing about 60 – 65% of anarcardic acid. Nevertheless, using solvent for

21

Figure 3.9: Thermal Polymerization of cardanol. (OLIVEIRA, LINCOLN DAVI M. DE ,2007)

cold extraction is possible to obtain approximately 90% of anacardic acid and 10% of

cardanol (PATEL et al., 2006).

1The percentages describe the lower and upper limits using different analytical techniques. New, dis-tilled and aged samples were analyzed. The dashed line represents a minimal percentage inconsiderable

22

Table 3.1: Difference between natural and technical CNSL (adapted from MAZZETTOet al. (2009))

Phenolic Compounds* Natural CNSL (%) Technical CNSL (%)

Anacardic Acid 71.70 – 82.00 1.09 – 1.75Cardanol 1.60 – 9.20 67.82 – 94.60

Cardol 13.80 – 20.10 3.80 – 18.862-Methyl Cardol 1.65 – 3.90 1.20 – 4.10

Minority Components 2.20 3.05 – 3.98Polymeric Material - - - 0.34 – 21.63

1

Table 3.2: Composition of natural CNSL measured by gas chromatography and massspectroscopy (GC-MS) (adapted from MAZZETTO et al. (2009))

Constituent Anacardic Acid (%) Cardanol (%) Cardol (%) 2-methyl cardol (%)

Saturated 2.2 – 3 3.9 - 4.4 0.2 - 2.7 0.9 - 1.3Monoene (8’) 25.0 - 33.3 21.6 - 32.2 8.4 - 15.2 16.3 - 25.3Diene (8’, 11’) 17.8 - 32.1 15.4 - 18.2 24.2 - 28.9 20.6 - 24.4

Triene (8’, 11’, 14’) 36.3 - 50.4 45.2 - 59.0 36.5 - 67.2 49.8 - 62.2

Anacardic acid has called attention due to its antimicrobial, anticoagulant, anti-

tumor, antifungal and molluscicide properties (DAVID et al. (2006), MAZZETTO et al.

(2009)). These properties are influenced by the side chain. Cardanol and its deriva-

tives have been used as plasticizers, surfactants, curing agents, doping agents, and in

phenolic and epoxy resins as well as in nano-composite films (LI et al., 2018b).

Table 3.2 presents the composition of natural CNSL in relation to number of unsat-

uration present in the side chain, being the triene the side chain more present in every

phenolic compound. Values change inside of an interval due to the regular variations

of CNSL.

Monoene anacardic acid can be obtained by boiling (with reflux condensate), the

acid mixture in a solution of isopropyl alcohol (or propan-2-ol, reducing agent) and

ruthenium (III) chloride (catalyst) in acetone. Under these conditions, the diene

and triene are hydrogenated to monoene (PERDRIAU et al., 2012). The temperature

reached () does not decarboxylate the anacardic acid to cardanol (MUBOFU, 2016).

Molecular weight, density, boiling point and viscosity for cardanol and anacardic

Table 3.3: Properties of cardanol and anacardic acid.

Property Value

Cardanol Anacardic AcidMolecular Weight (Da) 298.0 – 304.0 342.5- 348.5

Density (g/cm3) 0.927 – 0.934 (30°C ) 1.028Boiling Point (°C ) 228 – 235 496.4

Viscosity (cP) 95 (25°C ); 45 – 60 (30°C ) 150 – 600 (25°C )

23

Figure 3.10: Anacardic Acid: (i) 1-hydroxy-2-carboxy-3-pentadecyl benzene,(ii) 1-hydroxy-2-carboxy-3-(8-penta-decenyl)benzene,

(iii) 1-hydroxy-2-carboxy-3-(8,11-pentadecadienyl)benzene, and(iv) 1-hydroxy-2-carboxy-3-(8,11,14 pentadecatrienyl)benzene.

acid are presented in Table 3.3 (RODRIGUES et al. (2011), MAZZETTO et al. (2009).

3.3 Synthesis of polymers using CNSL

In the literature, there are reports of polymer synthesis from CNSL mainly by con-

densation with electrophilic agents (e.g. formaldehyde), polymerization of the side

chains using acid catalysts, or by the functionalization of the hydroxyl group to obtain

an oligomer as functionalized prepolymer (BALGUDE and SABNIS, 2014). The poly-

mers derived from the CNSL are known to have good flexibility because of the long

aliphatic chain, low wear, high resistance to friction and impact. LUBI and THACHIL

(2000) reviewed most of the reactions (including polymerizations) and applications

made with the CNSL until the late 90s.

However, instead of using the mixture of compounds (CNSL), isolated cardanol has

been the most studied and used substance obtained from the CNSL in polymer sci-

ence. In the literature, there are reports of oxidative (using acid catalysts), cationic, in

stages (polycondensation) and enzymatic polymerizations (IKEDA et al. (2000b), KIM

24

et al. (2003)) using cardanol directly. Furthermore, cardanol derivatives, i.e. monomers

synthesized from cardanol, have also been used (BHUNIA et al., 1999) as, due to the

hydroxyl moiety, many chemical transformations are possible with cardanol (Figure

3.11). A good review targeting the coating industry was made by BALGUDE and SAB-

NIS (2014). LOCHAB et al. (2014) also did a great review about application of natural

phenolic compounds in the area o polymers.

Figure 3.11: Chemical Transformations of Cardanol (BALGUDE and SABNIS, 2014)

About the oxidative polymerization, the synthesis of polycardanol has been made

using a Fe-salen (iron-N,N-ethylenebis(salicylideneamine) complex as catalyst. IKEDA

et al. (2000a) used Fe(II)-salen, which is posteriorly oxidized to Fe(III)-salen, the real

catalyst of the polymerization, by hydrogen peroxide. On a first study, he utilized

pyridine to increase the yield, and THF or 1,4-dioxane as solvents. After NMR and

FTIR analysis, the obtained polymers were found to be a mixture of oxyphenylene and

phenylene. These are considered a class of crosslinkable polyphones.

In a second publication, IKEDA et al. (2000a) studied the same polymerization in

bulk. It was investigated the concentration of catalyst, type of catalyst, concentration

of hydrogen peroxide, time and additives (water or pyridine). Thus, it was found that

the greatest yield (80%) was obtained with the Fe-salen complex when the concen-

tration of catalyst and hydrogen peroxide were high. However, the time was not very

25

Figure 3.12: Oxidative Polymerization of Cardanol using Fe-Salen as a catalyst (IKEDAet al., 2000a)

significant as from half an hour to two hours, the polymer yield did not vary largely.

Also, even using other metallic complexes, Fe-salen obtained the greatest result.

Figure 3.13: Structure of Fe-salen complex

Still using the oxidative route, OTSUKA et al. (2017) reported the polymerization of

cardanol in emulsion w/o, always using the weight proportion of cardanol bigger or

equal to the water. To keep the system stable, it was used polyamine and acetic acid.

After 24 hours, depending on the system conditions, 50% of conversion was achieved,

average molecular weight of 6800 Da e polydispersion index of 2.3.

Besides the oxidative route, poly(cardanol) was obtained by an enzymatic polymer-

ization using peroxidases. The product was a mixture of oxyphenylene and phenylene

units. Using soybean peroxidase, IKEDA et al. (2000b) obtained the highest yield (69%)

with isopropanol as solvent (Mn = 6100 and PI = 1.8). Doubling the amount of enzymes,

KIM et al. (2003) studied the same polymerization. They obtained 62% of yield, Mn =

3411 and Mw = 8221 Da (PI = 2.41). The structure of the obtained polymer is similar to

those obtained using the oxidative route.

Another route is the polycondensation in which oligomerization takes place, yield-

ing functionalized pre-polymers. CNSL step-growth polymerization can be performed

26

with electrophiles for reactions with hydroxyl group or, very commonly, an aldehyde

(formaldehyde) is used. In the last, it is obtained a formaldehyde-cardanol resin, form-

ing a phenolic thermosetting resin. For example, MISRA and PANDEY (1984) (or MISRA

and PANDEY (1985)) studied the polymerization using a alkaline catalysis (NaOH).

Moreover, BISANDA and ANSELL (1992) evaluated that this composited have adequate

strength for roofing applications.

In general, these resins are more flexible and soluble in organic solvents relative to

others. In addition, they are hydrophobic and have better resistance to acids and bases.

However, they have lower tensile strength and thermal stability than the traditional

phenol-formaldehyde resin (PF) (LOCHAB et al., 2014). This may be attributed to the

side chain that imparts steric hindrance and reduction in intermolecular interactions.

Thus, an optimum percentage replacement of phenol by cardanol is required. For ex-

ample, phenolic resins containing less than 15 wt% cardanol have distinctly improved

chemical resistance and mechanical properties compared to neat phenolic resins.

Moreover, polycondensation using acid catalysis has been reported using oxalic,

succinic, citric and other acids (SATHIYALEKSHMI (1993), YADAV and SRIVASTAVA

(2007a), LOUREIRO et al. (2017)). For example, SOUZA JR. et al. (2008a) prepared the

resin one step with H2SO4. FTIR and XPS confirmed the polymerization. It was also

noticed the presence of sulfur in the polymer, which influences its acidity and is advan-

tageous for mixtures with polyaniline (conductive polymer). In another work, SOUZA

JR. et al. (2008b) used the cationic route to produce the same resin applying H2SO4.

Cationic polymerization can also be used to synthesize poly(cardanol) with boron

trifluoride etherate (BF3O(C2H5)2) as initiator (SCARIAH (1990), LOUREIRO et al.

(2017)). Sulfuric or phosphoric acids can be used as catalysis (MANJULA et al., 1992)

The polymerization occurs on the side chain (Figure 3.14).

Figure 3.14: Cationic Polymerization of Cardanol

27

BESTETI et al. (2014) and GALVÃO (2016) studied the direct use of cardanol as

comonomer using several polymerization methodologies – bulk, solution, suspension

and emulsion. However, due to the poor reactivity of the C15 unsaturations, conver-

sion did not get very high, but their products have shown to be attractive for some

applications, inspiring the present study.

Figure 3.15: Synthesis of poly(cardanyl acrylate) and its film (MANJULA et al., 1992)

As previously mentioned, cardanol is used to prepare new monomers. Acrylates

and methacrylates derived from cardanol have proven to be more successful as poly-

merized monomers in radical polymerization (LI et al., 2018a). An example is the syn-

thesis of cardanyl acrylate or cardanyl methacrylate in which polymers, synthesized

via free radical polymerization, form a lattice structure when exposed to air or ultravi-

olet light, obtaining a thermoplastic and practically colorless polymer (MANJULA et al.,

1992). Figure 3.15 shows the chemical structure of cardanyl acrylate and its polymer.

Several companies have been using CNSL, and mainly cardanol, to produce car-

danol derivatives, polycarbonate, epoxy resins, detergents, foams, surfactants and

other products. Table 3.4 presents some patents that uses CNSL or cardanol as reagent.

A company that calls attention amoung the assignee is Cardolite Inc. This American

company is the world’s largest cardanol producer. Besides cardanol (commercially

available under the trade name Cardolite NC-700 or NX-2026), their most relevant

products are cardanol derivatives. To cite a few: reactive diluents and flexible resins

(e.g. NC-513, NC-514, NC-514 LV); epoxy novolac resins (e.g. NC-547) – see Figure

3.16; phenalkamine curing agents – condensation product of cardanol, formaldehyde

and a polyamine - (e.g. NC-540, NC-541, NC541LV, NC-556, NC-558, NC-559, NC-560,

NX 2015); polyols (NX-9001LV, GX-9005, GX- 9007, GX-9201, NX-4670)

Plus, polyurethanes resins are also been made with cardanol derivative polyols with

a wide range of polyisocyanates (Suresh et al, 2005; WO2006003668A1). There are

several patents discussing it (Table 3.5). A company that has been growing with the

cardanol polyols and polyurethanes business is Covestro, a Germany company and a

28

Table 3.4: Patents using cardanol as a reactive molecule

Co

de

Pri

ori

tyYe

arSt

atu

sC

urr

ent

Ass

ign

eeN

ame

Des

crip

tio

nA

pp

licat

ion

WO

2000

0342

19A

119

98G

ran

tin

US

Car

do

lite

Co

rp.

Der

ivat

ive

ofc

ard

ano

lan

du

ses

ther

efo

r

Am

eth

od

for

the

hyd

roxy

alky

lati

on

of

card

ano

l

Co

atin

gs,a

dh

esiv

es,

seal

ants

,ru

bb

er,p

last

ic,

elas

tom

er,c

om

po

site

s,an

das

am

od

ifier

inth

ein

kco

mp

osi

tio

n

WO

1998

0360

15A

119

97G

ran

tin

US,

EP,

JP,C

N,

DE

Sab

ic

Red

istr

ibu

tio

no

fp

oly

ph

enyl

ene

eth

ers

and

po

lyp

hen

ylen

eet

her

sw

ith

nov

elst

ruct

ure

Rea

ctio

nw

ith

PP

E,

resu

ltin

gin

aso

lf-s

egm

entP

PE

wh

ich

imp

rove

sfl

owp

rop

erti

esas

wel

las

com

pat

ibil

ity

wit

hn

on

-po

lar

com

po

nen

ts.

Ad

hes

ives

,sea

lan

ts,

ther

mo

sets

,sty

ren

icre

sin

s,p

oly

ole

fin

es,

po

lym

erb

len

ds,

pro

vid

ing

bet

ter

flow

,ta

ck,r

eact

ivit

y,ox

idat

ive

stab

ilit

yan

dth

erm

alp

rop

erti

es

WO

2003

0102

18A

120

01G

ran

tin

US,

EP,

KR

,CN

,D

E

Gen

eral

Ele

ctri

cC

om

-p

any

Met

ho

do

fpo

lyca

rbo

nat

ep

rep

arat

ion

Use

ofr

awca

rdan

olt

op

rep

are

po

lyca

rbo

nat

esw

ith

mo

reh

ydro

xyl

gro

up

s

Po

lyca

rbo

nat

esca

nb

eb

len

ded

wit

had

dit

ives

such

ash

eat/

UV

stab

iliz

ers,

mo

ldre

leas

eag

ents

,or

wit

ho

ther

po

lym

eric

mat

eria

lsan

db

em

old

ed(e

.g.o

pti

cal

dis

ks,o

pti

call

ense

s,au

tom

ob

ile

lam

pco

mp

on

ents

).

WO

2008

1377

06A

120

07G

ran

tin

CN

and

US

Car

do

lite

Co

rp.

Car

dan

olb

ased

dim

ers

and

use

sth

eref

or

Th

eca

rdan

old

imer

sar

efo

rmed

by

hyd

rosy

lyla

tio

nw

ith

sila

nes

.Car

dan

olb

ased

dim

ers

may

be

furt

her

reac

ted

tofo

rmep

oxy

curi

ng

agen

tsan

dep

oxie

so

rfr

icti

on

par

ticl

eso

rp

hen

oli

cre

sin

s

An

ti-f

ou

ling

coat

ings

on

ship

hu

llsan

dm

arin

est

ruct

ure

s;fr

icti

on

par

ticl

eso

rp

hen

oli

cre

sin

s

29

Co

de

Pri

ori

tyYe

arSt

atu

sC

urr

ent

Ass

ign

eeN

ame

Des

crip

tio

nA

pp

lica

tio

n

WO

2008

1319

18A

120

07A

pp

licat

ion

(200

8)B

asfS

eR

eact

ive

surf

acta

nts

and

thei

ru

se

Use

ofa

nio

nic

card

ano

lo

rca

rdan

ol

(po

ly)a

lkox

ylat

ed

eriv

ativ

esas

po

lym

eriz

able

surf

acta

nts

.Th

ere

acti

vity

oft

he

inve

nte

dp

oly

mer

izab

lesu

rfac

tan

tsca

nb

efu

rth

erim

pro

ved

by

con

juga

tio

no

fth

ese

do

ub

leb

on

ds.

Surf

acta

nts

,sta

bil

izer

san

d/o

rd

isp

ersi

ng

agen

tsin

the

pre

par

atio

no

fp

oly

mer

dis

per

sio

ns.

An

dfo

rco

atin

gsan

db

ind

erco

mp

osi

tio

ns.

Exe

mp

les:

man

ufa

ctu

reo

fpap

ero

rp

aper

bo

ard

,p

ain

ts,n

on

-wov

enfa

bri

csan

db

uil

din

gm

ater

ials

.Cu

rab

leep

oxy

resi

nco

atin

gco

mp

osi

tio

nh

avin

ga

low

visc

osi

tyw

hic

his

use

fuli

nco

atin

gap

pli

cati

on

s.

WO

2014

1173

51A

120

13A

pp

licat

ion

(201

4)

Dow

Glo

bal

Tech

-n

olo

gies

Llc

An

epox

yre

sin

com

po

siti

on

,an

dit

sap

pli

cati

on

s

Pro

du

ctio

no

faca

rdan

ol

mo

difi

edep

oxy

resi

n

Ap

pli

cati

on

inlo

wvo

lati

leo

rgan

icco

mp

ou

nd

s(V

OC

)h

igh

soli

ds

coat

ings

,an

dw

ater

bo

rne

coat

ing

WO

2014

1799

75A

120

13A

pp

licat

ion

(201

4)

Dow

Glo

bal

Tech

-n

olo

gies

Llc

Ep

oxy

resi

nco

mp

osi

tio

ns

Form

atio

no

fan

ovel

epox

yre

sin

by

reac

tin

gep

oxy

resi

ns

wit

hca

rdan

ol

Co

atin

gfi

lms

wit

hb

alan

ced

pro

per

ties

incl

ud

ing

goo

dfl

exib

ilit

y,go

od

chem

ical

resi

stan

ce,s

ho

rtd

ry-h

ard

tim

e,d

esir

able

har

dn

ess,

and

goo

dan

ti-c

orr

osi

on

per

form

ance

;an

dth

ep

roce

sso

fpre

par

ing

the

com

po

siti

on

s.

30

Co

de

Pri

ori

tyYe

arSt

atu

sC

urr

ent

Ass

ign

eeN

ame

Des

crip

tio

nA

pp

lica

tio

n

WO

2016

0864

01A

120

14A

pp

lica

-ti

on

(201

6)

Blu

eC

ub

eIp

Llc

or

Dow

Glo

bal

Tech

no

lo-

gies

Llc

Cu

rab

leep

oxy

resi

nco

mp

osi

tio

n

Form

atio

no

f"an

adva

nce

dep

oxy

resi

n"

by

reac

tin

gan

epox

yre

sin

wit

ha

CN

SLan

da

carb

oxyl

icac

id

Cu

rab

leep

oxy

resi

nco

atin

gco

mp

osi

tio

nh

avin

ga

low

visc

osi

tyw

hic

his

use

fuli

nco

atin

gap

pli

cati

on

s.

WO

2015

0543

60A

120

13G

ran

tin

US

Car

do

lite

Co

rp

CN

SL-b

ased

hyd

roca

rbo

nre

sin

s,p

rep

arat

ion

and

use

sth

ereo

f

Res

ins

mad

efr

om

CN

SLan

dvi

nyl

hyd

roca

rbo

ns

(e.g

.al

ph

a-m

eth

ylst

yren

e)u

sin

ga

Lew

isac

idas

anac

idca

taly

st(e

.g.

par

a-to

luen

esu

lfo

nic

acid

).

Ad

dit

ives

asn

on

-rea

ctiv

ed

ilu

ents

for

solv

ent-

free

coat

ing

form

ula

tio

ns;

tack

ifier

sfo

rst

ruct

ura

lad

hes

ive,

pre

ssu

rese

nsi

tive

and

ho

t-m

elta

dh

esiv

es;s

tab

iliz

ers

for

lub

rica

nts

,fu

elan

dp

oly

mer

form

ula

tio

ns;

pla

stic

izer

sfo

rth

erm

op

last

icp

oly

mer

san

dp

roce

ssin

gai

dfo

rru

bb

erco

mp

ou

nd

ing

and

stab

iliz

ers

for

resp

ecti

veru

bb

erar

tifa

cts.

Th

ese

resi

ns

are

also

valu

able

pre

curs

ors

for

the

man

ufa

ctu

reo

fep

oxy

resi

ns

and

po

lyo

lsfo

rco

atin

g,ad

hes

ive

and

com

po

site

form

ula

tio

ns

exh

ibit

ing

amel

iora

ted

per

form

ance

inw

ater

rep

elle

ncy

,an

ti-c

orr

osi

on

,an

dfa

sth

ard

nes

sd

evel

op

men

tdu

rin

gcu

re.

31

Figure 3.16: An epoxy novolac resin structure sold by Cardolite

spin-off of Bayer. Also Dow Chemical Company (now DowDuPont after merging with

DuPont in 2017) in which was an American company that worked with the plastic busi-

nesses.

Finally, very few studies can be found regarding the use of anacardic acid as a

monomer. Even when the mixture of CNSL is used in polymerizations, normally it

is the technical CNSL – the one with a small percentage of anacardic acid. CHE-

LIKANI et al. (2009) polymerized it through a free radical oxidative coupling process

using soybean peroxidase for 24 hours (Figure 3.17). It was used a redox mediator

(phenothiazine-10-propionic acid, PPA) and hydrogen peroxide (as oxidant agent) in

methanol or propanol in pH 7. 1H-NMR and FTIR showed that the polymer had a mix-

ture of oxyphenylene and phenylene meres with some decarboxylation that took place

during the polymerization. Therefore, the polymerization did not happen on any un-

saturation of the pentadecyl alkyl side chain, being very selective.

Still, when methanol was used as solvent, the average molecular weight obtained

was 5000 Da with 43% of yield. When propanol was used, the molecular weight was

3900 Da with 61% of yield. Also, the polymers were crosslinked, films were produced

and submitted to biofouling tests for gram-negative or gram-positive bacteria. It was

more effective against gram-negative bacteria. In a similar study, YOON and KIM

(2009) showed that both polymeric films can reduce the number of adherent cells to

the surface, but they do not have antimicrobial activity.

PHILIP et al. (2007) aiming to obtain molecularly imprinted polymers (MIPs),

utilized the anacardic acid as monomer in a toluene solution in presence of azo-

bisisobutyronitrile (AIBN) and ethylene glycol dimethacrylate (EGDMA) or divinyl-

benzene (DVB) as crosslinking agent in (60°C ). To synthesize the MIP, it was added

(R,S)-propranolol hydrochloride, (R)-propranolol hydrochloride or (S)-propranolol

hydrochloride as template. The copolymerization was confirmed by FTIR. However,

they also chemically modified the anacardic acid by acylation and methacrylation, ob-

taining anacardanyl acrylate and anacardanyl methacrylate respectively. New copoly-

merizations were made in the same system with these new monomers.

Inspired by the work of Philip et al., KAEWCHADA et al. (2012) copolymerized anac-

32

Table 3.5: Patents using cardanol to produce polyurethanes

Co

de

Pri

ori

tyYe

arSt

atu

sC

urr

ent

Ass

ign

eeN

ame

Des

crip

tio

nA

pp

lica

tio

n

WO

2006

0036

68A

120

04G

ran

tin

EP

and

DE

Co

un

cil

Of

Scie

nti

fic

An

dIn

-d

ust

rial

Res

earc

h

Pro

cess

for

pre

par

ing

po

lyu

reth

ane

po

lyo

lan

dri

gid

foam

sth

eref

rom

fro

mca

rdan

ol

Pro

du

ctio

no

fa)

card

ano

lpo

lyo

lan

db

)p

oly

ure

than

eT

her

mal

insu

lati

on

US2

0080

1396

85A

120

06G

ran

tin

US

and

CN

Cov

estr

oLL

C

Nov

elp

oly

eth

erp

oly

ols

bas

edo

nca

shew

nu

tsh

elll

iqu

id,a

pro

cess

for

the

pro

du

ctio

no

fth

ese

po

lyet

her

po

lyo

ls,

flex

ible

foam

sp

rod

uce

dfr

om

thes

ep

oly

eth

erp

oly

ols

,an

da

pro

cess

for

the

pro

du

ctio

no

fth

ese

foam

s

Pro

du

ctio

no

fa)

po

lyet

her

po

lyo

lsb

yth

eal

koxy

lati

on

ofC

NSL

and

b)

pro

du

ctio

no

ffl

exib

lep

oly

ure

than

efo

ams

Ind

ust

rial

and

con

sum

erap

pli

cati

on

ssu

chas

seat

s,d

ash

bo

ard

and

oth

erca

bin

inte

rio

rp

arts

ofa

uto

mo

bil

es

WO

2011

0301

97A

220

09G

ran

tin

US,

EP,

CN

,DK

,E

S,F

R

Co

atex

SAS

Car

dan

ol-

bas

edas

soci

ativ

ep

oly

ure

than

es,c

orr

esp

on

din

gas

soci

ativ

eth

icke

ner

san

du

ses

ther

eof

Syn

thes

iso

faw

ater

-so

lub

lep

oly

ure

than

ere

sult

ing

fro

mth

eco

nd

ensa

tio

no

fan

oxya

lkyl

ated

card

ano

l,a

po

lyal

kyle

ne

glyc

ola

nd

po

lyis

ocy

anat

e

Th

icke

nin

gag

ents

for

aqu

eou

sfo

rmu

lati

on

s.

