George Wypych-PVC Degradation and Stabilization, Third Edition-ChemTec Publishing (2015), Notas de estudo de Cultura

Baixe George Wypych-PVC Degradation and Stabilization, Third Edition-ChemTec Publishing (2015) e outras Notas de estudo em PDF para Cultura, somente na Docsity! PVC Degradation & Stabilization 3nd Edition George Wypych Toronto 2015 Published by ChemTec Publishing 38 Earswick Drive, Toronto, Ontario M1E 1C6, Canada © ChemTec Publishing, 2008, 2015 ISBN 978-1-895198-85-0 Cover design: Anita Wypych All rights reserved. No part of this publication may be repro- duced, stored or transmitted in any form or by any means with- out written permission of copyright owner. No responsibility is assumed by the Author and the Publisher for any injury or/and damage to persons or properties as a matter of products liabil- ity, negligence, use, or operation of any methods, product ideas, or instructions published or suggested in this book. Library and Archives Canada Cataloguing in Publication Wypych, George PVC degradation & stabilization/George Wypych.--Third Edition Includes bibliographical references and index. ISBN 978-1-895198-85-0 (bound) 1. Polyvinyl chloride--Deterioration. 2. Stabilizing agents. I. Title. II. Title: PVC degradation and stabilization. TP1180.V48W96 2015 668.4’236 C2014-907418-2 Printed in Australia, United States and United Kingdom 1.1 Repeat structures and their basic organic chemistry 3 A consequence of the crowded structure of the carbon atom in the transition state is that bulky groups introduced at the reaction center are generally associated with a decreased reaction rate. Comparing the C−H and C−Cl bonds, we may expect that the C−Cl bond should dissoci- ate more easily for two reasons: first, because the energy barrier is lower, and second, because the electric charge dislocation facilitates the chemical reaction. To better understand the subject under discussion and the complexities of the bonds' nature, one should review typical mechanisms of reactions. Elimination of the E1 type proceeds through the intermediary of the carbonium ion. The sequence of reactions could be written for the VC monomer in the following form: Several factors are known to favor this reaction mechanism. One is the presence of polar solvents in the reacting mixture. Reactions in solution might follow different routes without polar solvents, as there is a substantial change in energy during reaction and because of participation of solvation energy. Sometimes these changes in energy are more important for the reaction course than is the strength of forming and breaking bonds. If we also consider collision mechanics in the solution, the probability of such collisions, and the number of encounters, we may conclude that the presence of solvent might play the most important role for the reaction mechanism and for the reaction rates.6 On the other hand, the solvent might also contribute to the energy transfer from the molecular fragment in vibrational mode which otherwise would lead to a transitional state, and therefore, it can act in collisional deactivations. Table 1.2. Bond dissociation energy. Bond ΔH, kJ/mol CH3−H 435 (CH3)2CH−H 395 CH2=CHCH2−H 356 C6H5−CH2−H 356 CH3−Cl 351 (CH3)3C−Cl 331 C6H5CH2−Cl 285 CY X B + CH2CH Cl B + CH2CH + Cl BH + CH=CH + Cl 4 Chemical Structure of PVC The elimination reaction of the E1 type will also be favored by a good leaving group. Chlorine might be considered as such a group. Finally, the presence of any group at the position which stabilizes the carbonium ion would contribute to the E1 mechanism's prev- alence. This reaction mechanism is especially characteristic for the tertiary and secondary alkyl halides. The elimination reaction according to the E1cB mechanism includes carboanion as the intermediate: The presence of an acidic β−proton is thought to increase the rate of carboanion forma- tion. A poor leaving group should slow down E2 elimination and favor carboanion forma- tion. Among halogens, fluorine forms the poorest leaving group, and chlorine is the sec- ond in sequence. The E2 mechanism depends on an energy imbalance introduced in the transitional states described below: The character of the transitional state determines the degree of C−H and C−Cl bond break- ing. In the course of the elimination reaction, any group formed or any change in reaction conditions which might affect the charge distribution or activation process favors the par- ticular direction of the elimination process. In the case of the PVC monomeric fragment, the leaving group, that is, HCl, will stabilize the carboanion structure in relation to the car- bonium ion and thus increase C−H bond breaking. For the carboanion transition state the β-substituents of the electron-withdrawing character (e.g, double bonds) should increase the carboanion character and favor C−H bond breaking. Reacting base B might also con- tribute to the reaction mechanism due to its nucleophilicity. The E2C transition states have the following structure: E2C elimination is favored by dipolar aprotic solvents such as DMF and any good leaving group (e.g., Cl). Also, substrates, having tertiary carbon atoms, will undergo changes according to this mechanism more easily. B + CH2CH Cl BH + CHCH Cl BH + CH=CH + Cl CH H B δ CH Cl CHCH H B Cl central character CH H CH Cl δ carboanion carbocation B H B Cl HC CH 1.1 Repeat structures and their basic organic chemistry 5 Another mechanism, known in the literature under the name of four-center HX elim- ination, might help explaining the course of the PVC degradation process. This mecha- nism was suggested on the basis of studies of unimolecular elimination of HX from alkyl halides. The possibility of partial charge separation was proven quite early.7 Further stud- ies of the reaction and its activation energy parameter for many alkyl monohalides con- firmed the polar nature of reaction and the existence of surface catalysis.8 The essence of the above mechanism can be explained by the following equation: It has been established that methyl and other n-alkyl groups bound to the α-posi- tioned carbon atom participate in the stabilizing effect. The chlorine atom affects the for- mation of the positive charge on the carbon atom to which it is bound, but it does not contribute to the formation of a negative charge on the neighboring β-carbon. The effect just described contributes to C−X bond energy, which, together with H−X bond dissocia- tion energy, determines the rate of initiation. The presence of HX can promote autocata- lytic reactions. Overall, one might expect a dehydrochlorination reaction of the first order to take place. Molecular elimination of HX is usually preferred in the case of primary and second- ary carbon atoms, while the radical route might prove to be essential when halides are bound to tertiary carbon atoms. The activation energy in the four-center elimination pro- cess is drastically decreased when methyl, vinyl, aromatic groups or oxygen are bound to a carbon atom which forms a C−X bond at the same time. If one of these groups is bound to a carbon atom forming a C−H bond, the effect is negligible or not as strongly pro- nounced as in the other case. This might suggest that if one considers the regular distribu- tion of the C−X bonds at each second carbon atom, which is generally the case in PVC, one may expect reaction to proceed along the chain until the first irregularity appears, which might be why reaction is terminated for that center. Having completed this brief listing of the major mechanisms believed to be essential for explaining the nature of the HX elimination process, we still need to describe the reac- tion which involves the substitution of chlorine by other groups in the course of the same reaction. Two well-known mechanisms of substitution have been used in organic chemis- try for years: the SN1 and SN2 types. The mechanism SN2 resembles some features of elimination mechanism, E2, and mechanism SN1 competes with elimination of the E1 type because they share carbocation intermediate. For our studies, it is essential to notice the differences between both mechanisms in respect to reactions which may occur if segments of the PVC chain are thermodynamically unstable and therefore able to react. As we con- cluded above, the carbon atom, because of its four-valent character and its need to form a crowded transition state, including accommodation of five groups, is not the easiest one to undergo substitution reaction. Referring that to the mechanism of SN2 type, which C C R1 R2 R3 R4X H C C R1 R2 R3 R4X H C C R1 R2 R3 R4 δ δ δ δ + HX 8 Chemical Structure of PVC jugated double bonds which ultimately leads to polyacetylene − a conductive polymer. On the other hand, a volatile product of reaction − HCl − is used for in situ doping of emeral- dine base of polyaniline to make it conductive. In a similar experiment,11 PVC was used for lithographic purposes. HCl formed in a similar process was used to change the color of methyl violet. These two of many examples show unusual applications of major reaction in polyvi- nylchloride degradation process which is a subject of many studies and practical applica- tions of stabilization. The Bronsted acid principle is also associated with the catalysis of PVC dehydro- chlorination. Owen pointed out that is the most effective catalytic moiety in PVC dehydrochlorination reaction.12 Later Gerard13 confirmed, by electronic and reso- nance Raman studies, that polyene-metal salt complexes are formed, even when zinc chlo- ride is replaced by stannous chloride. Later, Tran et al. included this mechanism in their polaron mechanism of thermal stabilization.14 In summary, the concept of Bronsted acid-base is one of the fundamental principles governing chemistry of PVC degradation and stabilization. 