SYNTHESIS OF CHAIN-END FUNCTIONALIZED POLYOLEFINS AND FLUOROPOLYMERS AND APPLICATIONS IN NANOCOMPOSITES

Transcription

1 The Pennsylvania State University The Graduate School Department of Materials Science and Engineering SYNTHESIS OF CHAIN-END FUNCTIONALIZED POLYOLEFINS AND FLUOROPOLYMERS AND APPLICATIONS IN NANOCOMPOSITES A Thesis in Materials Science and Engineering by Zhiming Wang 2005 Zhiming Wang Submitted in Partial Fulfillment of the Requirements for the degree of Doctor of Philosophy August 2005

2 The thesis of Zhiming Wang was reviewed and approved* by the following: Tze-Chiang Chung Professor of Materials Science and Engineering Thesis Advisor Chair of Committee Evangelos Manias Associate Professor of Materials Science and Engineering Thesis Co-advisor Co-chair of Committee Ian R. Harrison Professor of Materials Science and Engineering Ayusman Sen Professor of Chemistry James P. Runt Professor of Materials Science and Engineering Associate Head of Graduate Studies * Signatures are on file in the Graduate School

3 iii ABSTRACT In this thesis, we have demonstrated a very useful and simple method (one-pot polymerization process) for synthesis of chain end functionalized polypropylene. The chemistry involves a chain transfer reaction to a styrenic derivative (St-f), with or without hydrogen during propylene polymerization, using an Exxon-Hoechst C 2 -symmetric catalyst (rac-me 2 Si[2-Me-4-Ph(Ind)] 2 ZrCl 2 /MAO complex) or a Mitsubishi C 1 - symmetric catalyst (Me 2 Si(2-Me-Benz[e]Ind(2-Me-4-Ph-4HAzu)HfCl 2 with MAO or trialkylaluminum-treated clay). In the presence of the Exxon-Hoechst catalyst, the propylene propagating chain-end engages in a facile consecutive chain transfer reaction, reacting with St-f and then with hydrogen, with high catalytic activity under the proper reaction conditions. The polymer molecular weight is proportional to the molar ratio of [propylene]/[st-f]. A silane protecting group in St-NSi 2 or St-OSi unit can be hydrolyzed in an acidic solution during the sample work-up step to obtain desirable i-pp polymers, such as i-pp with a terminal NH 2 or OH group, in one pot. Despite the low concentration, the terminal functional group is very reactive and can serve as an active site for many applications. One example was shown in a chain extension reaction (coupling reaction) with polycaprolactone (PCL) in solution to form PP-b-PCL diblock copolymers that are very effective compatibilizers in PP/PCL polymer blends. Unexpectedly, a Mitsubishi C 1 -symmetric catalyst exhibits significant polymerization activity even in the absence of hydrogen, indicating that the trialkylaluminum may participate in chain transfer to p-ms (p-methylstyrene) terminated propagating chains. In the case of polymerization using MAO as a cocatalyst at 55 o C, the

4 iv addition of hydrogen increases the activity and regulates the polymer molecular weight. The chain-end structure is solely terminal p-ms. When TEA (triethylaluminium) -treated clay is adopted as an activator and carrier at the optimal polymerization temperature of 75 o C, the high concentration of hydrogen suppresses catalytic activity. The chain ends consist of predominately terminal p-ms and a small amount of unsaturated end groups. A higher p-ms concentration or introduction of hydrogen eliminates the undesirable unsaturated chain ends. Furthermore, we also study a new chemical route to prepared side chain functionalized polyolefin, especially the desirable MA (maleic anhydride) -modified PE and PP polymers with well-controlled molecular structures. The chemistry involves a post-polymerization process using borane/o 2 stable radical initiators to create polymeric radicals that are simultaneously stabilized by in situ formed *O-BR 2 stable radicals. The dormant polymeric radicals do not undergo undesirable side reactions (crosslinking and degradation, etc.), but can react with maleic anhydride. Some MAH-modified PP polymers with high molecular weight and controlled MAH content have been obtained. They have been proven to be the effective compatibilizers to improve the interfacial adhesion in the PP/Nylon 11 blends. In a different approach, when dealing with fluoropolymers, a modified iodine transfer polymerization (ITP) method was also developed, based on the combination of a specific radical initiator (2,2'-azobisisobutyronitrile, AIBN) and a reversible additionfragmentation chain transfer (RAFT) process involving two iodo-compounds, i.e., α,ωdiiodoperfluoroalkane (I-R f -I) and mono-iodoperfluoroalkane (R f -I). We take advantage of the inactive radicals, created by the decomposition of AIBN, which readily react with

5 v the iodo-compounds (chain transfer agents). Pure telechelic fluoropolymers with almost all the polymer chains containing two terminal iodo groups have been synthesized. Using diiodoperfluoroalkane as the chain transfer agent is more effective than monoiodoperfluoroalkane. In turn, the reactive terminal CF 2 I groups can undergo facile ethylenation to convert to relatively stable CH 2 I group or readily transform to imidazolium ions that are very effective in forming fluoropolymer/clay nanocomposite. One major application of functional polyolefin and fluoropolymers is the preparation of exfoliated polymer/clay nanocomposites. The process involves melt or solution blending using functional polymer as a surfactant, namely, chain-end functionalized polypropylene containing a terminal hydrophilic functional group (NH + 3 ) and a high molecular weight hydrophobic polymer chain. The chain-end functionalized polypropylene exhibits very high surface activities and results in an exfoliated clay interlayer structure, even with pristine clay minerals without any organic treatment. Furthermore, this exfoliated clay structure maintains its disordered state even after further mixing with neat (unfunctionalized) polypropylene that is compatible with the backbone of the chain end functionalized polypropylene. Mechanical property evaluation + shows the addition of PP-t-NH 3 in the system remarkably enhanced flexural modulus. These experimental results demonstrate the advantage of chain-end functionalized PP in the formation of an exfoliated clay layer structure and lead to the proposition that the terminal hydrophilic NH + 3 functional groups anchors the PP chains on the inorganic surfaces via ion exchange, and that the hydrophobic high molecular weight and semicrystalline PP chains are repelled from the inorganic surfaces and exfoliate the clay platelets.

6 vi This chain-end functional polymer technology was extended to fluoropolymers, and we have prepared a new family of fluoropolymer/clay nanocomposites that exhibits an exfoliated and uniformly dispersed clay structure in a polymer matrix. The process involves a specific interfacial reagent-chain end functionalized fluoropolymer containing an unperturbed hydrophobic and oleophobic fluoropolymer chain and a terminal functional group, such as Si(OEt) 3, OH, imidazolium, or sulfonium ions. The terminal functional group can anchor fluoropolymer chain to the clay surfaces between interlayers, either by a chemical bond (such as a Si-O-Si bond), strong interaction (such as hydrogen bonding and ion-ion interaction) or ion-exchange with cation (Li +, Na +, etc) located on the surfaces between the clay interlayers. On the other hand, the rest unperturbed high molecular weight hydrophobic and oleophobic fluoropolymer chain, disliking the hydrophilic clay surfaces, exfoliates the clay layer structure and maintains this disordered clay structure even after further mixing with a neat (unfunctionalized) polymer that is compatible with the backbone of the chain end functionalized fluoropolymer.

7 vii TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES ACKNOWLEGMENTS Chapter I Metallocene Catalyst and Polymerization of Olefins 1.1 Introduction 1.2 C 2v -symmetric Metallocenes 1.3 Bridged C 2 -symmetric Metallocenes xi xii xvi Bridged C s -symmetric Metallocenes Bridged C 1 -symmetric Metallocenes Metallocenes with oscillating structure Bridged half-metallocenes Cocatalyst 1.9 Supported metallocene catalysts 1.10 Outlook References 21 Chapter II Progress in Functional Polyolefins and Fluoropolymer and Their 25 Application in Polymer/clay Nancomposite 2.1 Introduction of Polyolefins Three Approaches in the Functionalization of Polyolefins Functionalization of Polyolefins via Reactive Chain Transfer Agent Progress in Functionalization of Fluoropolymer 32

8 viii 2.5 Progress in Polymer/clay Nanocomposites Outlook 47 References 49 Chapter III Synthesis of Chain-end Functional i-pp via Exxon-Hoechst C 2 - symmetric Catalyst and Mitsubishi C 1 -symmetric catalyst Introduction Experimental Results and Discussion Synthesis of PP-t-St-NH 2 Polymer via Exxon Hoechst C 2 -symmetric Catalyst Synthesis of PP-t-p-MS via Mitsubishi C 1 -symmetric Catalyst/MAO system Synthesis of PP-t-p-MS via Mitsubishi C 1 -symmetric Catalyst/ clay system Synthesis of PP-b-PCL and PP-b-PS-diblock Copolymers Summary 92 References 93 Chapter IV Preparation of MAH Modified PP and Application in PP/nylon Blend Introduction Experimental Methods Results and Discussion Oxidation of trialkylborane Synthesis of MAH modified PP Application of MAH Modified PP in PP/nylon Blend Summary 115 References 116

9 ix Chapter V Synthesis of Chain-end functionalized Fluoropolymer Introduction Experimental Methods Results and Discussion Synthesis of Iodo-terminated Fluoropolymer Functional Group Transformation of Iodine Terminal Fluoropolymer Summary 153 References 153 Chapter VI Preparation and Characterization of Functional Polymer/clay Nanocomposites Introduction Experimental Methods Results and Discussion Application of Chain-end Functionalized Polyolefins in Polymer/clay 166 Nanocomposite Mechanical Property of Polypropylene/clay Nanocomposite Application of Chain-end Functionalized Fluoropolymers in Polymer/clay Nanocomposite Summary 191 References 192 Chapter VII Conclusion and Future work Conclusion Future work 198

10 x LIST OF TABLES Table 1.1 Table 2.1 Liquid Propylene Polymerization with some representative ansazirconocene catalysts Reactivity ratio of Functional Fluorinated comonomers with TFE or VDF 8 36 Table 3.1 A summary of PP-t-Cl, PP-t-St-OH and PP-t-St-NH 2 polymers 67 Table 3.2 Table 3.3 Table 3.4 PP-t-p-MS Polymers Prepared by Mitsubishi C 1 -symmetric catalyst /MAO (H 2 effect) PP-t-p-MS Polymers Prepared by Mitsubishi C 1 -symmetric catalyst/mao (p-ms effect) PP-t-p-MS Polymers Prepared by Mitsubishi C 1 -symmetric Hf Catalyst/TEA-treated Clay system Table 4.1 Summary of Maleation reaction of Polyolefin 109 Table 5.1 Assignment of 19 F NMR peaks 134 Table 5.2 Polymerization of VDF in the presence of diidoperfluorobutane (I- R f -I) 138 Table 5.3 Assignment of 19 F NMR peaks 145 Table 5.4 Polymerization of VDF in the presence of AIBN and iodoperfluorobutane (R f -I) 148 Table 6.1 Properties of clay minerals used 158 Table 6.2 Summary of PVDF-t-Si/Na + -mmt hybrids with different compositions 188

11 xi LIST OF FIGURES Figure 1.1 Model of a (R,R)-Me 2 C(1-Ind) 2 Mt(iso-butyl)]+ cation (Mt=Zr) with a re η 2 -coordinated propylene molecule. 7 Figure 1.2 Models of a [Me 2 C(Cp)(9-Flu)Mt(iso-butyl)] + (Mt = Zr) with a η 2 - coorinated propylene molecule. 10 Figure 1.3 A computer model of the cationic catalytic species [Me 2 C(3-Me- Cp)(9-Flu)Zr(iso-butyl)]+ with growing polymer chain simulated by an iso-butyl group 12 Figure 2.1 Schematic representation of various methods (solution blending, melt blending, and in situ polymerization) used to prepare polymer-layered- silicate nanocomposites. The delaminated (or exfoliated) and intercalated morphologies are shown. 40 Figure H NMR of 4-{2-[N,N-bis(trimethylsilyl)amino]ethyl}styrene 58 Figure 3.2 Figure 3.3 Figure 3.4 Figure 3.5 Figure 3.6 Figure 3.7 The plots of number average molecular weights (Mn) of PE-t-St-f polymers versus the mole ratio of [propylene]/[st-f] using (a) St- Cl, (b) St-OSi and (c) St-NSi 2, respectively. 1 H NMR spectra of (a) a PP-t-St-OSi polymer and its corresponding PP-t-St-OH. (M n = 22,000 g/mol, M w /M n = 2.0). (solvent: C 2 D 2 Cl 4 ; temp.: 110 o C). 1 H NMR spectra of (a) a PP-t-St-NSi 2 polymer and (b) its corresponding PP-t-St- NH 2 (M n = 24,200 g/mol, M w /M n = 2.3). (solvent: C 2 D 2 Cl 4 ; temp.: 110 o C). 1 H NMR spectra of two PP-t-p-MS samples (run T0 and ) prepared by Mitsubishi C 1 -symmetric catalyst (top) with p-ms but no H 2 and (bottom) with both p-ms and H 2. The plot of polymer molecular weight of PP-t-p-MS vs. [propylene]/[p-ms] ratio 1 H NMR spectra of pure PP (run C15) (bottom) and two PP-t-p- MS samples (run C23 and C27) prepared by Mitsubishi C 1 - symmetric catalyst/clay system (top) with p-ms but no H 2 and (middle) with both p-ms and H

12 xii Figure H NMR spectra of PP-t-NH 2 (a); PP-b-PCL (b); and PP-b-PS (c) 87 Figure 3.9 DSC curves of (a) PCL; (b) PP-b-PCL; and (c) PP-t-NH 2 88 Figure 3.10 SEM microgrphs of (a) two homopolymer blend with i- PP/PCL=70/30 (2,000x) and (b) two homopolymer with PP-b- PCL, i-pp/pp-b-pcl/pcl=70/10/30 (2,000x) 91 Figure 4.1 Figure 4.2 Figure B NMR spectra of Tributylborane(TBB)/O 2 in the presence of a- PP. (a) TBB;(b) TBB/O 2 =1/1; (c) TBB/O 2 = ½. FTIR of pure PP (a), PP-g-MA entry 3 with 0.4 wt% MA (b) and entry 5 with 1.0 wt% MA (c). DSC curves of pure PP (a), PP-g-MA entry 4 (b) and commercial PP-g-MA Aldrich (c) Figure 4.4 FTIR of pure s-ps (a) and s-ps-g-ma (entry 10) (b) 110 Figure 4.5 Figure 5.1 SEM graphs of polymer of blend PP/Nylon. (a) PP/Nylon=30/70wt%; (b) PP/PP-g-MA/Nylon=20/10/70wt%. 1 H NMR and 19 F NMR spectra of telechelic PVDF with two terminal iodine Figure H NMR and 19 F NMR spectra of PVDF prepared from AIBN 133 Figure 5.3 Plot of M v of PVDF vs. Conversion of VDF monomer. 137 Figure 5.4 Plot of M v of PVDF vs. 1/[AIBN]. 137 Figure H NMR (a) and 19 F NMR (b) Spectra of PVDF-t-I (571) 143 Figure 5.6 Plot of M v of PVDF vs. Conversion of VDF 146 Figure 5.7 Figure H and 19 F NMR spectra of the product from ethylene treatment H NMR spectra of PVDF-t-Imm 152 Figure 6.1 X-ray diffraction patterns of PP-t-NH 3 + Cl-/Na + -mmt (90/10 weight ratio): (a) physical mixture by simple powder mixing at ambient temperature and (b) the same mixture after static meltintercalation (PP-t-NH 3 + /mmt hybrid). 167 Figure 6.2 XRD patterns of PP-NH 3 + Cl - /2C18-mmt (90/10) weight ratio): (a) physical mixture by simple powder mixing at ambient temperature 167

13 xiii Figure 6.3 and (b) the same mixture after static melt-intercalation X-ray diffraction patterns of the 50/50 mixture by weight of exfoliated PP-t-NH3+/mmt structure (90/10 weight ratio) and neat-unfunctionalized-i-pp. the XRD traces shown correspond to (a) the physical mixture of PP-t-NH3+/mmt and i-pp and (b) the same mixture after static melt-intercalation 169 Figure 6.4 TEM image of exfoliated PP/PP-t-NH 3 + /mmt nanocomposite 170 Figure 6.5 TEM images of PP-NH /Chlorite hybrid (50/50 wt%). Magnification: 10,000 (left) and 50,000 (right). 172 Figure 6.6 X-ray diffraction patterns of 2C18-mmt clay and four nanocomposites with 6 wt% of alkylammonium-mmt and 94 wt% of three side-chain-functionalized PPs containing (a) 1 mol% p- methylstyrene, PP-r-MS; (b) 0.5 mol% maleic anhydride, PP-r- MA; and (c) 0.5 mol% hydroxyl, PP-r-OH. (d) a 6 wt% C18-mmt nanocomposite of a PP-b-PMMA block copolymer, with 5 mol% of methyl methacrylate 173 Figure 6.7 Figure 6.8 Figure 6.9 Bright-field TEM image of PP-r-MA/6 wt% 2-C18 mmt nanocomposite structure. Magnification: 200,000. Illustration of the molecular structure of (a) chain-endfunctionalized polyolefin and (b) side-chain-functionalized polyolefin located between clay interla Flexual moduli of PP and PP/clay nanocomposites. a. Chlorite, PP-NH 3 + and i-pp mixed together; b. Chlorite and PP-NH 3 + mixed first and then mixed with i-pp; PP-NH : M w = 70,000 g/mol, NH 3 + =3% per chain; PP-NH : M w =28,400 g/mol, NH 3 + = 39% per chain. i-pp: MFR= Figure 6.10 XRD patterns of PVDF/Na+-mmt hybrid (top) and PVDF/C20A hybrid (bottom) before heating (a) and after heating-annealing (b). 181 Figure 6.11 XRD of PVDF/2C18-mmt (2.6 wt%) from in-situ polymerization initiated by potassium sulfate. (a.) In-situ polymerization product;. (b). After heating-annealing at 200 o C for 3hr. 183 Figure 6.12 XRD of PVDF/Na + -mmt (3.3 wt%) from in-situ polymerization initiated by hydrogen peroxide 184

14 xiv Figure 6.13 X-ray diffraction patterns of in situ prepared PVDF composite with Na+-mmt (4 wt%) (a) and C20A (3.5 wt%) (b) in the presence of benzoyl peroxide. 185 Figure 6.14 X-ray diffraction patterns of (a) physical mixture of PVDF-t-Si (M n =12000 g/mol) and Na+-mmt (90/10 weight ratio), (b) the same mixture after static melt-intercalation, and (c) the 50/50 mixture by weight of exfoliated PVDF-t-Si/Na+-mmt structure (from b) and neat PVDF 186 Figure 6.15 X-ray diffraction patterns of PVDF-t-Si (M n = 30,000 g/mol) / Na + - mmt hybrids with different compositions. PVDF-t-Si/Na+-mmt wt% : 95/5 wt%, before heating (a) and after annealing (b); 90/10wt%, before heating (c) and after annealing (d); 85/15wt%, before heating (e) and after annealing (f); 80/20wt%, before heating (g) and after annealing (h). Figure 6.16 XRD patterns of PVDF-t-Imm/Na+-mmt (5 wt%) hybrid. before heating (a) and after heating-annealing (b)

15 xv ACKNOWLEDGMENTS I would like to express my gratitude to Professors T.C. Chung and E. Manias, for their guidance, encouragement and many helpful suggestions throughout the course of this thesis. I would also like to thank Mr. Hiroyoshi Nakajima for doing the XRD of PP/2C18-mmt nanocomposite, Ms. Missy Hazen for taking most of the TEM images, researchers of Mitsubishi Co. for measuring mechanical property of nanocomposites, and Dr. Hong for providing PVDF-t-Si polymer as well as other colleagues in both two groups for their cooperation and friendship. In addition, I would like to acknowledge National Institute of Standards and Technology, Petroleum research foundation and Mitsubishi Co. for financial support, and the Polymer Science Program of Penn State University for helpful support. The proofread work done by Mr. Bill Saxton and Mr. Justin Langston is highly appreciated. Lastly, special thanks are also due to my family and the church for their help and encouragement.

16 1 Chapter I Metallocene Catalysts In α-olefin Polymerization 1.1. Introduction Metallocene is a class of organometallic compounds that contains a transition metal sandwiched by one or two cyclopentadienyl ligands. Metallocene catalysts used in α- olefin polymerization are usually prepared by the combination of a metallocene (catalyst) from a group of IVB metals (titanium, zirconium, and hafnium) and a co-catalyst to form an active metal cation. This cation contains both M-C species and an empty d-orbital suitable for α-olefin coordination/insertion reaction. The co-catalysts can be divided into two categories, including organoaluminium compounds (alkylalumiums and alkylaluminoxanes) and anionic counterions (borates or fluorinated borates), which form a weak coordination with the cationic active site. As early as the 1950s, 1-3 the metallocene/alkylaluminum system was reported for mediating ethylene polymerization. However, the polymer obtained was of low molecular weight and had extremely low catalytic activity, compared with those prepared by a heterogeneous Ziegler-Natta catalyst system (TiCl 4 /AlEt 3 ), discovered a few years earlier. Despite the ill-defined active sites, most of research activities had focused on Ziegler- Natta catalysts and the resulting polyolefins in the following two decades. This includes polyethylene (PE), polypropylene (PP), poly(4-methyl-1-pentene), ethylene-propylene elastomer (EPR), and ethylene-propylene-diene rubber (EPDM). The combination of commercial importance of polyolefin products and furious competition among the major producers had led to technological advancement in developing Zigeler-Natta catalysts

17 2 with superior activity and stereospecificity, as well as having led to economically viable production processes and product developments. On the other hand, the research activity in homogeneous metallocene catalyst system was relatively dormant until the discovery of methylaluminoxane (MAO) cocatalyst by Kaminsky and Sinn 4 in the late 1970s. The combination of metallocene, such as Cp 2 ZrCl 2, and MAO produces high polyethylene with an unprecedented catalytic activity (10 6 g PE/g Zr.hr.bar) that is much higher than the commercially used Ziegler- Natta catalysts. This historic discovery opened up another golden age in polyolefin research with intense competition across the world both in industrial and academic laboratories. In 1984, Ewen 5 made the first stereo-specific bridged metallocene catalyst (racethylidenebis(indenyl)titanium dichloride/mao) that produced isotactic polypropylene. He also discovered ansa-ziroconium metallocene (Me 2 C(Cp-9-Flu)ZrCl 2 ) for the production of highly syndiotactic polypropylene 6 a couple of years later. Around the same time, Kaminsky 7 synthesized highly isotactic polypropylene using a chiral ansa zirconocene/mao catalyst. In the late 1980s, Dow 8 discovered constrained geometry metallocene catalysts, which are titanium-based catalysts with monocyclopentadienyl (mono-cp) and a donor ligand stabilizing the metal center. The resulting open active site allows the incorporation of high α-olefins with comparative rates similar to that of ethylene. This brings in the significant advantage of metallocene catalysis to obtain polyolefin copolymers with narrow composition and molecular weight distributions, which can not be realized by heterogeneous Ziegler-Natta catalysts. Since then, worldwide industrial and academic research in metallocene catalysis for olefin

18 3 polymerization has been rapidly accelerating, and this advanced technology has been brought to commercialization. The important features of metallocene catalysts, that have been stimulating the rapid development of this technology to commercialization, include (a) a tunable single catalytic site to produce well-defined polymer with extremely high catalyst activity and predictable microstructure, (b) an excellent ability to incorporate high α-olefins, (c) producing polyolefin copolymers with narrow composition and molecular weight distributions, and (d) well-controlled polymerization mechanism allowing the incorporation of functional groups in the side chains and/or polymer chain end. With the appropriate choice of metallocene catalysts and reactive comonomers, a broad compositional range of functional polyolefin copolymers have been prepared with controlled molecular structures. On the other hand, with the combination of metallocene catalysts and reactive chain transfer agents, the formed polyolefin contains a terminal functional group (OH, NH 2, etc.), with narrow molecular weight distribution (M w /M n ~2) and controlled polymer molecular weight. Metallocene catalysts can be classified into several categories based on ligand symmetrical structures. This includes C 2, C 2v, C s, C 1, oscillating structures, and bridged half-metallocenes. The main features of these catalysts and polymerization mechanisms, as well as their applications in olefin polymerization, will be discussed in detail in the following section C 2v -symmetric Metallocenes The Cp-sandwiched Cp 2 ZrX 2 /MAO complexes (Cp=cyclopentadienyl; X= halogen or alkyl) with C 2v -symmetry is the earliest metallocene system that has been used for

19 4 ethylene homopolymerization with remarkably high activity. The application in propylene polymerization is less useful. They produce polypropylene with low activity and low molecular weight. 9 Unexpectedly, the Cp 2 TiPh 2 /MAO system can be used for the synthesis of predominately isotactic polypropylene below room temperature. The stereochemical structure of the resulting polymer indicates a chain-end controlled model in the polymerization. The structure of two typical C 2v -symmetric metallocenes is shown as follows: Me Me Cl Zr Cl Si Cl Zr Cl Cp 2 ZrCl 2 Me 2 Si(9-Flu) 2 ZrCl 2 By means of 13 C and 1 H NMR, mainly propyl and 2-methyl-prop-1-enyl (vinylidene) end-groups are detected in the polymers prepared by the Cp 2 MtX 2 (Mt = Zr, or Ti) catalyst. This indicates that the predominant monomer insertion is 1,2-insertion and that β-h elimination from a last-inserted 1,2 unit is the major chain transfer pathway. 10 In polypropylenes prepared with related catalyst systems (Cp*) 2 ZrX 2 /MAO (Cp* = η 5 - pentamethylcyclopentadienyl), iso-butyl and prop-1-enyl (vinyl) groups are found to be the major end groups, which suggests that the chain transfer consists of predominant β- methyl elimination. 11

20 5 Bridged C 2v -symmetric metallocenes, such as Me 2 Si(9-Flu) 2 ZrCl 2 (Flu = Fluorenyl), provide a convenient access to high molecular weight atactic polypropylene which performs like an elastomer Bridged C 2 -symmetric Metallocenes Chiral C 2 -symmetric bridged metallocenes are the most successful highly enantioselective polymerization catalysts. The catalysts produce polypropylenes with microstructures ranging from almost atactic to almost perfectly isotactic, and often contain a small amount of isolated region irregularities. The basic structures of these catalysts are shown as follows: The bridge (X), such as -Me 2 Si-, -Me 2 C-, or -CH 2 -CH 2 - connecting two Cp rings, prevents their rotation and locks them in a chiral configuration. In type (I), highly enantioseletive catalysts can be obtained when preferably bulky substituents are present at position 3,3 or 4,4. In type (II), the role of such substituents is played by the phenyl of the 1-indenyl moieties and additional substitution. Particularly at position 4,4 is highly beneficial. In both cases, ancillary substituents at positions 2,2 result in producing higher molecular weight polymers, due to the reduction of β-h elimination in the propagating chain end. 13

21 6 The stereoselectivity of catalysts during the α-olefin polymerization is well understood (10). Figure 1.1 shows the model for an active species representative of type (II), i.e. rac-me 2 C(1-Ind) 2 ZrCl The catalytic complex is pseudo-tetrahedral and cationic, with an iso-butyl group simulating the growing polymer chain and a propylene molecule at the remaining coordination site. The aromatic ligand is in the (R,R) configuration. The growing polymer chain must adopt a conformation that minimizes the steric interaction with one of the two phenyl rings of the bis-indenyl moiety. The first C- C bond has to be bent to one side. This favors the 1,2-insertion of propylene with the enantioface that brings the methyl substituent anti to that C-C bond (re-face for the (R,R)-catalytic complex). As the result, 2,1-insertion is always difficult due to direct steric interactions of the CH 3 group with the aromatic rings. At the completion of each insertion step, the growing chain will always reside at the coordination site previously occupied by the monomer (chain-migratory insertion mechanism). However, this orderly coordination/insertion process may not be always the case, dependent on the reaction condition (catalyst, temperature, etc.). There are some stereoerrors generated during the chain propagation. Generally, the C 2 symmetry ensures the equivalence of the two active sites, chain propagation is expected to be predominantly isotactic and site-controlled with occasional mmmmrrmmmm stereodefects.

22 7 Figure 1.1 Model of a (R,R)-Me 2 C(1-Ind) 2 Mt(iso-butyl)]+ cation (Mt=Zr) with a re η 2 -coordinated propylene molecule. 14 Table 1.1 shows the data 15 of liquid propylene polymerization in the presence of some representative C 2 -symmetric metallocene catalysts. The flexibility of the bridge between two indenyl rings plays an important role in the microstructure and properties of the resulting polymer. As comparing in No.1-3, stereoselectivity and molecular weight increase in the order Me 2 C< Et < Me 2 Si. The introduction of an alkyl substituent of the 2 position (No.4) in the ansa-bisindenyl zirconium complexes increases both the stereoregularity and molecular weight of the produced polypropylene. Combining substitutions in the 2- and 4- positions lead to some of the most successful isospecific zirconocene catalysts (No 5 and 6), which provide increased catalyst activity, isotacticity, and molecular weight. These catalysts have also been successfully supported on a carrier, retaining their high activity at much lower Al/Zr ratios, and have met the requirements for industrial scale production of i-pp. In general, increasing the temperature to a certain extent results in the decrease of stereoselectivity and molecular weight of the products, especially using ansa-metallocenes that contain flexible ethylene or silane bridge.