WO

2011

0301

97A

220

09G

ran

tin

US,

EP,

CN

,DK

,E

San

dF

R

Co

atex

SAS

Ass

oci

ativ

ep

oly

ure

than

esb

ased

on

card

ano

l,co

rres

po

nd

ing

asso

ciat

ive

thic

ken

ers

and

use

sth

ereo

f

Syn

thes

iso

fp

oly

ure

than

esfr

om

po

lyal

kyle

ne

glyc

ol,

po

lyis

ocy

anat

esan

do

pti

on

ally

oxye

thyl

ated

card

ano

l.

Th

icke

nin

gag

ents

33

Co

de

Pri

ori

tyYe

arSt

atu

sC

urr

ent

Ass

ign

eeN

ame

Des

crip

tio

nA

pp

lica

tio

n

WO

2015

0779

44A

120

13G

ran

tin

US,

JP

Dow

Glo

bal

Tech

-n

olo

gies

Llc

Car

dan

olm

od

ified

epox

yp

oly

ol

Syn

thes

iso

fap

oly

ure

than

ere

sin

by

mix

ing

card

ano

l,a

card

ano

l-m

od

ified

epox

yp

oly

ol,

and

anis

ocy

anat

eco

mp

on

ent

Co

mp

osi

tes

form

edb

yfi

lam

entw

ind

ing

and

coat

ings

WO

2014

2097

04A

120

13G

ran

tin

US,

EP,

CN

Bay

ero

rC

oves

tro

inU

San

dE

P

Po

lyu

reth

ane

pu

ltru

sio

nfo

rmu

lati

on

sfo

rth

ep

rod

uct

ion

ofa

rtic

les

wit

him

pro

ved

coat

ing

adh

esio

nan

dar

ticl

esp

rod

uce

dth

eref

rom

Pro

du

ctio

no

fap

oly

ure

than

eco

mp

osi

teb

ya

reac

tio

no

fap

oly

iso

cyan

ate

and

aca

shew

oil

-bas

edp

oly

eth

erp

oly

ol

Use

inp

ult

rusi

on

pro

cess

es

WO

2015

0779

44A

120

13G

ran

tin

US

and

JP

Dow

Glo

bal

Tech

-n

olo

gies

Llc

Car

dan

olm

od

ified

epox

yp

oly

ol

Form

atio

no

fap

oly

ure

than

ere

sin

syst

emth

atin

clu

des

anis

ocy

anat

e-re

acti

veco

mp

on

entt

hat

has

afi

rstc

ard

ano

lco

mp

on

enta

nd

aca

rdan

ol-

mo

difi

edep

oxy

po

lyo

l

Co

mp

osi

tes

form

edb

yfi

lam

entw

ind

ing

and

coat

ings

.

WO

2015

1533

99A

120

14A

pp

lica

tio

n(2

015)

Dow

Glo

bal

Tech

-n

olo

gies

Llc

Nat

ura

loil

der

ived

blo

cked

pre

po

lym

ers

and

acry

lic

pla

stis

olc

om

po

siti

on

sh

avin

gth

eb

lock

edp

rep

oly

mer

s

Use

ofC

NSL

asa

blo

ckin

gag

entt

ofo

rma

pre

po

lym

er.A

nis

ocy

anat

ete

rmin

ated

pre

po

lym

eris

also

use

d.

Sub

stit

ute

for

PV

C-

bas

edp

last

iso

ls

WO

2017

0444

02A

120

15A

pp

lica

tio

n(2

017)

Dow

Glo

bal

Tech

-n

olo

gies

Llc

Blo

cked

po

lyu

reth

ane

tou

ghen

ers

for

epox

yad

hes

ives

Use

ofc

ard

ano

las

cap

pin

ggr

ou

pfo

rp

rod

uct

ion

ofa

nep

oxy

adh

esiv

esas

tou

ghen

er

Tou

ghen

ers

for

epox

yad

hes

ives

34

Co

de

Pri

ori

tyYe

arSt

atu

sC

urr

ent

Ass

ign

eeN

ame

Des

crip

tio

nA

pp

licat

ion

WO

2016

1678

68A

120

15A

pp

lica

tio

n(2

016)

3MIn

no

-va

tive

Pro

per

-ti

esC

om

-p

any

Fir

esp

read

pre

ven

tio

nm

emb

eran

dfi

resp

read

sup

pre

ssio

nm

eth

od

Car

dan

oli

sp

rese

nta

sa

po

lyo

lto

pro

du

cea

po

lyu

reth

ane

bin

der

Stru

ctu

res

use

din

afi

rep

reve

nti

on

faci

lity

WO

2005

0829

67A

120

04G

ran

tin

US,

EP,

CN

,DE

,E

SC

oves

tro

Hyd

rop

ho

bic

low

-vis

cou

sp

oly

ols

Th

ein

ven

tio

nre

late

sto

new

hyd

rop

ho

bic

po

lyo

lso

flow

visc

osi

ty,

toa

pro

cess

for

pre

par

ing

them

and

toso

lven

t-fr

eeb

ind

erm

ixtu

res

bas

edth

ereo

n

Pri

mer

so

fflo

or

coat

ings

.

US2

0080

1396

85A

120

06G

ran

tin

US

and

CN

Cov

estr

o

Nov

elp

oly

eth

erp

oly

ols

bas

edo

nca

shew

nu

tsh

elll

iqu

id,a

pro

cess

for

the

pro

du

ctio

no

fth

ese

po

lyet

her

po

lyo

ls,

flex

ible

foam

sp

rod

uce

dfr

om

thes

ep

oly

eth

erp

oly

ols

,an

da

pro

cess

for

the

pro

du

ctio

no

fth

ese

foam

s

Pro

du

ctio

no

fpo

lyet

her

po

lyo

lsw

hic

har

ep

rep

ared

by

alko

xyla

tio

no

fCN

SL.A

nd

also

the

pro

du

ctio

no

fflex

ible

po

lyu

reth

ane

foam

sp

rep

ared

fro

mth

ese

po

lyo

ls

Co

atin

gs,a

dh

esiv

es,

seal

ants

or

mo

ldin

gco

mp

ou

nd

s.

WO

2016

1830

73A

120

15G

ran

tin

US

Cov

estr

oF

ilam

entw

ind

ing

pro

cess

esu

sin

gp

oly

ure

than

ere

sin

san

dsy

stem

sfo

rm

akin

gco

mp

osi

tes

Th

ep

rese

nti

nve

nti

on

isd

irec

ted

to,a

mo

ng

oth

erth

ings

,fila

men

twin

din

gp

roce

sses

that

use

alo

wvi

sco

sity

po

lyu

reth

ane-

form

ing

com

po

siti

on

,sys

tem

sfo

rm

akin

ga

com

po

site

and

the

com

po

site

sso

mad

e.

Pip

e,p

ress

ure

bo

ttle

san

dta

nks

,sto

rage

con

tain

ers,

amo

ng

oth

erth

ings

35

Figure 3.17: Reaction scheme of anacardic acid polymerization catalyzed byperoxidase (adapted from CHELIKANI et al. (2009))

ardic acid with divinylbenzene (DVB) utilizing benzoyl peroxide (BPO) and toluene for

48h in (60°C ). As the idea was to synthesize a MIP, they added rice brain oil as template.

At the end, a rigid mass of polymer was obtained.

Figure 3.18: Synthesis of anacardanyl acrylate (a) and anacardanyl methacrylate (b)(adapted from PHILIP et al. (2007))

36

3.4 Polymer Functionalization

Functionalization of polymers means the introduction of desired functional groups

in its structure to create specific chemical, physical, biological, pharmacological, or

other properties. The functionalization is important to enable further reactions, like

bioconjugation, or to enable selective binding of particular species; transport of drugs;

catalytic activity and other applications (PICHOT (2004), TIAN et al. (2012)). These

interactions happen due to physical adsorption by hydrophobic, hydrophilic or elec-

trostatic interactions and hydrogen and Van der Walls bounds. But, it can also be done

due to covalent coupling by the presence of chemical groups in the biomolecules like

amine, thiol, carboxyl and hydroxyl groups.

The non-specific (physical) adsorption is the weakest one since the molecule can

be easily desorbed by changes in medium conditions like pH, temperature and ionic

strength (WELSCH et al., 2013). The conjugation by means of covalent attachment is

attractive since it provides a route for the formation if a biofunctional layer, being the

molecule irreversibly and stable bounded (RAMOS, 2018).

Therefore, as defined by IUPAC, a functional polymer is one that:

• Bears specified chemical groups or;

• Has specified physical, chemical, biological, pharmacological, or other uses

which depend on specific chemical groups.

PICHOT (2004) specified the most used methods for fabrication of functionalized par-

ticles. Two of them are:

• Heterogenous polymerization: usually in aqueous medium and initialized by

free-radical polymerization. This enables the synthesis of dispersions of parti-

cles with a wide size range depending on the technique used.

• Self-organization of block or grafted copolymers by means of molecular or elec-

trostatic interactions.

HERMANSON (2013) confirms that reactive or functional groups can be created

at the surface of particles for the subsequent coupling of binders with the aid of the

monomer mixture in the polymerization. Moreover, multiple factors, such as size and

curvature of the particle surface and the chemical nature of the surface and the parti-

cle’s nucleus, determine the arrangement of the biomolecules MAHON et al. (2012).

PEIXOTO (2013) studied the functionalization of poly(methyl methacrylate) parti-

cles by copolymerizing it with acrylic acid, methacrylic acid, 2-dimethylamino ethyl

methacrylate and 2-hydroxyethyl methacrylate. She observed that the co-monomers

were concentrated in the particle surface due to their affinity be the aqueous medium,

37

functionalizing the particles. Absorption tests with bovine serum albumin (BSA) indi-

cated that from 15 to 35% of the protein was immobilized.

WAY (2017) studied the click chemistry. Successful reactions were the thiol-ene re-

action between styrene and 3-mercaptopropionic acid and between maleimide and

functionalized polymers produced by RAFT polymerization. She detected that the use

of copolymer systems can lead to cell uptakes of nanoparticles of bioconjugated poly-

mers up to 8 times higher, constituting promising alternatives for use in targeted drug

delivery application.

Thus, in this study, suspension polymerization, in which is a polymerization

in heterogeneous medium, is used to functionalize polymer particles. Besides the

use of cardanol, employment of anacardic acid is interesting due to the carboxyl

group as it can react with an amine group of the biomolecule, forming a pep-

tide bound. This can be easily done using a carbodiimide, such as 1-ethyl-3-(3-

dimethylaminopropyl)carbodiimide (EDC), in parallel with N-hydroxysuccinimide

(NHS). The formation of a amide group can be detected by FTIR (RAMOS, 2018).

3.5 Concluding Remarks

Based on the literature review, there isn’t a detailed study about the free-radical

polymerization of cardanol or acid anacardic. GALVÃO (2016) started investigating

the copolymerization kinetics of cardanol with styrene in homogenous and heteroge-

neous systems, confirming that cardanol can be incorporated into the polymeric chain

through the one of the unsaturations. Also, cardanol seemed to be a great stabilizer for

suspension polymerization. However, it also inhibits the polymerization, reducing the

overall polymerization rate. Another work was the investigation of copolymerization

in suspension by ?. She produced a core-shell polymer particle containing cardanol

by semibatch combined suspension-emulsion polymerization aiming enzyme immo-

bilization with good characteristics. However, she was not successful copolymerizing

cardanol and styrene in usual suspension conditions.

Regarding anacardic acid, anything besides the synthesis of MIPs (cited in sec-

tion 3.3) have been reported about its use as monomer in free radical polymerization.

Therefore, it is exciting to see how it is going to behavior and the differences between its

kinetics and those of cardanol as the only thing that they have different is the addition

of the carboxyl group.

Moreover, the use of any of these renewable monomers for polymer functional-

ization is very interesting as, if attached successfully to the particle surfaces, they will

introduce the long aliphatic side chain – perhaps with one or two unsaturations, en-

abling further polymerizations -, and the reactive groups hydroxyl and carboxyl.

38

Kinetic Study of Bulk Copolymerization

1st Section

39

Chapter 4

Bulk Copolimerization: Methodology

4.1 Materials

Cardanol (Taian Health Chemical, minimum purity of de 99.5% m/m, Tai’an,

China), natural CNSL (donated by Embrapa, Brazilian Agricultural Research Corpora-

tion), acrylic acid (AA) (VETEC, Rio de Janeiro – Brazil, minimum purit of 99%), methyl

methacrylate (MMA) (VETEC, Rio de Janeiro – Brazil, minimum purity of 99.5%),

styrene (STy) (Sigma-Aldrich, Rio de Janeiro – Brazil, minimum purity 99%) and vinyl

acetate (VAc) were used as monomers. Hydroquinone (Tedia, Rio de Janeiro – Brazil,

minimum purity of 99%) was used to cease the polymerization in a solution of ethanol

(VETEC, Rio de Janeiro – Brazil, minimum purity 99.8%). Benzoyl peroxide (BPO)

(VETEC, Rio de Janeiro – Brazil, minimum purity of 99%, containing 25% of humidity)

was used as a initiator for the polymerizations. Tetrahydrofuran (THF) (Tedia, Rio de

Janeiro – Brazil, minimum purity of 99%) was used as a solvent for gel permeation chro-

matography analyzes. Deuterated chloroform (Cambridge Isotope Laboratoies Inc.,

United States, minimum purity 99.8%) was used for NMR analysis.

4.2 Set up for bulk polymerizaion

Reactions were performed in glass tubes. The reaction mixture was purged with ni-

trogen for one minute and placed on a silicone bath kept under constant stirring at the

desired reaction temperature. The sample on time zero was taken when the tempera-

ture of the reaction mixture (measured by a thermometer inside of the tube) reached

the desired one. At specified sampling times, tubes were removed, cooled quickly and

the reaction was ceased by adding a solution of hydroquinone (1wt %).

40

Figure 4.1: Set Up for Bulk Copolymerization

4.3 Kinetic Experiments

Every reaction mixture (independently of the co-monomer) used 1% w/w of ben-

zoyl peroxide (BPO). For each system, only the concentration of cardanol or natural

CNSL (0, 2.5, 5, 10% w/w) and the temperature changed (85 - 110°C ). Depending on

the co-monomer and the concentration of cardanol/CNSL, the reaction time changed.

Moreover, if the system had a smaller reactivity, longer sampling times were taken. Ta-

ble 4.1 summarizes the experiments done. In total, at least 38 bulk polymerizations

were done. X corresponds to cardanol Y corresponds to methyl methacrylate (MMA),

styrene (Sty), vinyl acetate (VAc) or acrylic acid (AA). And T corresponds to the tem-

perature used. The same reactions with MMA were repeated using natural CNSL. The

obtained polymer samples were used for purposes of gravimetric analyzes (for deter-

41

Table 4.1: Kinetic Experiments Done

ExperimentsMonomers (wt %)

Initiator (wt %) TemperatureCardanol Y

X.Y.T.1 0.0 99.0

1 85 and 110X.Y.T.2 2.5 96.5X.Y.T.3 5.0 94.0X.Y.T.4 10.0 89.0

mination of overall monomer conversions) and polymer characterizations.

4.4 Gravimetric Analysis

Monomer conversion was obtained by gravimetric analysis. This method is based

on the theory that every monomer reacted becomes part of a polymer, admitting that

there is not formation of sub-products. For this, the glass tube is weighted before the

addition of the sample, and the mass added of hydroquinone added in each tube is

known. The tubes containing the polymer were dried in the recirculating followed by

a vacuum oven at the ambient temperature, and weighted several times until the mass

became constant. The equations used were the following:

mmonomer = mi ni t i al −mtube −mBPO (4.1)

mpol ymer = mdr y −mtube −mBPO −msol .hydr oqui none ∗mhydr oqui none (4.2)

conver si on (%) = mpol ymer

mmonomer∗100% (4.3)

For mBPO , it is considered that it corresponded to 1% w/w of the mmonomer ), as this

was the amount added during the preparation of the reaction mixture. Csol .hydr o was

0.01 as the reactions were stopped using an alcoholic solution of hydroquinone (1%

w/w).

42

4.5 Characterizations

4.5.1 Gel Permeation Chromatography (GPC)

Gel permeation chromatography (GPC) was used to determine the molecular

weight distribution and average molecular weight of polymers. The sample is dis-

solved on a solvent (tetrahydrofuran, THF) and injected in a porous column filled with

a polymer. Smaller molecules take longer times to leave the column as they penetrate

the interior of the gel phase pores. Bigger molecules are detected first at the end be-

cause they do not percolate the pores. As the column is calibrated with a standard

with known molecular weight, it is possible to the determine molecular weight of the

polymer to be analyzed depending on its retention time inside the column,

The analysis were done in EngePOL using a GPC chromatograph (Viscotek GPC-

max, Malvern, Malvern, United Kingdom) equipped with a Shodex GPC HFIP-803,

Shodex GPC HFIP-804 and Shodex GPC HFIP-805 (Showa Denko, Tokyo, Japan) plus

Viscotek refractive detector model VE3580 and Viscotek model 2500 UV detector, set

to 255 ηm and equipped with deuterium lamp. The equipment was calibrated with

polystyrene standards having molecular weight ranging from 500 and 1.8 ∗ 106 Da.

Sample preparation consisted of solubilizing about 5 mg of the polymer in about 5

mL of THF. The solution obtained was filtered through a Teflon filter with 0.45 µm

pores. This filtered solution was then injected into the chromatograph. The temper-

ature of the columns was 40°C . It was injected 200 µl with a volumetric flow rate of 1

mL/minute.

GPC data were also used to analyze the conversion for homopolymers of cardanol

or natural cashew nut shell liquid. The monomer spectrum was used to compare to

those of the copolymers. When a should appeared in the direction of higher molec-

ular weight, indicating formation of oligomers or small polymers, the mass fraction

reported by the GPC of that fraction is said to correspond to the conversion. However,

as cardanol or natural cashew nut shell liquid are not purified, they also contain higher

molecules, creating a little shoulder. So, this mass fraction was also used subtracting

the value of the copolymer from this one.

4.5.2 Nuclear Magnetic Resonance (NMR)

The chemical structure of a material can be detected by nuclear magnetic reso-

nance. This technique is based that an energy absorption occurs at discrete frequen-

cies when a compound is placed in a strong magnetic field and irradiated with a radio-

frequency signal. The frequency is very sensitive to the atomic environment of the

proton.

Samples for 1H-NMR were prepared solubilizing 15mg in 0.8 mL of deuterated

43

chloroform. Analysis were done in the Chemistry Institute of Federal University of

Rio de Janeiro at room temperature in a Varian Mercury, model VX300, 300.2 MHz

of frequency and 5 mm of probe. Other analysis were done in LAMAR - Laboratório

Multiusuários de Análises por RMN in the Centro de Ciências da Saúde of Universi-

dade Federal do Rio de Janeiro. In addition, for same samples (when announced in

the discussion), the dry polymer material was redissolved in toluene, reprecipitated in

methanol, filtered and dried a second time in drying oven under vacuum at 50°C until

it reached constant weight again.

4.5.3 Fourier-transform infrared spectroscopy (FTIR)

Fourier-transform infrared spectroscopy (FTIR) is based on the fact that atoms vi-

brate in specific frequencies of the electromagnetic spectra depending on their chem-

ical structure. The vibration changes in relation to the atom neighborhood. This tech-

nique is used to detected chemical compounds and detect the appearance or disap-

pearance of functional groups.

FTIR analysis were done in the Chemistry Institute of Federal University of Rio de

Janeiro in the range of 4000 – 400 cm−1, with resolution of 4 cm−1, in a Thermo equip-

ment, model Nicolet 6700 equipped with a ATR Smart Orbit in reflectance mode, that

ensures application of a constant pressure in the samples and dispensing the prepara-

tion of KBr pellets.

4.5.4 Thermogravimetric analysis (TGA)

Thermogravimetric analysis (TGA) is a destructive analysis used to determine the

thermal stability of polymeric samples in an inert or not environment. It can also be

used to evaluate presence of solvent, residual monomer or any other volatile com-

pound present. During the analysis, temperature increases in a predefined rate and

the mass of the sample decreases along the whole time. Then, the equipment gives a

mass graph (thermogram) as a function of the temperature.

Thus, analysis were done in EngePOL, laboratory of polymerization of Federal Uni-

versity of Rio de Janeiro. Each sample had roughly 10 mg. It was used a Perkin Elmer

STA 600 equipment ((Massachusetts, United States) with temperature ranging from 30

to 700°C with a heating rate of 10°C /min under inert nitrogen atmosphere.

4.5.5 Differential Scanning Calorimetry (DSC)

For Differential Scanning Calorimetry (DSC) analyzes, polymer samples of 5 mg

were placed in aluminum pans under nitrogen atmosphere. Analyzes were performed

in the range from 0 to 280°C at heating rate of 10°C .min−1. As usual, the second heating

44

scan was used to record thermal transition temperatures in order to erase the thermal

history of the sample.

45

Chapter 5

Bulk Copolymerization: Experimental

Results

5.1 Introduction

Kinetics studies of copolymerization of cardanol with a vinyl monomer had never

been reported before the work of GALVÃO (2016). She studied the copolymerization of

cardanol with styrene in bulk, suspension and emulsion. As previously detected, sev-

eral problems may arise when cardanol is used as a co-monomer. Firstly, there may

be steric hindrance due to the long side chain. Secondly, the hydroxyl group might in-

teract with another compound. Thirdly, quite stables radicals might be formed due to

internal unsaturations, retarding the polymerization (RODRIGUES et al. (2006)). Fig-

ure 5.1 shows the reactivity sites of cardanol.

Figure 5.1: Reactive sites of cardanol

Thus, the bulk copolymerization of cardanol with methyl methacrylate, vinyl ac-

etate, acrylic acid and styrene is studied in this first section. These monomers have

different functional groups, relativities and applications. Thus, it is interested to see if

copolymerization is possible in these systems.

However, free radical polymerization of anacardic acid have has never been clearly

46

reported. Thus, copolymerization of natural CNSL with methyl methacrylate is also

studied. This copolymerization can be even more challenging due to the presence of

a carboxyl group. Theoretically, as it is expected that the polymerization will occur in

one of the double bonds, the reactivity ratios should be the same than those obtained

when cardanol was used as a comonomer. Moreover, as a kinetic model is later pro-

posed, homopolymerizations of cardanol and anacardic acid were done to figure out

their own kinetic rates.

5.2 Homopolymerization of cardanol and anacardic acid

Cardanol and anacardic acid are very similar phenolic compounds - both have 15

carbons in their side chain with 0 to 3 unsaturations, but anacardic acid has a car-

boxyl group in the ortho position in relation to the hydroxyl group. Anacardic acid is

found in high weight fractions in natural cashew nut shell liquid (CNSL) - over 70wt %.

CNSL has to be natural or the extraction procedures that obtains technical CNSL would

decarboxylate anacardic acid. Although anacardic acid can be purified from natural

cashew nut sheel liquid (PARAMASHIVAPPA et al. (2001)), this increases the cost of the

reagent and, consequently, of the homo- and co-polymerizations performed with this

compound. Therefore to evaluate the effect of functional groups of CNSL in the poly-

merizations, it was used natural CNSL.

The homopolymerization of cardanol and natural CNSL were done using 1wt % of

BPO at 85 and 110°C and using the same procedure for copolymerization. However as,

even at low pressure, their boiling point is too high (225°C for 0.013 atm) (HARVEY) it

cannot be removed at an air or vacuum oven as it could gradate the polymer. Therefore,

the conversion for this polymerizations could only be obtained from gel permeation

chromatography (GPC) data. The conversions for cardanol are presented in Table 5.1.

Although the conversion data has high variations and the values are quite small, it

Table 5.1: Conversion for homopolymerization of cardanol with 1% w/w of BPO

85 °C 110 °CReaction Time (h) Conversion (%) Reaction Time (h) Conversion (%)

2

< 1

1 3.484 3 6.767 4 11.52

8.5 5 6.949 5.47 6 3.62

10.5 < 1 8 7.57

indicates that some oligomerization took place. Figure 5.2 show GPC chromatograms

for these reactions at 85 and 110°C . Oligomers of over 3100 Da were obtained at both

temperatures. There are little tiny shoulders for cardanol on the right (corresponding

47

Table 5.2: Conversion for homopolymerization of natural cashew nut shell liquid with1% w/w of BPO

85 °C 110 °C

Reaction Time (h) Conversion (%) Reaction Time (h) Conversion (%)2 3.00 1 10.254 2.90 3 9.257 3.40 5 9.45

8.5 3.90 6 5.659 2.90 8 5.85

10.5 11.90 - - - -

to lower molecular weight) that may be impurities of the reagent. This also appears for

natural CNSL as can be seen in Figure 5.3. Conversion for CNSL is on Table 5.2.

(a) (b)

Figure 5.2: GPC of cardanol oligomers at 85°C (a) and 110°C (b)

48

(a) (b)

Figure 5.3: GPC of natural CNSL oligomers at 85°C (a) and 110°C (b)

Nuclear magnetic resonance (1H NMR) analysis for cardanol was done to see the

chemical bonds present and detect a possible reaction mechanism. However, the

spectrum is quite complex as many signals appears (GALVÃO (2016), FERREIRA et al.

(2015)). Thus, Figure 5.4 shows the 1H NMR for cardanol with the peaks identified.