1.2 MOLECULAR WEIGHT AND ITS DISTRIBUTION Polyvinylchloride is synthesized by several different methods of polymerization including suspension, mass, and emulsion polymerization (details of polymerization reaction are included in Chapter 2). In the course of synthesis, a mixture of polymer chains is produced having a broad distribu- tion of molecular weights. Molecular weight and its distribution are related as Figure 1.1 shows. Polydispersity of PVC increases with increase in its molecular weight. Molecular weight is one of the most important properties of polyvinylchloride. There are many examples given in this book which show that molecular weight determines almost all processing and appli- cation characteristics. Currently produced polymers have molecular weight tailored to the processing methods and they are optimized to have optimal thermal stability and other properties important for processing and application. Table 1.3 shows molecular weights and polydispersity of commercial polymers.15,16 H+ZnCl3 - Figure 1.1. Polydispersity of suspension PVC vs. its weight average molecular weight. [Data from Pepperl, G., J. Vinyl Additive Technol., 6, 2, 88-92, 2000.] 1.2 Molecular weight and its distribution 9 1.2.1 KUHN-MARK-HOUWINK-SAKURADA The intrinsic viscosity can be related to a molecular weight with a power law expression known as the Kuhn-Mark-Houwink-Sakurada, KMHS, equation: [1.1] where: [η] intrinsic viscosity K, a constants Mv molecular weight calculated based on viscosity measurements. Molecular weight obtained from this equation is not an absolute measure because it depends on solvent and molecular weight distribution. 1.2.2 FIKENTSCHER K NUMBER Fikentscher derived the K number in 1932 from measurements of the relative viscosity of dilute polymer solutions. The k value is calculated from the following equation: [1.2] where: ηc viscosity of solution η0 viscosity of solvent k constant c concentration in g/100 mol. Fikentscher K number is given by equation: K = 1000k [1.3] The Fikentscher K number is very commonly used in polyvinylchloride. Commercial polymers have K number usually in the range of 50 to 90. Fikentscher K number should not be confused with the K constant of the Kuhn-Mark-Houwink-Sakurada equation. 1.2.3 CHAIN LENGTH The Flory-Fox equation gives dependence between chain length and glass transition tem- perature. The equation was modified O’Driscoll et al.17 to fit experimental data for major polymers with a fitting coefficient (2/3). The equation takes the following form: Table 1.3. Typical molecular weight and polydispersity range.15,16 Polymerization method Mn Mw Mw/Mn Suspension 20,32-69,141 38,611-179,123 1.90-2.59 Emulsion 27,173-49,540 61,650-131,191 2.14-2.65 Mass 26,351-37,772 52,683-77,829 2.00-2.06 η[ ] KMv a= ηc η0 -----log 75k 2 1 1.5kc+ ---------------------- k+ c= 10 Chemical Structure of PVC [1.4] where: Tg,n glass transition temperature asymptotic value of glass transition temperature K polymer-specific constant Xn the number-average chain length Figure 1.2 shows experimental data for polyvinylchloride.17 In the absence of a detailed, theoretical model for Tg the modi- fied Flory-Fox equation can be regarded as an empirical expression of the glass transi- tion dependence on chain-length of poly- mer. An inverse 2/3 dependence on chain length has also been experimentally observed and modeled for both the polymer surface tension and the interfacial tension of immiscible polymers. New research may help in clarification and better understand- ing of this equation. 1.3 PREDICTION OF FORMATION OF IRREGULAR SEGMENTS 1.3.1 AB INITIO Ab initio calculations were used to predict formation of irregular segments during PVC synthesis. Table 1.4 gives predicted concentrations of various irregular segments. Table 1.4. Results of ab initio calculation. [Data from Van Cauter, K.; Van Den Bossche, B. J.; Van Speybroeck, V.; Waroquier, M., Macromolecules, 40, 4, 1321-1331, 2007.] Irregular segment Concentration in number of groups per 1000 VC mers chloromethyl branch (tertiary H) 4.3 2,4-dichloro-n-butyl branch 0.14 1-chloro-2-alkene end group 1.4 1,2-dichloroalkane end group 1.4 1,2-dichloroethyl branch 0.020 Tg,n Tg,∞ K Xn 2 3⁄ ----------–= Tg ∞, Figure 1.2. Dependence of glass transition temperature of PVC on its number-average chain length. [Data from O'Driscoll, K.; Sanayei, R. A., Macromolecules, 24, 4479-4480, 1991.] 1.4 Irregular segments 13 ethyl butyl The long branches are formed according to the same mechanism but by propagation from a different (more distant from the end of chain) radical location. In literature, the short branch term includes methyl, ethyl, and butyl, and long branches have up to 100 mer units.21 The length of a branch thus depends on the location of a radical. Methyl branches can be formed via a possible 1,2-intra- or intermolecular hydrogen abstraction reaction. Methyl branches containing a tertiary Cl may also be formed via an intermolecular hydro- gen abstraction from the β-carbon of the dead polymer. Methyl and ethyl branches having tertiary carbon can be formed by head-to-head addition and subsequent Cl shifts and prop- agation. 