23 8 Table 1.1. Liquid Propylene Polymerization with some representative ansa-zirconocene catalysts 15 No Cat. Al/Zr Temp ( o C) Activity a mmmm (%) % 2,1 T m ( o C) M v ( 10 3 ) 1 rac-et(1-ind)zrcl rac-et(1-ind)zrcl rac-me 2 C(1-Ind)ZrCl rac-me 2 C(1-Ind)ZrCl (M n ) 3 rac-me 2 Si(1-Ind)ZrCl rac-me 2 Si(1-Ind)ZrCl 2 nr nr rac-me 2 Si(2-Me Ind)ZrCl 2 4 rac-me 2 Si(2-Me nr Ind)ZrCl 2 5 rac-me 2 Si(2-Me-4-Ph Ind)ZrCl 2 5 rac-me 2 Si(2-Me-4-Ph nr Ind)ZrCl 2 6 rac-me 2 Si(2-Me Benz[e]Ind)ZrCl 2 6 rac-me 2 Si(2-Me nr Benz[e]Ind)ZrCl 2 a. Activity: kg PP/mol.hr. In general, C 2 -symmetric metallocenes are not exceedingly regioselective in propylene polymerization. The resulted polypropylene samples usually contain regioirregular enchainments in the range of mol%. The end groups in the polymer chain are typical of saturated propyl groups and 2-methyl-prop-1-enyl (vinylidene) groups, due to the chain transfer by intramolecular β-h elimination. Additional saturated end-groups, such as iso-butyl and unsaturated but-2-enyl, are also observed, due to the chain transfer to AlMe 3 and β-h elimination, respectively, at a last inserted 2,1 unit. 16 One major drawback of rac-c 2 -symmetric ansa-metallocene catalysts is that either during the synthesis of the precursors or by subsequent epimerization, the meso-form with C s -symmetry can also be obtained. The active sites of catalytic species formed from meso-cs-symmetric precursors are obviously nonchirotopic, and cannot exert any stereocontrol on the chain propagation. As a result, some atactic polypropylene is invariably obtained. 17

24 Bridged C s -symmetric Metallocenes The first catalyst in producing highly syndiotactic polypropylene (s-pp) is C s - symmetric Me 2 C(Cp)(9-Flu)ZrCl 2 (Flu = Fluorenyl). 18 The structure is shown as follows: Although s-pp polymers in the experimental measurements always show lower melting points (T m ) than corresponding i-pp polymers, the theoretical prediction of a fully syndiotactic s-pp shall have T m of 214 o C, notably higher than that (186 o C) of i-pp. It s clear that a scientific challenge and industrial important area to develop is new metallocene catalysts in order to prepare s-pp polymers with high syndiotacticity. The invention of C s -symmetric metallocenes has marked a major turning point in understanding the stereocontrol mechanism of metallocene catalysis. Figure 1.2 shows the computer models 19 of the cationic catalytic species [Me 2 C(Cp)(9-Flu)Mt(iso-butyl)] + (Mt = Zr) with a η 2 -coorinated propylene molecule. The mutual arrangements of the first C-C bond of the iso-butyl group (simulating a growing polypropylene chain) and of the monomer are the ones which minimize the non-bonded interactions at the 1,2-insertion step. According to the chain migratory polymerization mechanism, 20 situations (a) and (b) alternate regularly during chain propagation. In two successive insertion steps opposite propylene enantifaces are inserted, therefore syndiotactic polypropylene results. The presence of isolated insertion errors, showing rrrrmmrrrr stereodefects, indicates a site control mechanism.

25 10 Figure 1.2 Models of a [Me 2 C(Cp)(9-Flu)Mt(iso-butyl)] + (Mt = Zr) with a η 2 -coorinated propylene molecule. 15 At low polymerization temperatures with liquid propylene, Me 2 C(Cp)(9- Flu)ZrCl 2 /MAO catalyst system can produce syndiotactic polypropylene with very high stereoregularity ([rrrr] > 0.95) 21, 22 and a melting point close to that of highly isotactic polypropylene (T m >150 o C). Generally C s -symmetric metallocene catalysts are highly regioselective and no regioerrors are found in the typical syndiotactic polypropylene. 23 The end groups are predominantly propyls and 2-methyl-prop-1-enyls (vinylidenes). At a lower monomer concentration, small amounts of iso-butyls are found due to the chain transfer to AlMe Bridged C 1 -symmetric Metallocenes C 1 -symmetric metallocenes are complexes lacking any symmetry element, which can be divided into two categories: a) those with one (substituted or nonsubstituted) cyclopentadienyl ligand having two homotopic faces (type I), and b) those with two asymmetric cyclopentadienyls (type II). The structures of the prototype C 1 -symmetric metallocenes are shown as follows:

26 11 C 1 -symmetric metallocenes present a synthetic advantage over C 2 -symmetric ones in terms of their application for isospecific polymerization of propylene. As mentioned previously, a problem associated with the synthesis of ansa-c 2 -symmetric metallocenes is that they are almost invariably generated along with their meso-isomers, which are difficult to remove from the catalyst mixture and often produce undesirable low molecular weight atactic polypropylene with certain polymerization activity. The synthesis of pure C 2 -symmetric catalysts usually requires multiple purification steps with low yields. 25, 26 In contrast, a meso-form does not exist in C 1 -symmetric system, and therefore the synthesis of ligand and catalyst are simple. The common feature of C 1 -symmetric metallocenes is that their two coordination sites are diastereotopic. On the size basis of the substitution on the cyclopentadienyl ligand, C 1 -symmetric catalysts can vary in stereoselectivity to form PP with hemiisospecific (amorphous hi-pp) to partially isospecific (amorphous or low crystallinity PP) to isospecific (i-pp with high T m and high crystallinity).

27 12 Figure 1.3 A computer model of the cationic catalytic species [Me 2 C(3-Me-Cp)(9-Flu)Zr(iso-butyl)]+ with growing polymer chain simulated by an iso-butyl group 15 The steroselectivity of C 1 -symmetric metallocenes in propylene polymerization can be predicted based on rationalization of the ligand structure. In the case of typical type I metallocene, Me 2 C(3-Me-Cp)(9-Flu)ZrCl 2, a computer model of the cationic catalytic species with a growing polymer chain simulated by an iso-butyl group is shown in Figure 1.3. Clearly, the model is enantioselective when the propylene molecule is inward coordinated, because structure A with a re-coordinated propylene is favored, compared to structure B with a si-coordinated propylene. This is due to repulsive interactions between the growing chain and the fluorenyl ligand in structure B. On the contrary, the model is nonenantioselective when the growing chain is σ-bonded in the inward coordination position, since structures C and D are equally hampered by repulsive interactions of the

28 13 growing chain, either with the 9-Flu ligand (structure C) or with the methyl group of the 3-MeCp ligand (structure D). According to a regular chain migration insertion mechanism, every other insertion is enantioselective, while the other is nonenantioselective. Therefore, the model is hemiisospecific. 27 In fact, polypropylenes with almost ideal hemiisotactic structure can be obtained with Me 2 C(3-Me-Cp)(9- Flu)ZrCl 2 /MAO under conditions of chain migratory propagation (relatively low temperature and high propylene concentration). 28 The stereoselectivity, however, strongly depends on the structure of the ligand. In particular, catalysts of type I with a bulky R substituent (e.g. R = t-butyl or trimethylsilyl) 29, as well as those of type II, 30 are predominantly isotactic-selective. The presence of the t-qwbutyl group or phenyl ring forbids the growing chain to be located in the inward position. Particularly, in the absence of the monomer molecule, probably occurring at the end of each insertion step, the steric stress of the ligand skeleton could force the growing chain to skip back to the less crowded outward position. Thus, insertion always occurs with the same relative disposition of the monomer and the growing chain, and the model turns out to be isospecific. Experimental results show that Me 2 C(3-t-Bu-Cp)(9-Flu)ZrCl 2 /MAO produces i-pp with relatively high mmmm pentad contents, 31 from 77.5% at 60 o C in liquid monomer to 87.8% at 50 o C and a pressure of 2 bar. 32 Spaleck and co-workers 33 reported the remarkable performance of C 1 -symmetric systems, illustrated below:

29 14 These catalysts provide both high steroeregularity and high molecular weights. For example, Me 2 Si(Ind)(2-Me-4-Ph-Ind)ZrCl 2 gives i-pp with mm=96%, 2,1 = 0.4%, T m = 155 o C and M w = 530,000, at the relatively high polymerization temperature of 70 o C. 33 Heterocene catalysts prepared by Ewen and coworkers Dithienocyclopentadineyl catalysts 34, 35 Recently, Ewen and coworkers have designed C 2, C s and C 1 -symmetric catalysts bearing heterocycle-condensed Cp ligands, as illustrated above, which add electronic effects to the stereocontrol. The complexes, containing isopropylidene-bridged cyclopentadienyl and cyclopentyl thiophene ligands, show activity comparable to those of fluorenyl analogues, and obey the same symmetry rules for polypropylene tacticity versus catalyst symmetry. Such catalysts can produce i-pp with an even higher melting temperature (161 o C) and molecular weight (M w > 1 million g/mol). However, when dithienocyclopentadiene is applied to substitute fluorenyl, the catalytic properties suffer greatly. Although low amounts of MAO are used to activate the metallocenes, only polypropylenes with low isotacticity and melting point are produced. 36 Dialkoxyl substituted C 1 -symmetric metallocene complexes, such as rac-[1-(5,6-dialkoxy-2-methyl- 1-η 5 -indenyl)-2-(9-η 5 -fluorenyl)ethane]zirconium dichloride activated with triisobutylaluminium and Ph 3 C[B(C 6 F 5 ) 4 ], also show few advantages compared with their

30 15 counterparts without heteroatom substitutes, in terms of catalytic activity, isotactity, 37, 38 molecular weight and melting temperature of the products. It should be noted that C 1 -symmetric metallocenes show a stereoselectivity increase with elevated polymerization temperature and lowered monomer concentration, a behavior opposite to that displayed by the C 2 -symmetric metallocenes. The regioregularity of propylene samples prepared with most C 1 -symmetric metallocene catalysts is fairly high. The predominant monomer insertion mode is 1,2. Isolated 2,1 and/or 3,1 units can be observed with <0.5 mol% Metallocenes with oscillating structure In unbridged metallocenes, like Cp 2 MtX 2 or Ind 2 MtX 2, the η 5 -coordinated aromatic rings rotate freely even at very low temperatures. However, the rotation may become hindered in the presence of a bulky substituent, and the rate of rotation "oscillating" may be dependent on the temperature. Of special interest are (2-Ar-Ind) 2 MtX 2 complexes (with Mt = Zr or Hf, and Ar = phenyl or a substituted aromatic moiety), which give rise to an equilibrium between two rotational isomers 40, 41 : one with quasi-c 2 -symmetry (raclike form, structure a) and the other with quasi-c s -symmetry (meso-like form, structure b), as illustrated below. Metallocenes with Oscillating structure

31 16 Polypropylenes prepared by (2-Ar-Ind) 2 MtX 2 /MAO catalysts contain both isotactic and atactic sequences, although with a complicated structure revealed by studies of solvent fractionation. These polymers perform as thermoplastic elestomers involving atactic sequences imparting a moderately elastic response and with isotactic sequences acting as physical crosslinks. 42 The study on their regioregularity shows the predominant 1,2 propylene insertion with low amounts of detectable 2,1 insertion ( mol%) only in predominately isotactic environment. 43 This indicates that not only the stereoselectivity, but also the regioselectivity, of the rac-like and meso-like forms of oscillating catalysts resembles that of the corresponding bridged bis-indenyl complexes Bridged half-metallocenes Many bridged and unbridged half-metallocenes have interesting catalytic behaviors. For example, CpTiCl 3 /MAO is a highly active catalyst for syndiotactic polymerization of 44, 45 styrene. However, such a catalyst cannot perform specific polymerization of propylene. Also important half-metallocenes are constrained geometry catalysts with the general formula Me 2 Si(Me 4 Cp)(NR)MtX 2 (Mt = Ti or Zr, R = alkyl, X = halogen or alkyl), 46, 47 illustrated below. This is well-known as Dow "Insite" catalyst.

32 17 The absence of a second Cp ring and the short bridge result in a very open environment of the transition metal, allowing an easier insertion of bulky monomers compared with bis-cp systems. This special feature endows such catalysts with excellent performance in the copolymerizations of ethylene with higher α-olefins and even with styrene. In the case of propylene polymerization with such catalysts, the resulting atactic polymers have poor regioregularity (with up to 5 mol% of 2,1 insertions). 48 The replacement of Cp by a 9-fluorenyl moiety, the metallocene promotes syndiotactic polymerization of propylene under site control. Syndiotactic polypropylene with [rrrr] as high as 0.77 was obtained with Me 2 Si(9-Flu)(N-t-Bu)ZrCl 2 /MAO catalyst. 49, 50 The stereoselectivity is due to the (pseudo-) C s -symmetry of the catalytic complex, and the stereocontrol mechanism is analogous to that for the C s -symmetric ansa-metallocenes Cocatalyst Cocatalysts play an important role in the single-site polymerization of olefins. The relationship between the structure of cocatalysts and their activation in the metalcatalyzed olefin polymerization has been reviewed recently. 51 MAO is the most widely used cocatalyst, since it is able to activate a large number of metallocenes and other soluble transition complexes. It is usually synthesized by controlled hydrolysis of trimethylaluminum (TEA). The techniques for the preparation of MAO have been reviewed by Reddy. 52 It is generally believed that MAO consists of oligomers involving many [-O-Al(CH 3 )-] units with the molecular weight of g/mol. However, the precise chemical structure and composition of MAO are still not clear as yet. A three-

33 18 dimensional cage structure with four coordinate aluminum centers, resembling a half open dodecahedron, was proposed by some researchers. 53, 54 The cage can seize an anion from the metallocene to form a stable AlX 4 - anion due to the delocalization of the electron over the whole cage. The utilities of MAO mainly include alkylation of halogenated metallocene complexes, serving as scavengers for moisture or other impurities, and stabilization of active species. Usually, a large excess amount of MAO (Al/Mt = ) is required to reach desirable activity. This is a major drawback of using MAO as a cocatalyst. In addition, conventional MAO has very low solubility in aliphatic solvents, as well as poor storage stability in solution. Therefore, commercially modified methylaluminoxane (MMAO) is prepared by controlled hydrolysis of a mixture consisting of trimethylaluminum and triisobutylaluminum. This MMAO exhibits improved solution storage stability and solubility, and can be produced at a lower cost while providing good polymerization efficiency. The other important cocatalyst is perfluoroaryl boranes 55, 56, such as B(C 6 H 5 ) 3, and highly Lewis acidic B(C 6 F 5 ) 3, NR 3 H + B(C 6 F 5 ) 4- and Ph 3 C + B(C 6 F 5 ) - 4, which can serve as substituents for MAO in the combination with metallocene dialkyls. A relatively small amount of boranes (borane/mt=1:1) is needed to form an effective catalyst for olefin polymerization. Unfortunately, this catalyst is rather sensitive to moisture and other impurities, and the stability of the resulting metallocene complexes is much lower than that formed with MAO. Usually alklyaluminum (such as triisobutylaluminum) is added to the systems to scavenge the impurities and to stabilize the complexes. 57

34 Supported metallocene catalysts As discussed, homogeneous metallocene catalysts show special characteristics, such as single active site, extremely high catalytic activity, a capability to incorporate a high content of monomer, a narrow molecular weight distribution, and control of stereoregularity. However, the products prepared from homogeneous metallocenes exhibit relatively poor apparent properties, compared with those from heterogeneous Ziegler-Natta catalysts. They show an irregular particle shape and low bulk density, and contain a large amount of fine powder. Because such problems are the major concerns for the industrial application in slurry or gas-phase polymerization, immobilization of singlesite metallocene on certain supports may improve the morphology of the final products and make it more adaptable to the industrial process. Supported catalysts also show additional advantages in cost due to a less amounts of MAO required, and resulting in products with a higher molecular weight, a higher melting temperature, and better stereoregularity. 58,59 Recently, heterogeneous metallocene catalysts for olefin polymerization have been reviewed. 60, 61 The carriers include inorganic supports, such as silica, 62 Al 2 O 3, 63 MgCl 2, 64, 65 clay, zeolite, 69 CaCO 3, 70 and organic polymers, such as porous powdered high-density polyethylene, isotactic polypropylene, nylon granules 71 and copolymers of styrene and divinylbenzene. 72 One of the great breakthroughs in heterogeneous metallocene catalysts was made by investigators at Mitsubishi. 73 Acidic clays are formed by reacting AlR 3 with clay, such as montmorillonite, hectorite, and mica, before contacting with metallocenes to form active metallocenes/alr 3 complexes supported on clay. Overall, the AlR 3 -treated clay minerals serve as both activators and carriers. The resulting catalysts show high activity comparable to those with MAO as

35 20 cocatalyst, producing polyethylene (co)polymers with improved morphology and molecular weight distribution. This system has also been commercialized for propylene polymerization to form highly isotactic polypropylene Outlook Metallocene catalysts have been emerging as a promising catalyst technology for the preparation of polyolefins, both in the research laboratories and industrial practice. Several important features, centered on tunable single catalytic site and well-controlled polymerization mechanism, have presented to polymer chemists with unprecedented flexibility to design and synthesize new polymers and to improve the existing polyolefin commodities with higher performance. 75 One long term challenge in the polymer research community has been the preparation of functional polyolefins containing polar groups, which are very difficult to prepare by other methods. This has been a constant scientifically challenging and industrially important area, since the discovery of HDPE and i-pp about a half century ago. The constant interest, despite lack of effective functionalization chemistry, is due to the strong desire to improve polyolefin s poor interactive properties. The hydrophobicity and low surface energy of polyolefin has limited its applications, especially in the areas of coating, blends and composites, in which adhesion, comparability, dispersion, and paintability are paramount. In Chapter II, I will discuss the current synthesis approaches in the preparation of functional polyolefins and their application in nanocomposites.

36 21 References: 1 Wilkinson, G.; Briminham, I. M. J. Am. Chem. Soc. 1954, 76, Breslow, D.S.; Newburg, N.R. J. Am. Chem. Soc. 1957, 79, Natta, G.; Pino, P.; Corradini, P.; Danusso, F.; Namtica, E.; Mazzanti, G.; Moraglio, G. J. Am. Chem. Soc. 1955, 77, Sinn, H,; Kaminsky, W. Adv. Organomet. Chem. 1980, 18, Ewen, J.A. J. Am. Chem. Soc. 1984, 106, Ewen, J.A.; Jones, R.L.; Razavi, A.; Ferrara, J.D. J. Am. Chem. Soc. 1988, 110, Kaminsky, W.; Kulper, K.; Brintzinger, H.H.; Wild, F.R.W.P. Angew, Chem., Int. Ed. Engl. 1985, 24, Stevens, J.C. Stud. Surf. Sci. Catal. 1994, 89, Kulper, K.; Niedoba, S. Makromol. Chem. Macromol. Symp. 1986, 3, Tsuisui, T.; Mizuno, A.l Kashiwa, N. Polymer 1989, 30, Resconi, L.; Piemontesi, F.; Franciscono, G.; Abis, L.; Fiorani, T. J. Am. Chem. Soc. 1992, 114, 1025 Resconi, L.; Jones, R.L.; Rheingold, A.L.; Yap, GPA. Organometallics 1996, 15, Busico, V.; Cipullo, R. Prog. Polym. Sci. 2001, 26, Corradini, P.; Cavallo, L.; Guerra, G. In: Schevis, J.; Kamisky, W. editors. Metallocene-based polyolefins, vol.2 Chichester: Wiley, 2000, p Resconi, L.; Cavallo, L.; Fait, A.; Piemontesi, F. Chem. Rev. 2000, 100, Brintzinger, H.H.; Fischer, D.; Mulhaupt, R.; Rieger, B.; Waymouth, R.M. Angew. Chem. Int Ed. Engl. 1995, 34, Kamisky, W. Macromol. Chem. Phys. 1996, 197, Ewen, J.A.; Jones, R.L.; Razavi, A.; Ferrara, J.D. J. Am. Chem. Soc. 1988, 110, Cavallo, L.; Guerra, G.; Vacatello, M.; Corradini, P. Macromolecules 1991, 24, van der Leek, Y.; Angermund, K.; Reffke, M.; Kleinschmidt, R.; Goretzki, R.; Fink, G. Chem. Eur. J. 1997, 3, 585.

37 22 20 Farina, M.; Di Silvestro, G.; Terragni, A. Macromol. Chem. Phys. 1995, 196, De Rosa, C.; Auriemma, F.; Vinti, V.; Galimberti, M. Macromolecules, 1991, 24, Gureea, G.; Longo, P.; Cavallo, L.; Corradini, P.; Resconi, L. J. Am. Chem. Soc. 1997, 119, Longo, P.; Proto, A.; Grassi, A.; Ammendola, P. Macromolecules, 1991, 13, Speleck, W.; Kuber, F.; Winter, A.; Rohrmann, J.; Bachmann, B.; Antberg, M.; Dolle, V.; Paulus, E. Organometallics 1994, 13, Stehling, U.; Diehold, J.; Kirsten, R.; Roll, W.; Brintzinger, H. H.; Jungling, S.; Mulhaupt, R.; Langhauser, F.; Organometallics 1994, 13, Guerra, G.; Cavallo, L.; Moscardi, G.; Vacatello, M.; Corradini, P. Macromolecules 1996, 29, Ewen, J.A.; Elder, M.J.; Jones, R.L.; Haspeslagh, L. Atwood, J.L.; Bott, S.G.; Robinson, K. Macromol. Symp. 1991, 48, Razavi, A.; Peter, L.; Nafpliotis, L.; Vereecke, D.; Den Dauw, K.; Atwood, J.L.; Thewald, U. Makromol. Chem. Macromol. Symp. 1995, 89, Rieger, B.; Jany, G.; Fawzi, R.; Steimann, M. Organometallics 1994, 13, Razavi, A.; Vereecke, D.; Peters, L.; Den Dauw, K.; Nafpliotis, L.; Atwood, J.L. In Ziegler Catalysts; Fink, G.; Mulhaupt, R.; Britzinger, H.H. Eds.; Springer-Verlag: Berlin, 1995, p Kleinschmidt, R.; Reffke, M.; Fink, G. Macromol. Rapid Commun. 1999, 20, Spaleck, W.; Kuber, F.; Bachmann, B.; Fritze, C.; Winter, A. J. Mol. Catal. A. Chem. 1998, 128, Ewen, J.A., Jones, R.L., Elder, M.J.; Rheingold, A.L.; Liable-Sands, L. M. J. Am. Chem. Soc. 1998, 120, Ewen, J.A., Jones, R.L., Elder, M.J.; Rheingold, A.L.; Liable-Sands, L. M.; Sommer, R.D. J. Am. Chem. Soc. 2001, 123, Nifant ev, I.E.; Laishevtsev, I.; Ivchenko, P.V.; Kashulin, I.A.; Guidotti, S.; Piemontesi, F.; Camurati, I.; Resconi, L.; Klusener, P.A.A.; Rijsemus, J.J.H.; de Kloe, K.P.; Korndorffer, F.M. Macromol. Chem. Phys. 2004, 205, Schlogl, M.; Rieger, B. Z. Naturforsch 59b, Kukral, J.; Lehmus, P.; Klinga, M.; Leskela, M.; Bieger, B. Eur. J. Inorg. Chem. 2002, Busico, V.; Cippullo, R.; Talarico, G. Macromolecules, 1997, 30, 4786

38 23 39 Coates, G.W., Waymoath, R.M.; Science 1995, 267, Witte, P.; Lal J.K.; Waymouth, R.M. Organometallics 1999, 18, Hu, Y.; Krejchi, M.T.; Shah, C.D.; Myers, C.L.; Waymouth, R.M. Macromolecules 1998, 31, Lin, S.; Waymouth, R.M. Macromolecules 1999, 32, Ishihara, N., Kuramoto, M.; Uoi, M. Macromolecules 1988, 21, Yokata, K.; Inoue, T.; Naganuma, S.; Shozaki, H.; Tomotsu, N.; Kuramoto, M.; Ishihara, N. Metalorganic catalysts for synthesis and polymerization. Berlin: Springer, 1999 p Okuda, J. In: Kaminsky, W., editor. Metalorganic catalysts for synthesis and polymerization. Berlin: Springer, 1999 p Stevens, J.C., Timmers, F.J.; Rosen, G.W.; Lai, S.Y., Eur. Patent Appl EP A2, Mcknight, A.L.; Masood, MdA.; Waymouth, R.M.; Straus, D.A. Organometallics 1997, 16, Shiomura, T.; Asakura, T.; Inoue, N. Macromol. Rapid Commun. 1996, 17, 9 49 Shiomura, T.; Asakura, T.; Sunaga, T. Macromol. Rapid Commun. 1997, 29, Chen, E.Y.; Marks, T.J. Chem. Rev. 2000, 100, Reddy, S.S.; Sivaram, S. Prog. Polym. Sci. 1995, 20, Sinn, H. Macromol. Symp. 1995, 97, Koide, Y.; Bott, S.G.; Barron, A.R. Organometallics 1996, 15, Yang, X. Stern, C.L.; Marks, T.J. J. Am. Chem. Soc. 1994, 116, Bochmann, M.; Lancaster, S. J. Organometallics 1992, 12, Tsai, W. Rausch, M.D.; Chien, J.C.W. Appl. Organomet. Chem. 1993, 7, Kaminsky, W. Macromol. Symp. 1995, 80, Sacchi, M.C.; Zucchi, D.; Tritto, I.; Locatelli, P. Macromol. Rapid Commun 1995, 16, Hlakty, G. Chem.Rev. 2000, 100, Wang, W.; Wang, L. J. Polym. Mater. 2003, 20, 1 61 Takahashi, T.; Yamamoto, K.; Hirakawa, K. US Patent 5,474, Kaminaka, M.; Soga, K. Polymer 1992, 33, Soga, K.; Arai, T.; Uozumi, T. Polymer 1997, 38, Sensarma, S. Sivaram, S. Macromol. Chem. Phys. 1999, 200, 323

39 24 65 Yano, A.; Santo, M. US Patent 5,380,820, Kaneko, T.; Yano, A. Eur. Pat. Appl. 849,292, Suzuki, T.; Suga, Y. Polym. Prepr. 1997, 38(1), Woo, S.; Ko, Y.; Han, T. Macromol. Rapid Commun. 1995, 16, Lee, D.; Yoon, K. Macromol. Symp. 1995, 97, Herrmann, H.F.; Bachmann, B.; Hierholzer, B.; Spaleck, W. U.S. Patent 5,942,586, Sangokoya, S.A. US Patent 5,308,815, Suga, Y.; Suzuki, T.; Sugano, T.; Tayano, T.; Shimizu, F. Kobunshi Ronbunshu, 2002, 59(4), Sugano, T.; Iwama, N.; Isobe, E.; Suzuki, T.; Maruyama, Y. US Patent Appl. 2002/ Tullo, A. H. Chem.& En. News, 2001, 79(43), 35

40 25 Chapter II Progress in Functional Polyolefin and Fluoropolymers and Application in Polymer/clay Nanocomposites 2.1. Introduction Nowadays, the most widely used commercial polymers belong to the polyolefin family, which range from thermoplastics, including polyethylene (PE), polypropylene (PP), poly(4-methyl-1-petent), and polystyrene (PS), to elastomers, such as ethylenepropylene copolymer (EPR), and ethylene-propylene-diene rubber (EPDM). 1 They provide an excellent combination of mechanical and chemical properties, low cost, superior processability, and recyclability. Despite this great success, some inherent deficiencies in polyolefin, such as a lack of reactive functional groups in the polymer structure, have limited their further applications, particularly those in which adhesion, dyeability, paintability, printability or compatibility with other polar polymers is paramount. Logically, this problem can be overcome by introducing suitable functional groups in polyolefin, which provide interactive properties with other materials, such as pigments, paints, carbon black, glass fiber, and other polymers Three Approaches in the Functionalization of Polyolefin Theoretically, there are three possible approaches to the functionalization of polyolefin, which include (a) direct copolymerization of α-olefin with functional monomer, (b) chemical modification of the preformed polymer, and (c) reactive copolymer approach by incorporating reactive comonomers that can be selectively and effectively interconverted to functional groups. 2,3

41 26 The direct process could be an ideal one-step reaction if the copolymerization reaction with functional monomers would be as effective and straightforward as the corresponding homopolymerization reaction. Unfortunately, some fundamental chemical difficulties, i.e., poisoning of conventional catalyst containing Ti, Zr, Hf, or V component and other side reactions, bring about serious problems of low catalytic activity and inhibited polymerization. 4,5 In recent years the combination of a less oxophilic late transition metal catalyst, such as Fe, Ni, Co or Pd complexes, and an electronic-protected functional group, may remarkably enhance catalyst activity. 6 Such late transition metal polymerizations, however, usually afford some polymers containing a branched molecular structure with relatively low or no melting temperature and crystallinity. There is no example of steric specific polymerization by using these catalysts to prepare functionalized i-pp so far. Most of the commercial functionalization processes are based on postpolymerization reactions, which usually required vigorous reaction conditions, due to the inert nature of PE, PP, and EP polymers. The functionalization reaction involves free radical activation of the stable (saturated) polymer chain by breaking some high energy C-H bonds to form free radicals along the polymer chain. Ideally, these polymeric radicals then react with chemical reagents, such as maleic anhydride, 7 presented in the system. However, many side reactions, such as crosslinking, degradation, and homopolymerization, can also happen, and the resulting functional polymer is far from ideal, usually containing low functional group content and by-products. The third functionalization approach is a relatively new one developed in our laboratory. 8,9,10 The basic idea is to design a reactive copolymer intermediate that can

42 27 be effectively synthesized and subsequently interconverted to a functional polymer, thus overcoming the chemical difficulties in both direct and post-polymerization processes. 11,12 Several new reactive comonomers can be effectively incorporated into polyolefins in the side chain or at the chain end via metallocene catalysis polymerization. 13,14 In turn, these reactive groups incorporated in the polymer open up a lot of possibility for producing new polyolefin products, including functional graft and block copolymers, which would be very difficult to prepare by other methods. Three Reactive Comonomers Borane monomer p-methylstyrene (p-ms) Divinylbenzene (DVB) H 2 C CH H 2 C CH H 2 C CH (CH 2 ) 4 B CH 3 H 2 C CH The key factors in designing a reactive comonomer include that (a) the reactive group must be stable to metallocene catalysts and soluble in hydrocarbon polymerization media, (b) the reactive monomer should have good copolymerization reactivity with α- olefins, and (c) the reactive group must be facile in the subsequent interconversion reaction to a functional (polar) group or form a stable initiator for the polymerization of functional monomers. Three reactive comonomers, including borane monomers, p- methylstyrene (p-ms), and divinylbenzene (DVB), have been effectively incorporated into polyolefins with narrow molecular weight and composition distributions. 15,16 One example is shown in scheme 2.1, involving direct metallocene copolymerization of ethylene and 5-hexenyl-9-BBN that is completely stable with the

43 28 catalytic site, and can be effectively incorporated into polyolefin. The solubility of borane monomers and polymers in the reaction media (hexane and toluene) during metallocene polymerization is essential to assure the optimum reaction condition for high polymer. + H 2 C CH 2 H 2 C CH 2 Metallocene catalyst CH 2 -CH 2 x CH 2 -CH y (CH 2 ) 4 (CH 2 ) 4 B B NaOH/H 2 O 2 1. O 2 2.MMA CH 2 -CH 2 x CH 2 -CH y CH 2 -CH 2 x CH 2 -CH y (CH 2 ) 4 (CH 2 ) 4 OH O PMMA Scheme 2.1 Reactive comonomer approach to prepare functional polyolefins In turn, the formed polyolefin contains reactive borane groups that can be selectively interconverted to the desirable functional groups (such as OH, NH 2, and halides) under mild reaction conditions. This conversion can also be accomplished during the melt process (reactive extrusion). In addition, the borane side groups in polyolefin can then be selectively oxidized at the aliphatic C-B group for graft-form polymerization. The formed peroxyborane (B-O-O-C) initiates a control radical polymerization in the presence of free radical-polymerizable monomers, such as methacrylates, vinyl acetate, acrylonitrile, etc., at ambient temperature. This control radical polymerization process minimizes the undesirable chain transfer reaction and termination (coupling and disproportional) reaction between the two growing chain ends, which result in the

44 29 formation of homopolymers and crosslinked material. In most cases, the resulting graft copolymers are completely soluble and processible. The graft length (PMMA side chain) is controlled by the MMA concentration and reaction time. Some interesting graft polymers - including PE-g-PVA, PE-g-PMMA, PP-g-PMMA, PP-g-PMA, PP-g-PVA, and EP-g-PMMA - have been synthesized with controllable compositions and molecular microstructures. In the other two cases, p-methylstyrene and divinylbenzene units have been incorporated in polyolefin copolymers with a broad compositional range and narrow molecular weight and composition distributions. The resulting pendent reactive sites, i.e. benzylic protons, and styrene units, located along the polyolefin backbone, are versatile in the subsequent transformation reactions, both in functionalization and graft reactions. The transformation reactions are selectively taking place at the reactive sites. In other words, the concentration on functional groups is basically proportional to the concentration of the reactive sites, and the resulting functional polyolefin has a welldefined molecular structure Functionalization of Polyolefins via Reactive Chain Transfer Agent The reactive polyolefin approach has also been expanded to the preparation of polyolefin containing a reactive terminal group (borane 17,18 or p-methylstyrene 19 ) and polyolefin diblock copolymers. As shown in Scheme 2.2, polyolefin containing a terminal borane group (alkyl-9-bbn) can be prepared by two routes, including an in situ chain transfer reaction to a B-H group during metallocene polymerization and hydroboration of the chain-end unsaturated polymer. After spontaneous oxidation of the

45 30 terminal borane groups, the formed polymeric radical initiates living radical polymerization in the presence of polar monomer. Diblock copolymers, such as PE-b- PMMA and PP-b-PMMA, with a well-defined structure and narrow molecular weight distribution have successfully been synthesized. Hydroboration Reaction Chain-end Unsaturated Polyolefin Chain-Transfer Reaction a-olefin H-BR 2 metallocene catalyst H-BR 2 or p-methylstyrene/h 2 Polyolefin Containing a Reactive Terminal Group (Borane or p-ms group) Free radical or anionic graft-from polymerization Polyolefin Diblock Copolymer Scheme 2.2 [2] On the other hand, p-methylstyrene terminated polyolefins have been prepared by the combination of metallocene catalysis and chain transfer reaction to p-ms/h 2. The terminal p-ms group provides an active site for lithiation and subsequent living anionic polymerization. Diblock copolymers, such as PP-b-PS and PP-b-PMMA, have been synthesized via this transformation process from metallocene to living (radical and anionic) polymeriztion. In addition, the lithiated p-ms end group can be converted to a terminal OH, or COOH group in the presence of ethylene oxide, or CO 2, respectively, to form PP with a terminal OH, or COOH group.