To figure out which reaction may have happened for the formation of higher molec-

ular weight compounds, a peak at δ = 2.5 ppm (f ) is used as a standard to compare

the intensity between the peaks and, posteriorly, between spectrums. Hydrogens (2H)

of peak f were chosen because they are present in every cardanol molecule, being it

alone (as monomer) or as part of a polymeric structure. Figure 5.5 presents the spec-

trum for 4h of "polymerization" at 110°C . For most of the copolymerizations, the reac-

tion is not stopped before this time. Comparing the integration values, any significant

change can be detected. But for 8h of reaction at 110°C (Figure 5.6), an increase in

peaks c, e, g and j can be seen. However, almost none change can be seen for peaks

b or d in which correspond to the terminal double bond. This is quite interesting as

it seems like the proportion of saturated cardanol decreased, while the mono and di-

ene increased, and none reaction happened in the terminal double bond. Peaks a’s did

not change, showing that a mechanism similar to the oxidative or enzymatic polymer-

ization of cardanol did not happen. What can be also detected is the appearance of a

very broad peak within the range of 3 to 5 ppm. This area corresponds to the one that

R-O-R (ether) peaks appear. Therefore, some reaction with the hydroxyl group may

have happened. This analysis was repeated to make sure the spectrum was correct,

but the result became the same. Thus, although there were formation of higher molec-

ular weights molecules, the mechanism is still unknown looking only the 1H NMR.

49

0.00.51.01.52.02.53.03.54.04.55.05.56.06.57.07.58.0chemical shift (ppm)

87.7

9

677.

99

115.

78

164.

41

100.

00

113.

49

55.3

7

171.

81

19.9

7

91.9

347

.24

46.5

77.

26 C

DCl

3

b

cd

ef

g h

i

j

6.6 6.6

7.146.7

h

f i

i

i

i

g

c

c cce

e

b

d

f i

h i

i

i

i

i

i

i

i

i

i

i

j

OH

OH

aa

a

Figure 5.4: 1H NMR of cardanol

0.00.51.01.52.02.53.03.54.04.55.05.56.06.57.07.5chemical shift (ppm)

88.1

7

674.

61

118.

27

167.

04

100.

00

113.

54

48.6

4

169.

64

19.7

3

92.0

447

.14

46.4

57.

26 C

DCl

3

Figure 5.5: 1H NMR of the reaction mixture for synthesize polycardanol by radicalpolymerization using BPO after 4 hours

50

0.51.01.52.02.53.03.54.04.55.05.56.06.57.07.58.0chemical shift (ppm)

103.

80

813.

48

138.

60

199.

98

100.

00

147.

69

55.3

6

215.

81

25.5

1

91.2

754

.19

45.8

87.

26 C

DCl

3

a

a

a

b

c

d

e

f

g

h

i

j j

Figure 5.6: 1H NMR of cardanol after 8h for reaction with BPO

51

For natural CNSL, the composition of the mixture cannot be known only looking at

the 1H NMR as it is even more complex than the one for cardanol - although the highest

composition is for anacardic acid. The 1H NMR is presented at Figure 5.7 with a similar

classification of the peaks that was made for cardanol. However, for 8h of reaction at

110°C , a reference peak could not be choosen as the intensities of every peaks seem to

have changed, mainly those with the arrow pointing (Figure 5.8. Also, the peak around

4.0 and 4.9 ppm has also appeared.


4.81

36.5

2

6.49

9.68

1.81

12.6

5

2.85

10.8

8

1.80

2.97

6.88

0.24

2.43

7.26

CD

Cl3

j'

i'

h'

g'

f'

a'a'

a'e'

b'

c'

d'

Figure 5.7: 1H NMR of CNSL

52


4.88

37.5

1

7.08

9.95

4.44

10.2

7

3.11

10.7

5

1.39

1.80

6.57

1.33

0.92

7.26

CD

Cl3


4.81

36.5

26.

499.

681.

8112

.65

2.85

10.8

81.

802.

97

6.88

0.24

2.43

7.26

CD

Cl3

j'

i'

h'

g'

f'

a'a'e'a'

b'

c'

d'

CNSL

PolyCNSL (8h)

Figure 5.8: 1H NMR of CNSL after 8h of reaction

Both 1H NMR for cardanol and CNSL can be used to obtain the percentage of the

triene component(s) comparing correlating peaks f or h with peaks b or d. Therefore,

44.41±7.8 % corresponds to cardanol; and 49.69±6.0 % to natural CNSL.

FTIR analysis was also performed to detected functional groups present and which

group could be the responsible for the reactions. FTIR of cardanol reaction samples at

110°C (Figure 5.9), for 8h the peaks have reduced their intensity in relation to the spec-

trum of 4h. These reduction happened mainly at 3300, 1250 and 1150 cm−1, which

are related to the hydroxyl moiety. Therefore, reactions with this group may have hap-

pened. Moreover, peaks of double bonds also have reduced like the one for the ter-

minal double bond at 3050−1 and those at the region 1000 - 600−1 that corresponds

to internals and terminal unsaturations. The FTIR spectrums are similar to the one of

pure cardanol obtained by YADAV and SRIVASTAVA (2007b). Therefore, cardanol does

have multiple reactive sites - hydroxyl, internal and terminal unsaturations. All of them

can react in the presence of a polymerization initiator (such as BPO) in different mag-

nitudes. If one consideres that a possible polymer of cardanol only forms by reaction of

the unsaturations, the polymer conversion will never correspond to 100% of cardanol

conversion as some of the reagent is inhibit by the phenolic ring.

The reaction with the hydroxyl moiety happens because phenolic compounds are

donor radical inhibitors by giving an hydrogen atom (the one from the hydroxyl)

(Fig. 5.10). This reaction is more rapid than the transfer reaction on an alkyl radical

(LARTIGUE-PEYROU (1996)). Moreover, the slightest trace of oxygen will very rapidly

53

Figure 5.9: FTIR for cardanol after 4 and 8 hours of reaction

form a peroxy radical from an alkyl radical, reforcing the need for nitrogen purging

before the reaction.

Figure 5.10: Inhibition Mechanism for Cardanol

Moreover, although unsaturated C=C bonds permits radical addition for subse-

quently polymerization; the radical formed is stable retarding and trapping free rad-

icals. Therefore, a way to improve molecule reactivity could be by activating, function-

54

alizating or conjugating them. Acrylation and epoxidation some options. Cardolite NC

513 or NC 2513 (from Cardolite, Inc), for example, are epoxidized cardanol (ECL) used

as reactant (CHISHOLM et al.).

Thermogravimetric analysis were the last characterization performed to see the

thermal stability of the studies compounds. CNSL has a high thermal stability than

cardanol. They start to degrate at the same temperature - close to 150°C . However,

while cardanol only has one degradation stage, CNSL has three. The first stage of

CNSL degradation may be related with the descarboxylation of anacardic acid which

happens at temperatures above 180°C and transforms it into cardanol while liberat-

ing carbon dioxide. The second stage of degradation might be cardanol that is present

in small amounts in natural CNSL (< 10wt %), but also some amount had come from

anacardic acid. Thus, while cardanol is fully degradated at 325°C , natural CNSL is still

degrading. The presence of impurities in CNSL may also have degraded or be reason

for the presence of ashes at temperatures over 500°C . The TGA analysis for CNSL is

similar to the obtained by ??.

Figure 5.11: TGA and DTG analysis for cardanol and natural cashew nut shell liquid

5.3 Copolymerization of cardanol and styrene

Polystyrene is a very known polymer, being transparent and used in a variety of

applications, such as food packaging and appliances. Polystyrene also is made into a

55

foam material, called expanded polystyrene (EPS) which is used for thermal insulation

and protective packaging.

Figure 5.12: Chemical structure of polystyrene

Figure 5.13: Photo for synthesized polystyrene or poly(styrene-co-cardanol) at 110°C

Reactions between cardanol and polystyrene were done at 2.5, 5 and 10% w/w at

85 and 110°C using 1 wt % of BPO. Figure 5.13 is a photo of the copolymers produced

at 110°C , showing that the color of material changes with increase in cardanol con-

centration and that copolymerization may have happened. Conversion and molecular

weights for the reaction at 85°C is present in Figure 5.14. Although pure polystyrene

achieved fast conversion in less than three hours, copolymers with cardanol takes a

longer time to achieve high conversions. However, 10 wt % reached a plateau just after

30 minutes. Therefore, cardanol acts retarding and then inhibiting the polymeriza-

tion. At the same time, the molecular weight of the polymer is reduced. This may have

happened because the phenolic group and the internal unsaturations form radicals

very stable, retarding or stopping the polymerization, respectively. Therefore, as radi-

cals were been formed by BPO, cardanol was consuming these living radicals, reducing

their concentration in the reaction medium and translating in a lower PDI. For 10% of

cardanol, there is not data of molecular weight because only oligomers were formed.

For 110°C , a similar behavior happened. In spite of the fact that cardanol still acts

retarding the polymerization, the overall reaction rate was bigger. 10% w/w, for ex-

ample, reached more than 50% of conversion, and its growth did not stop before 330

minutes. Molecular weights were also reduced, but the PDI is higher because of the

56

Figure 5.14: Conversion and molecular weights for poly(Sty-co-C), 85°C

faster decomposition of initiator and the thermal initiation of styrene, which becomes

more relevant at this temperature.

Thermogravimetric analyzes were used to compare the thermal stability of the

copolymers produced with pure polystyrene and cardanol samples. For reactions per-

formed at 85°C (5.16), the addition of cardanol shifts the TGA curve towards the TGA

curve of pure cardanol below 400°C , showing that weaker chemical bonds are formed

or indicating that the decomposition of the side chain of cardanol takes place first.

This suggests the presence of cardanol reduces the thermal stability of the copoly-

mer, which can indicate the occurrence of the copolymerization. However, at higher

temperatures, the presence of cardanol increases the residual thermal stability of the

copolymer, which can also indicate the occurrence of the copolymerization. This can

be due to the addition of the phenolic group of cardanol to the copolymer chains. For

10%, there is a high amount of residual monomer (styrene boils at 145°C ) in the sam-

ple. However, almost 20wt % still correspond to the copolymer in which has the high-

est thermal stability between the other polymers as it fully degraded only at 550°C . In

relation to 85°C , a similar behavior can be noticed in the polymerizations conduced

at 110°C (5.17), although the shifting effect towards the curve of cardanol is less pro-

nounced, maybe due to the slight increase in conversion. However, as 1H −N MR later

57

Figure 5.15: Conversion and molecular weights for poly(Sty-co-C), 110°C

shown, the fraction of cardanol in the copolymer do not differ much from one reaction

temperature to the other.

58

Figure 5.16: TGA and DTG curves for poly(Sty-co-C), 85°C

Figure 5.17: TGA and DTG curves for poly(Sty-co-C), 110°C

59

On the other hand, although some effects on the thermal stability of the ob-

tained copolymers could be noticed with the introduction of cardanol, the observed

effects were very subtle, probably indicating the low incorporation of cardanol into

the copolymer chains. For this reason, DSC analyzes were carried out, as shown in

Table 5.3. The obtained results show the significant reduction of the glass transition

temperature (Tg ) with the increase of the cardanol feed content, indicating the incor-

poration of cardanol molecules into the copolymer chains. The reduction of the (Tg )

values is related to the large sizes of cardanol molecules and incorporation of long pen-

dant groups, which enhance the mobility of the polymer chains. As previously shown,

data for 10% at 85°C has a smaller conversion and quite a great amount of residual

monomers; therefore it was not possible to obtain the Tg for the copolymer, which is,

probably, smaller than 0°C .

In addition, 1H-NMR analyzes were carried out to evaluate if cardanol had been in

fact incorporated into the copolymer chains. Figure 5.18 shows the 1H-NMR spectrum

of the polymer sample produced at condition of 5% cardanol at 110°C as an example

and indicates the occurrence of styrene-cardanol bonds. It is important to emphasize

that residual styrene was removed from polymer samples through vacuum drying, and

residual cardanol was removed by reprecipitation in methanol . The 1H-NMR spectra

of the other polymer samples are not presented because they were qualitatively similar

to the 1H-NMR spectrum of polymer sample produced at condition A2. Based on the1H-NMR spectrum, it is possible to calculate the mol and weight cardanol copolymer

compositions respectively as:

ϕC (mol %) = ( f /2)

( f /2)+q(5.1)

ωC (w t %) =ϕC

MWST y

1MWC

+ϕC ( 1MWSt y

− 1MWC

)(5.2)

where f and q are the areas of the corresponding peaks. The obtained degrees of

cardanol incorporation are shown in Table 5.4. Samples collected after 3 and 5 hours

Table 5.3: Glass transition temperatures for P(STy-co-C)

Tg (°C )

85 °C 110°C0% C 97.5 92.1

2.5% C 78.4 83.05.0% C 44.3 66.610% C - - - 57.2

60

of reaction were selected for modeling purposes. In order to evaluate the possible effect

of the purification procedure, 1H-NMR analyzes were performed for samples obtained

after 5 hours of reaction before and after reprecipitation in methanol. As one can see,

most results were similar and within the expected experimental fluctuation of 1%. In

one case, however, the observed difference was significant, indicating the incomplete

removal of cardanol during the initial drying step or loss of copolymer during the sec-

ond reprecipitation step. In another case, for 10 wt% at 85°C , it is clear from TGA and

DSC that the amount of residual monomers, mainly cardanol, was high. Therefore, it

is very possible that the value reported does not correspond to what is trully incorpo-

rated. Moreover, there is not a value for this point purified because nothing precipi-

tated with methanol. Nevertheless, it seems clear that cardanol was incorporated into

the polymer chains during the polymerization trials.

Figure 5.18: 1H-NMR spectrum of poly(styrene-co-cardanol) prepared with 5 wt% ofcardanol at 110°C .

When the 1H-NMR spectrum of Figure 5.18 is compared with the 1H-NMR spec-

trum of cardanol, shown in Figure 5.4, one can see that characteristic polystyrene peaks

are present at 1.8, 2.2 and 7.1 ppm, although confounded with peaks that are character-

istic of cardanol. Moreover, the relative intensities of peaks corresponding to protons

b and d decreased due to the incorporation of cardanol through the terminal double

bond. It is important to observe that the peaks placed between 6.5 and 7.0 ppm in the

cardanol spectrum became much broader, which may indicate the reaction of the hy-

61

Table 5.4: Cardanol contents of copolymer samples (wt%) for P(STy-co-C)

Exp. Cond. 85°C 110°C

2.5% C 5% C 10% C 2.5% C 5% C 10% C3h 6.83 5.78 - - - 6.13 8.24 8.675h 5.60 20.16 74.32 5.60 8.24 22.98

5h (purified) 5.54 5.38 - - - 5.33 7.35 7.26

droxyl group and the occurrence of inhibition and chain transfer reactions. Finally, the

presence of additional peaks in the spectra of cardanol and poly(styrene-co-cardanol)

probably indicates the existence of other impurities in the comonomer feed (which is

provided as a complex mixture of cardanol derivatives, as already explained), which

may also affect the kinetics of the copolymerization reactions. Nevertheless, it is im-

portant to emphasize that additional purification of available commercial feed streams

is not feasible, due to both economical and technical constraints, imposing the use of

similar commercial feeds to perform the polymerizations.

Furthermore, one may arise the question ’if the terminal unsaturation is most re-

active group, then this might be the one responsible for the copolymerization and only

40% of the cardanol added should be accounted as this is the weight fraction for car-

danol triene’. However, in addition to the changes detected in the copolymer spec-

trum, the styrene-cardanol polymer produced with 2.5wt % at 110°Creached a high

conversion while, simultaneously, the composition data for 5 hours shows that 5.33wt

% is present on the copolymer, which does not correspond to 40% of 2.5% cardanol

added at the beginning and, in fact, correspond to over 200%. Therefore, the whole

cardanol stream can be reactive, although the mechanism are unknown. Another point

is that for the conversion of this last point, which is almost 90%, the mass balance

does not close for 5.33wt % of cardanol composition. However, it should be point out

that both techniques have errors. Gravimetric analysis may not evaporate all the unre-

acted styrene and the scale, although analytical, also have errors that can accumulate

throughout the analysis. At the same time, calculating the integrals in a NMR spectrum

is a very delicate analysis and even more for the present study case in which cardanol

has a very small peak in comparison with the one used for styrene, that also has other

peaks overlapping it. Therefore, at least, gravimetric analysis has 10% of error, while

every NMR has 5 % (depending on the conditions used for the analysis) but should be

a higher error for the present case.

For a few samples, 13C-NMR analysis were done to evaluate the possible presence

of a carbonyl group (C=O) that may have be formed after hydrogen donation by the

phenolic moiety. However, this peak was not detected.

Finally, FTIR was done to verify the presence or not of functional groups. As both

62

cardanol and styrene have an aromatic ring, many peaks are overlapping. Regions

identified in green are present majority in cardanol, while those in black are present

in both meres. In Figure 5.19 is possible to see the region of a hydroxyl present in

cardanol between 3600 and 3200 cm-1, which is also appears with smaller intensity

between 1350 and 1050 cm-1. However, for 2.5wt%, the hydroxyl peak desapeared at

3400cm-1. Although the concentration of cardanol is small, should appear something.

Therefore, it is possible that it reacted, inhibiting living radicals. In Figure 5.20, which

presents the FTIR spectrum for reactions performed at 110°C , one may see this peak

for 2.5wt% and for other concentrations. Therefore, besides the thermal initiation of

styrene, at 85°C the highest inhibition may have happened due to more prone hydroxyl

attacks. Also, as the polymer with 10.0wt% has a higher conversion, it is more similar

to the spectrum of polystyrene.

Figure 5.19: FTIR spectrum of poly(styrene-co-cardanol) prepared with at 85°C .

As it can be noticed that the peak of hydroxyl changes its intensity, probably caus-

ing a decrease in propagation rates, other samples at earlier reaction times (1, 3 and 5

hours) for the same reaction were also analyzed. For the condition 2.5wt % at 85°C (Fig.

5.21), for example, from one to five hours, the band at 3600-3500cm−1 (O-H bond) dis-

appeared completely, while there is also a reduction from 1250-1180cm−1 also related

with the phenolic moiety. However, one might notice that for, this condition, the rate

slowed down after 180 min (3h), not reaching 50%. And the main difference between 3

and 5 is the difference at 3600 - 3500cm−1, so the phenolic ring may have suppressed

63

Figure 5.20: FTIR spectrum of poly(styrene-co-cardanol) at 110°C .

new radicals. Moreover, the peak close to 900cm−1 decreased, which may indicate that

the terminal double bond was being reactive. The peak at 1700cm−1, not present in

styrene, is reduced as well and, as the conversion increased throughout 1-5 hours, this

reduction may be related with styrene incorporation. However, this is interesting as the

formation of a C=O bond due to hydrogen donation from the phenolic group would ap-

pear around this same range. Nonetheless, the intensities of the peaks between from

3100 to 2800 are much higher than for styrene alone, showing that there is cardanol in

the molecules.

For the condition 2.5wt % at 110°C (5.22), the reduction in the hydroxyl peak

around 3500 cm−1 almost does not occur. However peaks from 3200 - 2800 cm−1 ap-

pears more with what is presented for pure polystyrene, indicating styrene incorpo-

ration in the polymeric molecules. The increase in intensity for peaks between 2000

- 1650 cm−1 related to aromatic ring also shows that more meres of styrene are being

incorporated. But it cannot be notice a increase in peak 1700cm−1 that would corre-

spond to a carbonyl bond that would be formed by phenolic inhibition. Increases at

750 and 550 cm−1 makes the spectrum more different from the polystyrene spectrum,

showing that there is a different structure present (notice that, for 10wt % at 85°C , there

is excess of cardanol and this peaks do not appear). Therefore, this peaks might be re-

lated with the formation of a new, but unknown, bond with cardanol. However it was

expected that 1000-800 cm−1 should decrease their intensities are they are related with

64

Figure 5.21: FTIR spectrum of poly(styrene-co-cardanol) at condition 2.5wt % and85°C .

the terminal double bond and cardanol concentration, however there is a strange in-

crease in their intensities from 3h to 5h. This increase may indicate incorporation of

cardanol, but the mechanism of reaction is still unknown. A similar behavior occured

for the condition 10wt % at 110°C (5.23) and for 5wt % at the same temperature. But

for these cases there is a slightly reduction at 3600 cm−1, and also at 1250 and 1190

cm−1. Also, it cannot be notice a increase in peak 1700cm−1 that would correspond to

a carbonyl bond. So, at the same time that 5 and 10wt % had this reduction at these

peaks, they were also more inhibited in relation to 2.5wt %, indicating, again, that the

hydroxyl is a radical inhibitor.

65

Figure 5.22: FTIR spectrum of poly(styrene-co-cardanol) at condition 2.5wt % and110°C .

Figure 5.23: FTIR spectrum of poly(styrene-co-cardanol) at condition 10wt % and110°C .

66

5.4 Copolymerization of cardanol and methyl methacry-

late

Methyl methacrylate (MMA) (Fig. 5.24) is a vinyl monomer used for the synthe-

sis of PMMA with many application as, because it is amorphous, PMMA presents high

strength and excellent dimensional stability due to its rigid polymer chains. Addition-

ally, PMMA has exceptional clarity, very good weatherability, good impact resistance,

can be machined and is resistant to most chemicals, although it can be attacked by

different organic solvents.

Figure 5.24: Methyl methacrylate, MMA, structure

Copolymerization with MMA were performed for cardanol and natural CNSL (re-

sults are discussed in the next section) at 85 and 110°C . Figure 5.25 presents conver-

sion and molecular weights for P(MMA-co-C) at 85°C . Although completly conversion

occurs for PMMA in less than one hour, it takes almost 2 and 4 for 2.5 and 5wt % of

cardanol added, respectively. For 10wt %, it does not even reach 50%. The molecu-

lar weights are also reduced when cardanol is added due to cardanol’s retardant effect.

However, this effect might have implied in the decrease in polydispersion, acting a bit

as chain transfer agent.

For 110°C , the retardant effect can be also noticed, although rates were increased

in relation to 85°C . For this condition, 10% w/w of cardanol reached over 60% of con-

version; but it seems like, even on longer reaction periods, the conversion would not

be fully reached because there is a decrease on the overall kinetic rate. This may have

happened because of the higher amount of hydroxyl groups present in the reaction

mixture from the phenolic moiety as they may have reacted with living radicals, ter-

minating them. Also reactions with the internal unsaturations are another cause for

retardation as they form a secondary and stable radical. Plus, at this temperature, ben-

zoyl peroxide (BPO) is decomposed before 40 min (notice that is at 40 min that the

conversion start to become constant). Therefore, cardanol also acted as an inhibitor at

10wt %.

67

Figure 5.25: Conversion and molecular weights for P(MMA-co-C), 85°C

Figure 5.26: Conversion and molecular weights for P(MMA-co-C), 110°C

68

Despite the decrease in reaction rates, the addition of cardanol increased in ther-

mal stabilities detected by thermogravimetric analysis (TGA). Figures 5.27 and 5.28

shows that cardanol degrades over 250°C , which is, roughly, the same temperature

PMMA begins to degrade. However, for copolymers, the degradation starts afterwards

and fully degrades around 450°C (and even a bit further for the 10wt % copolymer).

Therefore, TGA curves are shifted to the right. The degradation of PMMA is similar to

others reported in the liturature (FERRIOL et al., 2003). Although roughly two degrada-

tion stages can be seen, there are in fact three. The first step (close to 165°C is initiated

by scissions of head-to-head linkages (H–H) as bond dissociation energy of these link-

ages are less that of a C–C backbone bond due to a large steric hindrance and inductive

effect of vicinal ester groups, promoting facile homolytic scission of the chain (KASHI-

WAGI et al., 1986). This stage occurs the same for copolymers. However, for 10wt % at

110°C , it starts a bit early, equaling to the start of cardanol degradation; so, probably,

there is residual cardanol for this samples. The second step for PMMA (around 270°C )

is due to at unsaturated ends (resulting from termination by disproportionation) in-

volving a homolytic scission β to the vinyl group. This would start due to formation

of radicals from impurities, H–H bonds, scission b to the vinyl group at unsaturated

ends and/or random scission (MANRING et al., 1989). This also means that the unsat-

urations of cardanol side chain are beeing degradate too. In fact, the whole extension

of the side chain is probably degradating as the phenolic ring is more thermally sta-

ble. And these scissions require more energy for the copolymer. For PMMA, the las

step (around 350°C ) is initiated by random scission within the polymer chain. But for

copolymers, it also corresponds to the degradation of the phenolic ring.

In addition, due to the long side chain of cardanol, it has some plastificant effect,

reducing the glass transition temperature (Tg ) (Table 5.28). For 10wt % of cardanol, the

data could not be detected as it probably is less than 0°C (the polymer is completely

flexible at room temperature).