2-chloroethyl and 2,4-dichloro-n-butyl branches are formed by intramolecular abstraction of hydrogen from a backbone CHCl moiety (back-biting reaction) and subse- quent head-to-tail propagation.22 Depending on the position of a radical, the resultant branches can be attached to a carbon atom which has either tertiary hydrogen or tertiary chlorine atom.19 Further details on these reactions, especially historical aspects of their discovery, can be found in the review papers.23,24 The first data on branching came from studies of the viscosity of PVC solutions25 and GPC.26 One can easily understand that these methods can give only very approximate results, and hence further studies were conducted to find possible application of spectral methods in quantitative determination of the number of branches in PVC. Two methods of polymer preparation for direct observation by spectral methods are used. These methods include PVC reduction, either by LiAIH428-33 or by Bu3SnH, as proposed by Starnes.27 Reduction by LiAIH4 is incomplete, requires long reduction time (1–3 weeks), and pro- duces undesirable side reactions; therefore, the Bu3SnH reduction is used more frequently. Reduced PVC is further studied by 13C NMR using methods developed for polyethylene. Historical details on development of the methods of study can be found in a review CH2 CH CH Cl CH2 Cl VC CH CH2 ClCHCl CH2 Cl CH2 CH CH2 C Cl Cl Cl CH2 CH2 CH2 VC CH2 CH2 CH Cl C ClCH2 CH2 CH2 Cl CH CH Cl 14 Chemical Structure of PVC paper.24 The concentrations of different types of branches are given in Table 1.6 based on the available literature. 1.4.2 TERTIARY CHLORINE Table 1.6 and discussion in the previous section show that some concentrations of tertiary chlorine are present and that they may come from ethyl (only in PVC synthesized at ele- vated temperature), butyl, and long branches as is commonly accepted. Considering that long branch concentration is usually low, most tertiary chlorine comes from butyl branches.44 The 1-2 intermolecular hydrogen shift would result in a fraction of the methyl branches carrying a tertiary chlorine, and they may therefore be more significant for the thermal stability of PVC (especially at high monomer conversions) than was described earlier.19 Rogestedt and Hjertberg have studied branched structures and they concluded that out of all tertiary carbons only 2-2.5% are associated with chlorine.35-36,38 In their exten- sive studies of tertiary chlorine, their concentration was estimated to be between 0.7 and 2.1 groups per 1000 VC.36 The energy of C-Cl is given below:37 It is pertinent from this comparison that lower energy is required to break a chlorine bond with tertiary carbon as compared to secondary carbon. But, allylic chlorine bond is still easier to break. At the same time, it should be noted that concentrations of tertiary chlorine are usually larger than the concentrations of allylic chlorines. For this reason the influence of tertiary carbons on degradation rate is believed to be prevalent.35,37 The enthalpy Table 1.6. Branches in PVC Irregular segment Concentration in number of groups per 1000 VC mers Refs. chloromethyl branch (tertiary H) 3.3-4.8 18 chloromethyl branch 3.0-4.8 34-36 1,2-dichloroethyl branch 0.14-0.55 18 ethyl branch 0.1-0.55 35-36 2,4-dichloro-n-butyl branch 0.5-1.7 18 butyl branch 0.5-1.7 35-36 short chain branch 0.8 37 long chain branch (tertiary Cl) 0.07-0.14 18 long chain branch (tertiary H) 0.03-0.07 18 long chain branch 0.1-0.2 18, 35-37 C H H C H Cl 322 kJ C H H C H C H C H Cl 243 kJ C Cl CH2 280 kJ 1.4 Irregular segments 15 required for reaction to occur is larger for chlorine attached to the secondary carbon (195 kJ/mol) than tertiary carbon (174 kJ/mol).39 Experimental studies show that if the content of allylic and tertiary chlorine is elimi- nated by reaction with trialkylaluminum or tin derivatives, PVC has a decreased rate of degradation.40 Hjertberg at al. regard tertiary chlorine to be the most important defect of PVC synthesis from the point of view of its impact on PVC thermal degradation.41 Studies on synthesis by Hamielec et al.42 lead to the same conclusion based on a good correlation between degradation rate and concentration of tertiary chlorine. The concentration of tertiary chlorines increases with an increase in monomer con- version beyond critical conversion (~60%). At the same conversion level, the concentra- tion of tertiary chlorine atoms increases with an increase in polymerization temperature.42 The theoretical investigation of PVC structural irregularities, using MNDO and AM1 calculations, show that chlorine atoms bound to the tertiary carbon atoms are responsible for the low thermal stability of PVC.43 1.4.3 UNSATURATIONS (END CHAIN – VINYL, IN-CHAIN – VINYLENE) The presence of internal double bonds in PVC has been studied by UV and Raman spec- trophotometry, 13C NMR, 1H NMR, and degradative methods. It is important for these studies to evaluate the quantity of the double bonds present in polymers, but it is also expected that single internal double bonds, sequences of conjugated double bonds, double bonds in conjugation with oxygen-containing groups, and double bonds at chain ends can be determined separately. Several methods can be used to study the unsaturated structures present in PVC. The main method is based on bromination, which gives the total number of double bonds pres- ent in polymer. Ozonolysis, in turn, coupled with molecular weight change determination, gives the number of double bonds