46 31 A - L Zr CH 3 (CH 2 -CH) x CH 2 =CH (St-NSi 2 ) L (II) (CH 2 ) 2 CH 3 k 12 2,1-insertion N R 3 Si SiR 3 CH 3 x CH 2 =CH k 11 A - L Zr -CH-CH 2 - (CH 2 -CH) x L L A - L Zr H (I) Cat: rac-me 2 Si[2-Me-4-Ph(Ind)] 2 ZrCl 2 /MAO (CH 2 ) 2 N R 3 Si SiR 3 (III) k 2 H 2 CH 3 (CH-CH 2 ) x -CH 2 -CH 2 - (CH 2 ) 2 -N(SiR 3 ) 2 (IV) HCl (in work-up step) CH 3 (CH-CH 2 ) x -CH 2 -CH 2 - (CH 2 ) 2 -NH 2 (V) Scheme 2.3 Chain transfer reaction mechanism during metallocene catalyzed polymerization of propylene It is interesting to point out that this chain transfer approach has also been applied to prepare amino-terminated polypropylene. Scheme 2.3 illustrates the general reaction mechanism. During the course of propylene 1,2-insertion, the propagating M + -C site (II) eventually reacts with a St-NSi 2 unit (k 12 ) (via 2,1-insertion) to form a St-NSi 2 -capped propagating site (III) with an adjacent phenyl group interacting with metal cation. The new propagating site (III) is incapable of continuing the insertion of St-NSi2 (k 22 ) or propylene (k 21 ) due to the steric jamming. However, it can react with hydrogen to complete the chain transfer reaction. This consecutive reaction with St-NSi 2 and

47 32 hydrogen results in a PP-t-St-NSi 2 polymer chain (IV), and a regenerated Zr-H species (I) reinitiates the polymerization of propylene and continues the polymerization cycles. After the polymerization is complete, the desirable NH 2 terminal group in PP-t-NH 2 (V) can be easily recovered during the sample workup step using acidic solution. Because amino group is very sensitive to the metallocene cationic site, the silaneprotected 4-{2-[N,N-bis(trimethylsilyl)amino]ethyl}styrene (St-NSi 2 ) is used as a chain transfer agent in the preparation of PP-t-NH 2. The silane protecting group is very effective in preventing catalyst poison during the polymerization, and is easy to be removed by acidic solution during the sample workup procedure. Therefore, the additional protection-deprotection step does not bring about significant cost and deteriorating effects on the polymerization. The highly isospecific rac-me 2 Si{2-Me-4- Ph-(Ind)] 2 ZrCl 2 /MAO complex used in the commercial production of i-pp is suitable for this reaction. This catalyst system produces i-pp with a high molecular weight and high melting temperature, and the bulky ligands around its specific opening active site may further prevent the catalyst from interacting with the protected functional group. This catalyst has also shown a selective chain transfer reaction to styrene or p-ms during the propylene polymerization in the presence of hydrogen Progress in Functionalized Fluoropolymers Fluoropolymers are another hydrophobic polymer family with superior properties. Fluoropolymers exhibit a unique combination of desirable physical and chemical properties, including high thermal and chemical stabilities (in the presence of acids, solvents, and petroleum), low dissipation factor, low water absorptivity, excellent weatherability, good resistance to oxidation and ageing, low flammability, and very

48 33 interesting surface properties. In spite of their high price, their excellent performances allow their use in numerous high end applications, such as aerospace, aeronautics, optics, textile finishings, military uses, and microelectronics. 20, 21, 22 In recent years, some fluoropolymers, exhibiting high dielectric constants and large electromechanical responses, have received considerable attention in many applications, such as transducers, actuators, sensors, and capacitors. Among this fluoropolymer family, poly(vinylidene fluoride) (PVDF) and its copolymers or terpolymers, especially vinylidiene fluoride/trifluoroethylene (VDF/TrFE), are the most-investigated ferroelectric polymer system with high piezoelectricity and pyroelectricity. In general, there are two classes of fluoropolymers, based on structure, properties, and applications. The first class of polymers is hydrocarbon polymers which contain fluorinated groups in lateral positions along the polymer chain. These polymers are mainly used to enhance surface properties, but they cannot serve as thermostable materials resistant to oxidation and to chemicals. The second class of fluoropolymers contains structures with fluorinated polymer backbones. This class of fluoropolymers exhibit superior thermal properties and resistant to aggressive media, under UV radiation and aging. In our research, we focus on synthesis and applications of the latter fluoropolymers, which will be discussed in Chapter V. However, the lack of functional groups in the fluoropolymers also limits their further applications where adhesion, dyeability, paintability, printability or miscibility with other polar materials is a requirement. It is interesting to investigate new chemical routes to prepare functional fluoropolymer containing desirable functional groups, and to study the applications of these functional fluoropolymer, especially the improved

49 34 interfacial properties between fluoropolymer and other materials that are otherwise immiscible with fluoropolymer. In general, there are two synthetic approaches for the preparation of functional fluoropolymers containing either side-chain or chain-end functional groups. They include (i) the copolymerization of fluorinated monomer with functional comonomers to form functional fluro-copolymers containing pendant functional groups; and (ii) control polymerization using functional initiators or chain transfer agents to prepare fluoropolymers containing terminal functional groups Copolymerization with functional fluorinated monomer In general, the control in the copolymerization between fluorinated monomer and non-fluorinated monomer is very difficult, due to the opposite e values (the inductive effect of the monomer). In fact, several reports showed copolymerization reactions between fluorinated monomers and non-fluorinated monomers, such as ethylene, vinyl ether, and N-vinyl pyrroridone, with a strong tendency to form alternative copolymer structures. 24 In addition, the introduction of the non-fluorinated functional comonomers significantly degrades the properties of the final products. To preserve the desirable fluoropolymer properties, it is highly desirable to use the fluorinated functional comonomers containing a CF 2 =CF- vinyl group. In the past decades, 25,26 there are considerable interest in the copolymerization reactions between fluoromonomers, such as VDF, chlorotrifluoroethylene (CTFE), or tetrafluoroethylene (TFE), and the fluorinated functional comonomers. The resulting functional fluorinated copolymers are used in many applications, ranging from perfluorinated ionomers, and thermosetting coatings, to fluoroelastomers. The functional

50 35 group introduced in a lateral position along the backbone of the copolymer endows specific and complementary properties, such as carboxylic groups for adhesion, 27 hydroxyl or epoxide functional groups for crosslinking, and phosphonated groups for heat and chemical resistance. 28,29 The preparation of fluoropolymers typically requires special reaction conditions, due to the relatively low reactivity of fluoromonomers, which can be attributed to the electron deficiency and lack of resonance in the double bond. In the past decades, two processes, including suspension and emulsion processes in aqueous solution, are evolved in the polymerization of fluorinated monomer. Both processes enjoy bulk reaction conditions with good heat dispersion. Many catalyst systems have been employed, including the inorganic peroxides (potassium, sodium, or ammonium persulfate) and organic peroxides (dibenzoyl peroxide, diacetyl peroxides or di-tert-butylperoxide). Both inorganic and organic initiators have some disadvantages. The inorganic peroxy initiators produce polymer with less processability and somewhat thermal instability, while organic peroxide initiators require extreme conditions in polymerization, such as high pressure, high temperature, and long reaction time, to achieve a reasonable yield. Although a great variety of random fluorinated copolymers have been produced in industrial scales, mainly for coating applications, most of the functional comonomers are very expensive and not commercially available. In addition, the functional comonomers usually show relatively low reactivities, comparing with those of VDF and TFE (as shown in Table 2.1). 30 The mismatching in reactivity usually results in the formation of low molecular weight copolymers with low monomer conversions and low incorporation of functional comonomers, and the superior properties of fluoropolymers aforementioned

51 36 may suffer to certain extent. 31,32,33,34,35 It is not surprising that further reactions such as crosslinking are necessary for many fluorocopolymer applications. Table 2.1 Reactivity ratio of Functional Fluorinated comonomers with TFE or VDF Monomer r TFE r M Reference Monomer r VDF r M Reference F 2 C=CFCH 2 OH F 2 C=CFOCF F 2 C=CFC 3 H 6 OH F 2 C=CF(CH 2 ) 3 OAc F 2 C=CFCO 2 CH F 2 C=CF(CH 2 ) 3 SAc F 2 C=CFOR F CO 2 R ,29 F 2 C=C(CF 3 )COF F 2 C=CFOR F SO 2 F Chain-end Functional Fluoropolymer The other type of functional fluoropolymers has a telechelic structure, containing functional groups at both chain ends. In general, the chemistry to prepare telechelic fluoropolymer is very limited. In polymer chemistry, the chain end functionalized polymer is usually prepared by a combination of living polymerization and selective termination of the living polymers with suitable reagents. There is no example of living radical polymerization of fluoromonomers and functionalization at the polymer chain end. A few examples of controlling chain end structure include the use of functional initiators, which was pioneered by Rice and Sandberg at the 3M Company. 36,37 They reported the preparation of low molecular weight VDF/HFP elastomers containing two ester terminal groups by using a diester peroxide initiator. The average functionality of the resulting telechelic VDF/HFP elastomer was not reported. However, it is logical to expect some difficulties in achieving a perfect telechelic structure by using regular free radical polymerization involving many side reactions in the termination step. Recently, Saint-Loup et al 38 also attempted to prepare chain end functionalized VDF/HFP

52 37 elastomers containing two opposing hydroxy terminal groups by using hydrogen peroxide as an initiator. Several advantages of using hydrogen peroxide initiator include cost, high reactivity, and directly forming hydroxy terminal groups. However, the resulting fluoropolymers contain both hydroxyl groups and substantial numbers of COOH groups at both chain ends, because the instable CF 2 -OH group is further decomposed to a COOH end group. Consequently, further selective reduction with lithium hydride is required to convert the COOH end groups to hydroxyl groups. Telechelic diol fluoropolymers with low molecular weight ( ) and low glass transition temperatures (ranging between -77 and -41 o C) are obtained, which may serve as intermediates for further polycondensation or polyaddition reactions. Scheme 2.4 Polymerization Mechanism via RAFT process

53 38 The most useful method to prepare telechelic fluropolymers containing two terminal functional groups is iodine transfer polymerization (ITP), which was disclosed in the late 1970s by Tatemoto et al. 39,40 The chemistry is based on the combination of a reversible addition-fragmentation chain transfer (RAFT) process and an α, ω- diiodoperfluoroalkane (I-R f -I) chain transfer agent. As shown in Scheme 2.4, due to low dissociation energy, the CF 2 -I bond can be easily cleaved at certain temperatures and participate in the chain transfer reaction during the polymerization of the fluoromonomer. 41 It is interesting to note that ITP shows controlled or pseudo-living polymerization characteristics, i.e., the molecular weight increases with conversion in a linear fashion while molecular weight distribution maintains relatively narrow (M w /M n = ). Several companies have taken advantage of this ITP technology to produce fluoroelastomers with special properties. Daikin 42 has been able to conduct emulsion terpolymerization of VDF, TFE and HFP with peroxides as initiators to produce thermoplastic elastomers call Dai-El. These commercially available thermoplastic elastomers exhibit very interesting properties, such as a high specific volume (1.90 ml/g), high melting point of o C, high thermostability up to o C and good surface properties. These characteristics offer excellent resistance against aggressive chemicals and strong acids, fuels, and oils. In addition, tensile modulus is comparable to that of cured fluoroelastomers. A decade latter Du Pont 43 and Ausimont 44 companies have also been attracted by this concept and used IC 4 F 8 I and I(C 2 H 4 (TFE) n C 2 F 4 I, respectively, in the controlled radical polymerization of VDF, TFE or TrFE, and also in the VDF/HFP copolymerization. The molecular weight distribution can be as low as 1.87.

54 39 Recently, 2,2'-azobisisobutyronitrile (AIBN) was discovered in our group to serve as an effective initiator superior to other initiators such as peroxides in the ITP process. 45 Due to a low decomposition temperature of AIBN (at 70 o C, half-life = 4.8 hour), 46 some undesirable side reactions, such as elimination of HF in the polymer backbone, can be completely avoided. Most important of all, carbon radicals generated from the decomposition of AIBN are extremely sluggish in the polymerization of fluoromonmers. Therefore, the carbon radicals have to react with I-R f -I before polymerization, which avoid the non-functionalized chain end group. Several improved telechelic fluoropolymers, including copolymer of VDF/HFP and terpolymer of VDF/TrFE/CTFE, have been obtained with a well-defined molecular structure and controlled polymer molecular weight. 2.5 Progress in Polymer/clay Nanocomposites Polymer/layered inorganic nanocomposites have attracted strong interest in today s materials research. They were first reported in the literature as early as 1961, when Blumstein demonstrated polymerization of vinyl monomers in the presence of montmorillonite (mmt) clay. 47 The basic structure of clay consists of a layered silicate structure. The common phyllosilicates, including montmorillonite (mmt), talc, and mica, contain multi-layered (2:1) silicates, having negative charge centers on the layers ranging from 0.25 to 1.5 charge centers per formula unit and a commensurate number of exchangeable cations in the interlayer spaces. The mmt crystal lattice consists of 1 nm thin layers, with a central octahedral sheet of alumina fused between two external silica tetrahedral sheets (in such a

55 40 way that the oxygen atoms from the octahedral sheet also belong to the silica tetrahedral). These layers organize themselves in a parallel fashion to form stacks with a regular van der Waals gap between them, called interlayer or gallery. In their pristine form their excess negative charge is balanced by cations (Na +, Li +, Ca 2+ ) which exist hydrated in the interlayer. The cations can be easily exchanged to proton (H + ) by acid-treatment to form an acidic clay, or ion-exchanged to other cations by treating with cationic-organic surfactants, such as alkylammoniums, to form oragnophilic clay. One common commercially available clay is dioctadecylammonium-modified montmorillonite (2C18- mmt). Clay Polymer Clay + Monomer SOLUTION BLENDING MELT BLENDING IN SITU POLYMERIZATION Silicate layer Delaminated or Exfoliated Polymer Intercalated Figure 2.1 Schematic representation of various methods (solution blending, melt blending, and in situ polymerization) used to prepare polymer-layered- silicate nanocomposites. The delaminated (or exfoliated) and intercalated morphologies are shown. 66

56 41 In general, there are three methods used in the preparation of polymer/clay nanocomposites, including (i) in situ polymerization, 48,49 (ii) solution blending, 50 and (iii) melt blending. 51 All the methods are aimed to achieve single layer dispersion of the layered silicate in the polymer matrix, because a high surface area is directly associated with the enhanced properties of polymer/clay nanocomposites. For the in situ polymerization method, the initiator or catalyst is usually pre-fixed inside the clay interlayer via cationic exchange, then the layered silicate is swollen by a monomer solution. The polymerization occurs in the presence of these exchanged silicates to form the polymer between the interlayers, with intercalated and/or exfoliated structures. In solution blending, layered silicates (modified with organic surfactants) are dispersed in an appropriate solvent before dissolving the polymer in the same solvent. When the solvent is evaporated (or the mixture precipitated), the sheets try to reassemble, kinetically trapping the polymer between them to form a nanocomposite structure. In the melt blending process, the layered silicate is mixed with the polymer matrix in the molten state. If the layer surfaces are sufficiently compatible with the chosen polymer, the polymer enters into the interlayer space and forms either an intercalated or an exfoliated nanocomposite. No solvent is required in this technique, rendering it as the most desirable industrial method. It is well-known that pristine clay, having highly hydrophilic polar surfaces, is only miscible with hydrophilic polymers, such as poly(ethylene oxide) and poly(vinyl alcohol). 52,53 To render clay miscible with hydrophobic (non-polar) polymers, one must exchange the alkali counterions with cationic-organic surfactants, such as alkylammoniums to form oragnophilic clay 54. The organic surfactant not only changes

57 42 the clay from hydrophilic to hydrophobic, e.g. by cation-exchange of the cations (Li +, Na +, Ca 2+, etc.) between clay interlayers with onium ions in organic surfactants, but it also expands the clay interlayer by increasing the (001) d-spacing between the layers. Based on theoretical modeling (without any experimental results), Balazs et al. suggested the potential benefit of an increased length of the surfactant (such as an endfunctionalized chain with two terminal groups) that might promote the dispersion of bare clay sheets within the polymer matrix 55,56 Experimentally, it is very common in polymer/clay nanocomposites to have a mixed nano-morphology, with both intercalated and exfoliated structures coexisting in the system. Intercalated structures are self-assembled, well-ordered multilayered structures where the extended polymer chains are inserted into the gallery space between parallel individual silicate layers separated by 2-3 nm. On the other hand, an exfoliated structure results when the individual silicate layers are no long close enough to interact with each other. In the exfoliated cases the interlayer distances can be on the order of the radius of gyration of the polymer; therefore, the silicate layers may be considered to be well-dispersed in the organic polymer. The silicate layers in an exfoliated structure are typically not as well-ordered as in an intercalated structure, although in many cases the exfoliated structures still bear previous parallel registry. X-ray diffraction (XRD) measurements can be used to characterize these nanostructures if diffraction peaks are observed in the low-angle region; such peaks indicate the (001) d-spacing (basal spacing) of the ordered-intercalated nanocomposites. As expected, if the nanocomposites are completely disordered, no peaks are observed in the XRD, due to loss of the parallel registry of the layers. However, XRD of

58 43 nanocomposites has limited sensitivity to detect the morphology, with only few layers of silicate structure caused by partially delamination. In all cases, the combination of transmission electron microscopy (TEM) and XRD is usually used to accurately characterize these materials; 56 since the observation of a strong intercalated XRD peak in the nanocomposite does not guarantee the absence of exfoliated layers. As discussed, it is possible to achieve impressive concurrent enhancements of material properties in the nanocomposite material. Especially when these properties depend on the surface area of the filler particles, small amounts (typically less than 5 percent) of dispersed nm-thin layered silicates typically give rise to the same level of mechanical and thermal improvements as respective conventional composites with loadings of 30 to 50 percent of micron-sized fillers. In addition, some concurrent property enhancements include decreased permeability to gas and liquids, better resistance to solvents, increased thermal stability, and improved mechanical properties, while maintaining a light weight and optical transparency Progress in Polyolefin/clay Nanocomposites In general, the thermodynamic driving force for the mixing of polymer and clay determines the outcome of whether the clay will be dispersed intercalated or exfoliatedin the polymer matrix. 57 Dispersion of the clay in the polymer requires sufficiently favorable enthalpic contributions to overcome any entropic penalties. 58 Favorable enthalpy of mixing for the polymer/organic treated clay is achieved when the polymer/clay interactions are more favorable compared to the surfactant/clay interactions. For most polar or polarizable polymers, an alkylammonium surfactant is adequate to 59, 60, 61 offer sufficient excess enthalpy and promote the nanocomposite formation.

59 44 However, in the case of apolar polyolefins, the alkylammonium treated clay has surfactants with the same aliphatic-apolar-nature as polyolefins. Consequently, such systems are at theta condition, i.e., there is no favorable excess enthalpy to promote PP/alkylammonium-treated clay dispersion. Thus, the challenge with polyolefins (especially PP) is to design systems where the polymer/clay interactions are more favorable than surfactant/clay interactions. 69 Initial attempts to create PP/clay nanocomposites were based on the introduction of a modified polypropylene with polar groups to mediate the polarity between the clay surface and bulk polypropylene. 62, 63 Maleic anhydride (MA) modified polypropylene is the most popular candidate to impart substantial polarity to the polymer in order to improve the dispersion of the clay in the polypropylene matrix. Polypropylene oligomers modified with either maleic anhydride or hydrodxyl groups were mixed with octadecylammonium-exchanged montmorillonite (o-mmt), creating a master batch, which was subsequently blended with neat PP, usually assisted by strong mechanical shear in an extruder or mixer. In this way, the PP polymer and the PP-MA oligomer pretreated o-mmt are effectively at theta conditions, and the extrusion is only promoting mixing due to the effect of the mechanical shear. As a result, the structure and the properties of the resulting hybrid materials depended strongly on the processing conditions and ranged from very moderate dispersions and property improvements to good dispersions and better performing hybrids. 69 Obviously, the pretreatment with PP- MA oligomer containing very low maleic anhydride content did not promote the nanocomposite formation, and too high of a maleic anhydride content made the master batch so robust that mmt clay may not mix further with neat PP. Therefore, it is very

60 45 crucial to control the content of MA of the modified PP in order to render mmt clay to be well-dispersed in the hybrids. Researchers from Toyota reported that when the mixture of stearylammoniumexchanged mmt clay, maleic anhydride modified polypropylene oligomer (MA content =13 mmol/g) and polypropylene homopolymer was melted, the resulting clay nanolayer was found to be exfoliated. 64 Gilman et al. 65 prepared intercalated-delaminated PP/silicate nanocomposites using PP-g-MA (mass fraction of MA 0.4%), known to be miscible with pure PP, and dimethylbis(hydrogenated tallow)ammonium mmt. The resulting nanocomposites display improved flammability (75% lower peak in the heat release rate (HRR) plot than a PP-g-MA polymer). An alternative method to overcome the dispersion between PP and clay is to use semi-fluorinated surfactants to decrease enthalpies interactions between the surfactant and the mmt, which effectively will render the PP/mmt contacts favorable. 66 Although this route is rather counter-intuitive, semi-fluorinated surfactants have stronger unfavorable interactions with the clay than the polyolefins have, and if used appropriately to modify the mmt clay, they will promote PP/mmt miscibility. Semifluorinated alkyltrichlorosilane surfactant (CF 3 -(CF 2 ) 5 -(CH 2 ) 2 -SiCl 3 ) was adopted as an effective surfactant to prepare PP/mmt nanocomposites by melt intercalation. The resulting hybrid structures contain 40% of exfoliated/disordered mmt layers in coexistence with intercalated mmt tactoids. Small additions of such semi-fluorinated surfactant modified mmt (<6wt%) enhance several properties of polypropylene materials, including improved tensile characteristics, a higher heat deflection temperature, retained optical clarity, high barrier properties, better scratch resistance, and increased flame retardancy.

61 46 Therefore, polypropylene-based nanocomposites are particularly attractive for such applications as packing materials, where superior mechanical and physical properties, barrier properties, and flammability are paramount Progress in Fluoropolymer/clay Nanocomposite As aforementioned, fluoropolymer is another type of hydrophobic material with superior performance. From a thermodynamic view of point, the driving force for mixing is not strong enough to render the clay to be well-dispersed in the fluoropolymer matrix. It is not surprising that the research on the fluoropolymer/clay nanocomposite is much less ardent than that of polyolefins. Priya 67 first reported poly(vinylidene fluoride) (PVDF) /clay nanocomposite produced by melt intercalation with organophilic clay, Cloisite 6A. It is expected that the property enhancement achieved in the resulted nancomposite was marginal due to an obviously poor dispersion of the clay in the polymer. 68 Liu 69 prepared intercalated-exfoliated PVDF/clay nanocomposites using dimethylformamide (DMF) or ethylene carbonate/propylene carbonate as a cosolvent between PVDF and a low content of organophilic clay. When applied as electrolyte films, the materials exhibited better film formation, solvent-maintaining capability, and dimensional stability than pure PVDF. These electrolyte films also showed enhanced electrochemical stability and ionic conductivity, because the dispersed inorganic layer structure can eliminate the attraction between lithium and perchlorate ions and accelerate the transfer of the lithium ions. VDF/HFP copolymer also displayed similar behavior when it was mixed with organophilic clay and used as lithium polymer electrolyte. 70 However, the cyclic recharging effect of the lithium secondary battery on the polymer films has yet to be further investigated. Ellsworth 71 prepared

62 47 hexadecyltributylphosphonium hectorite, which can be stable at high temperatures (> 250 o C) and applied it in the fabrication of nanocomposites with ethylenetetrafluoroethylene copolymer (ETFE), perfluorinated ethylene-propylene copolymer, and tetrafluroethylene-perfluoro(propylvinyl ether) copolymer. These nanocomposites exhibited increased stiffness without a significant reduction in elongation at break, reduced vapor permeability, and improved heat stability without any noticeable change in the crystallinity of the thermoplastics caused by the filler. ETFE/hectorite nanocomposites showed improved solvent resistance and thermal stability, comparable to those of polytetrafluoroethylene, while they retained a reasonable melt processability feature. These properties enable them to be used in wire insulation, tubing, packaging, molded parts, seals, gaskets, O-rings and coatings. Ebrahimian 72 used an fluoropolymer/montmorillonite nanocomposite for coating wires and conductors employed in high-speed telecommunication data transmission cables. Overall, it seems that fluoropolymer/clay nanocomposites can be a potential material for a variety of practical applications. On the other hand, it is a scientific challenge to prepare fluoropolymer/clay nanocomposites with an exfoliated and uniformly dispersed clay structure in polymer matrix, in order to make full use of the unique performance provided by the nanocomposite structure. 2.6 Outlook Great efforts have been made in both functionalization chemistry and polymer/clay nanocomposites involving polyolefin and fluoropolymers. Two research areas are closely relative, and are both scientifically challenging and industrial important subjects. So far, there are only few functional polyolefin and fluoropolymers available for studying

63 48 clay/polymer nanocompsoites. Unfortunately, most of the functional polymers also show limited capability of forming the desirable exfoliated clay layer structure in polymer matrix. In fact, the mostly used functional polymers contain multiple pendent functional groups along the polymer chain, which can form multiple contacts with the clay surface. Such an interaction may also bridge the consecutive clay platelets to promote an intercalated structure, especially for mmt with large lateral dimension. Further delamination of the layer structure will become extremely difficult. It s clear that new functional polymer structures have to be designed and synthesized before we can achieve the objective thermodynamically-stable exfoliated clay/polymer nanocomposites for polyolefin and fluoropolymer systems. As will be discussed later, my research has been focusing on chain-end functional polymers. The idea is based on a hypothesis that the terminal hydrophilic functional group anchors on the inorganic surface while the hydrophobic polymer chain is expelled from the clay surface. Consequently, an exfoliated structure of the clay platelets (from pristine clay) is spontaneously formed without organic surfactant. In Chapter III, I will discuss a new chemical route, by a combination of metallocene catalysis and reactive chain transfer agents, to prepare chain end functionalized polypropylene. For comparison, I will describe another new method in Chapter IV, which involves borane/o 2 initiator to prepare side chain functionalized PP with well-defined molecular structure, i.e. MA content and high molecular weight. Chapter V will discuss the synthesis of chain-end functional fluoropolymers via the combination of AIBN initiator and diiodoperfluoroalkane chain transfer agent. Chapter VI will present the experimental

64 49 results of clay/polymer nanocomposites using chain-end functional polymers. Chapter VII will summarize the results and propose some ideas for future works. References: 1 Vasile, C.; Seymour, R.B. Handbook of polyolefins, New York: Marcel Dekker, Chung, T.C. Prog. Polym. Sci. 2002, 27, Chung, T.C. Functionalization of polyolefins, Academic Press, San Diego, Schneider, M.J., Schafer, R.; Mulhaupt, R. Polymer 1997, 38, Kesti, M.R.; Coates, G.W.; Waymouth, R.M. J.Am Chem Soc 1992, 114, Correia, S.G.; Marques, M.M.; Ascenso, J.R.; Ribeiro, A.F.G.; Gomes, P.T.; Dias, A.R.; Blais, M.; Rausch, M.D., Chien, J.C.W. J. Polym Sci, Part A: Polym Chem 1999, 27, Gaylord, N.G.; Mehta, M.; Mehta, R. J.Appl Polym Sci 1987, 33, Chung T.C. Macromol Symp 1995, 89, Chung, T.C. Polym Mater Encycl 1996, 8, Chung, T.C. Metallocene-based reactive polyolefin copolymers containing p-methyl styrene. In: Scheirs, J.; Kamisky, W. editors. Metellocene-based polyolefins. New York: Wiley, 1999, p Chung, T.C.; Jiang, G.J. Macromolecules 1992, 25, Chung, T.C.; Rhubright, D.; Jiang, G.J. Macromolecules 1993, 26, Chung, T.C.; Rhubright, D.; Lu, H.L. J.Am Chem Soc 1996, 118, Chung, T.C.; Lu, H.L.; Ding R.D. Macromolecules 1997, 30, Chung, T.C.; Dong, J.Y. US Patent 6,096,849, Lu, H.L.; Hong, S.; Chung, T.C. Macromolecules 1998, 31, Xu, G.; Chung, T.C. J. Am. Chem. Soc. 1999, 121, Xu, G.; Chung, T.C. Macromolecules 1999, 32, Chung, T.C.; Dong, J.Y. J.Am. Chem. Soc. 2001, 123, Smart, B.E. Prperties of fluorinated compounds, physical and physicochemical properties, in: Hudlicky, M.; Pavalath, S.E. (Eds), Chemistry of Organic Fluorine Compounds, vol. II, ACS Monograph 187, American Chemistry Society, Washington, DC, 1995, p979.