Table 5.5: DSC analysis for P(MMA-co-C)

85°C 110°C

0% 104.8 105.32.5% 84.92 84.13

5% 72.36 71.2510% - - - - - -

69

Figure 5.27: TGA and DTG curves for P(MMA-co-C), 85°C

Figure 5.28: TGA and DTG curves for P(MMA-co-C), 110°C

70

To confirm the presence of functional groups in the copolymer, FTIR analysis of

the last samples is used (Figs. 5.29 and 5.30). The classification of the peaks is done in

the Figures. Besides the unsaturations (3100-2800, 1750-1650 and 1000-600 cm−1), the

greatest difference between the spectrum of PMMA and a copolymer is the presence of

the peak around 3450 cm−1 of the phenolic group. It is not possible to see a presence of

carbonyl (C=O) that may have formed from hydrogen donation by the phenolic moiety

as this bond is overlapped with the bond of PMMA. For 85°C (Fig. 5.29), although some

reaction with this functional group may have taken place, it is still present for 5% (that

reached full conversion) and for 10%. For 2.5%, there is a small concentration. So this

functional group may still be used to functionalize the polymers. However, notice that

for 110°C (Fig. 5.29), when 2.5 and 5wt % reached full conversion, or 10% reached a

higher conversion, the presence of the phenolic group is a lot less significant. Plus as

quite a big amount of BPO was used (1wt %), and it is fully consumed by 84 min at 85°C

and less than 40 min at 110°C , the faster decomposition at 110 was enough to initiate

new chains, while some were being terminated by the -OH inhibition. Thus, although

the retardant effect was not so intense, many radicals were suppressed, which, again,

may have helped to control the dispersion index.

Figure 5.29: FTIR analysis for P(MMA-co-C), 85°C

71

Figure 5.30: FTIR analysis for P(MMA-co-C), 110°C

To find out the copolymer composition, it is used the NMR analysis by comparing

the peak m with the peak f of cardanol. Because the spectrums are quite similar, Figure

5.31, a spectrum of P(MMA-co-C) at 85°C with 5% of cardanol is an example of them.

As m represents three hydrogens, while f represents two; the following equations are

used to obtain the composition in mol and mass fractions:

ϕC (mol %) = ( f /2)

( f /2)+ (m/3)(5.3)

ωC (w t %) =ϕC

MWM M A

1MWC

+ϕC ( 1MWM M A

− 1MWC

)(5.4)

However, as some unreacted cardanol could still be present within the polymeric

samples, they were dissolved in toluene followed by precipitation in methanol. Then,

the values of copolymer composition before and after the purification can be analyzed

in Tables 5.6 and 5.7. There are difference mainly for the reactions with 10w % and for

those done at 110°C . Also, reactions at 110°C incorporated less cardanol than those at

85°C , even when the conversion was smaller (for the 10w % case). This could have hap-

72

Table 5.6: Copolymer composition for P(MMA-co-C), 85°C

Not Purified Purified

Mol (%) Massa (%) Mol (%) Massa (%)2.5 0.01 1.32 0.01 1.305 0.03 2.87 0.03 2.87

10 0.19 18.93 0.02 2.40

Table 5.7: Copolymer composition for P(MMA-co-C), 110°C

Not Purified Purified

Mol (%) Massa (%) Mol (%) Massa (%)2.5 0.02 1.79 0.01 1.325 0.04 4.38 0.01 1.44

10 0.12 13.29 0.01 1.20

pened due to a) faster kinetic rates of MMA with another molecule of MMA, meaning

less preference for a molecule of MMA to be with one of cardanol at 110°C ; or b) due

to increase concentration of radicals at 110°C , more radicals may have suppressed by

internal unsaturations, turning them unreactive. And if one goes back to Figure 5.30, it

can be seen that most of the phenolic group was consumed at 110°C , which may also

have suppressed the radicals and made those cardanol molecules unreactive.

What could be notice for every spectrum is that the proportion between peaks d/f,

c/f or b/f decreased, meaning the reaction in the unsaturations took place. Although

these peaks were still present in small amounts, they might be a reflect of retardation

due to a reaction with one of the internal unsaturations - which may have left a car-

danol pendant with a "dormant" radical.

What it is interesting to see, though, is that although 110°C did not incorporate

cardanol as much, the DSC analysis showed similar Tg values for each composition at

both temperatures.

73

0.51.01.52.02.53.03.54.04.55.05.56.06.57.07.58.0f1 (ppm)

8.89

300.

00

7.26

0 CD

Cl3

no

O On 6.6 6.6

7.146.7

h

f i

i

i

i

g

c

c cce

e

b

d

OH

m

o

m

aa dc e f g hi

j

Figure 5.31: NMR spectrum for P(MMA-co-C), 5% w/w, 110°C

5.5 Copolymerization of natural CNSL and methyl

methacrylate

Aiming to compare the reactivity between anacardic acid and cardanol as their only

difference is the presence or not of a carboxyl functional group, polymerizations with

natural cashew nut shell liquid (natural CNSL) and methyl methacrylate were done at

the same conditions than those for cardanol-methyl methacrylate. Natural CNSL was

used because it’s rich in anacardic acid. Therefore, Figure 5.32 show the conversion

behavior for pure PMMA and for 2.5, 5.0 and 10.0% w/w of natural CNSL at 85°C and

110°C , and they are quite similar from those obtained with cardanol at both tempera-

tures (Figs. 5.33 and 5.34). However the copolymerization with natural CNSL was faster

for 85°C - the temperature that had the highest retardation effect. This is interesting be-

cause natural CNSL also has cardol (< 21w t%) and 2-methyl cardanol (< 4w t%), and

both compounds have 2 hydroxylic group in the aromatic ring, which would increase

the inhibition. However, perhaps, the presence of the carboxylic group influence on

the inhibition effect of the phenolic molecule, decreasing it. This may happen because

the carboxyl group is an electron withdrawing group and even being in the ortho posi-

tion in relation to the hydroxyl, the hydroxyl may not desire to donate its hydroxygen

atom as much because the radical formed will be less stable. Nonetheless, both com-

74

pounds have similar concentrations of triene portions, which may not be influence in

the difference among their kinetics.

Figure 5.32: Conversion for P(MMA-co-CNSL), 85 and 110°C

Figure 5.33: Conversion for P(MMA-co-CNSL) and P(MMA-co-C), 85°C

75

Figure 5.34: Conversion for P(MMA-co-CNSL) and P(MMA-co-C), 110°C

When it comes to thermal stability, CNSL helps to improve even more thermal sta-

bility. Figure 5.35 and 5.36 presents TGA curves for the polymers at 85°C and 110°C

respectively. While PMMA, degradates at aproximatelly 400°C ; the copolymers degra-

dated only over 425°C , shifting the curve to the right. Also, the first stage of degra-

dation in relation to pure PMMA was decreased as the percentage of weight loss was

lower, although this happened at higher temperatures. However, for 85°C , it possible

that the sample analyzed for 10wt % still has some residual CNSl as its final degra-

dation matches the final degradation for CNSL. Nonetheless, it can be afirmed that a

copolymer was formed to the changes in the TGA curves mentioned.

76

Figure 5.35: TGA and DTG curves for P(MMA-co-CNSL) and P(MMA-co-C), 85°C

Figure 5.36: TGA and DTG curves for P(MMA-co-CNSL) and P(MMA-co-C), 110°C

DSC analysis (Table 5.8) shows that the glass transition temperature (Tg ) decreased

with the addition of CNSL as well because of the incorporation of anacardic acid and

the others CNSL compounds. This occured due to the presence of the side chain that

increases the mobility of the polymeric chains, reducing the Tg . The DSC data for 10%

w/w of CNSL at 85°C was not obtained as it is lower than 0°C . 1H-NMR was done for

only one sample - P(MMA-co-CNSL), 10% w/w, 110°C ; and using the same procedure

for cardanol with MMA, the CNSL composition in the copolymer corresponds to 2.75%

in mass - roughly the double of the obtained for cardanol -, but this sample was not

77

Table 5.8: DSC analysis for P(MMA-co-CNSL)

85°C 110°C

0% 104.8 105.32.5% 101.67 98.00

5% 68.37 88.0810% - - - 60.43

purified, meaning that it still may have some residual CNSL.

5.6 Copolymerization of cardanol and vinyl acetate

Poly(vinyl acetate) (PVA or PVAc) is a polymer known for its application as a glue.

It belongs to the polyvinyl esters family, it is a type of thermoplastic and partially sol-

uble in water. Copolymerizations of vinyl acetate and cardanol were performed at 70,

85 and 110 °C . However, at 110°C conversion and molecular weight data became very

bad because the boiling temperature of vinyl acetate is 72.7°C and the glass tubes, al-

though threaded, could not avoid evaporation. Some characterizations were also ir-

regular and not trustable. Some results of 85°C were also poor and were removed. For

70°C , the kinetic rates were way too small and conversion reached only 20% for all con-

centrations even at 5 hours of reaction. Therefore, for further reactions, the method-

ology should be changed. In Figure 5.38, it is shown the conversion and molecular

weight for the reaction at 85°C . As also previously notice for the copolymerization with

styrene and methyl methacrylate, cardanol also retards the polymerization of PVAc,

limiting it to 50% even when the smallest concentration is used. Values for molecular

weight were also decreased, and mainly small chain polymers (mainly oligomers) were

formed. Therefore, the cardanol retardation was more pronounced for vinyl acetate

than was for the others, inhibiting the polymerization after some minutes.

Figure 5.37: Structure of poly(vinyl acetate), PVAc

78

Figure 5.38: Conversion and molecular weight for P(VAc-co-C), 85°C

From NMR data, copolymer composition can be get using Equations 5.1 and 5.2

(5.9). However, different from what was done for methyl methacrylate and styrene, the

polymer was not washed - solubilized in a solvent and them precipitated. So it is possi-

ble that these concentrations are higher than what actually contains on the copolymer,

mainy for 10% of cardanol. A spectrum for the copolymer with 5% of cardanol added

initially is shown in Figure 5.39.

Table 5.9: Copolymer composition for P(VAc-co-C), 85°C

mol (%) mass (%)

2.5% C 0.04 3.855.0% C 0.12 12.9610% C 0.49 76.32

79

0.00.51.01.52.02.53.03.54.04.55.05.56.06.57.07.5f1 (ppm)

300.

00

26.9

1

7.26

CD

Cl3

OO

no

m

o

m

n

aa

bc d

f

e gh

i

j

6.6 6.6

7.146.7

h

f i

i

i

i

g

c

c cce

e

b

d

OH

Figure 5.39: 1H NMR for P(VAc-co-C), 5% w/w, 85°C

For TG analysis, the copolymers reduced the thermal stability, which was not the

case for the copolymers with methyl methacrylate or styrene (Fig. 5.40). PVAc ther-

mal degradation have two stages - one that occurs close to 350 and the other close to

475°C . The addition of 2.5% w/w did not change the thermal behavior, but 5 or 10%

w/w reduced the temperature of the first stage to 275, while maintaining the other at

475°C . For the polymerization at 110°C a similar result was obtained, meaning that

even it with high conversions due to high temperature, was not capable to synthesize

a stronger polymer than PVAc. This weakining might have happened due to the pres-

ence of an inductive effect between the electron-withdrawing groups of cardanol (C-C

bonds) and the ester group in PAA, lowering the thermal stability by weakining the es-

ter bond and letting it prone to scisson (CERVANTES-UC et al. (2006)). DSC analysis

were not performed, but every copolymer is a liquid or viscous liquid at room temper-

ature.

FTIR analysis show that both vinyl acetate and cardanol are present at 2.5 and 5%

w/w. However, at 10%, there is a high concentration of residual cardanol because the

peaks related to PVAc cannot be seen with the same intensity them those previous sam-

ples. Therefore, the value of 76.32% of cardanol presente in the copolymer obtained by

NMR is probably higher than what should be.

To summarize, copolymerization of cardanol with vinyl acetate are not very indi-

80

Figure 5.40: TGA and DTG curves for P(VAc-co-C), 85°C

cated. Cardanol has a strong inhibition effect on vinyl acetate and conversions are

small. Moreover, the polymer thermal stability was not increased and a right poly-

mer mass could not be obtained. Although other interesting monomers can be tested,

perhaps high temperatures can synthesize better a copolymer, but, for this, a reaction

under pressure will be performed, requiring a closed and strong reactor.

81

Figure 5.41: FTIR for P(VAc-co-C), 85°C

5.7 Copolymerization of cardanol and acrylic acid

Polyacrylic acid (PAA) is water soluble, hygroscopic, brittle and colorless in nature

with Tg at nearly 106°C . As PAA is an anionic polymer, it has the ability to absorb

and retain water and swell to many times their original volume. That is why PAA is

used in diapers, and menstrual pads. When it is dry, PAA is used in paints, cosmetic,

pharmaceutical and personal care products to as thickening, dispersing, suspending,

and emulsifying agent. However, there are many polymer contaning acrylic acid as a

copolymer, mainly synthesizing polymers that are responsive to alteration in temper-

ature and/or pH.

Figure 5.42: Polyacrylic acid, PAA

Reaction with PAA was performed at 85 and 110°C . However, as the reaction was

so fast that is not possible to analyze its kinetics. As Figure 5.43 shows, the monomer

82

is practically completely consumed instantaneously even when 10wt % cardanol was

added. Experimentally, the reaction mixture, just a few seconds after added in the hot

bath, starts to polymerize so fast that it seems like it is boiling. Besides TGA, there is

not characterization analysis for this copolymer. Therefore, Figure 5.44 presents TGA

curves for the copolymer acrylic acid-cardanol. As can be seen, the addition of car-

danol shifts a the curves of the copolymer to the left in direction of pure cardanol.

Therefore, cardanol does not increase the thermal stability of the polymer. Plus, it

seems like there are three degradation stages for the copolymers. The beginning of

the first stage matches with the boiling point of acrylic acid, which is 139°C . So this

first stage seems to be the loss of residual acrylic acid. The second stage matches with

the cardanol curve, suggesting that it corresponded to residual cardanol. And the third

stage is similar to PAA curve and might be the degradation of pure PAA without high

concentrations of cardanol. So, even at high conversions, perharps cardanol incorpo-

ration was not high.

Moreover, as the reaction was really fast even for 85°C , another reaction was per-

formed at 60°C . However, as the initiator decomposition rate also decreases, the ki-

netic became a lot slower, and the copolymer conversion reached was of roughly 40%

after 6h and stopped. Therefore, some how, the slower kinetics makes the inhibition

of cardanol more likely and significant. In addition, very little is found in the literature

about the polymerization of acrylic acid in bulk to understand how fast its kinetics is.

Therefore, perhaps, for further reactions, the polymerization of acrylic acid with car-

danol should be done in heterogeneous medium as this method is more commonly

used and there is more data.

83

Figure 5.43: Conversion for P(AA-co-C) at 85°C

84

Figure 5.44: TGA curves for for P(AA-co-C) at 85°C

85


In this chapter, it was shown that cardanol and other compounds of cashew nut

shell liquid truly reacts via free radicals polymerize. For the case of homopolymer-

ization of these compounds, they are very slow and reached very small conversions.

However, 1H-NMR, FTIR and TG analysis proved that there are differences between

the reaction product after 8 or 10 hours and the "raw" compound. However, they re-

tard and, most of the times (mainly at lower temperatures and high concentrations),

inhibits the polymerization. Nonetheless, besides the use of the synthesized copoly-

mers for functionalization of the molecules, the addition of cardanol or natural CNSL

can be also used to decrease the Tg or increase the thermal stability when required.

However, the copolymerization experiments with vinyl acetate was not satisfactory,

and cardanol does not increase the thermal stability of PAA.

Moreover, increasing the temperature of reaction does increase conversion, but

this does not translate in high differences in cardanol composition in the copolymer.

In fact, the composition remains almost the same or even smaller. Polymerization of

MMA with natural CNSL was very satisfactory, meaning that it obtained slight high

conversions and in thermal stabilities. Therefore, further tests with it may be interest-

ing.

Therefore, to summarize, copolymerization with cardanol or natural cashew nut

shell liquid happened with four different monomers. These phenolic compounds were

able to react with free radical through several reactive sites. The hydroxyl group was re-

sponsible for inhibition as its concentration was detected in FTIR analysis, forming a

carbonyl group. The last could not be seen in FTIR spectrums due to peak over leaping,

and it was not detected in 13C -NMR. Though not likely, the extra radical in the oxygen

atom (which was stabilized by the aromatic ring) might have formed an ether bond

if it propagated or terminated. Internal unsaturations, although had some steric hin-

drance, probably were attacked too. Due to the generation of a stable secondary radi-

cal, propagation or termination were not frequently. However, reduction in their inten-

sities were detected by 1H-NMR and FTIR analysis. The terminal unsaturation was the

functional group most likely to continue propagation and incorporate cardanol/ nat-

ural CNSL in polymeric chains. Although they are present in less than 50% in phenolic

lumps, it is considered that the whole lump is reactive because the compounds have

many reactive sites. The carboxylic group of anacardic acid in relation to cardanol did

not influence intensively on polymerization kinetics. On the contrary, it seems like the

carboxylic moiety helped decreasing the inhibition for 85°C. And although cardanol

or natural CNSL are provided as a complex mixture of natural compounds, possibly

including small amounts of other inhibitors, one must consider that it is not econom-

ically feasible to purify this natural stream any further prior to use.

86

Chapter 6

Bulk Copolymerization: Modeling

6.1 Introduction

In this chapter, it is proposed a mathematical dynamic model for the copolymer-

ization in bulk via free radical polymerization. It is used the terminal model and quasi-

steady state assumption. Based on experimental results, model parameters are esti-

mated to allow for appropriate description of monomer conversions, average molec-

ular weights and copolymer compositions of the produced copolymer. For methyl

methacrylate or styrene, classical thermal-kinetic parameters for the literate was used.

As cardanol behaved similar to natural CNSL, with the experimental data available the

result would also be similar. Therefore, along the chapter, the parameter estimation

will be referring to cardanol, but it can be valid for both. Regarding the parameters

related to cardanol, these parameters are not available in the literature, and either for

copolymerization. Plus, it was shown that those phenolic compounds exerts strong

inhibitory effects on polymerizations and this effect must be take into account. There-

fore it is estimated a parameter ( fi nh) that reduces the concentration of free living rad-

icals in the system.

Given the fact that other molecules may be eventually present in the naturally oc-

curring cardanol stream and that different reactive sites (such as the phenolic group)

can take part in the reaction mechanism, the whole cardanol stream was treated as

a lumped monomer stream for purposes of kinetic modeling. However, as cardanol

polymerizes at slow rates and only produces oligomers at the analyzed reaction con-

ditions through free-radical polymerizations, as observed experimentally; for the sake

of simplicity, some model parameters were initially assumed to be either null or equal

to the styrene polymerization parameters. Thus, the parameter estimation happened

using the computation package ESTIMA, implemented in Fortran, which combines

particle swarm optimization (PSO) and Gauss-Newton algorithms (SCHWAAB et al.,

2008). One hundred particles were used and two thousand iterations were carried out

87

with statistical confidence level of 95 %.

6.2 Kinetic Mechanism

The dynamic model proposed describe the copolymerization in bulk via free rad-

ical polymerization. For this, it uses the terminal model assuming that only the last

specie added to a living chain represents the reactivity of whole molecule that it be-

longs to. Therefore, the step of propagation is extended to represent the different reac-

tions between monomers and the reactive macro-radicals presented in the reactional

medium. Similarly, other steps including the reaction of a living species with another

molecule are also extended – termination, transfer to monomer, etc.

Every free radical polymerization initiated thermally by an initiator (BPO or AIBN),

includes the steps of initiation, propagation and termination. Depending on the

monomer used, thermal initiation may take place and should be included. Chain

transfer reactions (to one of the monomers or to impurities) do not influence largely

on the conversion; however, it changes the average molar mass. Therefore, a generic

kinetic mechanism is represented below by 19 equations. However, notice that the

thermal initiation is represented as it happens for styrene (Eq. 6.4).

Initiation:

Ikd−→ 2R (6.1)

R +M1kr1−−→ P1,0 (6.2)

R +M2kr2−−→Q0,1 (6.3)

3M1kdm−−−→ 2P∗

1,0 (6.4)

Propagation:

Pi , j +M1kp11−−→ Pi+1, j (6.5)

Pi , j +M2kp12−−→Qi , j+1 (6.6)

Qi , j +M1kp21−−→ Pi+1, j (6.7)

Qi , j +M2kp22−−→Qi , j+1 (6.8)

88

Monomer transfer:

Pi , j +M1ktm11−−−→ P1,0 +Γi , j (6.9)

Pi , j +M2ktm12−−−→Q0,1 +Γi , j (6.10)

Qi , j +M1ktm21−−−→ P1,0 +Γi , j (6.11)

Qi , j +M2ktm22−−−→Q0,1 +Γi , j (6.12)

Termination:

Pi , j +Pm,nktc11−−−→ Γi+m, j+n (6.13)

Pi , j +Qm,nktc12−−−→ Γi+m, j+n (6.14)

Qi , j +Qm,nktc22−−−→ Γi+m, j+n (6.15)

Pi , j +Pm,nktd11−−−→ Γi , j +Γm,n (6.16)

Pi , j +Qm,nktd12−−−→ Γi , j +Γm,n (6.17)

Qi , j +Qm,nktd22−−−→ Γi , j +Γm,n (6.18)

In the mechanism, I represents a molecule of initiator, R∗ is a dissociated initiator

radical, M1 is a molecule of styrene (Sty) or methyl methacrylate (MMA) depending

on the homo/co-polymerization. And M2 is a molecule of cardanol or the idea of a

molecule that simulates the mixture of components of the natural CNSL.

Also, P∗i , j represents a living chain with i units of monomer 1 (Sty or MMA) and j

units of monomer 2 (cardanol or another compound of CNSL) in which the active cen-

ter is present in a Sty/MMA-terminated living polymer chain. Q∗i , j represents a living

chain with i units of monomer 1 and j units of monomer 2 in which the active center is

present in a CNSL molecule-terminated living polymer chain.

Although dead chains are different from each other in composition and molar

mass, they are represented by one unique specie Γi , j with i units of monomer 1 and

j units of monomer 2.

In addition, it is used the reactivity ratios (r11 and r22) to relate the specific homo-

propagation reaction rate to the hetero-propagation reaction rate.

ri , j =kp,i i

kp,i j(6.19)

r j ,i =kp, j j

kp, j i(6.20)

89

Moreover, it is introduced the relative efficiency factor of hetero-termination. It

measures how fast is the hetero-termination in relation to a reference value (OECH-

SLER (2016)).

ψi i = kt i i√kt i i kt j i

(6.21)

ψ j j =kt j j√

kt i i kt j j

(6.22)

The gel (g t i j ) and glass (gpi j ) effects cannot be ignored either. They multiply the

propagation/termination specific rates (Equations 6.23 and 6.24). k0pi j and k0

t i j rep-

resent the specific rates of propagation and termination in the absence of polymer,

while gpi j and g t i j are the correlations for glass and gel effects, respectively. These can

be calculated by Equations 6.25 to 6.27 which are based on the Free Volume Theory.

kpi j = kpi j gpi j (6.23)

kt i j = kt i j g t i j (6.24)

gpi j = 1 (υ f ≥ v f cr ) (6.25)

gpi j = exp

[−A

(1

υ f− 1

υ f cr

)](υ f < υ f cr ) (6.26)

g t i j = exp

[−A

(1

υ f− 1

υ f 0

)](6.27)

While υ f cr is a tabulated value for each monomer. The free volumes calculated at

the beginning of the reaction (υ f 0) and at time t (υ f ) can be obtained by Equations

6.28 and 6.29, where σi is the volume fraction of specie i , time t and σ f i is the volume

fraction of specie i at the initial time.

υ f =σm1υ f m1 +σm2υ f m2 +σpυ f p (6.28)

90

υ f 0 =σ f m1υ f m1 +σ f m2υ f m2 +σ f pυ f p (6.29)

Finally, the free volumes (υ f i ) for the monomers and polymer were calculated by

Equation 6.30. αi is the thermal expansion coefficient and Tg i is the glass transition

temperature.

υ f i = 0.025+αi (T −Tg i ) (6.30)

6.3 Material Balances

Have pointed out a generic kinetic mechanism, it was important to make some

considerations aiming to solve the model after doing the material balances. Also, not

every parameter can be estimated, and a few hypotheses should be pointed out. First

of all, the glass and gel effects were only considered to styrene or methyl methacrylate

homo-polymerization steps as there is not any information about this effects for

cardanol/CNSL. Also, the phenolic compounds have strong inhibitory effects and do

not polymerize to high conversions at conditions utilized for this research. So, as their

viscosity do not highly increase, it is believed that these effects are not applied for

them. Moreover, validity of volume additivity and of the long chain assumption was

assumed to write mass balance equations.