which are internally present in the polymer chain. The difference in reading gives double bonds at the chain ends. Numerous investigators have studied the chemical structure of chain ends. The most often applied technique is based on the 1H NMR method. Petiaud45 and Schwenk46 were able to record signals assigned to the following struc- tures: Caraculacu,47 using Fourier transform 1H-NMR, found a series of five signals, three of which were assigned to the following structures: Hjertberg48 confirmed the presence of Schwenk's structure: and found that the other groups included at chain ends contain saturated groups: –CHCl- CH2Cl and –CH2CH2Cl. The presence of all these groups was discussed by Braun.49 CH2CH CHCH2Cl CHClCH CHCH2Cl CH2CHClCH CH2 CH2CHClCCl CH2 CH2CH CClCH2Cl CHClCH CHCH2Cl 18 Chemical Structure of PVC 1.4.4 OXYGEN CONTAINING GROUPS 1.4.4.1 Ketochloroallyl groups Minsker66,67 proposed that allylic chlorine does not exist in normal polymer because it is immediately oxidized during production, drying, and storage to form: Caraculacu68 found that 4-chlorohexene-2 and 2-chlorohexene-3 were not oxidized to ketoallylic structures even after 1 year at room temperature. Braun69 found that keto groups in alpha position to allylic chlorine decrease their rate of degradation rather than increase it because of their electron withdrawing character. Minsker66 shows that ketochloroallyl groups are at concentrations of 0.1/1000 VC which is below the detection limits of 13C NMR.70 For this reason, in a dedicated study70 the presence of ketochloroallyl groups could not be confirmed. At the same time, it was found that spray drying produces these groups only if PVC is dried without prestabilizer. Drying in a laboratory oven required at least 50 h exposure to 50oC to detect the presence of some groups which could increase initial degradation rate.70 Minsker66 suggested that the ketochloroallylic groups are 100 to 1000 more reactive than chloroallylic groups and for this reason they are predisposed to act as initiation cen- ters. It is a correct way of reasoning, but at the same time it should be considered that they are in such low concentration that even if they were proven to be present in commercial PVC, their impact on the PVC degradation rate would have been minimal. 1.4.4.2 α- and β-carbonyl groups If a C=C bond can weaken a C-Cl bond in the β-position, the same effect should be expected when a double bond is formed between the carbon atom and oxygen. The possi- bility of these groups serving as initiation centers for dehydrochlorination was discussed by Bauer71 and George.72 Svetly73 suggested that the initiation reaction proceeds through a cyclic transition state. Braun demonstrated that carbon monoxide can copolymerize with vinyl chloride, leading to unstable structures, as follows: 1.4.5 HEAD-TO-HEAD STRUCTURES (1,2-DICHLORO GROUPS) It is thought that the PVC chain consists mainly of a head-to-tail arrangement of monomer units. The first studies on irregular structures were conducted by Canterino,74 Bailey,75 and Murayama,76 who obtained head-to-head PVC by chlorination of 1,4-polybutadiene. The reaction used for synthesis of head-to-head PVC was later studied by several researchers.77-81 CH2CH CHCHCl + O2 CHCHClCCH O + H2O CH2 CCH CH2 CH OCl Cl 1.4 Irregular segments 19 The following is the chemical reaction of head-to-head structure formation:21 The propagation reaction is an important elemental reaction to determine molecular weight and structural defects. Monomer can add onto the polymer chain in two ways: head-to-head or head-to-tail but it grows mainly by head-to-tail addition.21 The activation energy of head-to-head addition is 6.28 kJ/mol higher than head-to-tail addition. The prop- agating radicals produced by head-to-head addition induce 1,2-chloro-atoms migration followed by addition of VC, leading to the formation of chloromethyl branches.21 Table 1.8 shows the frequency of head-to-head addition as determined by various researchers Table 1.8 shows that there is a good agreement between different sources. Murayama76 showed that head-to-head PVC degrades at a lower temperature than the head-to-tail form, which was also confirmed by Crawley.81 1.4.6 INITIATOR RESTS Selection of initiator affects the reaction rate, molecular weight, morphology of PVC par- ticles, uniformity of polymerization reaction, and thermal stability of resultant polymer. Bulk, solution and suspension polymerization use monomer soluble initiators, unlike emulsion polymerization, which uses a monomer insoluble initiator. Initiator is usually selected in such a manner that it has about 1 h half-life under reac- tion conditions (formulation, temperature).84 For typical suspension and bulk initiators, the temperature of half-life varies in the range of 54-80oC.84 Chemical structures and basic properties of main commercial initiators can be found elsewhere.84 A single initiator is seldom used in commercial polymerization. Commercial formulations usually use 2-4 dif- ferent initiators. The most reactive initiator is used to begin polymerization and when it is exhausted, the