65 50 21 Scheirs, J. Modern Fluoropolymers, Wiley, New York, Hougham, G.; Johns, K.; Cassidy, P.E. Fluoropolymers: Synthesis and Properties, Plenum Press, New York, Petchsuk, A. Ph.D. thesis, Penn State University, Boutevin, Macromol. Symp. 1994, 82, 1 25 Logothetis, A. Progr. Polym. Sci. 1989, 14, Ameduri, B.; Boutevin, B.; Kostov, G. Progr. Polym. Sci. 2001, 26(1). 27 Ameduri, B.; Boutevin, B. J. Fluorine Chem. 2000, 104, Ameduri, B.; Boutevin, B.; Kostov, G.; Petrova, P. J. Appl. Polym. Sci. 1999, 23, Watanabe, T.; Mormose, T.; Ishigaki, I.; Tabata, Y.; Okamoto, J. J. Polym. Sci. Polym. Lett. Ed. 1981, 18, Miyake, H.; Sugaya, Y.; Yamabe, M. J. Fluorine Chem. 1998, 92, Ukihashi, H.; Yamabe, M.; Miyabe, H. Progr. Polym. Sci. 1986, 12, Gierke, T.D.; Murn, G.E.; Wilson, F.C. J. Polym. Sci. Polym. Ed. 1980, 19, Manseri, A.; Ameduri, B.; Boucher, M. Proceedings of the Presentation at the 15 th Winter Fluorine Conference, Presentation no. 49, Saint Petersburg Beach, FL, Jan Ameduri, B., Bauduin, G.; Boutevin, B.; Kostov, G.K.; Petrova, P.; Macromolecules 1999, 32, Ameduri, B.; Boutevin, B.; Kostov, G.K.; Petrova, P. Designed Monomers Polym. 1999, 2, Rice, D.E.; Sandber, C.K. Polym. Prepr., Am. Chem. Soc., Div. Polym. Chem. 1971, 12, Rice, D.E. US Pat. 3,461,155, Saint-Loup, R.; Manseri, A.; Ameduri, B.; Lebret, B.; Vignane, P. Macromolecules 2002, 35, Tatemoto, M.; Tomoda, M.; Ueta, Y. German Patent DE29,401,35 40 Tatemota, M.; Morita, S. US Patent US4,361,678, Oka, M.; Tatemoto, M.; Contemp. Topics Polym. Sci. 1984, 4, Tatemota, M.; Morita, S. Eur. Patent Appl EP Hung, M.H. US Patent 5,231,154 (1993) 44 Arcella, V.; Brinati, G.; Apostolo, M. La Chimica e I Industria 1997, 79, Petchsuk, A. Ph.D. Dissertation, 2003.

66 51 46 Odian, G. Principle of polymerization, Wiley, New York, 1991, p Blumstein, A. Bull. Chim. Soc. 1961, Usuki, A.; Kojima, Y.; Kawasumi, M.; Okada, A.; Fukushima, Y.; Kurauchi, T.; Kamigaito, O. J. Mater. Res. 1993, 8, Lan, T.; Pinnavaia, T. J. Chem. Mater. 1994, 6, 2216; Zhao, et al. J. Polym. Sci.: Part A: Polym. Chem. 2004, 42, 916, Chen, et al. Polymeric Materials: Science and Engineering 2004, 91, 605] 50 Jeon, H.G.; Jung, H.T.; Lee, S.D.; Hudson, S. Polym. Bulletin 1998, 41, Giannelis, E. P.Adv. Mater. 1996, 8, Manias,,E.; Chen H.; Krishnamoorti, R.; Genzer, J., Kramer, E. J., Giannelis E. P. Chem. Mater. 2000, 12, Vaia, R.A. Giannelis, E.P. Adv. Mater. 1995, 7, Giannelis, E. Adv. Polym. Sci. 1998, 138, Balazs, A.C.; Singh, C.; Zhulina, E. Macromolecules 1998, 31, Gilman, J.W.; Jackson, C.L.; Morgan, A.B.; Harris, R.; Manias, E.; Giannelis, E.P.; Wuthenow, M.; Hilton, D.; Phillips, S.H. Chem. Mater. 2000, 12, Vaia, R.A.; Giannelis, E.P. Macromolecules 1997, 30, Manias, E.; Touny, A.; Wu, L.; Strawhecker, K.; Lu, B.; Chung, T.C. Chem. Mater. 2001, Vaia, R.A.; Giannelis, E.P. Macromolecules 1997, 30, Strawhecker, K.E.; Manias, E. Chem. Mater. 2003, 15, Strawhecker, K.E.; Manias, E. Macromolecules 2001, 34, Kurokawa, Y.; Yasuda, H.; Oya, A. J. Mater. Sci. Lett. 1996, 15, Nour, M.A. Polimery 2002, 47(5), Kawasumi, M.; Hasegawa, N.; Kato, M.; Usuki, A.; Okada, A. Macromolecules 1997, 30, Gilman, J.W.; Jackson, C.L.; Morgan, A.B.; Harris, R.Jr.; Manias, E.; Giannelis, E.P.; Wuthenow, M.; Hilton, D.; Phillips, S. Flame Retardants 2000, London, Inter. Science 2000, Manias, E.; Touny, A.; Wu, L.; Lu, B.; Strawhecker, K.; Gilman, J.W.; Chung, T.C. Polym. Mater. Sci. Eng. 2000, 82, Priya, L.; Jog, J. P.; J. Polym. Sci. Part B: Polym. Phys. 2002, 40, 1682

67 52 68 Priya, L.; Jog, J. P.; J. Polym. Sci. Part B: Polym. Phys. 2003, 41, Liu, H.J.; Hwang, J.J.; Chen-Yang, Y.W. J. Polym. Sci. Part A: Polym. Chem. 2002, 40, Wang, M.; Zhao, F.; Guo, Z.; Dong, S. Electrochimica Acta, 4004, 49, Ellsworth, M.W. US Patent US 5,962,553, Ebrahimian, S.; Jozokos, M.A.; Pavilionis, W. US Patent US 6,797,760, 2004

68 53 Chapter III Synthesis of Chain-end Functionalized Polypropylene Using C 1 and C 2 -symmetric Metallocene Catalysts 3.1. Introduction Since the discovery of HPDE and PP about a half century ago, the functionalization of polyolefin has been a scientifically challenging and industrially important area. 1 The ultimate objective for the functionalization of polyolefin is to improve the poor interactive properties of polyolefins and to broaden polyolefin applications to high value products, especially in polymer blends and composite (or nanocomposite) areas. As mentioned in Chapter II, an effective functionalization chemistry to prepare polyolefin (including PE, PP, EP, etc) with a terminal reactive group and obtain high catalytic activity has been developed in our laboratory. 2,3 The chemistry is based on a chain transfer reaction involving a reactive chain transfer agent, including dialkylborane (H- BR 2 ) and p-methylstyrene/hydrogen (p-ms/h 2 ), during the metallocene-mediated olefin polymerization. All polymers formed contain a terminal borane or p-ms group and have a relatively narrow molecular weight distribution (M w /M n ~2). The polymer molecular weight (from a few thousand to a hundred thousand) was basically controlled by the mole ratio of [chain transfer agent]/[olefin]. Both the terminal borane and p-ms groups are versatile, and can serve as the reactive sites for subsequent functionalization reactions or conversion to living initiators for chain extension reactions. This metallocene-mediated α-olefin polymerization/chain transfer scheme is applicable to all polyolefin homo- and copolymers. It is desirable to extend this chemical route to directly prepare polyolefin

69 54 with a desirable terminal group, such as Cl, OH, and NH 2, with good chemical reactivity in many applications. As will be discussed in this chapter, a direct (one-pot) polymerization process has been developed to prepare isotactic polypropylene (i-pp) with a terminal NH 2 group. 4 This chemistry involves metallocene-mediated propylene polymerization using a rac- Me 2 Si[2-Me-4-Ph(Ind) 2 ZrCl 2 /MAO complex in the presence of a styrene derivative carrying a silane-protected NH 2 (St-NSi 2 ), followed by hydrogenation. Apparently, the propylene propagating chain end engages in a facile consecutive chain transfer reaction, reacting with St-NSi and hydrogen concentrations. The polymer molecular weight was inversely proportional to the molar ratio of [St-NSi 2 ]/[propylene]. The silane protecting groups were hydrolyzed in acidic aqueous solution during the sample work-up step to obtain the desirable i-pp polymers with a terminal NH 2 group (PP-t-St-NH 2 ). The terminal functional group was confirmed by a chain extension reaction. Despite the high molecular weight, PP with a terminal functional group can still engage a coupling reaction with polycaprolactone (PCL) in solution, or melt to form PP-b-PCL diblock copolymers that are effective compatibilizers in PP/PCL polymer blends. As mentioned in Chapter I, only certain C 2 - and C 1 - symmetric catalysts can be used to prepare isotactic polypropylene with a high isotacticity, molecular weight, melting point, and catalytic activity. 5,6 Among these catalysts, the Exxon-Hoechst C 2 - symmetric catalyst (Me 2 Si[2-Me-4-Ph(Ind) 2 ZrCl 2 /MAO] is an ideal candidate to prepare isotactic polypropylene with the desirable structure and properties, because the 2-methyl substitute represses β-h elimination and the phenyl substitute facilitates stereocontrol in the insertion of propylene monomer into the growing polymer chain. 7 In addition, the C 2 -

70 55 symmetric catalyst is effective in incorporating styrene derivatives via 2,1-insertion, which is the key factor for promoting chain transfer reaction in the presence of hydrogen. Therefore, most of our previous research on the preparation of chain-end functional polypropylene was based on the Exxon-Hoechst C 2 -symmetric catalyst and that led to inspiring results. As aforementioned, certain C 1 -symmetric catalysts with bulky substitutes in the indenyl or fluorenyl rings also do well in the isospecific polymerization of propylene. However, the preparation of functional polypropylene using C 1 -symmetric catalysts has not been reported as of yet. Thus, we chose one of the commercial Mitsubishi C 1 -Catalysts, i.e., dimethylsilylene(2-methyl-4,5-benzoindenyl)(2-methyl-4- phenyl-4-hydroazulenyl) hafnium dichloride (Me 2 Si(2-Me-Benz[e]Ind(2-Me-4-Ph- 4HAzu)HfCl 2 ), to be used for the study of chain transfer reaction with p-methylstyrene. The catalyst is usually activated and supported on an acid-treated clay surface. As will be discussed in this chapter, chain end functionalized PP polymers have been prepared by this Mitsubishi C 1 -catalyst with ease in the presence of H 2 and a chain transfer agent, just as by the Exxon-Hoechst C 2 -symmetric catalyst. Furthermore, it was unexpected to find that chain end functional PP can also be obtained even under the condition without any hydrogen. The resulting polymer shows a somewhat higher molecular weight than that obtained in the presence of hydrogen. 3.2 Experimental Instrumentation and Materials All 1 H and 13 C NMR spectrum were recorded on a Bruker AM-300 spectrometer. Differential scanning calorimetry was measured on a Perkin Elmer DSC-7 instrument

71 56 controller, from -20 to 190 o C with a heating rate of 20 o C/min. The viscosity of the polymer solution (i-pp in decalin) inhibited with BHT was determined at 135 o C with a Cannon-Ubbelohde viscometer. The molecular weight was calculated by the Mark- Houwink equation: [η]=km a ν wherein K= and a=0.80 for i-pp. [6] Bulk morphology in the polymer films was examined by scanning electron microscopy (SEM), using a Topcon International Scientific Instruments ISI-SX-40 with secondary electron imaging. SEM samples were prepared from films cryofractured in liquid N 2. Samples were mounted on an aluminum stub and carbon coated to form a conductive coating. All oxygen and moisture sensitive manipulations were performed inside an argonfilled dry box, equipped with a dry-train. Cp*TiCl 3, Methylaluminoxane (10 weight % in toluene), lithium bis(trimethylsilyl)amide, magnesium chips, chloromethyl methyl ether, butylated hydroxytoluol (BHT), 9-borabicyclo[3.3.1]-nonane (9-BBN), allylmagnesium chloride, and polycaprolatone (PCL) were purchased from Aldrich and used without further purification. Styrene and 4-vinylbenzyl chloride (from Aldrich) were dried over CaH 2 before distillation. Toluene and THF were deoxygenated by argon spurge before refluxing with sodium anthracide for 48 hours, and then distilling from their respective green or purple solution under argon. Synthesis of 4-{2-[N,N-bis(trimethylsilyl)amino]ethyl}styrene (St-NSi 2 ) 4-{2-[N,N-bis(trimethylsilyl)amino]ethyl}styrene was prepared in two reaction steps. [5] Into a 500 ml flask equipped with a magnetic stirring bar, 100g (0.598 mol) of lithium bis(trimethylsilyl)amide was slowly dissolved in 200 ml of THF into a mixture of 50 ml (0.658 mol) of chloromethyl methyl ether and 50 ml of THF at 0 o C, under a nitrogen atmosphere. After the addition was complete, the solution was allowed to warm

72 57 to room temperature for 2 hours before evaporating the excess chloromethyl methyl ether and THF solvent. N,N-Bis(trimethylsilyl)methoxymethylamine (80% yield) was isolated by distillation. In the second step, 4-{2-[n,n-bis(trimethylsilyl)amino]ethyl} styrene was prepared by treating 4-vinylbenzylmagnesium chloride with N,Nbis(trimethylsilyl)methoxymethylamine. In a 500mL flask equipped with a magnetic stirring bar and a condenser, 15.2 g (0.62 mol) of magnesium was suspended in 50 ml of dry ether, and then 80 ml (0.57 mol) of 4-vinylbenzyl chloride diluted with 50 ml dry ether was introduced dropwise through the condenser. The solution was refluxed for 4 hours before the additon of 117 g (0.57 mol) of N,Nbis(trimethylsilyl)methoxymethylamine over a period of 2 hours. The reaction was allowed to proceed at room temperature for another 2 hours before adding 100 ml of aqueous NaOH solution (30%). The organic layer was separated and dried with magnesium sulfate, and the residual was then distilled over CaH 2 to obtain 4-{2-[N,Nbis(trimethylsilyl)amino]ethyl}styrene (St-NSi 2 ) in 70% yield. The 1 HNMR spectra of St- NSi 2 (shown in Figure 3.1): δ 7.4(d, 2H, arom H), 7.2 (d, 2H, arom H), 6.8 (q, CH 2 =CH), 5.8 and 5.2 (d, CH 2 =CH-), 3.2 (t, CH 2 N), 2.7 (t, Φ-CH 2 -), 0.15 (s, 18H, -Si(CH 3 ) 3 ).

73 58 Figure H NMR of 4-{2-[N,N-bis(trimethylsilyl)amino]ethyl}styrene Synthesis of PP-t-St-NSi 2 and PP-t-St-NH 2 Polymers PP-t-St-Si 2 polymers were prepared by using 4-{2-[N,N-bis(trimethylsilyl)amino] ethyl}styrene as a chain transfer agent in propylene polymerization. In a typical reaction, a Parr 450 ml stainless autoclave reactor equipped with a mechanical stirrer was charged with 50 ml of toluene and 4.5 ml of MAO (10 wt% in toluene) before purging with hydrogen (20 psi). Then 1.0 ml (0.125) of 4-{2-[N,N-bis(trimethylsilyl)amino] ethyl}styrene was injected into the reactor and 100 psi (3.24 M) of propylene was charged, bringing the total pressure to 120 psi at an ambient temperature (30 o C). About mol of rac-me 2 Si[2-Me-4-Ph(Ind)] 2 ZrCl 2 catalyst in toluene solution was then syringed into the rapidly stirred solution under propylene pressure to initiate the polymerization reaction. Additional propylene was fed continuously into the reactor to

74 59 maintain a constant pressure (120 psi) during the course of polymerization. After 15 min of reaction at 30 o C, the reaction solution was quenched with methanol, then filtered and washed extensively with THF to remove any 4-{2-[N,N-bis(trimethylsilyl)amino] ethyl}styrene homopolymer. The resulting polymer was dried under vacuum at 50 o C for 8 hours. PP-t-St-NH 2 polymers were prepared from PP-t-St-NSi 2 polymers by treating them with hydrogen chloride, which can be accomplished during the sample workup step. Alternatively, the isolated PP-t-St-NSi 2 (2 g) was suspended in 50 ml of toluene at 50 o C before adding dropwise 2 N hydrogen chloride ether solution. The mixture was stirred for 5 hours at 50 o C, and then poured into a 1 N methanolic NaOH solution. The polymer was collected by filtration, and washed with 1 M aqueous ammonia and water under a nitrogen atmosphere. The polymer was dried overnight at 50 o C under vacuum. The polymer yield was quantitative. Synthesis of PP-b-PCL Diblock Copolymer The coupling reactions between chain end functinalized PP and ε-polycaprolactone (PCL) were carried out in both solution and melt. In a typical solution process, 3 g PCL (M n = g/mol, M w /M n = 2.0) was first dissolved in toluene (200 ml) in a 500 ml flask equipped with a stirrer and a condenser installed on a sidearm (trap) containing a P 2 O 5 dry agent. About 2 g of PP-t-St-NH 2 (M n = g/mol, M w /M n = 2.0) inhibited with BHT (~1%) was then added into the stirring solution and contracted with P 2 O 5 to maintain anhydrous conditions. The hot polymer solution was slowly poured into cold acetone, and the precipitated polymer was isolated by filtration. The insoluble polymer was then subjected to a vigorous Soxhlet extraction by boiling acetone to remove any

75 60 untreated PCL homopolymer. The purification continued until the composition of the insoluble portion became constant. After drying, 2.85 g of PP-b-PCL diblock copolymer was obtained. Dehydration of clay mineral Acid mmt clay was used as an activator and support for the Mitsubishi C 1 - symmetric catalyst. 1.5 g of sulfuric acid/lithium sulfate treated mmt was introduced into a 100 ml round bottomed flask, and dried in vacuo (less than 2 mmhg) at room temperature, followed by heating up to 200 o C for 2 hours. After that, the clay was cooled down to room temperature and kept under nitrogen atmosphere before use. Preparation of Mitsubishi C 1 -symmetric catalyst/clay system The dehydrated mmt clay (50 mg) was mixed with toluene (1mL) under nitrogen to obtain a slurry solution. While vigorously stirring, 0.7 ml toluene solution of triethylaluminum (containing 0.03 g TEA) was slowly dropwise added into the stirring solution at room temperature. Stirring was continued for 1 hour, followed by washing with toluene (3-5 ml) three times until the concentration of TEA in the solution became less than a hundredth of the original concentration to obtain a mmt clay activator slurry. Separately, 0.03 ml triisobutylaluminum (TiBA) (1M) in toluene solution was diluted with 0.5 ml of toluene before adding to 1 µmole of Mitsubishi C 1 -symmetric catalyst in 0.7 ml of toluene. This solution was stirred for 10 minutes at room temperature, before introducing it into the slurry containing the mmt clay activator. The mixture was then stirred for 10 min before use.

76 61 Synthesis of p-ms terminated polypropylene by Mitsubishi C 1 -symmetric catalyst In a typical reaction, a Par 300 ml stainless autoclave equipped with a mechanical stirrer and a catalyst injector was charged with 50 ml of toluene and 0.2 g of TiBA/toluene solution (1M), as well as a prescribed amount of p-methylstyrene (p-ms). The reactor was then introduced with a prescribed amount of hydrogen (ranging from 0 to 20 psi) and 120 psi of propylene at ambient temperature. After the reactor was heated up to 75 o C, the previously catalyst slurry was injected into the rapidly stirred solution under propylene pressure to initiate polymerization. Additional propylene was fed continuously into the reactor to maintain a constant pressure during the course of the polymerization. To minimize mass-transfer and to maintain the constant feed ratio, the reactions were carried out by rapid mixing and a short reaction time. After 10 minutes of reaction at 75 o C, the polymer solution was quenched with methanol. The resulting p-ms terminated polypropylene (PP-t-p-MS) was washed with THF to remove excess p-ms. Then it was dried under vacuum at 50 o C overninght. 3.3 Results and Discussion Both the Exxon-Hoechst C 2 -symmetric catalyst (rac-me 2 Si[2-Me-4-Ph(Ind) 2 ]ZrCl 2 ) and Mitsubishi C 1 -symmetric catalyst (Me 2 Si(2-Me-Benz[e]Ind(2-Me-4-Ph- 4HAzu)HfCl 2 are commercial metallocene catalysts, specially designed for propylene polymerization to produce isotactic polypropylene with high isotacticity (mmmm>95%), high molecular weight (M w >200,000) and high melting temperature (T m >150 o C). Two catalyst structures are illustrated below:

77 62 5' Me 4' 1' Me Si ' Me Cl Zr 4 2' Me Cl 5 5' Me 4' 3' 2' 1' Me Si Me Cl Hf Cl Me 5 Exxon-Hoechst C 2 -symmetric Catalyst: rac-me 2 Si(2-Me-4-Ph(1-Ind) 2 ZrCl 2 Mitsubishi C 1 -symmetric Catalyst: Me 2 Si(2-Me-Benz[e]Ind(2-Me-4-Ph- 4HAzu)HfCl 2 The combination in 2- and 4- positions of the indenyls is attributed to the success of these two catalysts in the isospecific polymerization of propylene. For the Exxon-Hoechst C 2 -symmetric catalyst, the methyl groups at position 2,2 offer a high molecular weight of the polymers produced, while the phenyl rings at position 4,4 are highly beneficial to the stereocontrol in isospecific insertion of the propylene monomer and offer high isotacticity of the polymers. In the case of the Mitsubishi C 1 -symmetric catalyst, the methyl groups at position 2,2, similar to those in the Exxon-Hoechst C 2 -symmetric catalyst, account for producing a high molecular weight polymer. Due to the asymmetric nature of this catalyst, both the phenyl ring at position 4 and the fused benzene ring at position 4 and 5 concurrently contribute to the steric control of monomer insertion with only one

78 63 enantioselective-face site available for propylene coordination insertion. This stereoselectivity may be further strengthened by the non-planar seven member ring. It is well known that consecutive insertions of monomer involve a flip-flopmechanism, i.e., after monomer insertion, the polymer chain migrates to the coordination position of the inserted monomer, leaving the position open for the next monomer coordination. In the Exxon-Hoechst C 2 -symmetric catalyst system, the symmetric nature ensures the equivalence of these two coordination sites. The resulting polymer chain will give almost the same configuration, independent of the coordination site the monomer inserts from. However, this is not the case for the C 1 -symmetric catalyst system. During the chain propagation, the polymer chain is always located at the more open coordination position (inward), which forces propylene to always insert from the more crowded coordination site, with a specific steric orientation. Based on traditional wisdom, the catalytic activity of the Exxon-Hoechst C 2 - symmetric Zr catalyst should be higher than that of the Mitsubishi C 1 -symmetric Hf catalyst because zirconocene usually shows a higher reactivity than hafnocene. On the other hand, hafnocene should produce PP with higher molecular weight. In this research, we take advantage of these two commercial catalysts that can produce PP polymers with a high isospecificity and high molecular weight. The catalyst also engages a consecutive chain transfer reaction to a styrene derivative (pmethylstyrene or 4-{2-[N,N-bis(trimethylsilyl)amino]ethyl} styrene (St-NSi 2 )) in the presence (or absence) of hydrogen to synthesis chain-end functional isotactic polypropylene with desirable properties (high molecular weight and high melting point). This research stemmed from several intriguing observations 8 during the copolymerization

79 64 of propylene and p-ms using the Exxon-Hoechst C 2 -symmetric catalyst rac-me 2 Si[2- Me-4-Ph(Ind) 2 ]ZrCl 2 /MAO complex. The reaction was completely stalled in the beginning of the copolymerization process. The deactivation of the catalyst was speculated to be due to a steric jamming during the consecutive insertion of 2,1-inserted p-ms and 1,2-inserted propylene (k 21 reaction), as illustrated follows: The combination of an unfavorable 1,2-insertion of propylene (k 21 ) and a lack of p- MS homopolymerization (k 22 reaction) at the propagating site (III) drastically reduces catalytic activity. This hypothesis was supported by the effect of a small amount of ethylene dramatically improving the catalytic activity. The sluggish propagating chain end (III) (which poses a difficulty in both the k 21 and k 22 reactions) allows the insertion of ethylene, which reenergizes the propagation process. If the above hypothesis of catalyst deactivation proves correct, we might be able to take advantage of the dormant propagating site (III) to react with hydrogen, which not only recovers the catalytic site but also produces PP polymer with a terminal p-ms group. It is logical to think that this chemistry is also applicable to preparing PP containing other styrene derivatives, such as p-chlorostyrene, p-hydroxylstyrene, and p-aminoethylstyrene.

80 Synthesis of PP-t-St-NH 2 Polymer via Exxon-Hoechst C 2 -Symmetric Catalyst The basic idea in the direct (one-pot) preparation of the chain functionalized PP is to use a functionalized styrenic chain transfer agent and hydrogen that can engage in a metallocene-mediated propylene polymerization/chain transfer reaction under some reaction conditions. A - L Zr CH 3 (CH 2 -CH) x CH 2 =CH (St-NSi 2 ) L (II) (CH 2 ) 2 CH 3 k 12 2,1-insertion N R 3 Si SiR 3 CH 3 x CH 2 =CH k 11 A - L Zr -CH-CH 2 - (CH 2 -CH) x L L A - L Zr H (I) Cat: rac-me 2 Si[2-Me-4-Ph(Ind)] 2 ZrCl 2 /MAO (CH 2 ) 2 N R 3 Si SiR 3 (III) k 2 H 2 CH 3 (CH-CH 2 ) x -CH 2 -CH 2 - (CH 2 ) 2 -N(SiR 3 ) 2 (IV) HCl (in work-up step) CH 3 (CH-CH 2 ) x -CH 2 -CH 2 - (CH 2 ) 2 -NH 2 (V) Scheme 3.1 Chain transfer Reaction mechanism during metallocene catalyzed polymerization of propylene Scheme 3.1 illustrates the general reaction mechanism. During the course of propylene 1,2-insertion, the propagating M + -C site (II) eventually reacts with a St-NSi 2

81 66 unit (k 12 ) (via 2,1-insertion) to form a St-NSi 2 -capped propagating site (III), with an adjacent phenyl group interacting with metal cation. The new propagating site (III) is incapable of continuing the insertion of St-NSi 2 (k 22 ) or propylene (k 21 ) due to the steric jamming. However, it can react with hydrogen to complete the chain transfer reaction. This consecutive reaction with St-NSi 2 and hydrogen results in a PP-t-St-NSi 2 polymer chain (IV) and a regenerated Zr-H species (I) that reinitiates the polymerization of propylene and continues the polymerization cycles. After the polymerization is complete, the desirable NH 2 terminal group in PP-t-NH 2 (V) can be easily recovered during the sample workup step, using an acidic solution. 4-{2-[N,N-bis(trimethylsilyl)amino]ethyl}styrene (St-NSi 2 ) is used as a chain transfer agent to prepare functionalized PP with a terminal amino group. Because the amino group is very sensitive to the metallocene cationic site, a silane group needs to be used for the protection of the amino group during the metallocene catalysis, and subsequently deprotected by acidic solution during the sample workup procedure. The overall reaction benefits from the small quantity of the chain transfer agent needed in the preparation of high polymers. Therefore, the additional protection-deprotection step does not bring about any deteriorating effect on the polymerization conditions. The highly isospecific rac-me 2 Si{2-Me-4-Ph-(Ind)] 2 ZrCl 2 /MAO complex used in the commercial production of i-pp is relatively suitable for this reaction. This catalyst system produces i- PP with high molecular weight and high melting temperature. The bulky ligands around its specific opening active site may further prevent the catalyst from interacting with the protected functional group.