Mass balance for initiator I :

d I

d t=−kd I − I

V

dV

d t(6.31)

Mass balance for monomer 1, M1:

d(M1/V )

d t=−

[Rckr

M1

V+ (kp11 +ktm11)

M1

Vγ0,0 + (kp21 +ktm21)

M1

VΠ0,0

]V −2kdm

[M1

V

]3

V

(6.32)

Mass balance for monomer 2, M2:

d(M2/V )

d t=−

[Rckr

M2

V+ (kp12 +ktm12)

M2

Vγ0,0 + (kp22 +ktm22)

M2

VΠ0,0

]V (6.33)

91

Mass balance for radicals, R∗:

dR∗

d t=

[2 f kd

I

V−kr 1

M1

V

R

V−kr 2

M2

V

R

V

]V (6.34)

Assuming the quasi-steaty state assumption and that kr 1 = kr 2, it is possible to obtain

the normalized rate of initiation for kr 1 = kr 2, given by:

Rckr = krR

V= 2 f kdCi

Cm1 +Cm2(6.35)

Mass balance for living chains, P1,0:

dP1,0

d t=

[kr 1

(R

V

)(M1

V

)−kp11

(P1,0

V

)(M1

V

)−kp12

(P1,0

V

)(M2

V

)]V

+[−ktm11

(P1,0

V

)(M1

V

)−ktm12

(P1,0

V

)(M2

V

)]V

+[

ktm11

∞∑i=1

∞∑j=0

(Pi , j

V

)(M1

V

)+ktm21

∞∑i=0

∞∑j=1

(Qi , j

V

)(M1

V

)]V

+[

2kdm

(M1

V

)3]V

(6.36)

Considering:

i = 2, . . . ,∞j = 0, . . . ,∞

92

Mass balance for living chains, Pi , j :

dPi , j

d t=

[−kp11

(Pi , j

V

)(M1

V

)−kp12

(Pi , j

V

)(M2

V

)]V[

kp11

(Pi−1, j

V

)(M1

V

)+kp21

(Qi−1, j

V

)(M1

V

)]V

+[−ktm11

(Pi , j

V

)(M1

V

)−ktm12

(Pi , j

V

)(M2

V

)]V

+[−ktc11

(Pi , j

V

) ∞∑m=1

∞∑n=0

(Pm,n

V

)−ktd11

(Pi , j

V

) ∞∑m=1

∞∑n=0

(Pm,n

V

)]V

+[−ktc12

(Pi , j

V

) ∞∑m=0

∞∑n=1

(Qm,n

V

)−ktd12

(Pi , j

V

) ∞∑m=0

∞∑n=1

(Qm,n

V

)]V

(6.37)

Mass balance for living chains, Q0,1:

dQ0,1

d t=

[kr 2

(R

V

)(M2

V

)−kp21

(Q0,1

V

)(M1

V

)−kp22

(Q0,1

V

)(M2

V

)]V

+[−ktm21

(Q0,1

V

)(M1

V

)−ktm22

(Q0,1

V

)(M2

V

)]V

+[

ktm12

∞∑i=1

∞∑j=0

(Pi , j

V

)(M2

V

)+ktm22

∞∑i=0

∞∑j=1

(Qi , j

V

)(M2

V

)]V

(6.38)

Mass balance for living chains, Qi , j :

dQi , j

d t=

[−kp21

(Qi , j

V

)(M1

V

)−kp22

(Qi , j

V

)(M2

V

)]V

+[

kp12

(Pi , j−1

V

)(M2

V

)+kp22

(Qi , j−1

V

)(M2

V

)]V

+[−ktm21

(Qi , j

V

)(M1

V

)−ktm22

(Qi , j

V

)(M2

V

)]V

+[−ktc12

(Qi , j

V

) ∞∑m=1

∞∑n=0

(Pm,n

V

)−ktd12

(Qi , j

V

) ∞∑m=1

∞∑n=0

(Pm,n

V

)]V

+[−ktc22

(Qi , j

V

) ∞∑m=0

∞∑n=1

(Qm,n

V

)−ktd22

(Qi , j

V

) ∞∑m=0

∞∑n=1

(Qm,n

V

)]V

(6.39)

93

Mass balance for dead chains, Γi , j :

d

d t

[Γi , j

]= ktm11M1

VPi , j V +ktm12

M2

VPi , j V +ktm21

M1

VQi , j V +ktm22

M2

VQi , j V

+ ktc11

2

i−1∑r=1

j∑q=o

Pr ,q Pi−r , j−qV +ktc12

i−1∑r=1

j∑q=o

Pr ,qQi−r , j−qV

+ ktd11

2Pi , j

∞∑i=0

∞∑j=o

Pi , j V +ktd12

[Pi , j

∞∑i=0

∞∑j=o

Qi , j +Qi , j

∞∑i=0

∞∑j=o

Pi , j

]V

+[−ktc12

(Qi , j

V

) ∞∑m=1

∞∑n=0

(Pm,n

V

)−ktd12

(Qi , j

V

) ∞∑m=1

∞∑n=0

(Pm,n

V

)]V

+ ktd22

2Qi , j

∞∑i=0

∞∑j=o

Qi , j V + ktc22

2

i∑r=1

j−1∑q=o

Qr ,qQi−r , j−qV

(6.40)

Volume Balance:

dV

d t=

[(1

ρm1− 1

ρp

)Rm1 +

(1

ρm2− 1

ρp

)Rm2

]V (6.41)

We also have that:

γ0,0 =√

Rckr (M1/V )(M2/V )+2kdm(M1/V )3

kt11 +kt12Ks +kt22K 2S

(6.42)

π0,0 = Ksγ0,0 (6.43)

Being:

Ks =(kp12 +ktm12)(M2/V )

(kp21 +ktm21)(M1/V )(6.44)

Thus, monomer conversion (X) can be obtained by Equation 6.45:

X = ρpυP

ρm1υm1 +ρm2υm2 +ρpυp(6.45)

being ρi the density of the component i (g .L−1) and ϑi the molar partial volume of

component i (L).

94

6.4 Methodology and Numerical Solution

There are seven types of species (not counting the radical, R(∗)) and, therefore,

six species for mass balances. But you also need to count the volume variance in the

batch reaction. Moreover, eighteen specific reaction rates to be found. You can find

both hetero-propagation rates by relating them to their reactivity ratios. You can also

find the hetero-termination rate by relating with the relative efficiency factor of hetero-

termination. But, at the end, you want to know at least the Mn, Mw (and the PI) along

the polymerization. For this, the moment technique is very popular. It uses statistical

moments of the molar mass distributions. And a good advantage is that can solve a

finite number of equations fast, which is very interesting for co-polymerizations due to

the large number of equations. The moments for living and dead chains were defined

as (OECHSLER (2016)):

γk,l =∞∑

i=0

∞∑j=o

i k j mPi , j (6.46)

πk,l =∞∑

i=0

∞∑j=o

i k j mQi , j (6.47)

λk,l =∞∑

i=0

∞∑j=o

i k j mΓi , j (6.48)

However, if the interest is in the distribution of average molar masses, another

method needs to be used. Monte-Carlo, for example, can be a solution. It is a stochas-

tic method that works with probabilities in the molecular level. One can also use a de-

terministic and robust mathematical method to solve ordinary differential equations

(ODEs). Nonetheless, these methods require longer computational times.

As the aim of this section is the study of the kinetic reactions, development of the

model and estimation of parameters. The moment technique meets the needs, and

only until the 2nd moment. Therefore, considering that:

kt12 = (ktc12 +ktd12) (6.49)

kt11 = (ktc11 +ktd11) (6.50)

95

kt22 = (ktc22 +ktd22) (6.51)

We have the following moments for living chains: γ1,0 & π1,0:

dγ1,0

d t= kp21

(M1

V

)(π1,0

V

)V

−[(

kp12 +ktm12)(M2

V

)+ktm11

(M1

V

)+kt11

(γ0,0

V

)+kt12

(π0,0

V

)](γ1,0

V

)V

+[

kr 1

(R

V

)+ktm11

(γ0,0

V

)+ktm21

(π0,0

V

)+kp11

(γ0,0

V

)+kp21

(π0,0

V

)](M1

V

)V

(6.52)

dπ1,0

d t=−

[(kp21 +ktm21

)(M1

V

)+ktm22

(M2

V

)+kt22

(π0,0

V

)+kt12

(γ0,0

V

)](π1,0

V

)V

+kp12

(M2

V

)(γ1,0

V

)V

(6.53)

γ0,1 & π0,1:

dγ0,1

d t= kp21

(π0,1

V

)(M1

V

)V

−[(

kp12 +ktm12)(M2

V

)+ktm11

(M1

V

)+kt11

(γ0,0

V

)+kt12

(π0,0

V

)](γ0,1

V

)V

(6.54)

dπ0,1

d t=−

[ktm22

(M2

V

)+ (

kp21 +ktm21)(M1

V

)+kt22

(π0,0

V

)+kt12

(γ0,0

V

)](π0,1

V

)V

+kp12

(M2

V

)(γ0,1

V

)V

+[

kr 2

(R

V

)+ (

kp22 +ktm22)(π0,0

V

)+ (

kp12 +ktm12)(γ0,0

V

)](M2

V

)V

(6.55)

96

γ2,0 & π2,0:

dγ2,0

d t= kp21

(M1

V

)(π2,0

V

)V

−[

ktm11

(M1

V

)+ (

kp12 +ktm12)(M2

V

)+kt11

(γ0,0

V

)+kt12

(π0,0

V

)](γ2,0

V

)V

+[

kr 1

(R

V

)+ktm11

(γ0,0

V

)+ktm21

(π0,0

V

)+kp11

[2(γ1,0

V

)+

(γ0,0

V

)]](M1

V

)V

+kp21

[2(π1,0

V

)+

(π0,0

V

)](M1

V

)V

(6.56)

dπ2,0

d t=−

[(kp21 +ktm21

)(M1

V

)+ktm22

(M2

V

)+kt22

(π0,0

V

)+kt12

(γ0,0

V

)](π2,0

V

)V

+kp12

(M2

V

)(γ2,0

V

)V

(6.57)

γ0,2 & π0,2:

dγ0,2

d t= kp21

(M1

V

)(π0,2

V

)V

−[(

kp12 +ktm12)(M2

V

)+ktm11

(M1

V

)+kt11

(γ0,0

V

)+kt12

(π0,0

V

)](γ0,2

V

)V

(6.58)

dπ0,2

d t=−

[(kp21 +ktm21

)(M1

V

)+ktm22

(M2

V

)+kt22

(π0,0

V

)+kt12

(γ0,0

V

)](π0,2

V

)V

+kp12

(M2

V

)(γ0,2

V

)V

+[

kr 2

(R

V

)+ktm22

(π0,0

V

)+ktm12

(γ0,0

V

)+kp22

[2(π0,1

V

)+

(π0,0

V

)]](M2

V

)V

+kp12

[2(γ0,1

V

)+

(γ0,0

V

)](M2

V

)V

(6.59)

97

γ1,1 & π1,1:

dγ1,1

d t= kp21

(M1

V

)(π1,1

V

)V

−[

ktm11

(M1

V

)+ (

kp12 +ktm12)(M2

V

)+kt11

(γ0,0

V

)+kt12

(π0,0

V

)](γ1,1

V

)V

+[

kp11

(γ0,1

V

)(M1

V

)+kp21

(π0,1

V

)(M1

V

)]V

(6.60)

dπ1,1

d t=−

[ktm22

(M2

V

)+ (

kp21 +ktm21)(M1

V

)+kt22

(π0,0

V

)+kt12

(γ0,0

V

)](π1,1

V

)V

+kp12

(M2

V

)(γ1,1

V

)V

+[

kp22

(π1,0

V

)+kp12

(γ1,0

V

)](M2

V

)V

(6.61)

The moment balance for dead chains is:

d

d t

[V λk,l

]= ktm11M1

Vγk,l V +ktm12

M2

Vγk,l V +ktm21

M1

Vπk,l V

+ktm22M2

Vπk,l V +ktd11γk,lγ0,0V

+ktd12[γk,lπ0,0 +γ0,0πk,l

]V +ktd22πk,lπ0,0V

+ ktc11

2

k∑i=o

l∑j=0

((k

l

)(l

j

)γi , jγk−i ,l− j V

)

+ktc12

k∑i=o

l∑j=0

(k

i

)(l

j

)γi , jπk−i ,l− j V

+ ktc22

2

k∑i=o

l∑j=0

(k

i

)(l

j

)πi , jπk−i ,l− j V

(6.62)

Because of the quick dynamics of the living chains compared to the other species

present in the polymerization system (ie, monomers, dormant chains, etc.), the quasi-

stationary state hypothesis can be used in the balance for the Pi , j and Qi , j . Finally,

average molecular weights were calculated by Equations 6.63 and 6.64. Where Mn and

Mw are the number and weight average molecular weights and M1 and M2 are the

molecular weights of monomer 1 (Sty or MMA) and cardanol, respectively.

Mn =(γ1,0 +π1,0 +λ1,0

)M1 +

(γ0,1 +π0,1 +λ0,1

)M2

γ1,0 +π1,0 +λ1,0(6.63)

98

Mw =(γ2,0 +π2,0 +λ2,0

)M 2

1 + (γ1,1 +π1,1 +λ1,1)M1M2 +(γ0,2 +π0,2 +λ0,2

)M 2

2(γ1,0 +π1,0 +λ1,0

)M1 +

(γ0,1 +π0,1 +λ0,1

)M2

(6.64)

Cardanol copolymer composition and cardanol weight fraction in reaction

medium can by Equations 6.65 and 6.66.

CnM2 =λ0,1

λ1,0 +λ0,1(6.65)

ωM2 = ρm2υm2

ρm1υm1 +ρm2υm2 +ρpυp(6.66)

6.5 Parameters for homopolymerization

Methyl methacrylate and styrene have been used for decades and their kinetics are

well known and kinetics parameters very consolidated. Every parameter is a function

of the temperature and is correlated by an Arrhenius expression (Equation 6.67). Thus,

for each monomer, a set of parameters were used. Moreover, as every reaction was ini-

tiated by benzoyl peroxide (BPO), the same dissociation parameter was used in every

simulation. Gas constant (R) used was 1.987cal .g mol−1.K −1.

ki = A exp

(−E I

RT

)(6.67)

T (K ) = T (°C )+273.15 (6.68)

Table 6.2 shows the kinetic parameters and physicochemical properties used to de-

scribe the homopolymerization of styrene (m1) with benzoyl peroxide initiator (BPO).

These parameters were obtained from several works published in the literature.

Table 6.3 shows the kinetic parameters and physicochemical properties used to de-

scribe the homopolymerization of methyl methacrylate (also labeled as m1). Equa-

tions used for glass and gel effect are in Table 6.4 (PINTO and RAY, 1995).

Regarding the parameters related to cardanol, they are not available in the litera-

Table 6.1: Parameter used for benzoyl peroxide, BPO

Parameter Unit Reference

k0d = 6.120 ·1017 exp

(−30000RT (K )

)h−1 KALFAS and RAY (1993)

99

Table 6.2: Parameters for styrene


ρm1 = 924−0.918(T (K )−273.15) g ·L−1 SCORAH et al. (2006)ρp = 1080−0.6058(T (K )−273.15) g ·L−1 SCORAH et al. (2006)

kdm = 2.190 ·105 exp(−27440RT (K )

)L2 ·mol−2 · s−1 KIM and CHOI (1989)

kp11 = 3.816 ·1010 exp(−7067RT (K )

)L ·mol−1 ·h−1 TSOUKAS et al. (1982)

ktc11 = 1.26 ·109 exp(−1680RT (K )

)L ·mol−1 · s−1 KALFAS and RAY (1993)

ktd11 = 2.052 ·1018 exp(−153000

RT (K )

)L ·mol−1 ·h−1 MELO et al. (2014)

ktm11 = 2.463 ·105 exp(−10280RT (K )

)L ·mol−1 · s−1 KIM et al. (1992)

f = 0.70 - SCORAH et al. (2006)αm1 = 0.001 K −1 VILLALOBOS et al. (1993)αp = 0.00048 K −1 VILLALOBOS et al. (1993)Tg m1 = 185.0 K VILLALOBOS et al. (1993)Tg p = 370.0K K VILLALOBOS et al. (1993)υ f cr = 0.036 - VIVALDO-LIMA et al. (1998)Ag el = 0.32 - MELO et al. (2014)Ag l ass = 1.7 - MELO et al. (2014)

M Mm1 = 104.15 Da

ture. Therefore, they are going to be estimated.

100

Table 6.3: Parameters for methyl methacrylate, MMA


ρm1 = 0.9654− 1.09103 (T −273.15)−

9.7107 (T −273.15)2 g · cm−3 PINTO and

RAY (1995)

ρp = ρm1(0.754− 9.0104 (T −343.15))−1 g · cm−3 PINTO and

RAY (1995)

kdm = 0.0 L2 ·mol−2 · s−1 Considered inthis work

kp11 = 4.92 ·105 exp(−4353RT (K )

)L ·mol−1 · s−1

MAHABADIand

O’DRISCOLL(1977)

k0tc11 = 9.8 ·107 exp

(−701

RT (K )

)L ·mol−1 · s−1

MAHABADIand

O’DRISCOLL(1977)

k0td11 = 3.956 ·10−4 exp

(−4090RT (K )

)∗k0

tc11 L ·mol−1 · s−1 BEVINGTONet al. (1954)

ktm11 = 4.66 ·109 exp(

−762908.3144621∗T (K )

)L ·mol−1 · s−1 PEKLAK et al.

(2006)

f = 0.80 -LOUIE et al.

(1985)

αm1 = 0.001 K −1 PINTO andRAY (1995)

αp = 0.00048 K −1 PINTO andRAY (1995)

Tg m1 = 167.0 KPINTO andRAY (1995)

Tg p = 387.15 KPINTO andRAY (1995)

υ f cr = 0.04227 -Considered in

this workυ f tc =

0.01856−2.965 ·10−4(T −273.15)-

PINTO andRAY (1995)

M Mm1 = 100.121 Da -

Table 6.4: Equations for gel and glass effects for methyl methacrylate, MMA

gT ,M M A

0.10575 ·exp(17.15υ f −0.01715(T −273.15) υ f > υ f tc

2.3 ·10−6 exp(75υ f ) υ f ≤ υ f tc

gP ,M M A

1 υ f > υ f cr

7.1 ·10−5 exp(171.53υ f ) υ f > υ f tc

101

6.6 Parameter estimation

As discussed by SCHWAAB, M AND PINTO, J.C. (2007), parameter estimation con-

sists in utilizing a reference model and changing the parameters until the predictions

of the model fits the experimental data, respecting the uncertainties of the measure-

ment.

Thus, all balance equations were solved numerically using backward differentia-

tion formula (BDF) as available in the DASSL code with relative and absolute toler-

ances of 10−4. Model parameters were estimated using the computation package ES-

TIMA, implemented in Fortran, which combines particle swarm (stochastic) optimiza-

tion (PSO) and the deterministic Gauss-Newton algorithms (SCHWAAB et al. (2008))

with the accelerator of LAW and BAILEY (1963) (deterministic).

The Gauss-Newton method frequently fails when dealing with ill-posed problems,

making impossible the determination of parameter uncertainties (as the covariance

matrix of parameter uncertainties cannot be obtained). In these cases, first, particle

swarm optimization can be applied. Then, with the point of minimum obtained, an

identifiability procedure can be applied to determine which set of model parameters

can have their uncertainties evaluated. Afterwards, ESTIMA can be applied to estimate

the selected parameters and refine the minimum of the previously estimated model

parameters (BRANDÃO (2017)).

Figure 6.1: Employed parameter estimation procedure (BRANDÃO (2017))

102

One hundred particles were used and ten thousand iterations were carried out with

statistical confidence level of 95%. To avoid numerical problems, due to the differences

of orders of magnitude among the parameter values, some model parameters were

estimated in their logarithmic form.

For the sake of simplicity, each estimation of the model parameters corresponded

to a each experimental condition. Then, the estimated parameters were analyzed to-

gether in order to evaluate the consistency of the obtained results. The objective func-

tion used in the present work was defined as follows:

Fob j =(ye − ym (

xm ,θ))T V −1

y

(ye − ym (

xm ,θ))

(6.69)

where y and ym are the vectors of the measured and predicted dependent vari-

ables, respectively; Vy is the covariance matrix of the measured outputs (assumed to

be diagonal); and xm and θ are the vectors of the measured independent variables and

model parameters, respectively. The measured variables considered were conversion,

number and weight average molecular weights (Mn and Mw ).

6.7 Simulations

For poly(styrene-co-cardanol) simulations, besides data previously reported of

gravimetric analysis for determination of overall monomer conversions and average

molecular weights obtained by gel permeation chromatography, it was also used data

from GALVÃO (2016) to have more degrees of freedom and reliability during the esti-

mations. However, as she did not have data for 10wt %, only the experimental condi-

tions of Table 6.5 were analysed initially by the model.

Regarding the parameters related to cardanol or cashew nut shell liquid (CNSL),

they are not available in the literature, and either for copolymerization. However,

as cardanol and natural CNSL had a similar behaviour polymerizing with methyl

Table 6.5: Experimental conditions used for parameter estimation

ExperimentalCondition

Styrene(wt%)

Cardanol(wt%)

BPO(wt%)

T(°C )

ReactionDuration

(h)

SampleInterval

(min)

A1 96.5 2.5 1.0 110 5 30A2 94.0 5.0 1.0 110 5 30A3 96.5 2.5 1.0 90 5 30A4 94.0 5.0 1.0 90 5 30A5 99.0 0.0 1.0 90 3 15A6 99.0 0.0 1.0 110 3 15

103

Table 6.6: Parameters used for cardanol


ktm21 = ktm12 = kp22 = 0 L ·mol−1 ·h−1 Considered in this workktc22 = ktc11 L ·mol−1 ·h−1 Considered in this work

methacrylate (Figs. 5.33 and 5.34), their kinetic parameters would also be the same be-

cause the kinetic model considers that the unsaturations in the side chain are respon-

sible for the polymerization, and the only difference between anarcardic acid (main

component of cashew nut shell liquid) and cardanol is the presence of a carboxylic

acid in the phenolic ring, which does not influence in the radical polymerization but

can be used for polymer functionalization. Also, based on their structures, if one as-

sumes that only terminal double bonds can react through free-radical polymerization

mechanisms, one will conclude that just 40% of the phenolic stream can be available as

monomer for the polymerization. Despite that, given the fact that other molecules may

be eventually present in the naturally occurring cardanol or anacardic acid stream and

that different reactive sites (such as the phenolic group) can take part in the reaction

mechanism, the whole stream is treated as a lumped monomer stream for purposes of

kinetic modeling.

It is important to emphasize that cardanol polymerizes at slow rates at the analyzed

reaction conditions through free-radical polymerizations, as observed experimentally

(section 5.2 - homopolymerization of cardanol and anacardic acid) therefore, for the

sake of simplicity, some model parameters were initially assumed to be either null or

equal to the styrene polymerization parameters, as shown in Table 6.6. Particularly, one

must observe that kp22 was assumed to be equal to zero, imposing the modification of

the normal reactivity ratio scheme used to describe the copolymerization.

Therefore, the parameters to be found are kp12 and kp 21, which are dependent only

of kp11. Thus, it was estimated r1,2 and r2,1 defined as:

kp12 = kp11 · r1,2 (6.70)

kp21 = kp11 · r2,1 (6.71)

Although r1,2 and r2,1 resemble the classical reactivity ratios, they are not the re-

activity ratios, as kp22 is assumed to be equal to zero because cardanol practically

does not homopolymerize through regular free-radical polymerizations at the ana-

lyzed conditions. Despite that,r1,2 and r2,1 are parameters that indicate the relative

magnitudes of the rates of cross propagation in respect to the rate of styrene homo-

104

propagation. Moreover, it is also estimated fi nh . This constant correspond to the in-

hibition effect of cardanol and it is used to reduce the amount of living radicals in the

reaction media. Thus, the initial initiator concentration is multiplied by this inhibi-

tion factor, which implies that inhibitors can react very fast with the primary radicals

and interrupt the course of the polymerization, reducing the efficiency of the initiation

step.

6.7.1 Styrene Homopolymerization

Thus, using experimental conditions A5 and A6, the kinetic parameters of Table 6.2

can be tested. A correction factor δtm was estimated for the kinetic constant for chain

transfer to monomer (ktm11 = δtmkt m11) to account for possible presence of impuri-

ties. Table 6.7 presents the estimated parameter for reactions at both temperatures.

As shown in Table 6.7, estimated values of the kinetic constant for chain transfer to

monomer ktm11 were very close to the real reference literature parameter value ktm11,

indicating the adequacy of the proposed modeling approach. The quality of the model

fit was evaluated with the chi-square statistical test. According to this test, the model

can be considered suitable since the final objective function value, Fob j , lay between

the limits provided by the χ2 distribution. Therefore, as shown in Figure 6.2 and 6.3, the

proposed model fits the available experimental data adequately. Thus, the homopoly-

merization model is trustable and ready for copolymerization analyzes.

Thus, for copolymerization parameter estimation, except for the constant of chain

transfer to styrene that was fixed as ktm11 = 1.211k0tm11 in which corresponds to the

estimated value of this kinetic constant for experimental condition A6, all the other

kinetic parameters used are presented in Table 6.2.

Table 6.7: Estimated model parameters for styrene homopolymerization reactions.

Exp. Cond. Fob j χ2i n f χ2

sup δtm

A5 37.58 16.85 45.72 0.849±0.129A6 35.40 16.85 45.72 1.211±0.113

105

(a) (b)

Figure 6.2: Predicted and experimental results for styrene homopolymerization at90°C (condition A5). (MCI = model confidence interval; MP = model predictions and

ED = experimental data).

(a) (b)

Figure 6.3: Predicted and experimental results for styrene homopolymerization at110°C (condition A6). (MCI = model confidence interval; MP = model predictions and

ED = experimental data).

106

6.7.2 Styrene/Cardanol Copolymerizations at 110°C

Table 6.8 presents the estimated model parameters for styrene/cardanol copoly-

merization reactions. At both conditions, A1 and A2, the final objective function values

lay inside the most likely range defined by the χ2 distribution distribution, indicating

the adequacy of the proposed model. Figure 6.4 shows the joint confidence region for

parameters r1,2 and r2,1 when parameter fi nh was fixed at its optimal value shown in

Table 6.8. To generate this figure, a new estimation procedure was run but this time

fi nh was fixed at 0.731. The new objective function obtained was 21.34 and the esti-

mated values for r1,2 and r2,1 were 1.628 and 0.026 respectively.