next initiator takes a leading role. Peroxydicarbonate is commonly used as the primary initiator.84 A typical formulation will use 0.03 to 0.08 parts of initiators per 100 parts of vinyl monomer.84 The method of addition of initiator has a great effect on PVC particle uniformity as well as the size distribution.21 If initiator is dissolved in monomer, polymer particles are more uniform than if initiator is predispersed in a continuous phase of suspension polym- erization medium. This shows that diffusion of initiator into monomer droplets is time- Table 1.8 Head-to-head structures in PVC Source Head-to-head structures per 1,000 monomer units Year Mitani82 6-7 1975 Schwenk83 8 1979 Rabek64 6-7 1985 Endo21 6.4-6.9 2002 CH2CHCH2CH Cl Cl + CH Cl CH2 CH2CHCH2CHCHCH2 Cl Cl Cl 20 Chemical Structure of PVC controlled.85 Composition of polymerizing mixture may affect also other morphological features such as particle size distribution and porosity. Morphology, especially porosity, is a factor in dehydrochlorination because it affects diffusion of hydrogen chloride − a catalytic product of dehydrochlorination. Also, molec- ular weight and its distribution affect dehydrochlorination. In addition, initiator rests remain as a part of polymer composition. It is proposed that initiator rests may dissociate, as a result of which a free, low molecular radical is formed.86 This radical is mobile and thus may participate in initiation of dehydrochlorination. Effect of initiator rests is also explained by polaron formation from the eventual oxidoreduction reaction of polyene sequences by residues of initiators.87 Depending on the selected initiator, more or less irregular segments are produced, and this affects dehydrochlorination.88 It was estimated that about 20% chain ends are formed by initiator ends.65 This is in line with the mechanism of polymerization because initiators begin a chain of polymer reactions with involvement of monomer which is more likely to react due to its higher concentration in reaction mixture. Initiator may also end a reaction chain by reacting with a microradical, and this produces initiator rest in the chain ends. Table 1.9 gives concen- tration of initiator rests from literature data. Direct data on influence of initiators on dehydrochlorination are limited. Azo- bis(isobutyronitrile) was found to give less stable polymer than bis(4-t-butylcyclo- hexyl)peroxydicarbonate.91 Initiator residues also affect early color of polymer most likely due to increased defect structures.50 Unreacted initiator was more detrimental for PVC thermal stability because it was able to abstract hydrogen from polymer chain, which initi- ated chain reactions.50 1.4.7 TRANSFER AGENT RESTS 2-mercaptoethanol is a popular chain transfer agent for manufacture of PVC of lower molecular weight.92 Other transfer agents include: dodecylmercaptan, α-thioglicerol, and thioglycolicacid.93-94 1.4.8 DEFECTS INTRODUCED DURING PROCESSING Considering that the entire book is devoted to the formation (and prevention of) of defects during various types of processing, we will only acknowledge the problem here as opposed to an exhaustive discussion. Any form of degradation produces HCl, and as a result, a system of isolated and con- jugated double bonds. The presence of these bonds changes electron distribution along the chain and invites oxidative reactions which produce hydroperoxides and eventually carbo- Table 1.9. Initiator rests in PVC. Source Initiator rests per 1,000 monomer units Year Guyot89 0.8 1986 McNeill90 0.2-0.25 1995 Endo21 0.14-0.4 2002 Van Cauter18 negligible 2007 References 23 67 Minsker, K. S.; Lisitsky, V. V; Kolesov, S. V.; Zaikov, G. E., Vyssokomol. Soed., 23, 483, 1981. 68 Caraculacu, A. A., Pure Appl. Chem., 53, 385, 1981. 69 Braun, D.; Bohringer, B.; Ivan, B.; Kelen, T.; Tudos, F., Eur. Polym. J., 22, 299, 1986. 70 Behnisch, J.; Zimmermann, H.; Anders, H., Polym. Deg. Stab., 15, 335-344, 1986. 71 Bauer, J.; Sabel, A., Angew. Makromol. Chem., 47, 15, 1975. 72 George, M. M.; Garton, A., J. Macromol. Sci., Chem., A11, 1389, 1977. 73 Svet1y, J.; Lucas, R.; Micha1cova, J.; Kolinsky, M., Macromol. Chem., Rapid Commun., 1, 247, 1980. 74 Canterino, P. J., Ind. Eng. Chem., 49, 712, 1957. 75 Bailey, F. E; Henry, J. P.; Lundberg, R. D.; Whelan, J. M., J. Polym. Sci., B2, 447, 1964. 76 Murayama, N.; Amagi, Y., J. Polym. Sci., B2, 115, 1966. 77 Takeda, M.; Endo, R.; Matsura, Y., J. Polym. Sci., Part C, 23, 487, 1968. 78 Horhold, H.; Kuhrnstedt, R.; Hindersin, P.; Dawczynski, H.; Drefal, C., Makromol. Chem., 122, 145, 1969. 79 Inomoto, S., J. Polym. Sci., Part A, 7, 1225, 1969. 80 Ito, H.; Tsuge, S.; Okumoto, T.; Takeuchi, T., Makromol. Chem., 138, 111, 1970. 81 Crawley, S.; McNeill, I. C.; J. Polym. Sci., Polym. Chem. Ed., 16, 2593, 1978. 82 Mitani, K.; Ogata, T.; Awaya, H.; Tornari, Y., J. Polym. Sci., Polym. Chem. Ed., 13, 2813, 1975. 83 Schwenk, W.; Loemker, F.; Koenig, J.; Streitberger, M., J. Appl. Polym. Sci., 23, 1595, 1979. 84 Zimmermann, H., J. Vinyl Additive Technol., 2, 4, 287-294, 1996. 85 Zerfa, M.; Brooks, B. W., J. Appl. Polym. Sci., 65, 1, 127-134, 1997. 