82 67 Table 3.1 summarizes the experimental results involving St-Cl/H 2, St-OSi/H 2, and St-NSi 2 /H 2, respectively in the rac-me 2 Si[2-Me-4-Ph(Ind)] 2 ZrCl 2 /MAO catalyzed polymerization of propylene. In all control reactions, a small amount of any styrene derivative (St-f) effectively stops the polymerization of propylene. The introduction of hydrogen gradually restores the catalyst activity. Overall, the PP molecular weight is governed by the chain transfer agent - the higher the concentration of the St-f, the lower the molecular weight of the resulting polymer. Run A-1 A-2 A-3 A-4 B-1 B-2 B-3 B-4 C-l C-2 C-3 C-4 D-1 D-2 D-3 D-4 E-1 E-2 E-3 E-4 F-1 F-2 F-3 Table 3.1. A summary of PP-t-Cl, PP-t-St-OH and PP-t-St-NH 2 polymers a St-f b H 2 Yield Cat. Cl, OH or NH 2 M n PDI (mol/l) (psi) (g) Activity c in PP (mol%) ( 10-3 ) (M w /M n ) ~ 0 ~ , , , ~ 0 ~ , , , ~ 0 ~ , , , ~ 0 ~ ~ 0 ~ , , ~ ~ ~ 0 1,327 8,480 31,655 ~ 0 2,622 10, Tm ( o C) a) Reaction conditions: 50 ml toluene, Catalyst: rac-me 2 Si[2-Me-4-Ph(Ind)] 2 ZrCl 2 / MAO catalyst, [Zr] = moles, [MAO]/[Zr] = 3000, propylene = 100 psi, temperature = 30 o C, time = 15 min. b) St-f: St-Cl for runs A and B; St-OSi for Runs C and D, St-NSi 2 for Runs E and F. c) Catalyst activity = kg of PP/ mol of catalyst.h.

83 68 Figure 3.2 shows the plot of the polymer molecular weight (M n ) versus the molar ratio of [propylene]/[st-f], including three St-Cl/H 2, St-OSi/H 2, and St-NSi 2 /H 2 chain transfer agents. In general, the polymer molecular weight and molar ratio of [propylene]/[st-f] are linearly proportional. It is clear that the chain transfer reaction to St-f (with rate constant k tr ) is the dominant termination process, and that it competes with the propagating reaction (with rate constant k p ). The degree of polymerization (X n ) follows a simple comparative equation X n =k p [olefin]/k tr [p-ms] with a chain transfer constant k tr /k p of 1/21 for St-Cl/H 2, 1/48 for St-OSi/H 2, and 1/34 for St-NSi 2 /H 2, respectively. It is intriguing that the k tr /k p values are significantly lower than those seen in styrene and p-ms cases [11] under similar reaction conditions. The bulky protected functional groups may reduce the frequency of chain transfer reaction. Polymers with very low molecular weight (a few thousand) have been obtained, and the molecular weight distributions are quite narrow, which are generally consistent with single site polymerization processes. Some low molecular weight polymers with exceptionally low Mw/Mn < 2 (despite extra efforts to recover all polymers) may be associated with the sensitivity of GPC at the very low molecular range.

84 c b a Mn [propylene] / [St-f] (mol / mol) Figure 3.2 The plots of number average molecular weight (Mn) of PE-t-St-f polymers versus the mole ratio of [propylene]/[st-f] using (a) St-Cl, (b) St-OSi and (c) St-NSi 2, respectively.

85 70 Figure H NMR spectra of (a) a PP-t-St-OSi polymer and its corresponding PP-t-St- OH. (M n = 22,000 g/mol, M w /M n = 2.0). (solvent: C 2 D 2 Cl 4 ; temp.: 110 o C). The terminal functional group at the polymer chain end provides direct evidence for the chain transfer reaction. Figure 3.3 shows the 1 H NMR spectra (with inset of magnified region and chemical shift assigments) of PP-t-St-OSi polymer and the corresponding PP-t-St-OH. In addition to three major peaks (δ= 0.95, 1.35, and 1.65 ppm) for the CH 3, CH 2, and CH groups in the PP backbone, there are three minor chemical

86 71 shifts at 0.25, 2.61, and ppm (with an intensity ratio near 6/2/4) shown in Figure 3.2(a), corresponding to -OSi(CH 3 ) 2 (t-bu), CH 2 -φ, and CH 2 C 6 H 4 -OSi, respectively. The chemical shift for the silane protecting group completely disappears in Figure 3.2(b), indicating the occurrence of a very effective deprotection reaction. The equally split chemical shifts for the phenyl protons, combined with no detectable side product, further indicate the terminal p-alkylphenol moiety. Figure 3.4 compares the 1 H NMR spectra of a PP-t-St-NSi 2 polymer (M n = 24.2 x 10 3 ; Mw/Mn= 2.3) and the corresponding PP-t-St-NH 2. In addition to the chemical shifts for the PP polymer, Figure 3.4(a) shows all of the chemical shifts associated with the protecting bis(trimethylsily)amino terminal group that is connected to the symmetrical p- dialkylbenzene moiety. In fact, all four phenyl protons merge into a single chemical shift at 7.22 ppm. Figure 3.4(b) shows an almost identical spectrum, except for the disappearance of the silane protecting group at 0.24 ppm.

87 72 Figure H NMR spectra of (a) a PP-t-St-NSi 2 polymer and (b) its corresponding PP-t- St- NH 2 (M n = 24,200 g/mol, M w /M n = 2.3). (solvent: C 2 D 2 Cl 4 ; temp.: 110 o C). The existence of a terminal functional group in PP was further evidenced by a chain extension reaction (discussed later) using the terminal functional group as the reaction site for coupling reaction with polyester and polyamide. Overall, the combined

88 73 experimental results strongly indicate a clean and effective reaction scheme, as illustrated in Scheme 3.1. The combination of the facile in situ chain transfer to St-f/H 2 during the catalytic polymerization of propylene, and the subsequent complete deprotection reaction during the sample work-up step, affords a very interesting reaction scheme in the preparation of the chain end functionalized i-pp with a Cl, OH, and NH 2 terminal group, via a one-pot reaction process Synthesis of PP-t-p-MS via Mitsubishi C 1 -Symmetric Catalyst/MAO system In general, the consecutive chain transfer to p-methylstyrene, during propylene polymerization using Mitsubishi C 1 -symmetric catalyst Me 2 Si(2-Me-Benz[e]Ind(2-Me-4- Ph-4HAzu)HfCl 2, works in similar ways to the one using Exxon-Hoechst C 2 -symmetric catalyst. The mechanism is shown as follows: (II) Hf + A - k 11 k 12 2,1-insertion Hf Hf + A - H 2 Hf + A - Scheme 3.2 Chain transfer reaction mechanism during propylene polymerization using Mitsubishi C 1 -symmetric catalyst

89 74 During the polymerization of propylene (with 1,2-insertion method), the propagation Hf-C site (II) can also react with p-ms (with 2,1-insertion method) to form a dormant propagating site (III) at the terminal p-ms unit. Although the catalytic Hf-C site in compound (III) becomes inactive to both propylene and p-ms, the dormant Hf-C site (III) can react with hydrogen to form p-ms terminated polypropylene (PP-t-p-MS) (V), and regenerate a Hf-H species (I) that is capable of reinitiating the polymerization of propylene and of continuing the polymerization cycles. In other words, the ideal chain transfer reaction will not significantly affect the rate of polymerization, but will reduce the molecular weight of the resulting polymer. The molecular weight of PP-t-p-MS will be linearly proportional to the molar ratio of [propylene]/[p-ms], and basically independent of the [propylene]/[hydrogen] ratio.

90 Table 3.2 PP-t-p-MS Polymers Prepared by Mitsubishi C 1 -symmetric catalyst /MAO (H 2 effect) 75

91 76 Table 3.2 summarizes the experimental results of propylene polymerization by using the Mitsubishi C 1 catalyst: Me 2 Si(2-Me-Benz[e]Ind)(2-Me-4-Ph- 4HAzu)HfCl 2 /MAO (Al/Hf= 3000) in the presence or absence of p-ms and H 2 chain transfer agents. The polymerization reactions were carried out at 55 o C for 15 minutes with 120 psi propylene pressure. In the first control set (runs , 1229, , and ), without p-ms and changing the H 2 pressure, catalytic activity increases and molecular weight decreases with the increase of H 2 pressure. On the other hand, T m is quite constant at ~ 150 o C. Overall, the results are comparable to those with other C 1 - symmetric ansa-metallocene catalysts containing similar substituted indenyl rings as ligands (For example, Me 2 Si(Ind)(2-Me-4-Ph-Ind)ZrCl 2 gives i-pp with mm=96%, 2,1 = 0.4%, T m = 155 o C and M w = 530,000, at the relatively high polymerization temperature of 70 o C, as mentioned in Chapter I). 9 In the presence of p-ms (0.017 and M) in two comparative sets (runs T0-T4 and runs , , , and , and , respectively), the H 2 pressure increases with the increase of catalyst activity, but no significant change is found in the polymer molecular weight and T m. The polymer molecular weight is predominantly controlled by p-ms concentration. The higher the p- MS concentration, the lower the polymer molecular weight. Overall, the chain transfer reaction follows the reaction scheme shown in Scheme 3.2.

92 77 a * c b Figure H NMR spectra of two PP-t-p-MS samples (run T0 and ) prepared by Mitsubishi C 1 -symmetric catalyst (top) with p-ms but no H 2 and (bottom) with both p-ms and H 2.

93 78 It is unexpected to see a significant catalyst activity in run T0, with the presence of p-ms and without H 2. After introducing a much higher p-ms concentration in run , the catalyst activity reduces significantly, but does not completely stop the polymerization. These results significantly differ from those observed in the Exxon- Hoechst C 2 -symmetric catalyst, in that a small quantity of p-ms completely stops propylene polymerization. In other words, the p-ms capped PP chain (III) is completely deactivated Zr to form a "dormant" species. Apparently, the Hf active site is still reactive after inserting p-ms to form a p-ms capped PP chain (III). It is very interesting to know the precise reaction of the Hf active site after the p-ms capped PP chain (III). Figure 3.5 (top) shows the 1 H NMR spectrum of the resulting PP polymer (run T0). In addition to three major proton chemical shifts (δ=0.95, 1.35 and 1.65 ppm), corresponding to CH 3, CH 2 and CH groups in the PP backbone, there are three minor peaks at 7.1, 2.55 and 2.35 ppm, corresponding to the protons of phenyl ring, methine, and methylene, respectively, of p-ms located at the polymer chain end. And there is no peak associated with the p-ms comonomer unit. Apparently, the Hf active site in the p-ms capped PP chain (III) can engage chain transfer reaction without the assistance of hydrogen. As illustrated in Scheme 3.3, the chain transfer reaction may involve TMA (existed in MAO), and the formed Al-capped polymer chain (IV) is subsequently hydrogenated to form a p-ms capped PP chain (V). Chain transfer to alkylaluminum is common in heterogeneous Ziegler-Natta catalysts. 10 It also occurs in some metallocene cases using a high concentration of MAO cocatalyst. 11,12

94 79 L A - Hf + -CH-CH 2 PP (CH 3 ) 2 Al-CH-CH 2 PP L Al(CH 3 ) 3 H + (III) (IV) CH 3 - CH 2 CH 2 PP CH 3 + CH 3 L A - Hf + L CH 3 (V) Scheme 3.3 Chain transfer reaction of p-ms capped PP to alkylaluminum during propylene polymerization using Mitsubishi C1-symmetric Catalyst Similar 1 H NMR spectra, with the incorporated p-ms unit located at the PP chain end, were observed in all PP samples (Table 3.2) prepared by the Mitsubishi C 1 - symmetric catalyst in the presence of both p-ms and H 2. Figure 3.5 (bottom) shows a typical example (run ). Hydrogen may simply provide an additional (more reactive) route for the Hf active site in a p-ms capped PP chain (III) to complete the chain transfer reaction and to form PP-t-p-MS (VI). In fact, catalyst activity increases with hydrogen pressure in both comparative sets. This occurs especially for the set (runs , , , and , and ) with high p-ms concentration, and with no significant change in polymer molecular weight. The Hf active site in a p-ms capped PP chain (III) is incapable of inserting propylene or p-ms and has to chain transfer to TMA or H 2 (more effectively) to complete the polymerization cycle that releases PP-t-p-MS (IV) and regenerates the Hf active site for reinitiating a new polymer chain.

95 80 Figure 3.6. The plot of polymer molecular weight of PP-t-p-MS vs. [propylene]/[p-ms] ratio. [Propylene] is calculated based on the data in reference 13. Table 3.3 PP-t-p-MS Polymers Prepared by Mitsubishi C 1 -symmetric catalyst/mao (p-ms effect) Table 3.3 compares several runs under the same H 2 pressure (20 psi, which is sufficient for chain transfer reaction in most cases) and varying p-ms concentrations. Catalyst activity is proportionally depressed with the concentration of p-ms, which

96 81 reflects the competitive coordination at metallocene active sites between monomer and chain transfer agents. Figure 3.6 shows the plot of viscosity-average molecular weight (M v ) of the resulting PP-t-p-MS polymers (Table 3.3), versus the mole ratios of [propylene]/[p-ms]. A linear proportional relationship is observed between the polymer molecular weight and [propylene]/[p-ms] ratio, which is a clear indication that the chain transfer reaction to p-ms (with rate constant k tr ) is the dominant termination process, and that it competes with the propagating reaction (with rate constant k p ). The degree of polymerization (X n ) follows a simple comparative equation, X n =k p [propylene]/k tr [p-ms] with the chain transfer constant k tr /k p ~ 1/6.7, which is quite similar with that of the Exxon-Hoechst C 2 -symmetric Zr catalyst. It is remarkable to think that the large size of the p-ms monomer has no problem in coordination/insertion at the active site of the Mitsubishi C 1 -symmetric Hf catalyst having a bulky ligand structure, but after p-ms incorporation, the Hf active site has to engage in chain transfer reaction. The situation must be associated with the steric jamming during the consecutive insertion of 2,1- inserted p-ms and 1,2-inserted propylene. The 2,1-inserted p-ms unit also can not isomerize (such as in the case of 1,3-propylene insertion), therefore, chain transfer reaction is the only route to re-activate the active site. With a good understanding of the Mitsubishi C 1 -symmetric Hf catalyst/mao system in chain transfer reaction to p-ms, we are extending the chemistry to use clayactivated Mitsubishi C 1 -symmetric Hf catalysts and p-ms, or other styrenic chain transfer agents containing a functional group, such as OH and NH 2. The objective is to find the suitable condition to prepare chain end functionalized PP polymers with good control of polymer molecular weight, high yield, and high purity.

97 Synthesis of PP-t-p-MS via Mitsubishi C 1 -Symmetric Catalyst/clay system Clay mineral serving as a cocatalyst (activator), instead of MAO, in propylene polymerization offers several advantages: (i) elimination of using expensive and unstable MAO will lower the cost of the catalyst and avoid the storage problems caused by the poor stability of MAO; (ii) the clay mineral also serves as carrier support that can improve the morphology of the resulting product and facilitate the catalyst being applied to industrial processes; and (iii) the clay minerals change the molecular weight distribution and composition distribution (when applied in copolymer) of the product, while retaining the high activity paralleled with MAO. Table 3.4 summarizes the experiment results of propylene polymerization by using the Mitsubishi C 1 -symmetric catalyst and clay mineral (triethylaluminum treated mmt) as an activator (Al/Hf= ~100), in the presence or absence of p-ms and H 2 chain transfer agents. The polymerization reactions were carried out at o C for 10 minutes with 120 psi propylene pressure. The first two series (Runs C10 to C00 and Runs C20 to C21) show a similar trend in the effect of polymerization temperature on the propylene polymerization. When the polymerization temperature increases from 55 to 75 o C, the catalytic activity dramatically increases, while maintaining high molecular weight and high melting point of the products. At 75 o C, perhaps the optimal polymerization temperature, the activity is gpp/mol Hf.hr in the absence of p-methylstyrene, which is comparable to that using MAO as cocatalyst, as aforementioned ( gpp/mol Hf.hr in Table 3.2). In the presence of p-methylstyrene, the catalyst still shows high activity ( gpp/mol Hf.hr) and yields high molecular weight polypropylene. However, the chain-end structure of the polymer shows predominantly p-methylstyrene

98 83 and a small amount of unsaturated vinylidene and vinyl groups (<10%) caused by β-h elimination, due to the high temperature. 14,15 The effect of hydrogen is compared in two reaction sets (runs C00 to C17 and runs C21 to C27), with and without p-ms. It is expected that the addition of hydrogen can control PP molecular weight very well in the runs C00 to C17. However, the activity decreases with the concentration of hydrogen. This phenomenon was also observed by Resconi 16,17 in a few instances at high hydrogen concentrations, but the explanation is not clear. In the runs C21 to C27, the PP molecular weight changes little with hydrogen concentration. This indicates that p-ms dominates the chain transfer reaction and controls the molecular weight of PP. It is interesting to note that the addition of hydrogen also eliminates β- H elimination. As shown in Figure 3.7 (b), only the peaks of terminal p- methylstyrene are found in the chain-end structure of the product. The effect of p-methylstyrene without hydrogen is shown in the last series (from Run C21 to C23). The polymerization activity and molecular weight of PP dramatically decrease with p-ms concentration due to the chain transfer reaction dominated by p-ms. The chain-end structure is completely terminal p-ms at the condition of high p-ms concentration. Some PP-t-p-MS polymers with very low molecular weight have been obtained.

99 84 Run Table 3.4 PP-t-p-MS Polymers Prepared by Mitsubishi C 1 -symmetric Hf Catalyst/TEA-treated Clay system Temp H ( 2 p-ms p-ms Yield (g) Activity a Activity b p-ms in PP p-ms Convn C) (psi) (mmol) (M) ( 10 6 ) (mol%) (%) M v (k) T m ( o C) Effect of temperture C C C C C C C Effect of hydrogen C C C C C C C C Effect of p-ms C C C Polymerization condition: Propylene =120 psi; Time: 10min. C 1 cat = 2µmol, Toluene: 50 ml, clay: 50 mg. a. g PP/ mol Hf. hr; b. g PP/ g clay. Hr.

100 Figure H NMR spectra of pure PP (run C15) (bottom) and two PP-t-p-MS samples (run C23 and C27) prepared by Mitsubishi C 1 -symmetric catalyst/clay system (top) with p-ms but no H 2 and (middle) with both p-ms and H 2. 85

101 86 In summary, PP with terminal p-methylstyrene (PP-t-p-MS) has been successfully prepared with the combination of the Mitsubishi C 1 -symmetric Hf catalyst and p-ms chain transfer agent, using both MAO and mmt clay as the activator. It is interesting to point out that this chemistry can also be extended to prepare PP with other functional styrenic groups, such as an OH or NH 2 group. The research on the synthesis of these chain-end functional PPs by one pot process is still under investigation Synthesis of PP-b-PCL and PP-b-PS diblock copolymers The existence of a terminal functional group in PP can be further examined by a chain extension reaction using the terminal functional group as the reaction site. Specifically, we investigated the coupling reactions between the terminal NH 2 in PP and the terminal COOH group in polyesters and polystyrene in solution. The in situ formed diblock copolymers can be used as the compatibilizers in PP/polyester blend. A coupling reaction between PP-t-St-NH 2 (M n = g/mol, M w /M n =2.0) and polycarpolactone (PCL: M n = g/mol, M w /M n =2.0) containing a terminal COOH group was carried out in the refluxing toluene solution. The resulting PP-b-PCL diblock copolymer (with an amide linkage) was subjected to a vigorous Soxhlet extraction by boiling acetone to remove any unreacted PCL homopolymer. The purification was continued until the composition of the insoluble portion became constant. The insoluble fraction (soluble in 1,1,2,2-tetrachloroethane at elevated temperature) is PP-b-PCL diblock copolymer.

102 Figure H NMR spectra of PP-t-NH 2 (a); PP-b-PCL (b); and PP-b-PS (c). 87

103 Figure 3.9 DSC curves of (a) PCL; (b) PP-b-PCL; and (c) PP-t-NH 2. 88

104 89 Figure 3.8 compares the 1 H NMR spectra of the starting PP-t-St-NH 2 and the resulted PP-b-PCL diblock copolymer. In addition to the chemical shift at 1.9, 1.6, and 1.1 ppm, corresponding to methine, methylene, and methyl groups in polypropylene, the new chemical shifts at 4.1 and 2.3 ppm correspond to methylene groups (CH 2 -O) and (CH 2 -C=O) in the PCL block, respectively. The quantitative analysis of the copolymer composition is calculated by the ratio of two integrated intensities between δ = 4.1 and δ = ppm and the number of protons that both chemical shifts represent. These chemical shifts indicate about 30 mol % of PCL in the PP-b-PCL diblock copolymer. Under similar coupling reaction conditions, PP-t-St-NH 2 (M n = g/mol, M w /M n =2.0) also shows an effective coupling reaction with polystyrene containing a terminal COOH group (PS-t-COOH, M n = g/mol, M w /M n =1.2) that is prepared by living anionic polymerization with CO 2 termination. Figure 3.3c shows the 1 H NMR spectra of the formed PP-b-PS diblock copolymer containing 24 mol% of PS. The two broad peaks with chemical shifts at 6.7 and 7.2 ppm are attributed to the phenyl protons of the atactic polystyrene segments. The consistency of the copolymer composition between theoretical and experimental values clearly points to the effective coupling (amidation) reaction and the existence of a NH 2 group at each PP chain end. Figure 3.9 compares the DSC curves of the PP-b-PCL diblock copolymer containing 30% of PCL and two corresponding PP and PCL homopolymers. All samples were heat-treated by heating the samples to 200 o C before cooling quiescently, and DSC curves were recorded in the second heating cycle. Two distinctive crystalline structures

105 90 formed in the diblock copolymer and exhibited melting temperatures (159 and 56 o C) that were similar to those seen in the corresponding PP and PCL homopolymers. The resulting PP-b-PCL diblock copolymer (containing 30 mol% of PCL) was used as a compatibilizer in PP/PCL polymer blends. Two polymer blends a PP/PCL (70/30 weight ratio), homopolymer blend, and a blend comprised of a 70/30/10 weight ratio of PP, PCL, and PP-b-PCL, respectively - were prepared by homogeneous mixing in a dichlorobenzene solution at 180 o C before precipitating in hexane at ambient temperature. Films were then press-molded at 180 o C. Figure 3.10 compares SEM images of the cross section of two cryofractured films. In the homopolymer blend (Figure 3.10a), the polymers are grossly phase separated, as can be seen by a minor component PCL that exhibits nonuniform, poorly dispersed domains, and voids at the fracture surface. This ball and socket topography is indicative of poor interfacial adhesion between the PP and PCL domains, and represents PCL domains that are pulled out of the PP matrix. Such pull-out indicates that no stress transfer takes place between the phases during fracture. Upon blending PP and PCL with the PP-b-PCL compatibilizer, a drastic change in the morphology occurs. The compatibilized blend shown in Figure 3.10b no longer displays the distinct PCL globules and has a rather flat, featureless surface, indicating very small domain size. The addition of the diblock copolymer leads to stabilizing the interfaces, and increasing the interfacial adhesion between the PP and PCL microdomains.

106 Figure 3.10 SEM microgrphs of (a) two homopolymer blend with i-pp/pcl=70/30 (2,000x) and (b) two homopolymer with PP-b-PCL, i-pp/pp-b-pcl/pcl=70/10/30 (2,000x) 91

107 Summary Isotactic polypropylene with a terminal functional group, including NH 2, OH and Cl, has been prepared via metallocene-mediated propylene polymerization, using the Exxon-Hoechst C 2 -symmetric catalyst (rac-me 2 Si[2-Me-4-Ph(Ind)] 2 ZrCl 2 /MAO complex) in the presence of a styrene derivative (St-f). The propylene propagating chain end engages in a facile consecutive chain transfer reaction, reacting with St-f and then hydrogen, with high catalytic activity under the proper reaction conditions. The polymer molecular weight is proportional to the molar ratio of [propylene]/[st-f]. The silane protecting group in a St-NSi 2 or St-OSi unit can be hydrolyzed in acidic solution during the sample work-up step to obtain the desirable i-pp polymers with a terminal NH 2 or OH group. The terminal functional group can serve as an active site to take part in the chain extension reaction (coupling reaction) with polycaprolactone (PCL) in solution, to form PP-b-PCL diblock copolymers that are effective compatibilizers in PP/PCL polymer blends. This chemistry is also extended to prepare p-ms terminated polypropylene in the presence of the Mitsubishi C 1 -symmetric catalyst (Me 2 Si(2-Me-Benz[e]Ind(2-Me-4-Ph- 4HAzu)HfCl 2 ) with MAO or trialkylaluminum-treated clay as cocatalyst. The propagating chains are predominately transferred to p-ms, which is revealed by the linear proportional relationship between the polymer molecular weight and the concentration ratio of [propylene]/[p-ms]. It is unexpected to find significant polymerization activity even in the absence of hydrogen, indicating trialkylaluminum may participate in a chain transfer to p-ms terminated propagating chains. In the case of polymerization using MAO as cocatalyst at 55 o C, the addition of hydrogen increases the activity and regulates

108 93 polymer molecular weight. The chain-end structure is solely terminal p-ms. When TEAtreated clay is adopted as an activator and carrier at the optimal polymerization temperature of 75 o C, the high concentration of hydrogen suppresses catalytic activity. The chain ends consist of predominately terminal p-ms and small amount of unsaturated end groups. Higher p-ms concentration or introduction of hydrogen eliminates the undesirable unsaturated chain ends. References: 1 Chung, T.C. Prog. Polym. Sci. 2002, 27, 39 2 Chung, T.C.; Dong, J.Y. J. Am. Chem. Soc. 2001, 123, Dong, J.Y.; Chung, T.C. Macromolecules 2002, 35, Dong, J.Y.; Wang, Z.; Hong, H.; Chung, T.C. Macromolecules 2002, 35, Busico, V.; Cipullo, R. Prog. Polym. Sci. 2001, 26, Resconi, L.; Cavallo, L.; Fait, A. Piemontesi, F. Chem. Rev. 2000, 100, Spaleck, W.; Kuber, F.; Winter, A.; Rohrmann, J.; Bachmann, B.; Antberg, M.; Dolle, V.; Paulus, E. F. Organometallics 1994, 13, Lu, H.L.; Chung, T.C. J. Polym. Sci. Part A: Polym. Chem. 1999, 37, Spaleck, W.; Kuber, F.; Bachmann, B.; Fritze, C.; Winter, A. J. Mol. Catal. A. Chem. 1998, 128, Chien, J.C.W.; Kuo, C.I. J. Polym. Sci. A.: Polym. Chem. 1986, 24, Chien, J.C.W.; Wang, B.P. J. Polym. Sci. A: Polym. Chem. 1988, 26, Chien, J.C.W.; Wang, B.P. J. Polym. Sci. A: Polym. Chem. 1990, 28,15 13 Dariva, C.; Lovisi, H.; Mariac, L.C.; Coutinho, F.M.B.; Oliveira, J.V.; Pinto, J.C. Canadian J. Chem. Eng. 2003, 81, Carvill, A.; Zetta, L.; Zannoni, G.; Sachhi, M.C. Macromolecules 1998, 31, Carvill, A.; Tritto, I.; Locatelli, P.; Sacchi, M.C. Macromolecules 1997, 30, 7056

109 94 16 Resconi, L.; Piemontesi, F.; Jones, R.L. In Metallocene-catalyzed polymers. Properties, processing & markets; Benedikt, G.M., Goodall, B.L. Eds.; Plastics Design Library: New York, 1998; p43 17 Resconi, L. In Metallocene-Based Polyolefins. Preparation, Properties and Technology, Kaminsky, W., Scheirs, J. Eds.; Wiley: 1999; vol. 1, p467

110 95 Chapter IV Preparation of MA Modified PP and Application in PP/nylon Blend 4.1 Introduction Maleic anhydride (MA) modified polyolefins have been the most studied and commercially important functional polyolefins in the industry due to low cost of maleic anhydride and high activity of the anhydride group. [1] MA modified polyolefins effectively promote adhesion and dyeability of polyolefins, and their compatibility with other polymers and fillers. They have been found in many applications, such as glass fiber reinforced polyolefins, [2] anticorrosive coating for metal pipes and containers, [3] metal-plastic laminates for structured use, [4] multilayer sheets of paper for chemical and food packaging, [5] and polymer blends such as polyolefins/polyamide and polyolefins/polyester [6-9] as well as polymer/clay nanocomposites. [10-15] MA modified polyolefins normally are prepared by free radical post-reactor modification processes. [16-20] It is generally accepted that the modification reaction involves an initiation step of abstracting hydrogen atoms from polyolefin chain by free radicals that are in situ formed by the decomposition of peroxide initiator at elevated temperature. In polyethylene case, secondary macroradicals are formed, and then react with MA to produce MA grafted polyethylene. However, some of the secondary macroradicals also involve crosslinking by coupling. In the case of polypropylene, it is suggested that hydrogen abstraction mainly takes place at the tertiary carbons along the PP backbone because the stability of macroradicals is in the order of tertiary > secondary > primary. As illustrated in Scheme 4.1, the formed tertiary radical promotes rapid β-