Figure 6.4: Joint confidence region for parameters r1,2 and r2,1 when parameter fi nh

was fixed at 0.731. The red center dot is the estimated values of r1,2 and r2,1 (1.628,0.026).

Table 6.8: Estimated model parameters for cardanol/styrene copolymerizationreactions performed at 110°C .

A2 A1

Exp. Cond.Estimated

ValueConfidence

IntervalEstimated

ValueConfidence

IntervalFob j 21.34 15.31 - 44.46 41.39 16.79 - 46.98fi nh 0.731 0.721 - 0.740 0.888 0.814 - 0.962r1,2 1.627 1.589 - 1.666 r1,2estimated at A2r2,1 0.023 0.022 - 0.031 r2,1estimated at A2

107

As one can see in Table 6.8, fi nh decreased with the cardanol feed concentration,

which caused the simultaneous increase of the inhibitor concentration in the react-

ing medium, as already described. Therefore, the concentration of active radicals is

expected to diminish with the amount of cardanol in the feed. Regarding the esti-

mated values of r1,2 and r2,1, it must be considered that the incorporation of cardanol

molecules by propagating chains can cause the steric hindrance of the active centers

of growing macroradicals, justifying the low kp21 values (kp21 = 0.026 ·k0p11). Also, one

must consider that more stable radicals can be formed due to the presence of the phe-

nolic ring and internal unsaturation, retarding the polymerization rate. It is interest-

ing to observe that the incorporation of cardanol molecules seems preferable than the

incorporation of styrene at initial times because styrene forms radicals that are more

stable and less reactive because of the proximity of the aromatic ring, while cardanol

forms secondary radicals that are not stabilized by the phenolic ring. Moreover, al-

though only the terminal double bond seems responsible for the usual addition steps,

cardanol presents multiple reactive sites that are also prone to reaction and formation

of radicals (as suggested by the 1H−N MR spectra), which can also retard the polymer-

ization, leading to small kp21 values. Figures 6.5 and 6.6 present the available experi-

mental data and simulation results for monomer conversion and average molecular

weights, illustrating the very good quality of the obtained model fits.

(a) (b)

Figure 6.5: Predicted and experimental results for styrene/cardanol copolymerizationat condition A1. (MCI = model confidence interval; MP = model predictions and ED =

experimental data).

The obtained results are strongly connected to the relative decrease of radical con-

centration (P(i , j )andQ(i , j )) in the reacting system, due to the previously described

inhibitory effects of phenolic compounds. As one can see in Figures 6.5 and 6.6,

monomer conversion of copolymer samples was smaller at condition A2 (with 5 wt%

of cardanol) than at condition A1 (2.5 wt% of cardanol). Nevertheless, as illustrated

108

(a) (b)


experimental data).

in Figure 6.7-a, despite the inhibitory effect exerted by cardanol, it was possible to ob-

tain copolymers with relatively high amounts of cardanol incorporated into copolymer

chains at both conditions, due to the relatively high cardanol reactivity of the polyun-

saturated cardanol molecules. Since styrene thermal initiation is more pronounced at

110°C, reaction rates were kept almost constant (although at low values) after approxi-

mately 1 hour of reaction (Figure 6.5b and 6.6b). Regarding the cardanol composition,

it is possible to observe in Figure 6.7 - a that model simulations indicated that cardanol

copolymer compositions did not change very significantly during the reaction course.

Besides, Figure 6.7 suggests that cardanol was almost completely consumed at experi-

mental condition A1, while at condition A2 a residual amount of nonreacted cardanol

remained in the reacting medium. Despite the larger amount of cardanol incorporated

into copolymer chains at condition A2, higher inhibition effects could also be noticed

at this condition.

109

(a) (b)

Figure 6.7: Predicted and experimental results for cardanol compositions atconditions A1 and A2. (MP = model predictions and ED = experimental data).

6.7.3 Styrene/Cardanol Copolymerizations at 85°C

At 90°C, all three parameters were selected for estimation again. The constant of

chain transfer to styrene was fixed as ktm11 = 0.849 ·k0tm11 which corresponds to the es-

timated value of this kinetic constant for experimental condition A5. Table 6.9 presents

the estimated model parameters for styrene/cardanol copolymerization reactions per-

formed at 90°. As one can observe, cardanol exerted stronger inhibitory effects at the

lower temperature. At 90°, lower concentrations of radicals are expected due to the

smaller decomposition rates of BPO and thermal initiation rates of styrene. In ad-

dition, as radicals react with cardanol due to r1,2, they get stabilized and react more

slowly due to steric hindrance, making more difficult the propagation of cardanol ter-

minated radicals. Figure 6.8 shows the joint confidence region for parameters r1,2 and

r2,1 when parameter fi nh was fixed at its optimal value shown in Table 6.9. To generate

this figure, a new estimation procedure was run but this time fi nh was fixed at 0.628.

The new objective function obtained was 43.44 and the estimated values for r1,2 and

r2,1 were 3.172 and 0.015 respectively.

Table 6.9: Estimated model parameters for cardanol/styrene copolymerizationreactions performed at 90°.

A3 A4

Exp. Cond.Estimated

ValueConfidence

IntervalEstimated

ValueConfidence

IntervalFob j 43.44 15.31 - 44.46 36.29 16.79 - 46.98fi nh 0.628 0.609 - 0.649 0.572 0.552 - 0.593r1,2 3.210 2.356 - 4.372 r1,2estimated at A3r2,1 0.015 0.010 - 0.025 r2,1estimated at A3

110

Figure 6.8: Joint confidence region for parameters r1,2 and r2,1 when parameter fi nh

was fixed at 0.628. The red center dot is the estimated values of r1,2 and r2,1 (3.172,0.015).

As one can see in Table 6.9, the values estimated for the cross-propagation rate

constants at 90°were not much different from the ones estimated at 110°(Table 6.8),

mainly for r 2,1. This indicates that the reaction temperature does not seem to exert a

strong effect on the relative kinetic constants for cardanol incorporation. The apparent

increase of r 1,2 can possibly be related to more complex mechanistic effects, such

as penultimate cross-propagation and cross-termination effects, which can become

more evident at lower rates of reaction. Moreover, other effects such as less intense

thermal initiation of styrene and lower reactivity of benzene radicals can also lead to

lower rates of reaction. Again, the obtained objective function values lay in the chi-

square confidence interval, showing that the model could predict well the available

experimental data. Figures 6.9 and 6.10 illustrate the obtained experimental data and

simulation results for monomer conversion and average molecular weights.

As previously discussed, conversion and average molecular weights presented

lower values at condition A4 (with 5% of cardanol) than the ones obtained at con-

dition A3 (with 2.5 % of cardanol), due to inhibition effects. Regarding the cardanol

copolymer and medium compositions for conditions A3 and A4, the obtained model

indicates higher cardanol copolymer compositions at condition A4 than at condition

A3, as shown in Figure 6.11, despite the lower reaction rates at condition A4, as a conse-

111

(a) (b)


experimental data).

quence of the higher cardanol concentrations in the feed. However, cardanol copoly-

mer compositions obtained in copolymerizations performed at 90°were expected to

be lower than the ones obtained at 110°, due to the lower rates of radical generation by

thermal styrene initiation at 90°.

112

(a) (b)

Figure 6.10: Predicted and experimental results for styrene/cardanolcopolymerization at condition A4. (MCI = model confidence interval; MP = model

predictions and ED = experimental data).

(a) (b)

Figure 6.11: Predicted and experimental results for cardanol compositions atconditions A3 and A4 (MP = model predictions).

113

6.7.4 Styrene/Cardanol Copolymerizations at 110°C for 10wt %

Using the experimental values obtained for the copolymerization at 110°C for 10wt

% (Section 5.3 - copolymerization of cardanol and styrene), the kinetic model can be

validated once again to see if has a good behavior at this "extreme" condition of 10wt

%. The experimental values for 85°C won’t be used because they were not satisfactory

enough as the inhibition was high, and molecular weights and composition data could

not be obtained. Therefore, Figure 6.12 shows the conversion and molecular weights

obtained experimentally and the model prediction. It can be seen that the model was

able to fit the data correctly. Thus the parameters obtained are shown in Table 6.10.

(a) (b)

Figure 6.12: Predicted and experimental results for cardanol copolymerization at110°C for 10wt % (MP = model predictions).

For the conditions A1 and A2, r1,2 and r2,1 obtained were 1.627 and 0.023, respec-

tively. They are not far for the values obtained at this same temperature. Moreover,

as the cardanol is been treated as a lump, the difference could happen. Beside it, fi nh

corresponded to 0.888 for 2.5wt % (condition A1), 0.731 for 5.0wt % (condition A2) and

0.478 for this condition, which was expected since the inhibition is higher (i.e. de-

creasing the concentration of radicals produced) due to higher amount of cardanol.

Figure 6.13 shows the copolymer composition for this condition and that the model

Table 6.10: Estimated model parameters for cardanol/styrene copolymerizationreactions performed with 10wt % at 110°.

10wt %; 110°C

Exp. Cond. Estimated Value Confidence IntervalFob j 15.00 6.26 - 27.49fi nh 0.478 0.439 - 0.519r1,2 1.765 1.699 - 1.833r2,1 0.016 0.012 - 0.021

114

was also able to fit this data. Therefore, the model is capable to reproduce experi-

mental results in the range of concentrations and temperature studied, following well

conversion, molecular weights and copolymer composition.

Figure 6.13: Predicted and experimental results for cardanol compositions at 110°Cfor 10wt % (MP = model predictions).

6.7.5 Methyl Methacrylate/Cardanol Copolymerizations at 85°C

For both temperatures (85 and 110°C), methyl methacrylate-cardanol copolymer-

ization can have its parameters estimated. However, due to a small number of exper-

imental data, results may not be so accurate. Nonetheless, the same kinetic model

proposed for styrene-cardanol copolymerization is used for this case as well as the pa-

rameters estimated. Using the previous strategy, the estimation happened using ex-

ponentials. Initially, parameter ranges were free, being the only limitations a positive

value for r1,2 (as cardanol has multiple reactive sites and can form active sites more re-

active than those for MMA) and a negative value for r2,1 (due to cardanol small kinetic

rates) and for fi nh (to simulate the inhibition behavior). These considerations were

made for the copolymerization with styrene. However, for 2.5 and 10wt %, at least a

little good fit was not possible in this way. For 5 wt%, Fob j = 61.36(4.4 ≤χ2 ≤ 23.33) was

not good, but more reasonable than the other two concentrations. Thus MMA has a

higher reactivity with molecules of your same type than styrene does. That is kp12 is

smaller than kp11 for this copolymerization. Therefore, if r1,2 is set as a negative value,

the objective function become better for 2.5 and 10wt %. Fixing r1,2 for 10wt% around

115

Table 6.11: Estimated model parameters for cardanol/MMA copolymerizationreactions, 85°C .

Exp.Cond.

2.5w t % 5w t % 10w t %

Esti-matedValue

ConfidenceInterval

Esti-matedValue

ConfidenceInterval

Esti-matedValue

ConfidenceInterval

10r1,2 −0.275−0.354 ⇐⇒

−0.195Same for 2.5w t % Same for 2.5w t %

10r2,1 −4.362−5.606 ⇐⇒

−3.118−4.140

−4.139 ⇐⇒−4.141

−4.629−5.705 ⇐⇒

−3.553

10 fi nh −0.035−0.040 ⇐⇒

−0.030−3.83∗

10−6

−3.95∗10−6 ⇐⇒

−3.71∗10−6−0.358

−0.428 ⇐⇒−0.287

Fob j 27.143.25 ≤χ

2 ≤20.48

105.035.01 ≤χ

2 ≤24.74

3.962.18 ≤χ

2 ≤17.53

the value obtained for the other two concentrations, parameters estimated are shown

in Table 6.11.

Still, for 5wt%, although r2,1 was similar to the others r2,1 estimated, Fob j increased

and the value for fi nh does not represent cardanol inhibition behavior. The problem

with 5wt% is that if molecular weights are got, copolymer composition is not and vice-

versa. Although Fob j is higher than what is imposed by chi-square distribution, the

experimental values were represented well. Small degrees of freedom is one of the rea-

sons for this. However, the model is not linear in the parameters. Thus, the probability

distribution of the parameters is not normal even if the deviations between the ex-

perimental values and the values calculated by the model have a normal distribution.

Thus, the objective function has no obligation to be within the limits of the chi-square

distribution. Graphics using the values of Table 6.11 are presented below. For model

prediction, only the points that have an experimental value to compare to are shown.

116

Figure 6.14: Predicted and experimental results for MMA/cardanol copolymerconversion and molecular weights at conditions 2.5 wt % (MP = model predictions).

Figure 6.15: Predicted and experimental results for cardanol composition in aMMA/cardanol copolymer at 85°C and 2.5wt % (MP = model predictions).

Figure 6.16: Predicted and experimental results for MMA/cardanol copolymerconversion and molecular weights at conditions 5 wt % (MP = model predictions).

117

Figure 6.17: Predicted and experimental results for cardanol composition in aMMA/cardanol copolymer at 85°C and 5wt % (MP = model predictions).

Figure 6.18: Predicted and experimental results for MMA/cardanol copolymerconversion and molecular weights at conditions 10wt % (MP = model predictions).

Figure 6.19: Predicted and experimental results for cardanol composition in aMMA/cardanol copolymer at 85°C and 10wt % (MP = model predictions).

118

As molecular weight and copolymer composition could not be fitted simultane-

ously for 5wt% while conversion was good, this was a sign that perhaps the model have

to be adapted for MMA/cardanol copolymerization by adding a monomer transfer re-

action. This modification is chosen because transfer reactions do not affect on the

conversion, but are able to decrease molecular weights. km22 is considered to be 0 be-

cause radicals that present on a cardanol terminal unit are not reactive. km21 is also

maintained at 0 for the same reason. Thus, a fourth parameter is estimated to obtain

ktm12: t1,2. To begin the estimation, parameters boundaries are set as presented in Ta-

ble 6.12. fi nh smallest value was set to 0.01 to not reduce profoundly the concentration

of radicals, which would not correspond to real polymerization.

ktm,12 = ktm,11 ·10t1,2 (6.72)

Table 6.13 have the values obtained for the 4 parameters and the objective function.

It can be seen that r1,2, r2,1, and t1,2 are similar to all concentrations. t1,2 obtained is a

high value, showing that radicals present on MMA-terminal units are very reactive, be-

ing capable of reacting with themselves and MMA molecules, but also attacking hydro-

gens on cardanol molecules - and there is a vast amount of options to choose from. Af-

ter this reaction, polymers are terminated with that MMA unit and a cardanol molecule

stays with a radical. But, as it was already discussed, this radical might be not very re-

active, retarding the polymerization. So, MMA-terminated polymers do not propagate

anymore and reduce the molecular weight. fi nh decreases with the initial amount of

cardanol, increasing the inhibition effect while also reducing conversion and molecu-

lar weight. It is also interesting to observe that MMA/cardanol copolymers have less

cardanol amounts than styrene/cardanol copolymers does (e.g. less than 3wt% for

10wt% initially added). The formation of a high amount of cardanol molecules with

a radical present without being part of a polymer unit can be one of the reasons why

this is happens. Another point is that fi nh values obtained are smaller than what was

gotten for the styrene/cardanol system, reinforcing the idea that radicals present in a

molecule containing only the cardanol are formed, retarding the polymerization and

decreasing the concentration of cardanol in the final polymer.

Table 6.12: Boundaries for estimated model parameters considering ktm12.

Par Inf Par Sup

10r1,2 -4.0 010r2,1 -4.0 0.010 fi nh -2.0 0.010t1,2 -4.0 5.0

119

Table 6.13: Estimated model parameters for cardanol/MMA copolymerizationreactions considering ktm12, 85°C .

2.5w t % 5w t % 10w t %

10r1,2 -0.8010 -0.7331 -0.672010r2,1 -4.3125 -3.7979 -4.088710 fi nh -0.5000 -1.0018 -1.499710t1,2 3.6978 3.4369 3.7931Fob j 23.62 (2.70≤χ2 ≤19.02) 27.78 (3.82≤χ2 ≤21.92) 2.09 (1.69≤χ2 ≤16.01)

Table 6.14: Estimated model parameters (without the exponentials) forcardanol/MMA copolymerization reactions considering ktm12, 85°C .

2.5w t % 5w t % 10w t %

r1,2 0.158 0.185 0.213r2,1 4.870 ·10−5 1.592 ·10−4 8.153 ·10−5

fi nh 0.316 0.100 0.032t1,2 4986.089 2734.891 6210.549

Again, even with the new parameter, only the objective function for 10wt% is inside

the chi-square distribution boundaries. However, both objective functions for 2.5 and

5wt% decreased, showing that model predictions were able to fit experimental data

better. It is also important to remember that a small amount of experimental data is

available and the model is not linear. Graphics using the estimated model parameters

considering t1,2 are below. Now, notice that both cardanol composition and molecular

weights are well represented for 5wt%.

Because cardanol copolymer composition is almost the same for 5 and 10wt%, but

still higher for 5wt%, r2,1 has to be bigger than other two r2,1. However, at the same

time, fi nh still decreases with an increase in cardanol amount, while MMA molecules

tend to attack more cardanol molecules (due to higher amount cardanol amount

around them) when the cardanol concentration is also increased. But, perhaps be-

cause more cardanol molecules are present in a growing polymer (instead of a dormant

molecule with a stable radical), the transfer reaction does not occur as frequently for

5wt% than for the other concentrations - 10wt% has the higher t1,2 obtained value be-

cause of a higher cardanol concentration.

120

Figure 6.20: Predicted and experimental results for MMA/cardanol copolymerconversion and molecular weights at conditions 2.5 wt % considering a 4th parameter,

t1,2 (MP = model predictions).

Figure 6.21: Predicted and experimental results for cardanol composition in aMMA/cardanol copolymer at 85°C and 2.5wt % considering a 4th parameter, t1,2 (MP

= model predictions).

121

Figure 6.22: Predicted and experimental results for MMA/cardanol copolymerconversion and molecular weights at conditions 5 wt % considering a 4th parameter,


Figure 6.23: Predicted and experimental results for cardanol composition in aMMA/cardanol copolymer at 85°C and 5wt % considering a 4th parameter, t1,2 (MP =

model predictions).

122

Figure 6.24: Predicted and experimental results for MMA/cardanol copolymerconversion and molecular weights at conditions 10wt % considering a 4th parameter,


Figure 6.25: Predicted and experimental results for cardanol composition in aMMA/cardanol copolymer at 85°C and 10wt % considering a 4th parameter, t1,2 (MP =

model predictions).

123

6.7.6 Methyl Methacrylate/Cardanol Copolymerizations at 110°C

For 110°C, molecular weights for 5wt% are still not available. So, only parameters

for 2.5 and 10wt% were estimated. Not considering t1,2, objective functions are 38.24

and 35.21 for 2.5 and 10wt% respectively. Again, when cardanol composition is good,

molecular weights are not well represented. When t1,2 started to be estimated, objec-

tive function for both concentrations decreased and got inside chi-square distribution

(even though Fob j cannot be inside the distribution while the model is representing

well the experimental data, chi-square still can be a good parameter to see how good

the estimation procedure is). t1,2 are a bit lower than those obtained for 85°C, but

still higher. r1,2 decreased quite a lot (roughly 103). This is because MMA becomes

even more reactive at high temperatures, but cardanol still retards the polymerization

(maintaining small value for fi nh). However, r2,1 remains the same, showing that car-

danol reactivity truly does not increase with temperature - and even less if the poly-

merization is with MMA.

The parameters obtained also show that MMA attacks more than 2 times when

more cardanol is present (higher value for r1,2, because more cardanol molecules are

around waiting to react. However, because more stable radicals are formed, the poly-

merimerization is more retarded (smaller fi nh for 10wt%). And more cardanol in the

reaction medium does not translate in a higher cardanol composition in the copoly-

mer, therefore a higher r1,2 is obtained for 2.5wt% to obtain the necessary composi-

tion. But, because full conversion is reached so fast for 2.5wt% (similar for MMA ho-

mopolymerization) while the molecular weight is still small than the MMA homopoly-

mer, monomer transfer reaction is responsable for this. Molecular weight for 10wt% is

similar to the obtained for 2.5wt%, reflecting in a similar t1,2.

Table 6.15: Estimated model parameters for cardanol/MMA copolymerizationreactions considering ktm12, 110°C .

2.5w t % 10w t %

10r1,2 -3.956 -3.54010r2,1 -4.000 -4.27610 fi nh -0.730 -0.99510t1,2 3.267 3.094Fob j 11.46 (2.70≤χ2 ≤19.02) 21.87 (3.82≤χ2 ≤21.92)

124

Table 6.16: Estimated model parameters (without exponentials) for cardanol/MMAcopolymerization reactions considering ktm12, 110°C .

2.5w t % 10w t %

r1,2 1.107 ·10−4 2.885 ·10−4

r2,1 1.000 ·10−4 5.295 ·10−5

fi nh 0.1864 0.1012t1,2 1849.6945 1241.7953

Figure 6.26: Predicted and experimental results for MMA/cardanol copolymerconversion and molecular weights at 110°C and 2.5 wt % considering a 4th parameter,


Figure 6.27: Predicted and experimental results for cardanol composition in aMMA/cardanol copolymer at 110°C and 2.5wt % considering a 4th parameter, t1,2 (MP


125

Figure 6.28: Predicted and experimental results for MMA/cardanol copolymerconversion and molecular weights at 110°C and 10wt % considering a 4th parameter,


Figure 6.29: Predicted and experimental results for cardanol composition in aMMA/cardanol copolymer at 110°C and 10wt % considering a 4th parameter, t1,2 (MP



A kinetic mechanism was proposed and shown to represent adequately the poly-

merization system under study. Firstly, a correction factor for styrene homopolymer-

ization was estimated (ktm11 = 1.211 ·k0tm11), and experimental conversion and molec-

ular weights for styrene polymerization were well represented by the model. Then, it

was used this corrected kinetic parameter and the others of Table 6.7 to calculate car-

danol parameters. The kinetic model proposed for the copolymerization was able to

follow experimental data although some assumptions were used such as a) not consid-

ering that cardanol represents a mixture of similar components with a variable number

of unsaturations; b) considering that cardanol has many reactive sites - terminal unsat-

uration, internal unsaturations, and a hydroxyl group; c) considering that cardanol in-

126

hibits the polymerization due to its internal unsaturations and hydroxyl group, which

required the estimation of fi nh parameter to decrease the concentration of initiator;

d) considering that cardanol does not homopolymerize and it has a propagation con-

stant (kp22) equal to zero; e) considering that the living polymeric chains ending in

cardanol also terminates with the same rate of polystyrene; f) considering the quasi-

state hypothesis and the terminal model for the polymerization; g) considering that the

polymerization takes place at the external double bond; h) considering that kinetic ex-

periments were performed at truly constant temperature; i) considering that cardanol

does not have impurities.

Moreover, transfer reactions, although possible (mainly due to steric hindrance and

great availability of hydrogens), were not considered for the cardanol mixture. Also,

The terminal unsaturation was the functional group most likely to continue propa-

gation and incorporate cardanol. But, although cardanol only has less than 50% of

compounds with terminal unsaturation, it was considered that the whole lump was

reactive in the kinetic model proposed because, as explained, the compounds have

many reactive sites. Nonetheless, the copolymerization parameters r1,2 and r2,1 were

shown to depend little on the reaction temperature in the analyzed experimental range

and to be equal to 1.627 - 3.210 and 0.015 - 0.026, respectively. As detecting experimen-

tally, the inhibition of cardanol at 90°C is more intense than at 110°. This reflected in a

lower value for fi nh obtained for 90°C (0.628) than for 110°C (0.731), meaning that it re-

duced the concentration of available radical molecules more the last condition. From1H −N MR analysis, it was possible to obtain copolymer composition and, therefore,

predict cardanol compositions in the copolymer and in the reaction medium. For the

conditions studied, cardanol is incorporated at the begging quite fast (r1,2 > 1.0) during

early reaction times due to its multiple reactive sites.

For MMA/cardanol copolymerizations, without the monomer transfer reaction (a

MMA-ended living chain attacking a cardanol’s hydrogen), the model could not repre-

sent correctly experimental data. MMA is much more reactive than styrene, therefore

r1,2 became lower than 1.0. The parameter for chain transfer reaction (t1,2) turned out

to be quite high to represent correctly molecular weights while still maintaining car-

danol composition in the copolymer. With this adapted model in relation to the one for

styrene, for MMA copolymerization, the model was able to follow experimental data at

every studied composition (2.5, 5 and 10wt%) and temperature.

127

Kinetic Study of Suspension

Copolymerization

2nd Section

128

Chapter 7

Suspension Copolimerization:

Methodology

7.1 Materials

Cardanol (Taian Health Chemical, minimum purity of de 99.5% m/m, Tai’an,

China), methyl methacrylate (MMA) (VETEC, Rio de Janeiro – Brazil, 99.5%) and

styrene (STy) (Sigma-Aldrich, Rio de Janeiro – Brazil, 99%) were used as monomers.