86 Michell, E. W. J., J. Mater. Sci., 20, 3816-3830, 1985. 87 Hoang, T. V.; Guyot, A., Polym. Deg. Stab., 32, 93-103, 1991. 88 Endo, K.; Emori, N., Polym. Deg. Stab., 74, 1, 113-117, 2001. 89 Guyot, A., Macromolecules, 19, 4, 1090-1105, 1986. 90 McNeill, I. C.; Memetea, L.; Cole, W. J., Polym. Deg. Stab., 49, 1, 181-91, 1995. 91 Xie, T. Y.; Hamielec, A. E.; Wood, P. E.; Woods, D. R.; Chiantore, O., Polymer, 32, 9, 1696-1702, 1991. 92 Yamamoto, K.; Maehala, T.; Mitani, K.; Mizutani, Y., J. Appl. Polym. Sci., 54, 8, 1161-9, 1994. 93 Yamamoto, K.; Maehala, T.; Mitani, K.; Mizutani, Y., J. Appl. Polym. Sci., 51, 4, 755-9, 1994. 94 Yamamoto, K.; Maehala, T.; Mitani, K.; Mizutani, Y., J. Appl. Polym. Sci., 51, 4, 749-53, 1994. 95 Minsker, K. S.; Zaikov, G. E., J. Vinyl Additive Technol., 7, 4, 222-234, 2001. 2 PVC Manufacture Technology 2.1 MONOMER Vinyl chloride is the sole monomer used in the production of PVC. Its structure and prop- erties are given in Table 2.1. Table 2.1. Typical properties of vinyl chloride Property Description General Vinyl chloride is a flammable, colorless gas with a sweet odor at room temperature. It is mainly used for production of PVC but also as a refrigerant and comonomer Chemical structure Chemical formula C2H3Cl IUPAC name 1-chloroethene Alternate names chloroethylene, chloroethene, ethylene monochloride Identifiers CAS # − 75-01-4; EINECS − 200−831−0; UN # − 1086; RTECS − KU9625000 Molecular weight, g/mol 62.5 Odor mild, sweet Color colorless Melting point, oC -153.8 Boiling point, oC -13.37 Decomposition temperature, oC 450 Flash point, oC -78 Autoignition temperature, oC 472 Explosive limits in air, vol% 3.6 and 33 Density, kg/m3 2.56 (gas); 908.41 (liquid) Vapor density (air=1) 2.15 Vapor pressure at 25oC, mm Hg 2600 C Cl C H H H 26 PVC Manufacture Technology 2.2 BASIC STEPS OF RADICAL POLYMERIZATION 2.2.1 INITIATION Vinyl chloride has low reactivity but its radical is highly reactive. For this reason, PVC is manufactured by radical polymerization. Initiation is a two step process, as given by the following reactions: First, the initiator produces free radicals which then react with monomer, forming a radical ready for the next step − propagation. This method of initiation causes the chain end structure of PVC macromolecules to consist of initiator fragments. Initiators are important components of industrial polymerization formulations. Table 2.2 shows chemical structures of the most common representatives with their half-life temperatures, HLT. Solubility in water, wt% 0.95 Solubility of water in 100 g VC, ml 0.08 (20oC) Solubility in solvents alcohol, ether, carbon tetrachloride, benzene Henry’s law constant, atm m3/mol 2.71E-02 Critical temperature, oK 432 Critical pressure, MPa 5.67 Enthalpy of vaporization, kJ/mol 21.998-23.3 Enthalpy of fusion, kJ/mol 4.92 Enthalpy of formation, kJ/mol 28.45 Specific heat capacity, J/kg/g 0.8592 (gas), 0.9504 (solid) Heat capacity of liquid at 298.15K, J/molK 0.053625 Latent heat of vaporization, kJ/kg 333 Antoine equation parameters A − 3.98598, B − 892.757, C − -35.051 (165.2-259.3K) Octanol/water partition, logKow 1.36 Odor threshold, ppm 3000 Half-life in air, h few Table 2.1. Typical properties of vinyl chloride Property Description I 2I (I - initiator) I + M R (M - monomer; R - growing polymer radical) 2.2 Basic steps of radical polymerization 29 Polymerization rate is almost independent of initiator concentration because the presence of polymer in the reaction mixture exerts a catalytic effect, promoting further polymerization. The catalytic effect of polymer is proportional to its actual surface area. The propagation rate constant, kp, is given by the following equation:1 [2.1] where: R gas constant T absolute temperature. This equation confirms the previous statement that initiator concentration does not enter the relationship, which is only controlled by temperature. The rate of polymerization is not uniform throughout the mixture because it is higher inside a swollen particle of polymer. This is most likely due to the catalytic effect of poly- mer and reduction of radical termination rate because of reduced mobility of the growing radical. The rate of polymerization is thus auto-accelerating, which, in combination with its high exothermic effect (106 kJ/mol), puts emphasis on effective cooling which controls the process rate and quality of the product. 