111 96 scission that degrades the PP chain into two parts, a chain-end unsaturated PP and a radical-terminated PP macroradical. The terminal macroradical with good mobility then reacts with MAH molecule to form MA-terminated PP. Therefore, the PP molecular weight is inversely proportional to the extent of MA modification. In ethylene/propylene copolymer cases, both crosslinking and degradation take place, and the extent of each side reaction is dependent on the comonomer ratio. In addition, there are some other free radical side reactions, including oligomerization of MA molecules, chain transfer and disproportionation reactions, to result in a complex color mixture. The darkness of the product is also proportional to the extend of MA modification The modification reaction is usually carried out in solution, melt or slurry conditions. In a solution process, the polymer is dissolved in a suitable solvent at elevated temperatures, and MA is added together with a suitable initiator. The MA grafting content and final structure of the modified polymer depend on the reaction parameters, including the solvent, initiator and temperature. Priola [21] showed that in aromatic solvents (xylene) containing benzylic hydrogens most of the reacted MA was linked to

112 97 the solvent molecules. In solvents without benzylic hydrogens (t-butylbenzen) or o- dichlorobenzene), all the reacted MA was linked to the polymer. The efficiency of the - 2,2 > (BPO) free radical initiators follows dicumyl peroxide (DCP) > benzoyl peroxide azobiisobutyronitrile (AIBN). [22] The extent of side reactions depends on the reaction conditions. In the melt process, the modification reaction is usually carried out in a continuous mixing equipment like extruders. [23] This process has an economic advantage by combining both modification and processing of polymers in a single step. Melt modification also provides an opportunity to combine modification and reactive blending together. For example, polyolefin/polyamide and polyolefin/polyester compatibilized blends can be formed in situ. [24-26] Actually, most commercial MA modified polyolefin products have been manufactured by the melt process. However, the MA content in polymer and the structure of the products are affected by processing condition, such as the amount of the initiator and MA, temperature, and the nature of polyolefin. High temperature and shear force in melt modification can lead to intensification of side reactions (degradation, crosslinking, or oxidation). Great efforts have been made in optimizing the reaction conditions and extruder parameters to promote desired reactions while suppressing undesired ones. [27-39] Gaylord [36,37] reported that in the presence of electron-donating nitrogen, phosphorous or sulfur-containing compounds, the crosslinking of PE or degradation of PP in MA melt modification may be somewhat suppressed. MA modification was also tried in slurry condition at a temperature well below the melting point of the polymer. [38,39] Polypropylene, MA, BPO, small amount of solvent

113 98 (toluene or decalin) (no more than 20 wt% in slurry), and a special catalyst (promoting the formation and stabilization of the free radicals) were mixed at lower than 120 o C during the modification. [38] But a low graft content of MAH was observed in the product. In short, due to the inherent complexity of the free radical modification, it is difficult to obtain MA modified polyolefins with both well-defined structure and high MA graft content simultaneously. Usually, the commercial product is a complicated mixture with yellowish to brownish color, indicating a substantial amount of impurities. The MA content in the MA modified polyolefins can be analyzed by titration or FTIR. [40] Titration can be carried out in hot water-xylene solution mixture. However, the inhomogeneity of the polymer in the system may bring about great experimental errors because the polymer cannot be completely soluble in the water-xylene mixture solvent. On the other hand, FTIR has become the most common method to determine the MA amount in the polymer due to its quickness and convenience. FTIR quantification needs a calibration curve that can be measured from standard samples or similar model compounds. Unfortunately, 1 H-NMR is usually not sensitive enough to detect MA content in the graft polymer because of the very weak resonance intensity. This may be attributed to dipolar broadening of resonance near graft points that have restricted mobility of side chains. However, if 13 C-enriched MA was used in the graft reaction, it is very useful to determine the structure of the MA modified polymer with 13 C-NMR. It is usually suggested that the structure of grafted MA is predominately a single succinic anhydride unit because of the low reactivity of MA homopolymerizaiton. [41] Heiden showed that MA attached to HDPE and LDPE in the form of single succinic anhydride rings or short oligomers. In (co)polymers with abundant tertiary hydrogen atoms, such as

114 99 alt-epm and i-pp, MA grafted onto the polymer backbone mainly in the form of single succinic anhydride rings. In (co)polymers with high propylene content, a chain scission reaction can occur, yielding an anhydride ring attached to the chain end. [20,42] Scheme 4.2 Chain end modification mechanism In our group, we have developed several methods (including chain-end, side chain and backbone approaches) to prepare MA modified polyolefins with controllable molecular structure. As is illustrated in Scheme 4.2, [43] the chain-end unsaturated PP (I) can be prepared by metallocene polymerization or thermal degradation of high molecular weight PP. Following the hydroboration reaction with 9-BBN (9-borabicycoborane) to obtain the borane terminated PP (II), the borane group at the chain end was selectively oxidized [44,45] by oxygen to form a polymeric radical (III) that is associated with a stable borinate radical. [46] With the presence of MA, the polymeric radical in situ reacts with maleic anhydride to produce maleic anhydride terminated PP (PP-t-MA) (IV) with a single MA unit. In the presence of styrene, the polymeric radical initiates stable copolymerization [47,48] of styrene and maleic anhydride with an alternative manner. [49]

115 100 The resuling PP-b-SMA diblock copolymer (V) contains both PP and alternating styrenemaleic anhydride (SMA) segments. The content of MA in the product can be adjusted by the length of SMA block. However, the starting material, chain-end unsaturated PP, is not commercially available and shows low molecular weight and relatively poor properties. Scheme 4.3 Side chain modification mechanism The chemistry of the side chain method is based on the reactive polyolefin copolymers containing p-methylstyrene (p-ms) units that provide the selective reaction sites for free radical modification reaction. Scheme 4.3 illustrates the major differences between this approach and the current commercial routes (discussed before). Both poly(ethylene-co-p-methylstyrene) (PE-p-MS) and poly(propylene-co-p-methylstyrene) (PP-p-MS) copolymers with a broad range of copolymer compositions and wellcontrolled molecular structures can be prepared via metallocene polymerization. [50-52] Ideally, only the highly reactive benzylic protons (Φ-CH 3 ) in a p-ms unit involve hydrogen abstraction by alkoxyl radical. The formed relatively stable benzylic radical then reacts with facile maleic anhydride monomer. The overall process avoids any side reaction in the polymer backbone. Therefore, backbone degradation and crosslinking can

116 101 be largely prevented. Unfortunately, PE-p-MS and PP-p-MS are not yet commercially available and show a reduced melting point due to comonomer units. In this chapter, we will discuss a new MA functionalization approach by directly modifying commercially-available polyolefin using borane-based free radical initiators. The chemistry is advantaged by the combination of simplicity of post-reactor process and stable macroradicals by borane initiator to minimize side reactions. Some resulting MAH-modified PP with desirable MA content, high molecular weight, and colorless appearance, will be discussed in this chapter. 4.2 Experimental Materials. Isotactic and atactic polypropylene were received from Dow Chemicals. Tributyl borane, diethylmethoxyborane and benzoyl peroxide were purchased from Aldrich and used as received. Preparation of Maleic Anhydride Modified Polypropylene. Commercial pelletform isotactic polypropylene was purified by dissolving it in xylenes at 140 o C. The resulting homogeneous solution was then discharged into acetone to remove antioxidants. The precipitated flake-fashioned PP was washed with acetone twice before drying in a vacuum oven at 50 o C for 24 hr. In a 100mL flask equipped with a stirrer, 10 g of the purified PP was suspended in 80 ml of benzene at 25 o C under nitrogen. Then 2.5 g of maleic anhydride and 1.1 g of tributylborane were added. After stirring the mixture for 10 min, 390 ml of oxygen were introduced into the reactor over a period of 3 hr. The reaction was stirred for another 6 hr at 25 o C before precipitating the reaction mixture into 200 ml of acetone. The maleic

117 102 anhydride modified polymer was isolated by filtration, washed with acetone four times, and dried under vacuum at 50 o C for 24 hr. Polymer Blending. 0.3g of PP (M n =100,000 g/mol) (or PP-g-MA) and 0.7 g of Nylon 11 (M n = 24,800 g/mol) were dissolved in 50 ml of m-cresol/1,1,2,2 tetrachloroethane mixture (V/V= 1/1) at 120 o C. The hot polymer solution was slowly poured into cold acetone, and the precipitated polymer blend was isolated by filtration. The mixture was then melt pressed into a film which was cryofractured in liquid N 2 to obtain an undistorted cross-section representative of the bulk material. Characterization. 11 B NMR spectra were obtained on a Bruker AM-300 spectrometer using benzene-d 6 as solvent. The MA units incorporated into the polymer were determined by FTIR (Bio-Rad FTIR-60 spectrometer) using a polymer thin film (about 100 micron), which was prepared by compression-molding polymer powders between Teflon coated aluminum sheets at 190 o C and 2500 psi. The MA content was calculated from FTIR by the following equation: MA wt% =K(A 1780 /d), where A 1780 is the absorbance of carbonyl group at 1780 cm -1, d is the thickness (mm) of the film, and K is a constant (=0.25) detected by calibration of the known MA content of MA grafted PP. Although the correlation between the absorbance and MA content or film thickness may not be perfectly linear, especially for the samples with high MA contents, the general trends of this free radical MA grafting reaction are valid. The intrinsic viscosity of the polymer was measured in a dilute decalin solution at 135 o C with a Cannon-Ubbelohde viscometer. The viscosity molecular weight was calculated by the Mark-Houwink equation: [η]=km a, where K= dl/g and a=0.80 for PP in decalin. The melting point of the polymer was measured under nitrogen by differential scanning calorimetry

118 103 (Perkin-Elmer DSC-7) with a rate of 20 o C/min. Bulk morphology in the polymer films was examined by scanning electron microscopy (SEM), using a Topcon International Scientific Instruments ISI-SX-40 with secondary electron imaging. SEM samples were prepared from films cryofractured in liquid N 2. Samples were mounted on an aluminum stub and gold coated to form a conductive coating. 4.3 Results and Discussion Oxidation Reaction of Trialkylborane The chemistry for preparing PP-g-MA involves an in situ controlled oxidation reaction of trialkylborane (BR 3 ) in the presence of polyolefin and maleic anhydride. The oxidation mechanism of the trialkylborane by oxygen is relatively complicated due to the presence of three identical and equally reactive B-C bonds. In addition to the oxidation of multiple B-C bonds in each molecule, the intermolecular reaction between an oxidized B- O-O-C bond and an unoxidized B-C bond can also take place to form an inactive alkoxyl (B-O-C) group. Scheme 3.4 shows the general oxidation reaction of trialkylborane. After the first oxygen insertion into a tributylborane molecule, the formed R-O-O-BR 2 (I) can be oxidized further by oxygen to form (R-O-O-) 2 BR (II'), or it can readily react with an unreacted BR 3 to form two R-O-BR 2 (II) molecules. Both adducts are inactive in MAH graft reaction. The R-O-BR 2 compound can further be oxidized by oxygen to form an alkoxylperoxide (R-O-)(R-O-O-)BR (III), which, in turn, can react further with a B-R bond to form (R-O-) 2 BR (IV). After this stage of the oxidation process, the concentration

119 104 of unreacted BR 3 is significantly reduced, such that the intermolecular reaction becomes sluggish. R B R R O 2 R O O B R R O 2 R O O B R O O R (I) (II') R B R R R B R R R O B (II) R R O O O R 2 R O B R (III) R B R R or R O B R R R O B (IV) O R R Scheme 4.4. Oxidation Process of Trialkylborane and Initiation of Graft Reaction In situ 11 B NMR measurement is the most convenient method to detect the changing borane species during the oxidation reaction. As shown in Figure 3.1, with tributylborane(tbb)/o 2 = 1:1, the chemical shift at 86 ppm (BR 3 ) almost disappeared and three new peaks resulted, including a major peak at 58 ppm corresponding to monooxidized species (R-O-O-BR 2 ) (I) and two minor peaks at 35 and 33 ppm for two types of double oxidized O-BR-O species (II ) and (III). With increasing the oxygen concentration to TBB/O 2 = ½, the double oxidized O-BR-O species became the major product with a minor B(-O) 3 species in the product.

120 105 c b a Figure B NMR spectra of Tributylborane(TBB)/O 2 in the presence of a-pp. (a) TBB;(b) TBB/O 2 =1/1; (c) TBB/O 2 = ½. Among these oxidized products, the mono-oxidized adduct R-O-O-BR 2 (I) is the most reactive, and is largely responsible for the hydrogen-abstraction of polyolefin. Both alkoxied compounds, i.e., (R-O-)BR 2 (II) and (R-O-) 2 BR (IV), are incapable of initiation graft reactions. Although (R-O-)(R-O-O-C-)BR (III) may be capable of some other reactions, it is also too stable to react with the inert polyolefin chain at ambient temperature. In the presence of polyolefin and maleic anhydride, R-O-O-BR 2 (I) can cleave homolytically at the peroxyl bond to form (R-O* *O-BR 2 ) (V), and alkoxyl radical (R-O*) can activate the saturated polyolefin chain by means of hydrogen-abstraction of a secondary proton in a PE chain and a tertiary proton in a PP chain, respectively, as shown in Scheme 4.5.

121 106 Scheme 4.5 The formed polymeric radicals immediately associate with the oxidized borane moiety to form the protected species (C* *O-BR 2 ) (VI and VII), where C* denotes the polymeric carbon radical derived from the initial PE or PP. The protected species, (C* *O-BR 2 ) (VI and VII), are relatively stable, compared to the regular unprotected polymeric carbon radicals (C*), and are ready for reaction with maleic anhydride by an addition reaction without side reactions. On the contrary, the regular unprotected polymeric carbon radicals (C*) are unstable and immediately engage in an undesirable free radical coupling crosslinking reaction, in cases where PE is the polymer being modified, and in an undesirable polymer chain scission reaction in cases where PP is the polymer being modified. It is interesting to point out that the stability of the borinate radical (*O-BR 2 ) is due to the back-donating of electron density to the empty p-orbital of boron. [54] The borinate radical is comparable to nitroxide radicals, such as 2,2,6,6-tetramethyl-1-piperidinyloxy (TEMPO) radical [55-58], which have opposite stabilization mechanisms by electrondonating of the lone-pair electrons in the p-orbital of nitrogen to the free radical and have been used for free radical living polymerization.

122 107 To optimize the grafting efficiency, it is essential to enhance the key reaction step. In other words, it is very important to control the oxidation reaction condition so that the mono-oxidation species (R-O-O-BR 2 ) being formed immediately reacts with the polymer chain. Therefore, there is less possibility for R-O-O-BR 2 to engage in a side reaction, such as intermolecular reaction between oxidized and unoxidized borane compounds (discussed before). A favorable reaction condition maintains a high mole ratio of polymer repeating units/trialkylborane, and a low mole ratio of oxidizing agent/trialkylborane during the entire reaction process. So the amount of R-O-O-BR 2 species is kept in low concentration during the grafting reaction.

123 108 a b 1780 c 1780 Figure 4.2 FTIR of pure PP (a), PP-g-MA entry 3 with 0.4 wt% MA (b) and entry 5 with 1.0 wt% MA (c).

124 109 Table 4.1. Summary of Maleation reaction of Polyolefin Entry Starting Polymer Initiator Solvent Temp Product Polymer M ν ( 10 3 ) ( o C) MA (wt%) M ν ( 10 3 ) 1 a-pp 17 BBu 3 /O 2 benzene a-pp 17 BEt 2 (OMe) /O 2 benzene i-pp 270 BBu 3 /O 2 benzene i-pp 270 BBu 3 /O 2 benzene i-pp 270 BBu 3 /O 2 biphenyl i-pp 270 BBu 3 /O 2 biphenyl i-pp 270 BPO biphenyl Aldrich PP-g-MA 8 PP a 217 BBu 3 /O 2 benzene P(E-co-p-MS) b 217 BBu 3 /O 2 biphenyl s-ps 110 BBu 3 /O 2 biphenyl Reaction condition: Borane/O 2 mole ratio=1/1. a Copolymer contains 0.3 mole% of PE. b Copolymer contains 1 mol% of p-ms.

125 110 a b c Figure 4.3 DSC curves of pure PP (a), PP-g-MA entry 4 (b) and commercial PP-g-MA Aldrich (c). Figure 4.4 FTIR of pure s-ps (a), s-ps-g-ma (entry 10) with 1.8 wt% MA (b)

126 Graft Reaction of Maleic Anhydride onto Polypropylene Table 4.1 summarizes the experimental results of MA modified polyolefins by different initiators. We used an atactic polypropylene (Mv= 17 x 10 3 g/mol) in a control study to examine maleation reaction by borane initiators. Comparing entries 1 and 2, it is apparent that the mono-oxidation adducts of diethylmethoxyborane by O 2, resulting in namely [CH 3 -O-B(O-O-C 2 H 5 )(C 2 H 5 )] species, fail to react with the PP chain. Therefore, the reactive species in the MA graft reaction is only the mono-oxidized trialkylborane species, R-O-O-BR 2, which reacts with the PP chain by proton-extraction to the form a polymeric radical as illustrated in Scheme 4.5. From entry 3 to 6, it can be seen that trialkylborane/o 2 is an effective initiator for the maleation reaction of polypropylene. The modified PP may contain up to 1.0 wt% of MA, which was determined by FTIR (shown in Figure 4.2). In the case of MA grafting reaction of PP, the molecular weight of the product (entries 3 and 4) changes little when the reaction takes place in benzene at ambient temperature or 50 o C. DSC results show little change in melting temperature (T m = 162 o C), as shown in Figure 4.3. In contrast, a commercial PP-g-MA polymer (Aldrich) shows a lower melting point (157 o C) and relatively broad melting peak, which indicates the product is very inhomogeneous and contains large quantities of undesirable low molecular weight fractions. However, when the reaction is carried out in biphenyl at high temperature (125 o C), the molecular weight of the product decreases to a certain extent. Overall, the extent of degradation in the polymer is remarkably alleviated in the system of using tributylborane as an initiator compared with that of BPO (entry 7). This indicates that the in situ formed *OBR 2 radical does stabilize the tertiary carbon radical in the PP main chain and noticeably prevents the latter from participating in the degradation

127 112 reaction of PP. The mechanism for the meleation reaction of polypropylene at a high temperature still needs to be investigated in detail in the future. It is interesting to point out that this chemistry of using trialkylborane/o 2 as an initiator for the grafting reaction can be extended to other polyolefins, such as ethylpropylene copolymer, poly(ethyl-co-p-methylstyrene) and syndiotactic polystyrene (Figure 4.4). Entries 8-10 in Table 4.1 show that the MA modified polymers have a similar molecular weight as their corresponding starting materials, strongly indicating that there is no detectable crosslinking or other noticeable side reaction in all polymers during the grafting reactions. This provides another evidence that the in situ formed *O- BR 2 radical does stabilize the tertiary carbon free radical in the PP and s-ps and the secondary carbon free radical in the PE backbone Application of PP-g-MA in PP/Nylon Blend The PP-g-MA (entry 6 in Table 4.1, M v = 136k) with relatively well-defined structure was used as a compatibilitizer in the reactive PP/polyamide blend containing PP (M n =100,000 g/mol) and Nylon 11 (M n = 24,800 g/mol). The reactive mixing was carried out in m-cresol/1,1,2,2 tetrachloroethane solution at 120 o C. Ideally, the PP-g-MA would react with Nylon 11 in situ right at the interfaces between PP and Nylon 11 domains and produce a PP-g-Nylon 11 graft copolymer with brush-like microstructure, as illustrated in Scheme 4.6. This would serve as the interfacial agent to promote the formation of uniform micro-phase separated morphology.

128 113 CH 3 CH 3 (CH 2 CH) n -CH 2 C CH 3 CH 3 (CH 2 CH) n -CH 2 C CH 3 CH 3 (CH 2 CH) n -CH 2 C O O O O O O O O O H 2 N (OC(CH 2 ) 10 NH) n CH 3 CH 3 (CH 2 CH) n -CH 2 C CH 3 CH 3 (CH 2 CH) n -CH 2 C CH 3 CH 3 (CH 2 CH) n -CH 2 C O N O O N O O N O Nylon 11 Nylon 11 Nylon 11 Scheme 4.6 (a) (b) Figure 4.5 SEM graphs of polymer of blend PP/Nylon. (a) PP/Nylon=30/70wt%; (b) PP/PP-g-MA/Nylon=20/10/70wt%. The morphology of the polymer blend was examined by SEM micrograph, with the surface topography of cold fracture film edges representative of bulk morphology. Figure

129 shows SEM photographs of the polymer blend of PP and nylon with and without PPg-MAH in the ratio of 30/70 wt% of PP to nylon. In the simple homopolymer blend (Figure 4.5(a)), the minor PP components are grossly phase separated into poorly dispersed domains and voids at the fracture surface. This ball-and-socket topography is indicative of poor interfacial adhesion between PP and Nylon domains and represents minor component domain PP that is pulled out from the Nylon 11 matrix. Such pull-out indicates that no stress transfer takes place between phases during fracture. On the other hand, the introduction of 10 wt% of PP-g-MA (M n = 136 k, MA content: 0.8wt%) as a compatibilizer into the PP/Nylon 11 blend in situ forms a PP-g-Nylon 11 graft copolymer at the interface, which serves as a compatibilizer between PP and Nylon. A completely different morphology of the fracture cross-section of the polymer blend is observed in Figure 4.5(b). There are no distinct globules in the fracture surface. Rather, a relatively flat mesa-like fracture surface is shown, and the domain size of the disperse phase PP in this system is much smaller than that in the homopolymer PP/nylon blend. Overall, the ternary blend PP/PP-g-MA/Nylon 11 exhibits strong interface adhesive, and the cold fracture is mainly due to cohesive failure. The co-crystallization between PP-g-Nylon copolymer and PP homopolymer must take place, which overwhelms the tendency of PPg-Nylon to form a discrete domain by itself. On the other hand, the surge of hydrogen bonding provides the interactions between PP-g-Nylon copolymer and Nylon 11 homopolymer in the continuous matrix. Therefore, the high molecular PP-g-MA can form strong interactions with both PP and Nylon phases which secure both PP and Nylon interfaces.

130 Summary In this chapter, I have discussed a new chemical route to prepare MA-modified polyolefins with a controlled polymer molecular weight and desirable MA content. The chemistry is centered on the use of stable borane/o 2 initiators, which involves two steps. (a) An in situ controlled oxidation reaction of trialkyborane (BR 3 ) forms a mono-oxidized adduct (R-O-O-BR 2 ) that can undergo hemolytic cleavage to form (R-O* *O-BR 2 ) and activate a polypropylene chain by alkoxyl radical (R-O*) hydrogen-abstraction of tertiary carbon at ambient temperature. (b) The subsequent graft reaction of maleic anhydride onto the activated tertiary carbon radical of PP, which is associated with the oxidized borane moiety (*O-BR 2 ), prevents undesirable side reactions. From the study of the oxidation of trialkylborne and the control reaction of a-pp, it is found that the major species involving in the modification reaction is the mono-oxidized trialkylborane species, R-O-O-BR 2, which reacts with the main chain of polyolefin via proton abstraction to form a polymeric radical that is simultaneously stabilized by an in situ formed *O- BR 2 stable radical. Experimental results from MA modification of i-pp and other polyolefins using borane/o 2 as an initiator show little change in molecular weight during the grafting reaction, which indicates no detectable degradation for i-pp or crosslinking for P(E-co-p-MS). In addition, the MA modified PP exhibits a melting temperature as high as that of the starting material. In contrast, the grafting reaction of i- PP using traditional peroxide as the initiator leads to the formation of much lower molecular weight product with a lower melting temperature and broad melting range, compared with the starting materials.

131 116 In the polymer blending experiment, PP-g-MA (prepared from the grafting reaction using borane/o 2 initiator) shows effective compatability in the PP/Nylon 11 blend due to in situ formation of a PP-g-Nylon copolymer located at the interfaces between PP and Nylon. The introduction of only 10% of PP-g-MA in the system significantly improves the interfacial interaction between the PP and Nylon phases. As shown in the topography of a fracture cross-section examined by SEM, a relatively flat mesa-like fracture surface and smaller domain size of the dispersed phase occurred. In contrast, the simple PP and Nylon homopolymer blend shows a ball-and-socket topography that is indicative of poor interfacial adhesion between PP and Nylon domains. References: [1] Trivedi, B.C.; Culbertson, B.M. Maleic Anhydride, Plenum Press, New York, [2] Garagnai, E.; Marzola R.; Moro, A Mater. Plast. Elastomeri 1982, 5, 298 [3] Johnson, A.F.; Simms, G.D. Composites 1986, 17, 321 [4] Fukushima, N. et.al. Eur. Patent [5] Ashley, R.J. Adhesion 1988, 12, 239 [6] Jean-Marc, D. US Patent [7] Felix, J.M.; Gatenholm, P. J. Appl. Polym. Sci. 1991, 42, 609 [8] Myers, G.E. et. Al. J. Polym. Mater. 1991, 15, 21 [9] Majumdar, B.; Keskkula, H.; Paul, D.R. Polymer 1994, 35, 1386 [10] Manias, E.; Touny, A.; Wu, L.; Strawhecker, K.; Lu, B.; Chung, T.C. Chem. Mater. 2001, 13, 3516 [11] Gilman, J.W.; Jackson, C.L.; Morgan, A.B., Harris, R. Jr.; Manias, E.; Giannelis, E.P.; Wuthenow, M.; Hilton, D.; Phillips, S.H. Chem. Mater. 2000, 12, 1866 [12] Maiti, P.; Nam, P.H.; Okamoto, M.; Hasegawa, N.; Usuki, A. Macromolecules, 2002, 35, 2042 [13] Kato, M.; Usuki, A.; Okada, A.; J. Appl. Polym. Sci. 1997, 66, 1781 [14] Alexandre, M.; Dubois, P.; Mater. Sci. & Eng. 2000, 28, 1

132 117 [15] Ray, S.S.; Okamato, M.; Prog. Polym. Sci. 2003, 28, 1539 [16] Roover, R.D. et. al. J. Polym. Sci. Part A: Polym. Chem. 1995, 33, 829 [17] Cimmino, S. et.al. Polym. Eng. Sci. 1984, 24, 48 [18] Sammy, G.; Nagy, T.; White, J.L. J. Appl. Polym. Sci. 1995, 56, 1423 [19] Gloor, P.E.; Tang, Y.; Kostanska, A.E.; Hamielec, A.E. Polymer 1994, 35, 1012 [20] Heiden, W.; Rosenmoller, C.H.; Wenzel, C.B.; de Groot, H.J.M.; Lugtenburg, J. Macromolecules 1996, 29, 1151 [21] Priola, A.; Bongiovanni, R.; Gozzelino, G. Eur. Polym. J. 1994, 30, 1047 [22] Minoura, Y.; Ueda, M.; Mizunuma, S.; Oba, M. J. Appl. Polym. Sci. 1969, 13, 1625 [23] Lambla, M. Comprehensive Polymer Science, First Supplement (G.Allen Ed.), Pergamon Press, New York, 1982, p619 [24] Vanderbilt, J. U.S. Patent [25] Chen, P.N. U.S. Patent [26] Famili, A. U.S. Patent [27] Simmons, A.; Baker, W.E. Polym. Eng.Sci. 1989, 29, 1117 [28] Coutinho, F.M.B.; Ferfeira, M.I. Eur. Polym. J. 1994, 30, 9aa [29] Ostenbrink, A.J.; Gaymans, R.J. Polymer 1992, 33, 3086 [30] Ho, R.M.; Su, A.C. Polymer, 1993,34,3264 [31] Martinez, J.M.G.; Taranco, J.; Laguna, O.; Collar, E.P. Int. Polym. Process, 1994, 9, 346 [32] Chen, P.N. et.al. U.S. Patent [33] Campbell, J.R. U.S. Patent [34] Auda, R.S. U.S. Patent [35] Wong, C.S. U.S. Patent [36] Gaylord, N.G.; Mehta, M. J. Polym. Sci. Polym. Lett. Ed. 1982, 20, 481 [37] Gaylord, N.G.; Mishra, M.K. J. Polym. Sci. Polym. Lett. Ed. 1983, [38] Lee, S.; Rengarajan, R.; Parameswaran, V.R. J. Appl. Polym. Sci. 1990, [39] Borsig, E.; Heckova, L. J. Macromol. Sci. Pure Appl. Chem. 1994, A31, 1447

133 118 [40] Slavons, M.; Carlier, V.; Roover, B.; Franquinet, P.; Devaux, J.; Legras, R. J Appl. Polym. Sci. 1996, 62, 1205 [41] Russel, K.E.; Kelusky, E.C. J. Polym. Sci. Part A: Polym. Chem. 1988, 26, 2273 [42] Heinen, W.; Erkens, S. W.; Duin, M.V.; Lugtenburg, J. J. Polym. Sci. Part A: Polym. Chem. 1999, 37, 4368 [43] Lu, B.; Chung, T.C. Macromolecules 1999, 32, [44] Chung, T.C.; Janvikul, W.; Bernard, R.; Jiang, G.J. Macromolecules 1993, 27,26 [45] Chung, T.C.; Janvikul, W.; Bernard, R.; Hu, R.; Li, C.L.; Liu, S.L.; Jiang, G.J. Polymer 1995, 26, 3565 [46] Chung, T.C.; Lu, H.L.; Janvikul, W. J. Am. Chem. Soc. 1996, 118, 705 [47] Chung, T.C.; Jiang, G.J. Macromolecules 1992, 25, 4816 [48] Chung, T.C.; Jiang, G.J.; Rhubright, D. U.S. Patent [49] Bruah, S.D.; Laskar, N.C. J. Appl. Polym. Sci. 1996, 60, 649 [50] Chung, T.C.; Lu, H.L. J. Polym. Sci. Part A: Polym. Chem. 1997, 35, 575 [51] Chung, T.C.; Lu, H.L. J. Polym. Sci. Part A: Polym. Chem. 1998, 36, 1017 [53] Lu, H.L.; Hong, S., Chung, T.C. J. Polym. Sci. Part A: Polym. Chem. 1999, 37, 2795 [54] Chung, T.C.; Janvikul, W.; Lu, H.L. J. Am. Chem. Soc. 1996, 118, 705 [55] Georges, M.K.l Vergin, R.P.N.; Kazmaier, P.M.; Hamer, G.K. Macromolecules 1993, 26, [56] Holmberg, S.; Holmlund, P.; Nicolas, R.; Wilen, C.E.; Kallio, T.; Sundholm, G.; Sundholm, F. Macromolecules 2004, 37(26), 9909 [57] Schulte, T.; Knoop, C.A.; Studer, A. J. Polym. Sci. Part A: Polym. Chem. 2004, 42(13), 3342 [58] Ohno, K.; Tsujii, Y.; Miyamoto, T.; Fukuda, T.; Goto, M.; Kobayashi, K.; Akaike, T. Macromolecules 1998, 31(4), 1064.