Benzoyl peroxide (BPO) (VETEC, Rio de Janeiro – Brazil, 99%, containing 25% of hu-

midity) and azobisisobutyronitrile (AIBN) (donated) was used as an initiator for the

polymerizations. Poly(vinyl alcohol) (PVA) (Vetec Química Fina, Mw = 78 kDa, 86.5 –

89.5% of hydrolysis) was used as a suspension agent.

7.2 Set Up

For suspension polymerization, it was used a 1L glass reactor (Figure 7.1). The re-

actor has a jacket that receives a 50/50 mixture of hot water and ethylene glycol from a

controlled hot bath. It also has a metal cap that fits the mechanical stirrer, reflux con-

denser and a thermocouple that is connected to the bath; and has holes for monomer

feed and holes for aliquot removals.

7.3 Kinetic Experiments

For the suspension polymerization, different phases are prepared separately. The

aqueous one is prepared on the day before the reaction to solubilize the PVA. The

129

Figure 7.1: Set Up for Suspension Polymerizations

organic is prepared by mixing the monomer with the initiator only instants before

the polymerization starts. Every reaction was conducted at 85°C with agitation of

850 ± 50r pm. For methyl methacrylate polymerizations, samples were taken for 2

hours. For styrene for 4 hours. Each aliquot was every 15/30 minutes.

On the reaction day, the following procedure happened: (1) heating bath was

turned on with 85°C as set point; (2) PVA solution was added to the reactor; (3) stir-

ring was turned on and kept around 500 rpm; (4) when the temperature of the PVA

solution reached the desired one, the monomer was mixed with the initiator; (5) the

monomer mixture was added to the reactor and the stirring rate was changed to 850

rpm; (6) samples were taken in specifics intervals and a solution of hydroquinone was

added to stop the polymerization; (7) the heating bath was turned off after the planned

reaction time; (8) after cooling, some reaction samples were filtered under vacuum, but

other had very tiny particles that could not be filtered; (9) the samples were left at the

recirculating and vacuum oven until dry or were dried by freeze-drying.

Conditions for the suspension polymerizations are summarized in Table 7.1. Con-

centrations are presented in weight fraction. X stands for cardanol and Y for styrene

(Sty) or methyl methacrylate (MMA). Moreover, 1 wt% or 3 wt% of BPO was added

in relation to the total mass of monomer. Sometimes, BPO was substituted by AIBN.

Thus, for reactions in a 1 L reactor, the follow masses were the real values used: 180g of

monomer, 418.32g of water and 1.68g of PVA.

130

Table 7.1: Conditions for Suspension Polymerizations

ExperimentsOrganic Phase (30%) Aqueous Phase (70%)

T (°C)

StirringRate

X (wt%) Y (wt %)Water(wt%)

PVA(wt%)

(rpm)

HomoY 0 99 99.6 0.4 85 8502.5% X 2.5 96.5 99.6 0.4 85 8505% X 5 94 99.6 0.4 85 850

10% X 10 89 99.6 0.4 85 850

7.4 Gravimetric Analysis

Monomer conversion was obtained by gravimetric analysis. This method is based

on the theory that every monomer reacted becomes part of a polymer, admitting that

there is not formation of sub-products. For this, the crucible is weighted before the

addition of the sample, and the mass added of hydroquinone added in each tube is

known. The crucibles containing the polymer were dried in the recirculating followed

by a vacuum oven and weighted several times until the mass became constant.

To obtain the conversion, Equations 7.1 - 7.3 were used. However, mpol ymer had to

be multiplied by 0.3 to consider only the monomer/polymer fraction. Therefore, the

equations used to obtain the conversion of samples obtained with suspension poly-

merization were:

mmonomer = mi ni t i al −mcr uci bl e −mBPO (7.1)

mpol ymer = mdr y −mcr uci bl e −mBPO −Csol ,hydr oqui none mhydr o (7.2)

Conver si on (%) = mpol ymer

0.3mmonomer·100% (7.3)

For mBPO , it was considered that it corresponded to 1wt % of the mmonomer , as this

was the amount added during the preparation of the reaction mixture. Csol .hydr o was

0.01 as the reactions were stopped using an alcoholic solution of hydroquinone (1wt

%).

131

7.5 Characterizations

7.5.1 Contact Angle

To determine the degree of hydrophobicity of the polymeric materials produced, a

tensiometer Teclis was used. Initially, the particles were compressed in a disk model

using a press, forming tablets. Each sample was placed over a acrylic superficie, and

a drop of water as placed over to obtain contact angle between the surfaces under at-

mospheric conditions at 25°C using the equipment’s software. Such an angle is used as

an inference of the degree of hydrophobicity of the polymeric particles. The analyzes

were made in triplicate.

7.5.2 Interfacial tension

The interfacial tension was calculated between the organic phase (containing only

monomers, without BPO) and water (without PVA) at 25 and 85°C . This analysis were

performed to see the influence of cardanol as an suspension agent. Analysis were done

with pure MMA, MMA with cardanol, styrene, and styrene with cardanol. For this,

it was used a tensiometer (Kruss Process, model K100) and using a plaque that was

conected to equipment’s balance. Analysis occurred until the interfacial tension was

constant. Data points were obtained every 30 seconds.

7.5.3 Particles Size

Particle size distribution was obtained by a light scattering technique caused by

the presence of the polymer particles in aqueous medium. The equipment used was

the particle size analyzer Master Sizer Hydro 2000S (Malvern, United Kingdom). The

equipment has a detector system for front, side and back light scattering. The source

is a helium neon laser whose characteristic wavelength is 632.8 ηm. Its detection limit

ranges from 0.01 to 3000 µm. The samples were prepared dispersing roughly 200 - 500

mg of sample in an aqueous solution of ethanol. The dispersion medium used does

not dissolve the particles. Analyzes were done at room temperature.

7.5.4 MEV

The detailed visualization of the surface of the particles produced was possible with

the aid of SEM technique (Quanta 250, Fei Company, 30kV) using vacuum and sec-

ondary electrons for image construction. The photomicrographs were processed in an

image analyzer from the same manufacturer.

132

7.5.5 Optical Microscopy (OM)

Morphology of the particles produced by suspension were determined by optical

microscopy. The micrographs were obtained using the optical microscope Axiovert 40

MAT equipped with a 1.4 megapixel camera, model Axiocam MRc, both from manu-

facturer Carl Zeiss (Oberkochen, Germany), using the dark field technique.The micro-

scope is associated with a computer that has the Axiovision software version 4.8.1 from

the same manufacturer. This software enables acquisition of images, provides scales

and other measures.

133

Chapter 8

Suspension Copolimerization: Results

8.1 Introduction

As one of the main objectives of this study is the immobilization of enzymes and

functionalization of polymers using natural CNSL or cardanol, the synthesis of these

copolymers by a heterogeneous medium is interesting (PAIVA et al. (2018), CAMPOS

et al. (2016), MENDES et al. (2012)). Suspension, for instance, is a good choice be-

cause it has many advantages, including easy separation of the polymer particles, easy

removal of the heat of reaction, easy temperature control, and low levels of impuri-

ties and additives in the final polymer resin.Therefore, it is suitable for production of

polymer resins intended for many distinct applications, including biotechnological,

medical, and dental applications (PINTO et al., 2002).

Therefore, copolymer synthesis by suspension was done using only cardanol with

methyl methacrylate or styrene as they had the best results in bulk polymerizations.

The same procedure for natural cashew nut shell liquid may result in similar outcomes.

For initiator, benzoyl peroxide (BPO) was used in two different concentrations - 1wt

% and 3wt %. However, some polymerizations used azobisisobutyronitrile (AIBN) in-

stead. Although, molecular weights and copolymer compositions were not obtained,

other aspects of this reaction and its products are studied. An interesting aspect is that

after the reaction, the reaction mixtures were stable even when they were kept resting

for several weaks. Thus the addition of cardanol resulted in an increase in suspension

particles stability due to reduction of interfacial tension and combination of electrical

and geometric effects.

8.2 Styrene-Cardanol Copolymer Particles

First of all, the interfacial tension between the organic phase (styrene and 10wt %)

was studied to see if cardanol would interfere in particle sizes, and if some conditions

134

Table 8.1: Interfacial tension between styrene or styrene-cardanol with distilled water

Average Value [mN/m] Standard Deviation [mN/m]

Sty 18.0977 6.2072Sty + 10%C 11.0678 0.4502

for suspension polymerization should be changed such as PVA concentration and stir-

ring rate. In table 8.1, values for interfacial tension verify that the addition of cardanol

truly reduces the interfacial tension. This occurs because cardanol is an amphiphilic

molecule, thereby increasing the iteration between the organic phase and the aque-

ous one. However, the reduction is not as high as what happens with the addition

of PVA in the aqueous phase that corresponds to roughly 5 mN/m with 1wt% of PVA

(CARVALHO, 2011). GALVÃO (2016) obtained higher values for both styrene or styrene

with cardanol organic phases, but the addition of cardanol also decreased the interfa-

cial tension. As an example for the roughly good interaction between cardanol-water,

when the interfacial tension between cardanol and water tried to be measured by pen-

dant drop tensiometry, this could not be done due to high relation between the two

phases as cardanol was carried by the water phase - not remaining in the drop (Fig-

ure 8.1-a). For measures between styrene (or methyl methacrylate) and water by this

method, they remained in the drop, being this one with a perfect shape. It also likely

that cardanol moves to the interface leaving styrene in the nucleus of the particle or the

bottom of the crucible used to measure the interfacial tension (DE MARIA et al., 2005).

However this is a good aspect as cardanol deposition on the surfaces allows surface

modification through the functional groups, rendering the particles compatible with

different mediums and with the potential for various kinds of applications (BOHARA

et al. (2016)). Figure 8.1-b is a photo for a drop in which its interfacial tension was been

measured and there is a "snake" of cardanol running away.

Styrene-cardanol copolymers were synthesized using 5wt % of cardanol with 1 or

3wt %; and using 10wt % of cardanol with 1wt % of AIBN (Figure 8.2). The data for

2.5wt% with 1wt% BPO was obtained from GALVÃO (2016) for comparative purposes.

Polystyrene reached 100% of conversion around 200 min for 1wt% of BPO. However,

as was detected and discussed in the 1st section, cardanol inhibits the polymerization,

reaching only 53% for 2.5wt% (after 540 minutes) and 7 for 5wt% (GALVÃO (2016) ob-

tained 13.4% after 300min). Therefore, the kinetics was slower when the reaction was

performed in suspension due to differences in mass and heat transfers. But because

nitrogen purging did not occur for these reactions, this can also have an influence in

hydrogen donation by the phenolic group (LARTIGUE-PEYROU, 1996). Then, chang-

ing the initiator for AIBN, monomer conversion was improved extremely. For 5wt %,

conversion reached over 80%; and, for 10wt %, 45% was obtained - almost the value

got for 2.5wt% with BPO. GALVÃO (2016) obtained total conversion with 2.5wt % using

135

(a) (b)

Figure 8.1: Photos of pendant upward drop tensiometry used for some non-reportedinterfacial tension calculations. (a) was the test for cardanol-water and (b) for styrene

+ cardanol in water.

Figure 8.2: Conversion values for styrene-cardanol suspension polymerizations

AIBN. This initiator, AIBN, is known to be more reactive, and this difference is a signal

that, by changing the initiator, the inhibition by cardanol can be less effective.

136

Although, the copolymer composition still needs to be investigated in future stud-

ies. Analysis of contact angle were done by drop deposition of water on a copolymer

pastille. For styrene pastille, the drop stayed on top throughout the whole analysis, and

contact angle (measured by image analysis using a software) was also constant. How-

ever, the water drop was quickly absorbed by the pastille, and contact angle could not

be measured (see the discussion of contact angle measures for MMA-cardanol parti-

cles for photos of the phenomenon). This happened because the copolymer has some

affinity by water due to the presence of cardanol, being a sign of cardanol incorpora-

tion.

8.3 Methyl Methacrylate-Cardanol Copolymer Particles

As same as was done for styrene, the interfacial tension was also calculated for

methyl methacrylate (with or without cardanol) and water system. The values ob-

tained are shown in Table 8.2. It can be seen that cardanol reduces the tension, but not

as much as 1wt % of PVA does. Therefore, PVA could not be substituted by cardanol, as

it would increase inhibition and led to small conversions.

Figure 8.3: Conversion values for methyl methacrylate-cardanol suspensionpolymerizations with 1wt% of BPO

Table 8.2: Interfacial tension between styrene or styrene-cardanol with distilled water

Average Value [mN/m] Standard Deviation [mN/m]

MMA 9.404 0.627MMA + 10%C 6.555 1.015MMA + PVA 1.335 1.262

137

Suspension polymerization with 1wt % of BPO (Figure 8.3) was slower than the bulk

polymerization with the same BPO concentration probably due to same reasons re-

ported for styrene - differences in mass and heat transfers; and presence of oxygen

(LARTIGUE-PEYROU, 1996). Maybe due to small conversion, copolymer particle sizes

also became small and, mainly for 10wt% (Figure 8.4), very stick (PINTO et al. (2002))

Figure 8.5 how small, but smooth the particles for 5wt% got. Also, a high amount of

residual monomer was present. Therefore, these polymers could not be filtrated be-

cause they clogged the pores of the filter paper (55 or 205µm) and had to be freeze

dried.

Figure 8.4: SEM photo for methyl methacrylate-cardanol particles produced with10wt% of cardanol and 1wt% of BPO after 6 hours of polymerization

138

Figure 8.5: SEM photo for methyl methacrylate-cardanol particles produced with5wt% of cardanol and 1wt% of BPO after 4 hours of polymerization

It is also possible that the particles were small due to the amphiphilic characteris-

tic of cardanol. Nevertheless, the sizes were dispersed (Figure 8.6). One day after the

reaction, the particles were segregated by size - being the bigger and denser at the bot-

tom of the becker, while the smallest and less dense were on top. After getting samples

of each phase, optical microscopy images done for them detected the size difference

- in figure 8.8. The particle size segregation was also detected by GALVÃO (2016) for

styrene-cardanol copolymer particles (Fig. 8.7).

Using 3wt % of BPO (Figure 8.9), the conversion increased just a little bit (as can

be also predicted by the copolymerization model of chapter 6), however this lead to an

increase in particle size as it could be filtrated and the particles weren’t so stick. Figure

8.10 shows how the size got bigger for 3wt% BPO when 2.5wt% of cardanol was added.

It is interesting that even reducing interfacial tension, the presence of cardanol in the

copolymer particles could not avoid bigger sizes than for pure MMA.

Changing BPO by AIBN as initiator, conversions increased tremendously. This was

139

Figure 8.6: Particle size distribution for copolymer particles produced with 1wt% ofBPO

tested for 5 and 10wt% of cardanol, as can be seen in Figure 8.11 a and b, respectively.

Even with 10wt% of cardanol, conversion almost reached 100%. AIBN is more reactive

than BPO and its use may be a way to work around cardanol inhibition. However, the

increase of conversion, also increased the average particle size. Figure 8.12 presents

the difference for particle size distributions to the copolymer with 5wt% of cardanol

and 1wt% of BPO, and 5wt% of cardanol and 1wt% of AIBN. The average size for the

copolymer with AIBN (143µm) is even bigger than for pure PMMA (76µm). It is im-

portant to emphasize that these differences in particle sizes are only due to changes

in cardanol and/or BPO (or AIBN) concentration and initiator type (BPO or AIBN) -

stirring rate and concentration of PVA were kept constant.

140

Figure 8.7: Photo of a becker contaning the reaction mixture of methyl methacrylateand 5wt% of cardanol one day after the reaction

(a) (b)

Figure 8.8: Optical microscopy images for the center (a) and bottom (b) for methylmethacrylate-cardanol copolymer particles

141

Figure 8.9: Conversion values for methyl methacrylate-cardanol suspensionpolymerizations with 5wt% of cardanol and 1 or 3wt% of BPO

Figure 8.10: Particle size distribution for PMMA or copolymer particles made with 1 or3wt% of BPO respectively

142

(a) (b)

Figure 8.11: Conversion data for methyl methacrylate-cardanol suspensionpolymerizations with 5wt% (a) or 10wt% (b) of cardanol and 1wt% of AIBN

Figure 8.12: Particle size distribution for 5wt% cardanol copolymer particles madewith 1wt% of BPO or AIBN

143

Contact angle analysis was performed to evaluate the hydrophobicity of produced

copolymers in comparison to the PMMA homopolymer. For this, a drop of water was

deposited over a pastille of pressed particles and the contact angle could be measured

by an image analysis by the tensiometer software. For PMMA, this analyzis turned out

fine (results not reported). Figure 8.13 shows the drop on the polymer while contact

angle was being measured. However, for the copolymers, the drop of water was quickly

absorbed (less than 8 seconds). See how the contact angle changes while the drop is

being absorbed in Figure 8.14. This happened even for the smallest concentration of

cardanol (2.5wt%). (The copolymer with 2.5wt% was produced with 3wt% of BPO, and

reached over 80% of conversion). Therefore although NMR data is not available for

copolymer composition data, cardanol could be incorporated. It also interesting to

remember that cardanol is not soluble in water - its partition factor is small (DE MARIA

et al. (2005)).

Figure 8.13: Experimental set up for PMMA-water contact angle measure

144

Figure 8.14: Photos taken while a contact angle analysis was being performed for acopolymer of MMA and cardanol


Suspension polymerizations with cardanol were done with methyl methacrylate or

styrene as comonomers. Conversions were smaller than the ones obtained by bulk

polymerization with BPO. Although cardanol could reduce interfacial tension between

the organic and the aqueous phase, it was not enough to stabilize completely the sys-

tem. Therefore, PVA was still used. For some conditions, aiming to increase polymer-

ization conversion, BPO was changed for AIBN. This difference made a good improve-

ment in the kinetics, obtaining good values. For MMA, although cardanol molecules

might had reduced particle sizes, the size distribution was still very broad while the

average size changed with cardanol concentration, amount and type of initiator. Fi-

nally, it was interesting to see how the hydrophobicity of the particles changed with

cardanol incorporation using contact angle technique. However, it looked like the par-

ticles become more hydrophilic and absorbed water, not allowing the measure to be

145

completed. Therefore for applications that require water-soluble polymers, the ob-

tained particles may be a good option. Nonetheless, the synthesis of these particles in

the heterogeneous medium show that a broad range of applications for them is avail-

able, e.g. core-shell particles of cardanol-styrene or cardanol-MMA (BESTETI et al.,

2014).

146

Chapter 9

Conclusions

Although there is not any kinetic study verifying that CNSL does not polymerize via

free radicals, it is common to found in the literature claims that says that it cannot.

However, bulk polymerizations of cardanol and natural cashew nut shell liquid (CNSL)

showed that polymeric chains with more than 10 meres can be formed. The strong

limitations to maximize conversions arise from steric hindrance; addition of radical to

any of the internal unsaturations of the side chain, forming a stable radical that tends

to not propagate; and donation of a hydrogen by the phenolic group, remaining a free

radical in the oxygen atom that is stabilized by the aromatic ring due to resonance

structures. Despite this, bulk copolymerization of cardanol/natural CNSL with methyl

methacrylate, styrene, vinyl acetate and acrid acid occurred. The best results were ob-

tained by the first two cited monomers, although inhibitory effects and lower reactions

rates were still present. Their copolymers did not have high cardanol contents, but

even 1% was enough to improve thermal stability and had their glass transition tem-

perature reduced. Also, because the phenolic compounds have many reactive sites and

functional groups, analysis of FTIR and 1H-NMR could be more complex with peaks

overleaping each other, but bonds for both compounds were detected and it is clear

that phenolic unsaturations are reactive, i.e. their corresponding peaks decrease in in-

tensity. Nonetheless, cardanol content in the copolymers was determined by 1H-NMR.

For vinyl acetate, cardanol acted as a strong inhibitor and, even worse, its presence

in VAc polymeric chains decreased its strength and, thereby, decreasing thermal sta-

bility. For acrylic acid, practically instant copolymerization happened. But, because

of this, deep kinetic investigations could not be performed. Further reactions with

acrylic acid should be done in heterogeneous medium to have a better control of poly-

merization, as bulk kinetic parameters for PAA polymerization is hard to find to study

its kinetics.

Moreover, only mildly differences from cardanol kinetics to CNSL ones could be de-

tected. For homopolymerizations, CNSL had higher conversions, but considering the

high deviations of the obtained values, their oligomerization happened basically the

147

same. For MMA-CNSL copolymerization, the biggest gap between its conversion and

MMA-cardanol copolymerization conversion happened for the lowest temperature (85

°C). Therefore, CNSL had a lesser inhibition effect, but still a present one. This could

have happened due to induction effect between the carboxyl and hydroxyl around the

aromatic ring, decreasing the possibility of phenolic inhibition. However, even with

this slightly difference, they are both lump mixtures and can their kinetics can be con-

sidered the same. Furthermore, using CNSL, decreases one purification step to obtain

cardanol, reducing process costs.

Therefore, based on the bulk results, one may expect to face strong limitations to

maximize the cardanol/CNSL content of the copolymers. Despite that, for future stud-

ies, the production of copolymers with high CNSL compositions should be performed

at much higher temperatures as was shown that the increase in temperature led to a

decrease in CNSL retardation effect on kinetics. However, this might not translate in

high CNSL copolymer compositions. Other possibility is using alternative processes,

mainly polymerization under pressure to avoid traces of oxygen that can increase rad-

ical inhibition by the phenolics. This may lead to overall high conversions. Nonethe-

less, changing the initiator to a more reactive one, such as AIBN, can obtain high con-

version. A good start for this future studies was the validation of a proposed kinetic

model in which kinetic parameters for cardanol copolymerizations with styrene were

estimated (r1,2 and r2,1). To characterize the inhibition, an inhibition factor responsi-

ble for reducing the concentration of free radicals was also estimated. These constants

could be obtained for the studied range of cardanol concentrations and temperatures

used with low objective functions. Thus, even with many hypothesis, the proposed

model was able to follow monomer conversions, molecular weights and copolymer

compositions in a good range of conditions.

As polymerization in heterogeneous medium is more interesting for biomedical

and other noble applications, MMA/cardanol or styrene/cardanol copolymerization

were performed in suspension. Although worse rates for monomer conversion were

obtained using the same mass fractions of reactants in the organic phase, the intro-

duction of small amounts of cardanol into the copolymer backbone (below 2.5 wt%)

is sufficient to modify the hydrophobicity of the final polymer product, which can be

advantageous for biotechnological applications. From SEM photographies, the parti-

cles did not have a high porosity, and, in fact, their surface were smooth. Nonetheless,

cardanol concentration reduces the interfacial tension, but not as much to substitute

PVA in the aqueous phase. However, its amphiphilic characteristic affected the average

particle size, reducing it from the average size of 76µm for pure PMMA to 30~45µm for

10 or 5wt% of cardanol respectively. But particle sizes were influenced by differences in

initiator concentration and type - as well as monomer conversion. Finally, the addition

of chemical functionals groups on the polymeric chains (and mainly on the surface of

148

polymeric particles), renders them compatible and with the potential for various kind

of applications.

149

Bibliography

BALGUDE, D., SABNIS, A. S., 2014, “CNSL: an environment friendly alternative for

the modern coating industry”, Journal of Coatings Technology and Research,

v. 11, n. 2, pp. 169–183. doi: <10.1007/s11998-013-9521-3>. Availability:

<http://link.springer.com/10.1007/s11998-013-9521-3>.

BESTETI, M. D., SOUZA JR, F. G., FREIRE, D. M., et al., 2014, “Production of core-

shell polymer particles-containing cardanol by semibatch combined sus-

pension/emulsion polymerization”, Polymer Engineering & Science, v. 54,

n. 5, pp. 1222–1229.

BEVINGTON, J., MELVILLE, H., TAYLOR, R., 1954, “The termination reaction in radi-

cal polymerizations. Polymerizations of methyl methacrylate and styrene at

25°”, Journal of Polymer Science, v. 12, n. 1, pp. 449–459.

BHUNIA, H. P., NANDO, G. B., BASAK, A., et al., 1999, “Synthesis and characterization

of polymers from cashewnut shell liquid (CNSL), a renewable resource III.

Synthesis of a polyether”, European Polymer Journal, v. 35, n. 9, pp. 1713–

1722. doi: <http://dx.doi.org/10.1016/S0014-3057(98)00244-4>.

Availability: <http://www.sciencedirect.com/science/article/pii/

S0014305798002444>.

BILLMEYER, F. W., 1984, Textbook of Polymer Science. 3rd ed. United States of America,

John Wiley & Sons.

BISANDA, E. T. N., ANSELL, M. P., 1992, “Properties of sisal-CNSL composites”,

Journal of Materials Science, v. 27, n. 6, pp. 1690–1700. doi: <10.1007/

BF00542934>.

BOHARA, R. A., THORAT, N. D., PAWAR, S. H., 2016, “Role of functionalization: strate-

gies to explore potential nano-bio applications of magnetic nanoparticles”,

RSC Advances, v. 6, n. 50, pp. 43989–44012.

BRANDÃO, A., 2017, Mathematical modeling of chain branching reactions in olefin

coordination polymerizations. Tese de doutorado, Programa de Engenharia

Química (PEQ) - Universidade Federal do Rio de Janeiro.