134 119 Chapter V Synthesis of Chain-end Functionalized Fluoropolymers 5.1. Introduction Fluoropolymers are another family of hydrophobic polymers with superior physical properties. They exhibit high thermal stability, chemical inertness (to acid, solvent, and petroleum), low water absorption, excellent weatherability, and low surface energy. This unique combination of properties allows them to be used in many high-end applications, such as aerospace, aeronautics, engineering, optics, textile finishing, military use and microelectronics. 1 However, fluoropolymers also exhibit some drawbacks, such as a high processing cost, poor solubility in common organic solvents, and poor compatibility with other materials. The lack of a functional group in the fluoropolymer hampers their applications in polymer blends and composites where the miscibility with other materials is paramount. The chemistry for preparing a functional group terminated fluoropolymer is very limited. In polymer chemistry, the chain end functionalized polymer is usually prepared by a combination of living polymerization and selective termination of the living polymers with suitable reagents. There is no example in prior art which shows living free radical polymerization of fluoromonomers and functionalization at the polymer chain end. The few examples of a controlling chain end structure include the use of functional initiator, which was pioneered by Rice and Sanderg at the 3M Company. 2 They reported the preparation of low molecular weight vinylidene fluoride/hexafluoropropylene (VDF/HFP) elastomers containing two ester terminal groups by using a diester peroxide initiator. The average functionality of the resulting telechelic VDF/HFP elastomer was

135 120 not reported. However, it is logical to expect some difficulties in achieving a perfect telechelic structure with the functionality of 2 in each polymer chain, which requires all the propagating radicals to be involved in radical coupling reaction in the termination step. Recently, Saint-Loup et al 3 also attempted to prepare chain end functionalized VDF/HFP elastomers containing two opposing hydroxyl terminal groups by using hydrogen peroxide as an initiator. Several advantages of using the hydrogen peroxide initiator include low cost, high reactivity, and the direct formation of hydroxyl terminal groups. However, many side reactions also concurrently occur during the polymerization, and the final product contains not only hydroxyl terminal groups but also substantial amount of carboxylic acid terminal groups and some unsaturated terminal groups. The most useful method for the functionalization of fluoropolymer is iodine transfer polymerization (ITP), developed by Daikin Corp 4,5 in the later 1970 s and early 1980 s, to prepare fluoropolymers having two terminal iodine groups. The chemistry is based on the combination of a reversible addition-fragmentation chain transfer (RAFT) process and an α,ω-diiodoperfluoroalkane (I-R f -I) chain transfer agent, whereas R f are CF 2 CF 2, CF 2 CF 2 CF 2 CF 2, etc. The living characteristics are usually demonstrated by an increase in molecular weight with the conversion of the monomer and a relatively narrow molecular weight distribution (M w /M n < 2). The active CF 2 -I groups are always located at both ends of the polymer chain and maintain similar reactivity despite the growing polymer molecular weight. This reaction process has led to an important commercial product, i.e., diiodo-terminated VDF/HFP elastomers with the trade name Dai-El. This is a liquid rubber at room temperature and is readily curable via heating or radiation to form a 3-D network that has excellent heat, oil, solvent, chemical and ozone resistance, and a high

136 121 mechanical strength and low compression set. It is useful as a sealing material for O- rings, gaskets, tubes, valves and bellows, and linings for protective gloves and shoes. In the past few years, a new class of borane-based radical initiators has been developed in our group, which can mediate living radical polymerization at ambient temperature. The original research was focused on the functionalization of polyolefins by first incorporating borane groups into a polymer chain, which were then selectively oxidized by oxygen to form the mono-oxidized borane moieties that initiated a control radical graft-from polymerization at ambient temperature to form polyolefin graft and block copolymers. 6,7 Several relatively stable radical initiators were discovered, 8,9 which were based on the mono-oxidation adducts of trialkylborane (R-O-O-BR 2 ). In the presence of acrylates and methacrylates, this R-O-O-BR 2 moiety exhibits living radical polymerization characteristics, with a linear relationship between polymer molecular weight and monomer conversion and produces block copolymers by sequential monomer addition. 10, 11, 12 This stable radical initiator system was recently extended to the polymerization of fluorinated monomers, which can effectively occur in bulk and solution conditions. 13 Some interesting ferroelectric fluoroterpolymers, 14, 15 have been prepared with a high molecular weight and controlled polymer structure with relatively narrow molecular weight and composition distributions. In this chapter, I will discuss an improved iodine transfer polymerization (ITP) based on the combination of a specific radical initiator (AIBN) and a reversible additionfragmentation chain transfer (RAFT) process involving two iodo-compounds, i.e., α,ωdiiodoperfluoroalkane (I-R f -I) and mono-iodoerfluoroalkane (R f -I). We take advantage of the inactive radicals created by the decomposition of AIBN, which readily react with the

137 122 iodo-compounds (chain transfer agents). As a result, we can obtain pure telechelic fluoropolymer with almost all the polymer chains containing one or two terminal iodo groups. In turn, the terminal iodo-groups can undergo facile functional group transformation to form imidazolium ions that are very effective in forming a fluoropolymer/clay nanocomposite. 5.2 Experimental Materials: Vinylidene Fluoride (VDF) was obtained from the PCR, Inc, Pittsburgh, Pennsylvania. Iodoperfluorobutane, chlorotrifluoroethylene (CTFE), and conventional initiators, such as benzoyl peroxide (BPO), azobis(isobutylonitrile) (AIBN), and hydrogen peroxide, were purchased from Aldrich Chemical Company. All other fluoromonomer gases, obtained either from the PCR (Lancaster) Company or the Synquest Laboratory Company, were purified by the freeze-thaw process before being used. Diiodoperfluorobutane was kindly supplied by Daikin American, Inc. Polymerization initiated by AIBN/diiodide system In a typical example, 0.30g of AIBN (1.8 mmol) was dissolved in 30 ml of acetonitrile in a 75 ml stainless steel reactor. 0.4 g of a diiodoperfluorobutane chain transfer agent (0.84 mmol) was then added to the solution. After the reactor was vacuumed to remove air, 32 ml (~20g) of VDF was vacuum distilled into the reactor at liquid nitrogen temperature. The reactor was then rapidly heated up to 80 o C and maintained at this temperature for 4 hr. After the polymerization, the unreacted monomer was vacuum recovered, and the polymer solution was precipitated in a methanol/water

138 123 mixture and purified twice by solution-dissolution in a acetone and water/methanol mixture, respectively. Ethylenation of diiodo-terminal PVDF 0.5 g of diiodo-terminated PVDF polymer powder and 0.02 g of AIBN (0.12 mmol) were dissolved in 30 ml of DMF in a 75 ml stainless steel reactor. After the reactor was vacuumed to remove air, 8 ml of ethylene was vacuum distilled into the reactor at liquid nitrogen temperature. The reactor was then rapidly heated up to 80 o C and maintained at this temperature for 4 hr. After the reaction, the unreacted ethylene was vacuum recovered, and the polymer solution was precipitated in a methanol/water mixture and purified twice by solution-dissolution in acetone and a water/methanol mixture, respectively. Transformation to Imidazolium ion 0.5 g of diiodo-terminated PVDF polymer powder and 1 g of imidazole were dissolved in 20 ml of DMF in a 100 ml flask under a nitrogen atmosphere. The solution was heated up to 70 o C and maintained for 12 hr. After the reaction, the polymer solution was precipitated in a methanol/water mixture and purified twice by solution-dissolution in acetone and a water/methanol mixture, respectively. Characterization of fluoropolymer The compositions of the copolymers and terpolymers and their chain end structure were examined by a combination of 1 H NMR and 19 F NMR on a Bruker AM-300 specctrometer. Typically about 128 scans were recorded using 10 wt% of polymer solution in either d 7 -dimethylsulfoxide or d 6 -acetone. The chemical shifts were referenced to the external standards, tetramethylsilane for 1 H NMR, and

139 124 fluorotrichloromethane for 19 F NMR. The molecular weight and molecular weight distribution of polymers were measured by GPC using a Waters 710B delivery system with 410 refractive index detector, and compared with the monodisperse polystyrene standards. 5.3 Results and Discussion Synthesis of Iodo-terminated Fluoropolymer a. Chain transfer to diiodoperfluoroalkane The reaction scheme 5.1 shows the proposed reaction mechanism during the polymerization of VDF by using AIBN as the initiator and diiodoperfluoroalkane as the chain transfer agent. Firstly, AIBN is thermally decomposed at 70 o C to create a relatively stable carbon radical (I), which is a poor initiator for fluoromonomers. However, in the presence of the diiodo-chain transfer agent (II), the carbon free radical (I) abstracts an iodine atom from (II) and creates a new carbon free radical (III), which then can initiate the polymerization of fluoromonmers. The propagation process is continued by a repeated addition of monomers until the monomers are depleted. Several termination reactions can take place during the polymerization. The major one shall be chain transfer reaction of the propagating macro-radical (V) with diiodo-chain transfer agent (II) to from the diiodoterminated products (VI) and (VII). There are also some possible coupling reactions between two macro-radicals (V) or between macro-radical (V) and iodo-compounds (III), which leads to product (VIII) and (IX), respectively. All the polymer chains formed have both ends capped with the iodine groups. The only potential undesirable reaction is the

140 125 coupling reaction between the macro-radical (V) and the radicals (I) from thermal decomposition of AIBN, which leads to polymers with only one iodine terminal group. However, this reaction shall rarely occur due to the tiny amount of this radical (I) (usually the radical concentration ~ 10-9 M) 16 at any given time during the reaction. In addition, this reaction can be avoided by increasing the ratio of the concentration of the diiodocompound to AIBN. Based on the reaction Scheme 5.1, there are four possible terminal groups in the resulting polymer, including -CH 2 CF 2 I, CF 2 CH 2 I, CF 2 CF 2 I and C(CH 3 ) 2 CN. The end groups CH 2 CF 2 I, and CF 2 CH 2 I are caused by the combination of a chain transfer reaction to diiodoperfluoroalkane and two possible regio-addition modes of the VDF monomer during the polymerization. The other end groups CF 2 CF 2 I and C(CH 3 ) 2 CN are due to the fragment of the chain transfer agent diiodoperfluoroalkane and the initiator AIBN, respectively.

141 126 Initiation CH 3 CH 3 CH 3 C N=NCCH 3 CN CN CH 3 CH 3 C * + N 2 CN (I) CH 3 (I) + ICF 2 CF 2 CF 2 CF 2 I ICF 2 CF 2 CF 2 CF 2 * + ICCH 3 (II) (III) CN (IV) CH 3 (I) + CF 2 =CH 2 CH 3 CCH 2 CF 2 * (I') (Low reactivity) CN Propagation (III) + n CF 2 =CH 2 ICF 2 CF 2 CF 2 CF 2 (CH 2 CF 2 ) n CH 2 CF 2 * (V) Chain transfer (V) + (II) ICF 2 CF 2 CF 2 CF 2 (CH 2 CF 2 ) n CH 2 CF 2 I + (VI) (V') + (II) ICF 2 CF 2 CF 2 CF 2 (CH 2 CF 2 ) n CF 2 CH 2 I + + ICF 2 CF 2 CF 2 CF 2 (CH 2 CF 2 ) n CF 2 CH 2 * (V') (III) (III) (VII) Coupling reaction (V) + (V) or (V) + (V') or (V') + (V') ICF 2 CF 2 CF 2 CF 2 (CH 2 CF 2 ) 2n+2 CF 2 CF 2 I (VIII) (V) or (V') + (III) ICF 2 CF 2 CF 2 CF 2 (CH 2 CF 2 ) n+1 CF 2 CF 2 I (IX) CH 3 (V) or (V') + (I) ICF 2 CF 2 CF 2 CF 2 (CH 2 CF 2 ) n+1 CCH 3 CN (minor reaction) Scheme 5.1 VDF polymerization mechanism in the presence of AIBN and diiodoperfluoroalkane

142 127 Run 484 CF 2 CH 2 I CH 2 CF 2 I C(CN)(CH 3 ) 2

143 128 Run 484 CH 2 CF 2 I CF 2 CF 2 CF 2 I CF 2 CH 2 I

144 129 Run 495 CH 2 CF 2 I C(CN)(CH 3 ) 2 CF 2 CH 2 I

145 130 Run 495 CH 2 CF 2 I CF 2 CF 2 CF 2 I CF 2 CH 2 I

146 131 Run 513 CF 2 CH 2 I CH 2 CF 2 I C(CN)(CH 3 ) 2

147 132 Run 513 CH 2 CF 2 I CF 2 CF 2 CF 2 I CF 2 CH 2 I Figure H NMR and 19 F NMR spectra of telechelic PVDF prepared by the combination of AIBN and diiodoperfluorobutane.

148 133 -C(CH 3 ) 2 CN Figure H NMR and 19 F NMR spectra of PVDF prepared from AIBN.

149 134 Figure 5.1 and 5.2 show 1 H NMR and 19 F NMR spectra of two PVDF polymers prepared by the combination of the AIBN initiator and the diiodoperfluorobutane chain transfer agent (with 1/1 mole ratio) and AIBN only, respectively. The chain end structure assignments are based on the reported data. 17 Comparing two 1 H NMR spectra, two new multiple peaks are observed in Figure 5.1 (top) at a chemical shift of 3.6 and 3.8 ppm, corresponding to chain end CH 2 CF 2 I and CF 2 CH 2 I, respectively. The peak intensity for CH 2 CF 2 I is significantly larger than that of CF 2 CH 2 I, which indicates the dominating propagating site of CH 2 CF 2 * and addition reaction through CH 2 moiety of VDF monomer. A extremely small peak at 1.6 ppm is assigned to the fragment (- C(CH 3 ) 2 CN) of AIBN. On the other hand, comparing the two 19 F NMR spectra and peaks assignments in Table 5.1, the new peaks at 41 and 108 ppm in Figure 5.1 (bottom) correspond to the end groups CH 2 CF 2 I, and CF 2 CH 2 I, respectively, while the peak at 67 ppm is assigned to CF 2 CF 2 CF 2 I. From three peak intensity comparison, it is clear that chain transfer reaction to diiodo-chain transfer agent (II) is the dominate termination mechanism, and the major propagating radical is CH 2 CF 2 * species. Both 1 H NMR and 19 F NMR results are consistent and support the reaction mechanism in Scheme 5.1. Table 5.1 Assignment of 19 F NMR peaks Chemical Shift (ppm) Structure 41 CH 2 CF 2 I 67 CF 2 CF 2 CF 2 I CH 2 CF 2 CH 2 CF 2 (head to tail) 108 CH 2 CF 2 CF 2 CH 2 I CH 2 CF 2 CF 2 CH 2 I and -CF 2 CF 2 CF 2 CF CH 2 CF 2 CF 2 CH 2 (head to head) CF 2 CH 2 CF 2 CF 2 CF 2 CF 2

150 135 Table 5.2 summarizes the detailed experimental results of PVDF polymers using the AIBN initiator and the diiodoperfluorobutane chain transfer agent. In the presence of AIBN only (run 512), the polymerization of VDF gives a very low yield of 2% due to the poor reactivity of the carbon radical generated from AIBN toward fluoromonomers. The introduction of diiodoperfluorobutane as a chain transfer agent dramatically increases the yield of PVDF polymer. Comparing runs 484, 492, 495, and 493, the conversion of the monomer increases with the reaction time, and the molecular weight of the resulting polymer is almost linearly proportional to the monomer conversion. Figure 5.3 shows the linear plot of polymer molecular weight vs. monomer conversion, which indicates the living characteristic of this reversible addition-fragmentation chain transfer (RAFT) process by using diiodoperfluoroalkane chain transfer agent. It should be noted that the extrapolation of the molecular weight vs. the monomer conversion plot does not exactly go through the original. This slight deviation may be caused by the imperfect chain end containing fragments of AIBN, which cannot continuously grow with time during the polymerization. Based on the end group analysis from the combination of 1 H and 19 F NMR spectra, the content of imperfect AIBN chain end is around 5% in the reaction condition with a 1/1 mole ratio of AIBN/diidoperfluoroalkane, and slightly decreases with the reaction time. In detail, the amount of CF 2 CF 2 I and CH 2 CF 2 I chain end also decreases with time, while the amount of CF 2 CH 2 I end group increases. This indicates the high reactivity of the end groups with iodine adjacent to CF 2, which involves further reactions (chain transfer and propagation) and eventually transforms all CF 2 I groups to the relatively stable CH 2 I terminal group. Therefore, it can be deduced that the stability of the end groups decreases in the sequence: CF 2 CH 2 I > CH 2 CF 2 I> CF 2 CF 2 I.

151 136 Comparing runs 491, 513, 487, and 484, using the same amount of diiodoperfluoroalkane and the same reaction time and temperature, but altering the AIBN/ diiodoperfluoroalkane mole ratio from 1/4 to 1/1, the polymerization results are strongly dependent on the amount of AIBN used. The conversion of the monomer increases with the increase of AIBN concentration, and the molecular weight of the resulting the PVDF polymer is linearly proportional to the reciprocal of AIBN concentration, as shown in Figure 5.4. The combination clearly indicates the proportional relationship of propagating radical to the AIBN concentration, and living polymerization of VDF in the presence of diiodoperfluoroalkane. On the other hand, when the AIBN/diiodoperfluoroalkane mole ratio increases from 1/4 to 1/1, the imperfect chain end C(CH 3 ) 2 CN is also systematically increased from 1.5 to 5.4 %. The concentration of *C(CH 3 ) 2 CN radicals at a given time is too much for the diiodoperfluoroalkane molecules to completely capture all the radicals. With providing certain diiodoperfluoroalkane concentration, it is possible that almost all the *C(CH 3 ) 2 CN radicals are chain transferred to diiodoperfluoroalkane chain transfer agents before reacting with VDF monomers.

152 Molecular weight (X 10 3 ) Conversion (%) Figure 5.3 Plot of M v of PVDF vs. Conversion of VDF monomer M v of PVDF /[AIBN] (1/mol) Figure 5.4 Plot of M v of PVDF vs. 1/[AIBN].

153 138 Table 5.2 Polymerization of VDF in the presence of diidoperfluorobutane (I-R f -I) Run AIBN (mmol) I-R f -I (mmol) AIBN/ I(CF 2 ) 4 I (mol/mol) Reaction Time (hr) Yield (g) Conversion (%) Effect of reaction time M v (k) End group (%) (CH 3 ) 2 C CF 2 CH 2 I CH 2 CF 2 I CF 2 CF 2 I / / / / Effect of AIBN concentration / / / / Polymerization condition: temperature=80 o C, solvent: acetonitrile, 30 ml. VDF: 30 ml (~20g).

154 139 Overall, this approach using AIBN as an initiator is superior to that of the commercial method, in which a low concentration of peroxide initiator and a large quantity of diiodoperfluoroalkane (20 times over peroxide in mole ratio) must be adopted to prevent severe polymer chain end imperfections. In addition, a very long reaction time is required during the polymerization process due to the very low initiator concentration. In the present improved method, the extremely low reactivity of the carbon radical from AIBN toward fluoromonomers allows us to increase initiator concentration without causing too much of the undesirable chain end from AIBN, and the polymerization time is also significantly shortened. Hydrogen abstraction CF 2 CH H 2 C F 2 C H *CF2 CH 2 CF 2 C H* H 2 C CF 2 CH 2 HF2C CF 2 CH H 2 C CF 2 CH 2 HF 2 C (X) (XII) CH 2 CF 2 (XIII) CF 2 CH H 2 C F 2 C H *CH2 CF 2 CF 2 C H* H 2 C CF 2 CF 2 H3C CF 2 CH H 2 C CF 2 CF 2 H3C (XI) (XIV) CH 2 CF 2 (XV) Scheme 5.2 Reaction mechanism for producing short chain branching PVDF In many commercial PVDF polymers, the polymer chain normally contains a small amount of short chain branches that are very undesirable, because the tertiary C-H branch sites are unstable at elevated temperature. The formation of these branches is due to the hydrogen abstraction mechanism involving an intramolecular C 1, 5 hydrogen shift

155 140 analogous to the short chain branching mechanism happened in low density polyethylene. In the case of PVDF, this reaction cannot occur in a normal head to tail chain sequence, because the growing CF 2 * radical is always in a 1,4 or 1,6 relationship to CH 2 - groups. Therefore, the C 1, 5 shift comes from an irregular head to head sequence (as illustrated in Scheme 5.2), which is present at a concentration of about 3-5% in most VDF polymerization. 18 It s interesting to note that all PVDF-t-I polymers prepared by the combination of AIBN and diiodoperfluoroalkane show almost no short chain branch structure, below the sensitivity of the NMR measurement. The reversible iodo-capping of the propagating chain end provides control radical polymerization mechanism to prevent an intramolecular C 1, 5 -hydrogen shift.

156 141 b. Chain transfer to mono-iodoperfluoroalkane It is interesting to extend this RAFT process to the preparation of fluoropolymers with only one terminal iodine group. The logical thinking is to use monoiodoperfluoroalkane as the chain transfer agent, instead of diiodoperfluoroalkane. Based on the reaction mechanism (Scheme 5.3), the propagation of pure PVDF with a single iodo-terminal group may be much more difficult than anticipated. Initiation (CH 3 ) 2 CN=NC(CH 3 ) 2 2 (CH 3 ) 2 C* + N 2 CN CN CN (I) (CH 3 ) 2 C* CN + CH 2 =CF 2 (CH 3 ) 2 C-CH 2 CF 2 * CN (II) (sluggish) (CH 3 ) 2 C* CN + CF 3 CF 2 CF 2 CF 2 I (CH 3 ) 2 C-I + CF 3 CF 2 CF 2 CF 2 * (IV) CN (III) Propagation CF 3 CF 2 CF 2 CF 2 * + CH 2 =CF 2 CF 3 CF 2 CF 2 CF 2 ~PVDF~CH 2 CF 2 * (III) or CF 3 CF 2 CF 2 CF 2 ~PVDF~CF 2 CH 2 * (V) (VI) Coupling reaction a. Macroradical coupling 2 (V) or (V) + (VI) CF 3 CF 2 CF 2 CF 2 ~PVDF~CF 2 CF 2 CF 2 CF 3 (VII) (not desirable) or 2 (VI) b. Coupling with CF 3 CF 2 CF 2 CF 2 * (III) (V) + (III) CF 3 CF 2 CF 2 CF 2 ~PVDF~CF 2 CF 2 CF 2 CF 3 (VII) (not desirable) Chain transfer reaction to CF 3 CF 2 CF 2 CF 2 I (IV) (V) + (IV) CF 3 CF 2 CF 2 CF 2 ~PVDF~CH 2 CF 2 I or (VI) + (IV) CF 3 CF 2 CF 2 CF 2 ~PVDF~CF 2 CH 2 I (VIII) (IX) + + (III) (III) (desirable) Scheme 5.3 VDF Polymerization mechanism using AIBN and iodoperfluoroalkane

157 142 After thermal decomposition of AIBN to generate a *C(CH 3 ) 2 CN radical (I), which is inert to the VDF monomer, the carbon radical has to be transferred by iodoperfluoroalkane to form an active species CF 3 CF 2 CF 2 CF 2 * (III) that initiates the polymerization. The propagation reaction with monomers continues until the monomers are depleted. The propagating polymer chains can be terminated by several possible coupling and chain transfer reactions. The coupling reactions between the propagation chains themselves and between the propagating radical and CF 3 CF 2 CF 2 CF 2 * (III) will give undesirable chain ends without any iodine terminal group. However, most of the propagating chains can be chain transferred to CF 3 CF 2 CF 2 CF 2 I (IV) to form the desirable iodine-terminated polymers (VIII) and (IX) and an active site (III) for continuing the polymerization. Overall, there are four possible end groups in the resulting polymer, including - CH 2 CF 2 I, CF 2 CH 2 I, CF 2 CF 3 and C(CH 3 ) 2 CN. The end groups CH 2 CF 2 I, and CF 2 CH 2 I are caused by the combination of a chain transfer reaction to mono-iodoperfluoroalkane and the regioirregular addition of the VDF monomer during the polymerization. The other end groups CF 2 CF 3 and C(CH 3 ) 2 CN are due to the fragment of the chain transfer agent mono-iodoperfluoroalkane and the initiator AIBN, respectively. Figure 5.5 shows 1 H NMR and 19 F NMR spectra of a typical PVDF polymer prepared by the combination of AIBN and CF 3 CF 2 CF 2 CF 2 I. In the 1 H NMR spectrum, the chemical shifts at 1.6, 3.6 and 3.8 ppm correspond to the chain end -C(CH 3 ) 2 CN (from AIBN fragment), CH 2 CF 2 I, and CF 2 CH 2 I, respectively. In 19 F NMR spectrum, the peak at 41 ppm corresponds to the end group CH 2 CF 2 I, and the peak at 108 is

158 143 assigned to CF 2 CH 2 I, while the peaks at 81 and 126 ppm are attributed to different fluorine in the end group of CF 2 CF 2 CF 3. a * CH 2 CF 2 I C(CN)(CH 3 ) 2 CF 2 CH 2 I

159 144 b CH 2 CF 2 I CF 2 CH 2 I CF 2 CF 2 CF 3 CF 2 CF 2 CF 3 Figure H NMR (a) and 19 F NMR (b) spectra of PVDF-t-I (571)

160 145 Table 5.3 Assignment of 19 F NMR peaks. 19 Chemical Shift (ppm) Structure 41 CH 2 CF 2 I 81 CF 2 CF 2 CF CH 2 CF 2 CH 2 CF 2 (head to tail) 108 CH 2 CF 2 CF 2 CH 2 I CF 2 CH 2 CF 2 CF 2 CF 2 CF 3 and CH 2 CF 2 CF 2 CH 2 I CH 2 CF 2 CF 2 CH 2 (head to head) 123 CF 2 CH 2 CF 2 CF 2 CF 2 CF CF 2 CF 2 CF 2 CF 3 Table 5.4 summarizes the experiment results of VDF polymerization using AIBN and a mono-iodoperfluorobutane chain transfer agent. Compared with the corresponding case of diiodoperfluorobutane, the use of the AIBN/mono-iodoperfluorobutane with mole ratio of 1:1 significantly increases the undesirable C(CH 3 ) 2 CN terminal group. The reduction of chain transfer I-CF 2 groups allows more *C(CH 3 ) 2 CN radicals to initiate the polymerization. In most cases the resulting PVDF polymer chains contain more than 5% of -C(CH 3 ) 2 CN terminal groups and only ~80% of polymer chain containing a iodine terminal group. In contrast, the polymers, prepared from diiodoperfluorobutane, contain >95% of two iodine terminal groups and <5% of -C(CH 3 ) 2 CN terminal groups. In order to increase the amount of terminal iodine group, it is necessary to decrease the AIBN/iodoperfluorobutane mole ratio. When this ratio decreases to ½ (run 572), about 95% of the resulting polymers contain an iodine end group.

161 Mv of PVDF-t-I (X 10 3 ) Conversion of VDF (%) Figure 5.6 Plot of M v of PVDF vs. Conversion of VDF On the other hand, VDF monomer conversion increases with the reaction time. Figure 5.6 shows a good linear relationship between the resulting polymer molecular weight and the monomer conversion. This indicates living polymerization of VDF using iodoperfluoroalkane as the chain transfer agent. The extrapolation of the molecular weight vs. monomer conversion plot does not go through the original. Compared with the diiodoperfluorobutane case, the deviation from the original is even larger. This is because more imperfect chain ends containing fragments of AIBN are generated, which are inert and cannot continuously grow the polymer chain with time.

162 147 Overall, the iodine transfer polymerization using mono-diiodoperfluorobutane is not as well-controlled as with diiodoperfluorobutane. Generally, the resulting polymer chains contain higher amounts of undesirable end groups from AIBN fragments.

163 148 Run Table 5.4 Polymerization of VDF in the presence of AIBN and iodoperfluorobutane (R f -I) AIBN (mmol) R f -I (mmol) AIBN/ R f -I (mol/mol) Rxn time (hr) Yield (g) Covn of VDF (%) Covn. of R f -I (%) I chain end (%) MW End group (k) (CH 3 ) 3 C CF 3 CH 2 CF 2 I CF 2 CH 2 I Effect of reaction time / / / / / / Reaction condition: VDF: 30 ml (~20g), temperature: 80 o C.