150

10.1007/s11998-013-9521-3

http://link.springer.com/10.1007/s11998-013-9521-3

http://dx.doi.org/10.1016/S0014-3057(98)00244-4

http://www.sciencedirect.com/science/article/pii/S0014305798002444

http://www.sciencedirect.com/science/article/pii/S0014305798002444

10.1007/BF00542934

10.1007/BF00542934

CAMPOS, I. M., FERRAZ, H. C., PINTO, J. C., 2016, “Production and Functionaliza-

tion of P (MMA-co-AA) Nanoparticles by Miniemulsion Polymerization”. In:

Macromolecular Symposia, v. 368, pp. 70–77. Wiley Online Library.

CANEVAROLO JR., S., 2006, Ciência dos Polímeros. 2nd ed. São Paulo, Artliber Editora.

CARVALHO, E. P., 2011, Uso do persulfato de amônio para estabilização da polimeriza-

ção em suspensão do estireno. Tese de mestrado, Programa de Engenharia

Química (PEQ) - Universidade Federal do Rio de Janeiro.

CERVANTES-UC, J., CAUICH-RODRÍGUEZ, J., VÁZQUEZ-TORRES, H., et al., 2006,

“TGA/FTIR study on thermal degradation of polymethacrylates containing

carboxylic groups”, Polymer degradation and stability, v. 91, n. 12, pp. 3312–

3321.

CHELIKANI, R., KIM, Y. H., YOON, D. Y., et al., 2009, “Enzymatic polymerization

of natural anacardic acid and antibiofouling effects of polyanacardic acid

coatings”, Applied Biochemistry and Biotechnology, v. 157, n. 2, pp. 263–277.

doi: <10.1007/s12010-008-8284-2>.

CHISHOLM, B. J., ALAM, S., KALITA, H., et al. “Monomers and polymers derived from

natural phenols (W O2014197041A2)”. .

DAVID, J. P., FARMÁCIA, F. D., FEDERAL, U., et al., 2006, “Metabólitos Secundários de

Espécies de Anacardiaceae”, Quim. Nova, v. 29, n. 6, pp. 1287–1300.

DE MARIA, P., FILIPPONE, P., FONTANA, A., et al., 2005, “Cardanol as a replacement

for cholesterol into the lipid bilayer of POPC liposomes”, Colloids and Sur-

faces B: Biointerfaces, v. 40, n. 1, pp. 11–18.

FERREIRA, S. R., LOUZADA, H. F., DIP, R. M. M., et al., 2015, “Influence of the ar-

chitecture of additives on the stabilization of asphaltene and water-in-oil

emulsion separation”, Energy & Fuels, v. 29, n. 11, pp. 7213–7220.

FERRIOL, M., GENTILHOMME, A., COCHEZ, M., et al., 2003, “Thermal degradation

of poly (methyl methacrylate)(PMMA): modelling of DTG and TG curves”,

Polymer degradation and stability, v. 79, n. 2, pp. 271–281.

FINEMAN, M., ROSS, S. D., 1950, “Linear method for determining monomer reactivity

ratios in copolymerization”, Journal of Polymer Science, v. 5, n. 2, pp. 259–

262.

GALVÃO, L. D. A., 2016, Estudo da Copolimerização de Estireno e Cardanol em Meios

Homogêneo e Heterogêneo. Msc dissertation, Programa de Engenharia

Química (PEQ) - Federal University of Rio de Janeiro.

151

10.1007/s12010-008-8284-2

GEDAM, P., SAMPATHKUMARAN, P., 1986, “Cashew nut shell liquid: extraction,

chemistry and applications”, Progress in Organic Coatings, v. 14, n. 2,

pp. 115–157.

HARVEY, M. T. “Process of destructively distilling cashew nut shell liquid

(U S2098824)”. .

HERMANSON, G. T., 2013, Bioconjugate techniques. Academic press.

IKEDA, R., TANAKA, H., UYAMA, H., et al., 2000a, “Enzymatic Synthesis and Curing of

Poly(cardanol)”, Polymer Journal, v. 32, n. 7, pp. 589–593.

IKEDA, R., TANAKA, H., UYAMA, H., et al., 2000b, “New crosslinkable polyphenol

from renewable resource”, American Chemical Society, Polymer Preprints,

Division of Polymer Chemistry, v. 41, n. 1, pp. 83–84. doi: <10.1002/(SICI)

1521-3927(20000501)21:8<496::AID-MARC496>3.0.CO;2-G>.

KAEWCHADA, A., BORVORNPONGSAKUL, C., JAREE, A., 2012, “Synthesis of molec-

ularly imprinted polymers from AnAc for the separation of γ-oryzanol”, Ko-

rean Journal of Chemical Engineering, v. 29, n. 9 (sep), pp. 1279–1284. doi:

<10.1007/s11814-012-0021-4>. Availability: <http://link.springer.

com/10.1007/s11814-012-0021-4>.

KALFAS, G., RAY, W. H., 1993, “Modeling and experimental studies of aqueous sus-

pension polymerization processes. 1. Modeling and simulations”, Industrial

& engineering chemistry research, v. 32, n. 9, pp. 1822–1830.

KASHIWAGI, T., INABA, A., BROWN, J. E., et al., 1986, “Effects of weak linkages on the

thermal and oxidative degradation of poly (methyl methacrylates)”, Macro-

molecules, v. 19, n. 8, pp. 2160–2168.

KELEN, T., TÜDOS, F., 1974, “A new improved linear graphical method for determing

copolymerization reactivity ratios”, Reaction Kinetics and Catalysis Letters,

v. 1, n. 4, pp. 487–492.

KIM, K. J., CHOI, K. Y., ALEXANDER, J. C., 1992, “Dynamics of a continuous stirred

tank reactor for styrene polymerization initiated by a binary initiator mix-

ture. II: Effect of viscosity dependent heat transfer coefficient”, Polymer En-

gineering & Science, v. 32, n. 7, pp. 494–505.

KIM, K., CHOI, K., 1989, “Modeling of free radical polymerization of styrene cat-

alyzed by unsymmetrical bifunctional initiators”, Chemical engineering sci-

ence, v. 44, n. 2, pp. 297–312.

152

10.1002/(SICI)1521-3927(20000501)21:8<496::AID-MARC496>3.0.CO;2-G

10.1002/(SICI)1521-3927(20000501)21:8<496::AID-MARC496>3.0.CO;2-G

10.1007/s11814-012-0021-4



KIM, Y. H., AN, E. S., SONG, B. K., et al., 2003, “Polymerization of cardanol using soy-

bean peroxidase and its potential application as anti-biofilm coating mate-

rial”, Biotechnology Letters, v. 25, n. 18, pp. 1521–1524. doi: <10.1023/A:

1025486617312>.

KUMAR, S., DINESHA, P., ROSEN, M. A., 2018, “Cashew Nut Shell Liquid as a Fuel for

Compression Ignition Engines: A Comprehensive Review”, Energy and Fu-

els, v. 32, n. 7, pp. 7237–7244. doi: <10.1021/acs.energyfuels.8b00579>.

LARTIGUE-PEYROU, F., 1996, “The use of phenolic compounds as free-radical poly-

merization inhibitors”. In: Industrial Chemistry Library, v. 8, Elsevier, pp.

489–505.

LEMOS, T. S. M., 2014, Modelagem Estocástica da Formação de Produtos Copoliméri-

cos com Microestrutura Controlada em Reatores Contínuos. Ph.D. The-

sis, Programa de Engenharia Química (PEQ) - Federal University of Rio de

Janeiro.

LI, W. S. J., NEGRELL, C., LADMIRAL, V., et al., 2018a, “Cardanol-based polymer latex

by radical aqueous miniemulsion polymerization”, Polymer Chemistry, v. 9,

n. 18, pp. 2468–2477. doi: <10.1039/C8PY00167G>. Availability: <http:

//xlink.rsc.org/?DOI=C8PY00167G>.

LI, W. S. J., NEGRELL, C., LADMIRAL, V., et al., 2018b, “Cardanol-based polymer latex

by radical aqueous miniemulsion polymerization”, Polymer Chemistry, v. 9,

n. 18, pp. 2468–2477.

LOCHAB, B., SHUKLA, S., VARMA, I. K., 2014, “Naturally occurring phenolic sources:

monomers and polymers”, RSC Adv., v. 4, n. 42, pp. 21712–21752. doi:

<10.1039/C4RA00181H>. Availability: <http://xlink.rsc.org/?DOI=

C4RA00181H>.

LOUIE, B. M., CARRATT, G. M., SOONG, D. S., 1985, “Modeling the free radical so-

lution and bulk polymerization of methyl methacrylate”, Journal of applied

polymer science, v. 30, n. 10, pp. 3985–4012.

LOUREIRO, T., DIP, R. M. M., LUCAS, E., et al., 2017, “Cardanol Polymerization Un-

der Acid Conditions By Addition And Condensation Reactions”, Journal

of Polymers and the Environment, v. 0, n. 0, pp. 1–12. doi: <10.1007/

s10924-017-0969-6>.

LUBI, M. C., THACHIL, E. T., 2000, “Cashew nut shell liquid (CNSL) - a versatile

monomer for polymer synthesis”, Designed Monomers & Polymers, v. 3, n. 2,

153

10.1023/A:1025486617312

10.1023/A:1025486617312

10.1021/acs.energyfuels.8b00579

10.1039/C8PY00167G

http://xlink.rsc.org/?DOI=C8PY00167G

http://xlink.rsc.org/?DOI=C8PY00167G

10.1039/C4RA00181H

http://xlink.rsc.org/?DOI=C4RA00181H

http://xlink.rsc.org/?DOI=C4RA00181H

10.1007/s10924-017-0969-6

10.1007/s10924-017-0969-6

pp. 123–153. doi: <10.1163/156855500300142834>. Availability: <http:

//www.tandfonline.com/doi/full/10.1163/156855500300142834>.

MAHABADI, H. K., O’DRISCOLL, K., 1977, “Absolute rate constants in free-radical

polymerization. III. Determination of propagation and termination rate

constants for styrene and methyl methacrylate”, Journal of Macromolecular

Science—Chemistry, v. 11, n. 5, pp. 967–976.

MAHANWAR, P. A., KALE, D. D., 1996, “Effect of processing parameters on refining of

CNSL”, Indian Journal of Chemical Technology, v. 3, n. July, pp. 191–193.

MAHON, E., SALVATI, A., BOMBELLI, F. B., et al., 2012, “Designing the nanoparticle–

biomolecule interface for “targeting and therapeutic delivery””, Journal of

Controlled Release, v. 161, n. 2, pp. 164–174.

MANJULA, S., KUMAR, V. G., PILLAI, C. K., 1992, “Kinetics and mechanism of

oligomerization of cardanol using acid catalysts”, Journal of Applied Polymer

Science, v. 45, n. 2, pp. 309–315. doi: <10.1002/app.1992.070450213>.

MANRING, L. E., SOGAH, D. Y., COHEN, G. M., 1989, “Thermal degradation of

poly (methyl methacrylate). 3. Polymer with head-to-head linkages”, Macro-

molecules, v. 22, n. 12, pp. 4652–4654.

MAYO, F. R., LEWIS, F. M., 1944, “Copolymerization. I. A basis for comparing the be-

havior of monomers in copolymerization; the copolymerization of styrene

and methyl methacrylate”, Journal of the American Chemical Society, v. 66,

n. 9, pp. 1594–1601.

MAZZETTO, S. E., LOMONACO, D., MELE, G., 2009, “Óleo da Castanha de Caju: Opor-

tunidades e Desafios no Contexto do Desenvolvimento e Sustentabilidade

Industrial”, Quim. Nova, v. 32, n. 3, pp. 732–741.

MELO, C. K., SOARES, M., CASTOR, C. A., et al., 2014, “In Situ Incorporation of Recy-

cled Polystyrene in Styrene Suspension Polymerizations”, Macromolecular

Reaction Engineering, v. 8, n. 1, pp. 46–60.

MENDES, A. N., HUBBER, I., SIQUEIRA, M., et al., 2012, “Preparation and cytotoxic-

ity of poly (methyl methacrylate) nanoparticles for drug encapsulation”. In:

Macromolecular Symposia, v. 319, pp. 34–40. Wiley Online Library.

MISRA, A. K., PANDEY, G. N., 1984, “Kinetics of alkaline-catalyzed car-

danol–formaldehyde reaction. I”, Journal of Applied Polymer Science,

v. 29, n. 1 (jan), pp. 361–372. doi: <10.1002/app.1984.070290134>.

Availability: <http://doi.wiley.com/10.1002/app.1984.070290134>.

154

10.1163/156855500300142834

http://www.tandfonline.com/doi/full/10.1163/156855500300142834

http://www.tandfonline.com/doi/full/10.1163/156855500300142834

10.1002/app.1992.070450213

10.1002/app.1984.070290134

http://doi.wiley.com/10.1002/app.1984.070290134

MISRA, A. K., PANDEY, G. N., 1985, “Kinetics of alkaline-catalyzed car-

danol–formaldehyde reaction. II. Mechanism of the reaction”, Jour-

nal of Applied Polymer Science, v. 30, n. 3, pp. 969–977. doi:

<10.1002/app.1985.070300307>.

MUBOFU, E. B., MGAYA, J. E., 2018, “Chemical Valorization of Cashew Nut Shell

Waste”, Topics in Current Chemistry, v. 376, n. 2, pp. 1–15. doi: <10.

1007/s41061-017-0177-9>. Availability: <https://doi.org/10.1007/

s41061-017-0177-9>.

MUBOFU, E. B., 2016, “From cashew nut shell wastes to high value chemicals”,

Pure and Applied Chemistry, v. 88, n. 1-2, pp. 17–27. doi: <10.1515/

pac-2015-0603>.

ODIAN, G., 2004, Principles of Polymerization. 4th ed. New York, John Wiley & Sons,

Inc.

OECHSLER, B. F., 2016, “Análise Dinâmica de Modelos de Mistura Imperfeita em

Reatores de Polimerização via Radicais-livres em Solução”, p. 317.

OLIVEIRA, LINCOLN DAVI M. DE , 2007, Síntese, Caracterização e Funcionalidade de

Aditivos de Lubricidade, Derivados do LCC. Tese de mestrado, Universidade

Federal do Ceará.

OTSUKA, T., FUJIKAWA, S. I., YAMANE, H., et al., 2017, “Green polymer chemistry:

The biomimetic oxidative polymerization of cardanol for a synthetic ap-

proach to ’artificial urushi’”, Polymer Journal, v. 49, n. 3, pp. 335–343. doi:

<10.1038/pj.2016.118>. Availability: <http://dx.doi.org/10.1038/

pj.2016.118>.

PAIVA, T., VIEIRA, L., MELO, P., et al., 2018, “In Situ Incorporation of Praziquantel in

Polymer Microparticles through Suspension Polymerization for Treatment

of Schistosomiasis”, Macromolecular Reaction Engineering, p. 1800064.

PARAMASHIVAPPA, R., KUMAR, P. P., VITHAYATHIL, P., et al., 2001, “Novel method

for isolation of major phenolic constituents from cashew (Anacardium oc-

cidentale L.) nut shell liquid”, Journal of Agricultural and Food Chemistry,

v. 49, n. 5, pp. 2548–2551.

PATEL, R. N., BANDYOPADHYAY, S., GANESH, A., 2006, “Extraction of cashew (Anac-

ardium occidentale) nut shell liquid using supercritical carbon dioxide”,

Bioresource Technology, v. 97, n. 6, pp. 847–853.

155

10.1002/app.1985.070300307

10.1007/s41061-017-0177-9

10.1007/s41061-017-0177-9

https://doi.org/10.1007/s41061-017-0177-9

https://doi.org/10.1007/s41061-017-0177-9

10.1515/pac-2015-0603

10.1515/pac-2015-0603

10.1038/pj.2016.118

http://dx.doi.org/10.1038/pj.2016.118

http://dx.doi.org/10.1038/pj.2016.118

PEIXOTO, A. C. B., 2013, Funcionalização de Nanopartículas Poliméricas para Liber-

ação de Medicamentos Sítio-Dirigida. Ph.D. Thesis, Programa de Engen-

haria Química (PEQ) - Federal University of Rio de Janeiro.

PEKLAK, A. D., BUTTÉ, A., STORTI, G., et al., 2006, “Gel effect in the bulk reversible

addition–fragmentation chain transfer polymerization of methyl methacry-

late: Modeling and experiments”, Journal of Polymer Science Part A: Polymer

Chemistry, v. 44, n. 3, pp. 1071–1085.

PERDRIAU, S., HARDER, S., HEERES, H. J., et al., 2012, “Selective conversion of

polyenes to monoenes by RuCl3-catalyzed transfer hydrogenation: The case

of cashew nutshell liquid”, ChemSusChem, v. 5, n. 12, pp. 2427–2434. doi:

<10.1002/cssc.201200503>.

PHILIP, J. Y., BUCHWEISHAIJA, J., MKAYULA, L. L., et al., 2007, “Preparation of molec-

ularly imprinted polymers using anacardic acid monomers derived from

cashew nut shell liquid”, Journal of Agricultural and Food Chemistry, v. 55,

n. 22, pp. 8870–8876. doi: <10.1021/jf0718289>.

PICHOT, C., 2004, “Surface-functionalized latexes for biotechnological applications”,

Current Opinion in Colloid & Interface Science, v. 9, n. 3-4 (nov), pp. 213–

221. doi: <10.1016/j.cocis.2004.07.001>. Availability: <http://

linkinghub.elsevier.com/retrieve/pii/S1359029404000688>.

PINTO, J., RAY, W., 1995, “The dynamic behavior of continuous solution polymer-

ization reactors—VII. Experimental study of a copolymerization reactor”,

Chemical engineering science, v. 50, n. 4, pp. 715–736.

PINTO, M. C., SANTOS JR, J. G., MACHADO, F., et al., 2002, “Suspension polymeriza-

tion processes”, Encyclopedia of Polymer Science and Technology.

RAMOS, I. M. F. C., 2018, Nanopartículas Poliméricas Funcionalizadas para Liberação

de Fármaco no Sistema Nervoso Central. Ph.D. Thesis, Universidade Federal

do Rio de Janeiro (UFRJ).

RODRIGUES, F. H. A., FRANÇA, F. C. F., SOUZA, J. R. R., et al., 2011,

“Comparison between physico-chemical properties of the tech-

nical Cashew Nut Shell Liquid (CNSL) and those natural ex-

tracted from solvent and pressing”, Polímeros, v. 21, n. 2, pp. 156–

160. doi: <10.1590/S0104-14282011005000028>. Availability:

<http://www.scielo.br/scielo.php?script=sci_arttext&pid=

S0104-14282011000200016&lng=en&tlng=en>.

156

10.1002/cssc.201200503

10.1021/jf0718289

10.1016/j.cocis.2004.07.001

http://linkinghub.elsevier.com/retrieve/pii/S1359029404000688

http://linkinghub.elsevier.com/retrieve/pii/S1359029404000688

10.1590/S0104-14282011005000028

http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0104-14282011000200016&lng=en&tlng=en

http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0104-14282011000200016&lng=en&tlng=en

RODRIGUES, F. H. A., FEITOSA, J., RICARDO, N. M., et al., 2006, “Antioxidant activity

of cashew nut shell liquid (CNSL) derivatives on the thermal oxidation of

synthetic cis-1, 4-polyisoprene”, Journal of the Brazilian Chemical Society,

v. 17, n. 2, pp. 265–271.

RODRIGUEZ, F., COHEN, C., OBER, C. K., et al., 2015, Principles of Polymer Systems.

6th ed. New York, Taylor & Francis Group.

SATHIYALEKSHMI, K., 1993, “Studies on structure and properties of CNSL novolac

resins prepared with succinic acid catalyst”, Bulletin of Materials Science,

v. 16, n. 2, pp. 137–150. doi: <10.1007/BF02745329>.

SCARIAH, K. J., 1990, “GPC Studies on the Cationic Polymerization of Cardanol Initi-

ated by”, v. 41, pp. 1765–1775.

SCHWAAB, M., BISCAIA JR, E. C., MONTEIRO, J. L., et al., 2008, “Nonlinear parame-

ter estimation through particle swarm optimization”, Chemical Engineering

Science, v. 63, n. 6, pp. 1542–1552.

SCHWAAB, M AND PINTO, J.C., 2007, Análise de Dados Experimentais: I. Fundamen-

tos de Estatística e Estimação de Parâmetros.

SCORAH, M., DHIB, R., PENLIDIS, A., 2006, “Modelling of free radical polymerization

of styrene and methyl methacrylate by a tetrafunctional initiator”, Chemical

engineering science, v. 61, n. 15, pp. 4827–4859.

SOUZA JR., F. G., PINTO, J. C., RODRIGUES, M. V., et al., 2008a, “New polyani-

line/polycardanol conductive blends characterized by FTIR, NIR, and XPS”,

Polymer Engineering & Science, v. 48, n. 10 (oct), pp. 1947–1952. doi:

<10.1002/pen.21047>. Availability: <http://doi.wiley.com/10.1002/

pen.20921http://doi.wiley.com/10.1002/pen.21047>.

SOUZA JR., F. G., RICHA, P., DE SIERVO, A., et al., 2008b, “New in situ blends of

polyaniline and cardanol bio-resins”, Macromolecular Materials and Engi-

neering, v. 293, n. 8, pp. 675–683. doi: <10.1002/mame.200800077>.

TIAN, H., TANG, Z., ZHUANG, X., et al., 2012, “Biodegradable synthetic poly-

mers: Preparation, functionalization and biomedical application”, Progress

in Polymer Science (Oxford), v. 37, n. 2, pp. 237–280. doi: <10.1016/

j.progpolymsci.2011.06.004>. Availability: <http://dx.doi.org/10.

1016/j.progpolymsci.2011.06.004>.

157

10.1007/BF02745329

10.1002/pen.21047

http://doi.wiley.com/10.1002/pen.20921 http://doi.wiley.com/10.1002/pen.21047

http://doi.wiley.com/10.1002/pen.20921 http://doi.wiley.com/10.1002/pen.21047

10.1002/mame.200800077

10.1016/j.progpolymsci.2011.06.004

10.1016/j.progpolymsci.2011.06.004

http://dx.doi.org/10.1016/j.progpolymsci.2011.06.004

http://dx.doi.org/10.1016/j.progpolymsci.2011.06.004

TIDWELL, P. W., MORTIMER, G. A., 1965, “An improved method of calculating copoly-

merization reactivity ratios”, Journal of Polymer Science Part A: General Pa-

pers, v. 3, n. 1, pp. 369–387.

TIDWELL, P. W., MORTIMER, G. A., 1970, “Science of determining copolymerization

reactivity ratios”, Journal of Macromolecular Science—Reviews in Macro-

molecular Chemistry, v. 4, n. 2, pp. 281–312.

TSOUKAS, A., TIRRELL, M., STEPHANOPOULOS, G., 1982, “Multiobjective dynamic

optimization of semibatch copolymerization reactors”, Chemical Engineer-

ing Science, v. 37, n. 12, pp. 1785–1795.

VILLALOBOS, M. A., HAMIELEC, A. E., WOOD, P. E., 1993, “Bulk and suspension poly-

merization of styrene in the presence of n-pentane. An evaluation of mono-

functional and bifunctional initiation”, Journal of applied polymer science,

v. 50, n. 2, pp. 327–343.

VIVALDO-LIMA, E., WOOD, P. E., HAMIELEC, A. E., et al., 1998, “Kinetic model-

based experimental design of the polymerization conditions in suspension

copolymerization of styrene/divinylbenzene”, Journal of Polymer Science

Part A: Polymer Chemistry, v. 36, n. 12, pp. 2081–2094.

WAY, D. V., 2017, Uso de Click Chemistry para a Bioconjugação de Polímeros Visando ao

Tratamento do Câncer. Phd thesis, Programa de Engenharia Química (PEQ)

- Universidade Federal do Rio de Janeiro.

WELSCH, N., LU, Y., DZUBIELLA, J., et al., 2013, Adsorption of proteins to functional

polymeric nanoparticles. Ph.D. Thesis. Availability: <http://dx.doi.org/

10.1016/j.polymer.2013.03.027>.

YADAV, R., SRIVASTAVA, D., 2007a, “Kinetics of the acid-catalyzed cardanol-

formaldehyde reactions”, Materials Chemistry and Physics, v. 106, n. 1,

pp. 74–81. doi: <10.1016/j.matchemphys.2007.05.020>.

YADAV, R., SRIVASTAVA, D., 2007b, “Kinetics of the acid-catalyzed cardanol–

formaldehyde reactions”, Materials Chemistry and Physics, v. 106, n. 1,

pp. 74–81.

YOON, D. Y., KIM, D. S., 2009, “Molecular design of anti-biofouling materials from nat-

ural phenolic compounds”, Korean Journal of Chemical Engineering, v. 26,

n. 2, pp. 433–437. doi: <10.1007/s11814-009-0073-2>.

158

http://dx.doi.org/10.1016/j.polymer.2013.03.027

http://dx.doi.org/10.1016/j.polymer.2013.03.027

10.1016/j.matchemphys.2007.05.020

10.1007/s11814-009-0073-2

ZALDÍVAR, C., DEL SOL, O., IGLESIAS, G. D., 1998, “On the preparation of acrylic

acid/vinyl acetate copolymers with constant composition—1. copolymer-

ization reactivity ratios”, Polymer, v. 39, n. 1, pp. 245–246.

159

cashew nut shell liquid as a green alternative to ... - pantheon ufrj

Documents