164 Functional Group Transformation of Iodine Terminal Fluoropolymer a. Ethyleneation of iodine terminal PVDF From the iodine transfer polymerization of VDF, it is found that the chemical stability of the end groups decreases in the order of CF 2 CH 2 I > CH 2 CF 2 I> CF 2 CF 2 I. It is logical to think the transformation of all the reactive CH 2 CF 2 I and CF 2 CF 2 I groups to CH 2 CH 2 I group by reacting with ethylene in the presence of a tiny amount of AIBN. However, it s also possible more than one CH 2 CH 2 unit inserting into the CF 2 -I bond due to the oligomerization tendency of ethylene, as illustrated in Scheme 5.4. PVDF CH 2 CF 2 I + n CH 2 =CH 2 AIBN PVDF CH 2 CF 2 (CH 2 CH 2 ) n I PVDF CF 2 CF 2 I + n CH 2 =CH 2 AIBN PVDF CF 2 CF 2 (CH 2 CH 2 ) n I Scheme 5.4 Ethylenation of iodine terminal PVDF Figure 5.7 represents the NMR spectra of the polymer that is subjected to ethylene and AIBN treatment at 80 o C for 4hr, followed by the extraction with hexane. The peaks corresponding to CH 2 CF 2 I (3.4 ppm in 1 H NMR and 41 ppm in 19 F NMR), and that assigned to CF 2 CF 2 I (67 pmm in 19 F NMR) almost completely disappear. Although the expected peak at 3.2 ppm (based on the report from reference 20 ) in 1 H NMR spectrum for the resulting new CH 2 CH 2 I chain end is overlapped by the major peak assigned to the repeat units of PVDF (-CH 2 CF 2 CH 2 CF 2 -), a new peak at 1.3 ppm corresponding to repeated CH 2 CH 2 units emerges in the 1 H NMR spectra. From the intensity of this ethylene peak, it is found that each polymer chain in average contains about 2.3 ethylene repeat units. Overall, the combination indicates that the AIBN/ethylene treatment is a relatively effective approach to converting the reactive CF 2 I group into the stable CH 2 I group.

165 150 Figure H and 19 F NMR spectra of the product from ethylene treatment.

166 151 b. Transformation to Imidazolium ion On the other hand, the reactive CF 2 I end groups can react with imidazole to form the imidazolium cation, as illustrated in Scheme 5.5. Figure 5.8 shows the 1 H and 19 F NMR spectra of the imidazolium capped PVDF polymer. The peaks at 3.4 ppm in 1 H NMR and 41ppm in 19 F NMR, assigned to CH 2 CF 2 I, disappear. The peaks at 7-8 ppm in 1 H NMR are attributed to the protons in the imidazolium, while the peak at 85 ppm in 19 F NMR is assigned to CH 2 CF 2 N moiety. Apparently, almost all the CF 2 I groups have been transformed to imidazolium ions. It is interesting to point out that the PVDF with terminal imidazolium can serve as an effective surfactant in the preparation of a polymer/clay nanocomposite, which will be discussed in Chapter VI. PVDF CF 2 I + N PVDF CF 2 + N I - N H N H Scheme 5.5 Transformation to imidazolium ion

167 152 CF 2 CH 2 I CH 2 CF 2 N Figure H NMR spectra of PVDF-t-Imm.

168 Summary This chapter has discussed an improved iodine transfer polymerization (ITP), based on the combination of a specific radical initiator (AIBN) and a reversible additionfragmentation chain transfer (RAFT) process involving two iodo-compounds, i.e., α,ωdiiodoperfluoroalkane (I-R f -I) and mono-iodoerfluoroalkane (R f -I). We take the advantage of the inactive carbon radicals, created by the decomposition of AIBN, toward fluoromonomers, which readily react with the iodo-compounds (chain transfer agents) before initiating the polymerization. The combination of this two-step initiation and RAFT control radical polymerization mechanism affords the telechelic fluoropolymers with almost all the polymer chains containing one or two terminal iodo groups. Using diiodoperfluoroalkane as the chain transfer agent is more effective than monoiodoperfluoroalkane. In turn, the reactive terminal CF 2 I groups can undergo facile ethylenation to convert all polymer chain ends to stable CH 2 I groups. On the other hand, the terminal CF 2 I groups have also been successfully transformed to imidazolium ions that are very effective in forming a fluoropolymer/clay nanocomposite (will be discussed in next Chapter). References: 1 Ameruri, B.; Boutevin, B. J. Fluor. Chem. 2000, 104, 53 2 Rice, D.E., Sandber, C.K US Patent US 3,461,155 3 Saint-Loup, R.; Manseri, A.; Ameduri, B.; Lebret, B.; Vignane, P. Macromolecules 2002, 35, Tatemota, M.; Morita, S. US Patent US4,361,678, Tatemota, M.; Nakagawa, T. US Patent US4,158,678, Chung, T.C.; Jiang, G.J.; Rhubright, D. US Patent US5,286,800, 1994

169 154 7 Chung, T.C.; Jiang, G.J.; Rhubright, D. US Patent US5,401,805, Xu, G.; Chung, T.C. J. Am. Chem. Soc. 1999, 121, Lu, B.; Chung, T.C. Macromolecules, 31, Chung, T.C. US Patent 6,420,502, Chung, T.C. US Patent 6,515,088, Chung, T.C.; Janvikul, W.; Lu, H.L. J. Am.Chem. Soc. 1996, 118, Petchsuk, A. Ph.D. dissertation, Chung, T.C.; Petchsuk, A. US Patent US , Chung, T.C.; Petchsuk, A. Macromolecules 2002, 35, Odian, G. Principles of Polymerization, New York, Wiley Manseri, A.; Ameduri, B.; Boutevin, B.; Chambers, R.; Caporiccio, G.; Wright, A.P. J. Fluor. Chem. 1995, 74, Pianca, M.; Barchiesi, E.; Esposto, G.; Radice, S. J. Fluor. Chem. 1999, 95, Balague J.; Ameduri B.; Boutevin B.; Caporiccio, G. J. Fluor. Chem.1995, 70, Renn, J.A.; Toney, A.D.; Terjeson, R.J.; Gard, G.L. J. Fluor. Chem. 1997, 86, 113

170 155 Chapter VI Preparation and Characterization of Functional Polymer/clay Nanocomposites 6.1. Introduction Although it has long been known 1 that polymers can be mixed with appropriately modified clay minerals and synthetic clays, the field of polymer/clay nanocomposites has recently gained large momentum. Two major findings pioneered the revival of these materials. First, the report of a nylon-6/montmorillonite (mmt) material from Toyota research, 2,3 where it was shown that very moderate inorganic loadings resulted in concurrent and remarkable enhancements of thermal and mechanical properties. For instance, a polyamide-6 clay nanocomposite, containing 5% clay, showed improvement of 40% in tensile strength, 68% in tensile modulus, 60% in flexural strength, and 126% in flexural modulus, while the heat distortion temperature increased from 65 to 152 o C and the impact strength was lowered by only 10%. Second, Giannelis 4 found that it is possible to melt-mix polymers with clays without the use of organic solvents. Since then, the high promise for industrial applications has motivated vigorous research, which has revealed concurrent dramatic enhancements of many materials properties by the nano-dispersion of inorganic silicate layers. These improvements are generally applicable across a wide range of polymers in instances where the property enhancements originate from the nanocomposite structure. At the same time, also discovered were property improvements in these nanoscale materials that could not be realized by conventional fillers. Examples include increased tensile strength, flexural modulus, impact toughness, general flameretardant characteristics, 5 and a dramatic improvement in barrier properties. 6

171 156 Scheme 6.1 Structure of 2:1 phyllosilicates 7 The clay (layered silicates) commonly used in the nanocomposite belong to the structural family known as the 2:1 phyllosilicate. Most preferred are smectite clay minerals, such as montmorillonite (mmt). Scheme 6.1 illustrates the structure of 2:1 phyllosillicates. 8 The crystal lattice consists of 1 nm thin layers, with a central octahedral sheet of alumina fused between two external silica tetrahedral sheets (in such a way that the oxygen atoms from the octahedral sheet also belong to the silica tetrahedral). These layers organize themselves to form stacks with a regular van der Waals gap between them, called interlayer or gallery. In the pristine form their excess negative charge is balanced by cations (Na +, Li +, Ca 2+ ). The cations can be easily exchanged to protons (H + ) by acid-treatment to form acidic clay or to ion-exchange to other cations by treating them with cationic-organic surfactants, such as alkylalmmoniums, to form organophilic clay. One of the common commercially available clays is dioctadecylammonium-modified montmorillonite (2C18-mmt).

172 157 Oxygen OH group Silicone (Aluminum) Aluminum or Magnesium Scheme 6.2 Structure of Chlorite 8 Besides mmt, other clays, such as bentonite, and chlorite, have also been used in preparing polymer/clay nanocomposite with specific properties. Bentonite consists predominantly of mmt with a minor portion of illite, a clay mineral with a dioctahedral structure and higher layer charge. The introduction of bentonite usually provides the nanocomposite with better mechanical and barrier properties than that of mmt. On the other hand, Chlorite also has a basic 2:1 layer structure, but with a stable, positive charged octahedral sheet in the interlayer space (shown in Scheme 6.2). The positive charges of this octahedral sheet neutralize the negative charge of the 2:1 sheets. Because chlorite contains two octahedral sheets, it is called a 2:1:1 layer mineral. Sometimes, octahedral materials in chlorite neither totally fill the interlayer space between sheets nor completely neutralize the negative charge of the sheets. This unsatisfied charge is neutralized by various cations adsorbed to the particle surfaces from other materials. In

173 158 other words, chlorite undergoes an ion-exchange process more readily than other clays with a 2:1 layer structure. 9,10 Table 6.1 lists the clay minerals used in our study of polymer/clay nanocomposite, including pure Na + mmt, bentonite and chlorite. The cation exchange capacity decreases in the order of chlorite > bentonite > Na + mmt (Kunimine) > Na + mmt (Southern clay). As will be discussed, the mechanical property of the polymer/clay nanocomposite is related to the nature of the clay minerals. Table 6.1 Properties of clay minerals used Clay Structure Supplier/brand Characteristics Na+-mmt TOT (or 2:1) Southern Clay / Na Cloisite CEC: 90 meq/100g, d spacing: 1.2 nm Na+-mmt TOT(or 2:1) Kunimine Ind. /Kunipia-F 98-99% of mmt and <1% of quartz/chalcedony. CEC: 115 meq/100g Bentonite TOT(or 2:1) Mitsubishi / MC-C 75-85% of mmt, 5-10% of illite, and 5-10% of quartz, feldspar etc. CEC: 119 meq/100g Chlorite [TOT]O[TOT] (or 2:1:1) Iriki Kaolin /C-1 CEC: 145 meq/100g, d spacing: 1.4 nm On the basis of theoretical models, Bazlazs 11,12 suggested that an increase in the length of the organic molecules tethered on clays is a promising approach to further promoting dispersion of clay sheets in a polymer matrix. Toward this end, recent reports have employed a variety of polymers (such as polystyrene, 13 poly(methyl methacrylate), 14,15 and poly(ε-caprolactone) 16 grafted on montmorillonite surfaces, and have observed varied extents of clay dispersion. Furthermore, subsequent annealing of these polymer-bearing clays, or mixing them with the respective homopolymers, can lead to retention or collapse of the clay dispersion depending on the polymer. The differences

174 159 in these responses were attributed to the specific polymer-clay interactions in each case, indicating that where there exist strong interactions the clay dispersions collapse, whereas in those cases with weak polymer-clay interactions the dispersions can be retained. Recent advances in polymer/clay nanocomposites (mostly in polar polymer) have inspired researchers to investigate the systems involving polyolefin, the most important family of commercial polymers. Many research efforts 17,18 have focused on dispersing montmorillonite in polypropylene (PP), which is a fast growing thermoplastic that dominates industrial applications due to its attractive combination of properties and low cost. As expected, because of the absence of any strong interactions, it has been a scientific challenge to disperse silicate clays in the highly apolar polyolefins. The general approach for improving the compatibility of PP with organically modified clays has been the addition of polar functional groups to the PP polymer, typically resulting in PP/clay nanocomposites with a mixed nanomorphology, including both intercalated and exfoliated structures coexisting in the system. 19 Unfortunately, the availability of functional PP is very limited due to synthetic chemical difficulties. Most of PP/clay nanocomposite studies employed a commercially available maleic anhydride-grafted PP (PP-g-MAH) polymer, despite its shortcomings with a very complicated molecular structure due to many side reactions, 20 (including severe chain degradation) during the free radical grafting process, as well as impurities from oligomerization of MAH. As mentioned in Chapter II, the hurdle of obtaining appropriately functionalized PP can be overcome by a general funcitonalization approach involving the combination of metallocene catalysts and reactive comonomers, 21 which can yield a broad range of sidechain functionalized polyolefins with relatively well-defined molecular structures. As

175 160 discussed in Chapter III, we discovered a facile route using reactive chain transfer agents to prepare chain-end functionalized polyolefins 22,23,24 containing a terminal functional group (OH, NH 2, COOH, anhydride, etc.), while maintaining a well-controlled polymer molecular weight and narrow molecular weight and composition distributions. 25,26 The availability of many well-defined chain-end or side-chain functionalized polypropylenes allows us to study the most desirable functional PP structure that can disperse pristine montmorillonite (without any treatment of organic compound) into a polymer matrix. As will be discussed, the ammonium group terminated i-pp (PP-t-NH + 3 ) is one of the most effective macromolecular surfactants to result in an exfoliated PP/clay nanocomposite. Both X-ray diffraction and TEM image show the complete disorder clay layer structure in the PP/clay nanocomposite. Mechanical property evaluation shows the + addition of PP-t-NH 3 in the system remarkably enhanced flexural modulus. These experimental results demonstrate the advantage of chain-end functionalized PP in the formation of an exfoliated clay layer structure, and lead to the proposition that the terminal hydrophilic NH 3 + functional group anchors the PP chains on the inorganic surfaces via ion exchange, and that the hydrophobic high molecular weight and semicrystalline PP chains are repelled from the inorganic surfaces and exfoliate the clay platelets. On the other hand, fluoropolymers, such as poly(vinylidene fluoride) (PVDF), and poly(vinylidene-co-hexafluoropropylene (VDF/HFP elastomer), exhibit a unique combination of properties, including thermal stability, chemical inertness (acid and oxidation resistance), low water and solvent absorptivities, excellent weatherability, and very interesting surface properties. They are commonly used in many high-end

176 161 applications, such as aerospace and automotive technologies, and in microelectronics. However, fluoropolymers also have some drawbacks, including limited processability, poor adhesion to substrates, limited crosslinking chemistry, and inertness to chemical modification, which limit their applications when interactive and active properties are paramount. As expected, fluoropolymers with both hydrophobic and organophobic properties are the most difficult polymers in which to achieve a good dispersion in polymer/clay composites. The pristine clays with a highly hydrophilic property are immiscible with fluoropolymers, and there is no conceivable driving force to stabilize interfaces to form a thermodynamically stable exfoliated clay structure in a fluoropolymer matrix. The scientific challenge may have deterred some exploration of fluoropolymer/clay nanocomposites, despite many potential advantages of such a unique material with the combination of physical properties from both materials. Only very few prior arts 27 discuss the preparation of fluoropolymer/clay nanocomposites with very limited success. The general approach for improving the incompatible fluoropolymer blending problem has been the use of organophilic clay (pretreated with an organic surfactant). However, the fluoropolymer/clay nanocomposites thus formed exhibit an intercalated structure, with very limited exfoliated structure. Overall, there was no experimental result demonstrating the advantage of using the chain end functionalized fluoropolymer that can be directly mixed with neat (pristine) silicate clay (without pre-treatment with organic surfactants) to form the exfoliated fluoropolymer/clay nanocomposites, which, furthermore, maintains this disordered clay structure even after further mixing with a neat

177 162 polymer that is compatible with the backbone of the chain end functionalized fluoropolymer. In this chapter, I will discuss a new in situ process to prepare chain end functionalized fluoropolymers that firmly anchor onto clay surfaces between interlayers and exfoliate the clay interlayer structure during the polymerization. This process involves the use of a functional radical initiator, which allows one step polymerization to prepare fluoropolymer/clay nanocomposites. 6.2 Experimental Materials Na + - montmorillonite (Na + -mmt) with an exchange capacity of ca. 90 mequiv/100g (WM) and Cloisite 20A (2C18-treated mmt clay) was obtained from Southern Clay Product; Potassium persulfate, hydrogen peroxide, and solvents were purchased from Aldrich and used as received. Preparation of PP-t-NH + 3 Cl - /Na + -Montmorillonite Nanocomposite Static melt intercalation was employed to prepare a PP-t-NH + 3 Cl - /Na + - montmorillonite nanocomposite. PP-t-NH + 3 Cl - (M n = 58,900 and M w = 135,500 g/mol) dried powder and Na + -mmt with 90/10 weight ratio were first mixed and ground together in a mortar and pestle at ambient temperature. The XRD pattern of this simple mixture shows a (001) peak at 2θ~7, corresponding to a Na + -mmt interlayer structure with a d- spacing of 1.26 nm. The mixed powder was then heated at 190 o C for 2 hr under a nitrogen condition. The resulting PP-t-NH 3 + Cl - / Na + -mmt nanocomposite shows a featureless XRD pattern, indicating the formation of an exfoliated clay structure.

178 163 The binary PP-t-NH + 3 Cl - / Na + -mmt exfoliated nancomposite aforementioned was further melt mixed (50/50 weight ratio) with commercial neat i-pp (M n =110,000 and M w = g/mol). The PP-t-NH 3+ Cl - / Na + -mmt exfoliated nanocomposite and neat i- PP with a 50/50 weight ratio were ground together in a mortar and pestle at ambient temperature. This simple mixture shows a featureless XRD pattern. The mixed powder was then heated at 190 o C for 2 hours under nitrogen conditions. The resulting ternary PP/ PP-t-NH + 3 Cl - / Na + -mmt nanocomposite also shows a featureless XRD pattern, indicating that the stable exfoliated structure in the binary PP-t-NH + 3 Cl - /Na + -mmt exfoliated nanocomposite is clearly maintained after further mixing with PP, which is in turn compatible with the backbone of PP-t-NH + 3 Cl -. Preparation of PVDF/clay Nanocomposite by melt blending using PVDF-t-Si interfacial Agent The silane terminated PVDF polymers (PVDF-t-Si) were used as interfacial agents in the preparation of exfoliated PVDF/clay nanocomposites by the melt blending process. Whereby a PVDF-t-Si polymer containing a terminal C 2 H 5 OSi(CH 3 ) 2 - group (T m =168 o C, M n =40,000 g/mol) was mixed with Na + -mmt. The static melt intercalation method was employed by first mixing PVDF-t-Si dried powder and Na + -mmt with 90/10 weight ratio in a mortar and pestle at ambient temperature. The XRD pattern of this simple mixture shows a (001) peak at 2θ ~ 7, corresponding to a Na + -mmt interlayer structure with a d- spacing of 1.45 nm. The mixing powder was then heated at 190 o C for 3 hr under nitrogen condition. The resulting PVDF-t-Si/Na + -mmt nanocomposite shows a featureless XRD pattern, indicating the formation of an exfoliated clay structure. The resulting binary PVDF-t-Si/Na+-mmt exfoliated nanocomposite was further mixed by melting (50/50

179 164 weight ratio) with commercial neat PVDF (Mn =70,000 and Mw =180,000 g/mol). First, the PVDF-t-Si/Na + -mmt exfoliated nancomposite and neat PVDF with 50/50 weight ratio were ground together in a mortar and pestle at ambient tmepertature. The mixed powder was then heated at 200 o C for 3 hr under nitrogen condition. The resulting ternary PVDF/PVDF-t-Si/Na + -mmt nanocomposite also shows a featureless XRD pattern, indicating that the stable exfoliated structure in the binary PVDF-t-Si/Na + -mmt exfoliated nancomposite is clearly maintained after further mixing with PVDF that is compatible with the backbone of PVDF-t-Si. In situ Polymerization process to prepare PVDF/clay nanocomposite using potassium persulfate as initiator In a 25 ml vial, 0.10 g of Na+-MMT clay and 0.10 g of potassium persulfate were dispersed in 2 ml of de-ionized water. The obtained dispersion was charged into a 75 ml stainless steel reactor containing a magnetic stirring bar and 30 ml of acetonirtrile. Then, 20 g of the VDF monomer was condensed from a high-vacuum line into the reactor, which was subsequently warmed up to room temperature and put in a heated oil bath at 80 o C for 8 hrs. After the solvent was evaporated, 6.7 g of polymer (Mn=52,500 g/mol, PDI = 2.0) was obtained, with a yield of 34% and clay content of 1.5 wt%. The dried product was ground in a mortar and pestle at ambient temperature and cast into a pellet for XRD measurement. In situ Polymerization process to prepare PVDF/clay nanocomposite using hydrogen peroxide as initiator In a 25 ml vial, 0.15 ml of H 2 O 2 (50 weight % in water) was mixed with 0.20 g of Na+-MMT clay in 2 ml of de-ionized water. The obtained dispersion was charged into a

180 ml stainless steel reactor containing a magnetic stirring bar and 30 ml of acetonirtrile. Then, 20 g of VDF monomer was condensed from a high-vacuum line into the reactor, which was subsequently warmed up to room temperature and put in a heated oil bath at 110 o C for 3 hrs. After the solvent was evaporated, 6 g of polymer hybrid was obtained with a yield of 30% and clay content of 3.3 wt%. The dried product was ground in a mortar and pestle at ambient temperature and cast into a pellet for XRD measurement. X-ray Diffraction Measurements XRD data were collected by a diffractometer using Cu Kα radiation (λ = nm). Bragg s law (λ = 2d sinθ) was used to calculate the spacing. Transmission Electron Microscopy Ultra thin sections of the polymer/clay nanocomposite with a thickness of approximately 70 nm were prepared at room temperature using an ultra microtome equipped with a diamond knife. The sections were transferred dry to carbon-coated Cu grids of 200 mesh. The contrast between the layered silicates and the polymer phase was sufficient for imaging, so heavy metal staining of sections prior to imaging was not required. Direct observation of the polymer/clay nanocomposite structure was realized by a bright field TEM of nanocomposite films under strain in a JEOL 1200EX operating at 120kV.

181 Results and Discussion Application of Chain-end Functionalized Polyolefins in Polymer/clay Nanocomposite As aforementioned, the availability of a broad range of well-defined functionalized polyolefins provides us with a great advantage in evaluating their applications in polyolefin/clay nanocomposites. The most desirable ammonium group-terminated i-pp (PP-t-NH + 3 ) polymers 28 were prepared by the combination of an Exxon-Hoechst C 2 - symmetric catalyst (rac-me 2 Si[2-Me-4-Ph(Ind) 2 ] 2 ZrCl 2 /MAO system) and silaneprotected styrene (4-{2-[N,N-bis(trimethylsilyl)amino]ethyl}styrene) as chain transfer agent, as discussed in Chapter III. For comparison, both pristine Na+-montmorillonite clay (Na+-mmt) and a dioctadecylammonium-modified montmorillonite organophilic clay (2C18-mmt) were used. Static melt intercalation 29 was employed to prepare all PP/clay nanocomposites. The advantage of this method is being free of mechanical shear or solvent. The dispersions achieved are thermodynamically favored, rather than kinetically trapped by the processing conditions.

182 167 Figure 6.1 X-ray diffraction patterns of PP-t-NH 3 + Cl-/Na + -mmt (90/10 weight ratio): (a) physical mixture by simple powder mixing at ambient temperature and (b) the same mixture after static melt-intercalation (PP-t-NH3+/mmt hybrid). PP-NH 3 + /10 wt% 2C18-mmt Relative Intensity 2.41 nm (b) Nanocomposite (a) physcial mixture θ Figure 6.2 XRD patterns of PP-NH 3 + Cl - /2C18-mmt (90/10) weight ratio): (a) physical mixture by simple powder mixing at ambient temperature and (b) the same mixture after static melt-intercalation

183 168 Both PP-t-NH + 3 Cl - /Na + -mmt and PP-t-NH 3+ Cl - /2C18-mmt nanocomposites were prepared and evaluated under the same conditions. Figure 6.1 compares the X-ray diffraction (XRD) patterns before and after static annealing of a physical mixture (90/10 weight ratio) of an ammonium-terminated i-pp (PP-t-NH + 3 Cl - : M n =58,900 and M w = g/mol; T m = o C) and a pristine Na + -mmt clay. Simple mixing of dried PP-t- NH 3+ Cl - powder and Na + -mmt, ground together by mortar and pestle at ambient temperature, created the XRD pattern in Figure 6.1a, with a (001) peak at 2θ ~ 7 o, corresponding to the characteristic Na + -mmt d spacing of ca nm. The mixed powder was then heated-annealed under static conditions at 190 o C for 2 hr under nitrogen. The resulting PP-t-NH + 3 /mmt hybrid shows a featureless XRD pattern in Figure 6.1b, indicating the formation of an exfoliated clay structure, which corresponds to the thermodynamically stable state, as the ammonium-terminated PP exchanged with the alkali (Na + ) cations at the mmt surfaces. Similar results were also observed in the PP-t-NH + 3 Cl - /2C 18 H 37-2CH 3 -N + -mmt nanocomposite case. The mixture of the powders gives a peak at 2θ ~ 3.58, indicating the interlay distance of this 2C18-mmt is around 2.41nm, as shown in Figure 6.2(a). After being heated-annealed at 190 o C for 2 hr, the sample shows a featureless XRD pattern in Figure 6.2(b), suggesting an exfoliation structure was formed in the PP/2C18-mmt nanocomposite. It is clear that an organic surfactant is not needed to promote compatibility between the PP-t-NH 3 + Cl - and pristine Na + -mmt clay. Besides the economic benefits, the elimination of organic surfactant also offers some significant advantages. For example, it removes two major concerns relating to the thermal stability of the surfactant during a high-temperature melt processing, and to the long-term stability

184 169 of the small organic surfactant in the polymer/clay nanocomposite under various application conditions. Figure 6.3 X-ray diffraction patterns of the 50/50 mixture by weight of exfoliated PP-t- NH3+/mmt structure (90/10 weight ratio) and neat-unfunctionalized-i-pp. the XRD traces shown correspond to (a) the physical mixture of PP-t-NH3+/mmt and i-pp and (b) the same mixture after static melt-intercalation.

185 Figure 6.4 TEM image of exfoliated PP/PP-t-NH 3 + /mmt nanocomposite. Magnification: 30,000 (top) and 120,000 (bottom) 170

186 171 The binary PP-t-NH 3 + /mmt hybrid was further mixed /blended (50/50 weight ratio) with a neat-unfunctionalized-i-pp (M n = and M w = g/mol). Figure 6.3 shows the XRD patterns of (a) the physical mixture and (b) melt blending of the exfoliated PP-t-NH 3 + /mmt structures with neat i-pp. The exfoliated structure is maintained after further mixing with i-pp, which is compatible (cocrystallizable) with the backbone of the PP-t-NH + 3 polymer. The exfoliated structure was directly observed in the TEM image (Figure 6.4), being well-dispersed by layers in the PP matrix. Apparently, the i-pp polymer chains largely serve as diluents in the ternary PP-t-NH3+/mmt/i-PP system, with the thermodynamically stable PP-t-NH + 3 /mmt exfoliated structure dispersed in the i- PP matrix. It is interesting to study the efficiency of the PP-t-NH + 3 in the PP/clay composite having a high clay content. Figure 6.5 shows two TEM images of a PP-NH + 3-1/Chlorite (50/50 wt%) hybrid, where the darker features correspond to the Chlorite layers dispersed in the bright polymer matrix, which display the distribution of clay clusters (tactoids and agglomerates) at a lower magnification (left image), and also large numbers of single clay layers dispersed in the polymer at a higher magnification (right image). Even in this composite containing 50% of clay and a very low concentration of NH + 3 chain end, the substantial amount of an exfoliated clay structure again verifies the efficiency of PP-t- NH + 3 in the polymer nanocomposite. Although these two images do not display the desirable structure of the clay layers, a better dispersed structure would be possible to + form in the hybrid with a high content of nanofiller, if high molecular weight PP-NH 3 with a high content of NH + 3 chain ends would be used.

187 Figure 6.5 TEM images of PP-NH /Chlorite hybrid (50/50 wt%). Magnification: 10,000 (left) and 50,000 (right). 172

188 173 Figure 6.6 X-ray diffraction patterns of 2C18-mmt clay and four nanocomposites with 6 wt% of alkylammonium-mmt and 94 wt% of three side-chain-functionalized PPs containing (a) 1 mol% p-methylstyrene, PP-r-MS; (b) 0.5 mol% maleic anhydride, PP-r- MA; and (c) 0.5 mol% hydroxyl, PP-r-OH. (d) a 6 wt% C18-mmt nanocomposite of a PP-b-PMMA block copolymer, with 5 mol% of methyl methacrylate. For comparison, several functionalized PP polymers, containing randomly distributed functional groups in the side chains, or lumped together in a block copolymer microstructure, were also evaluated in PP/montmorillonite nanocomposites. Similar static melt-intercalation procedures were followed, except for employing alkyl-ammoniummodified montmorillonites (2C18-mmt for all random copolymers and C18-mmt for the block copolymer). Figure 6.6 shows the XRD patterns of four nanocomposites made with a 6 wt% of 2C18-mmt clay and 94% of three side-chain-functionalized PPs (M n =

189 174 and M w = g/mol) containing (a) 1mol % p-methylstyrene, (b) 0.5 mol% maleic anhydride, and (c) 0.5 mol% hydroxyl side groups and the original 2C18-mmt clay. All the functionalized PPs were derived from the same random PP copolymer synthesized by metallocene catalysis, which contained a 1 mol% p-methylstyrene (p-ms) comonomer. Subsequently, the p-mss were selectively functionalized toward hydroxyl (OH) and maleic anhydride without changing the PP backbone. Moreover, we show the XRD pattern of PP-b-PMMA block copolymker/6 wt % C-18mmt, which contains a 5 mol% methyl methacrylate block (Figure 6.6). These XRD patterns clearly show that there is a definite intercalated structure for all the cases of side-chain-functionalized PPs, manifesting itself through an interlayer d spacing increase about 1 nm compared to that of the parent alkylammonium-mmt. 13 Figure 6.7 Bright-field TEM image of PP-r-MA/6 wt% 2-C18 mmt nanocomposite structure. Magnification: 200,000.