2.2.3 TERMINATION Termination occurs by disproportionation and combination, as shown by the following reactions: The rate constant of termination, kt, is expressed by the following relationship: The ratio of disproportionation to combination is difficult to determine but the com- bination is known to be a prevalent mechanism of termination.1,5 One peculiarity of PVC polymerization is that polymer is not soluble in its monomer and it precipitates when formed. Monomer is capable of swelling polymer, which results in a porous structure of polymer grains formed, as well as the possibility that reaction of polymerization continues in a portion of monomer used for swelling the already formed grains. This gives PVC peculiar morphological structures which can be manipulated by process. kp 3.3 10 6 3700 RT -----------–exp×= R1CH2CHCl + R2CH2CHCl R1CH CHCl + R2CH2CH2Cl R1CH2CHClCHClCH2R2R1CH2CHCl + R2CH2CHCl kt 1.3 10 12 4200 RT -----------–exp×= 30 PVC Manufacture Technology 2.2.4 CHAIN TRANSFER TO MONOMER A growing radical may react with monomer, forming polymer and a monomer radical, which further grows, due to reaction with monomer. These reactions, called chain transfer reactions, may occur according to one of the schemes below:1 The third reaction scheme is considered to be the main reaction of a new radical genera- tion based on activation energies of reactions.1 Chain transfer constant for PVC is given by the following equation: [2.2] The values of Cm are large due to the high reactivity of vinyl chloride. This reaction con- trols the molecular weight of polymer. With temperature of reaction increasing, the value of constant increases, which results in an increased number of chain transfers and lower molecular weight of the resultant polymer. Chain transfer to polymer results in formation of long branches as discussed in Chapter 1.3 Various chain transfer agents, such as dodecylmercaptan, 2-mercaptoethanol, α-thio- glycerol, or thioglycolic acid, may be used to decrease molecular weight of the resultant polymer.4 2.3 POLYMERIZATION TECHNOLOGY PVC production has grown by about 4% per year since 1990. The largest growth is experi- enced in Asia (especially in China), followed by North America and Europe.6 There are four major methods of production of PVC, including bulk, emulsion, solution, and suspen- sion polymerizations. Suspension polymerization is the most common since it accounts for 80% of world production of PVC (75% in USA and 93% in Japan).6 It is followed by production of paste-type resins by emulsion (12% in the world) and microsuspension polymerization (3% USA and 5% Japan). Bulk polymerization contributes 8% of world production of PVC. Solution polymerization is not a commercial scale method. CH2 Cl CH + CH2 CH Cl CH2 CH2 Cl CH2 CHCl Cl CH2 CH2 Cl CH CH Cl + + + CH2 C Cl CH2 CH+ CH CH Cl CH3 CH Cl Cm 125 7300 RT -----------–exp= 2.3 Polymerization technology 31 2.3.1 SUSPENSION Suspension polymerization is a batch process which includes four unitary operations: • polymerization • stripping residual vinyl chloride • centrifugation of water suspension to remove water • drying. The typical formulation of reactor charge is given in Table 2.3. Types of initiators used are presented in Table 2.2 and their role is discussed in the previous section. A suspending agent (frequently called a protective colloid) prevents agglomeration of vinyl chloride droplets changing later into sticky particles of formed polymer. Without a suspending agent, uniformity of suspension will suffer and particle size will have a very broad range, with large lumps which will diminish the quality of PVC. Water soluble natural and synthetic polymers are known to prevent coalescence of droplets and agglomeration of particles. Partially hydrolyzed polyvinyl acetate (polyvinyl alcohol) and hydroxypropyl methylcellulose are the most popular suspension agents. Two types of commercial products are used: primaries and secondaries.6 Primaries, which have a degree of hydrolysis in the range of 71 to 82%, are used to control particle size of PVC. Secondaries, which have a degree of hydrolysis in the range from 45 to 57%, are known to increase porosity. Depending on application and, thus, required morphology, different combinations of both suspending agents are used. It is quite obvious from formulation that suspension PVC contains limited amounts of additives which may complicate its further processing and its thermal and UV stability. Polymerization is performed in large reactors (usually having a volume of up to 200 m3). Considering that vinyl chloride is carcinogenic, most reactors are operated without opening. Several hundred batches can now be run without opening the reactor. This is due to excellent coating which prevents deposition of reacting mass on the reactor walls. It is also very important to control the temperature of reaction which is exothermic. The rate of heat removal is usually the rate controlling factor. After reaction is complete, walls of the reactor are rinsed with water to remove any particles. If such particles are left in the reactor for the next batch, a monomer will be absorbed by them and polymerized within particles, resulting in non-porous particles which will cause formation of so-called “fish-eyes”. In addition, these particles will hold more vinyl chloride and increase its residual concentration in the final product. The reac- tor is coated with adhesion-preventive compound before a new batch begins. The reactor is charged again and reaction mixture is brought to a desired temperature to begin a new Table 2.3. Typical formulation for suspension polymerization.6 Component Concentration, parts Vinyl chloride 100 Demineralized water 120 Suspending agent 0.05-0.1 Initiator 0.03-0.16 34 PVC Manufacture Technology In microsuspension polymerization a mixture of vinyl chloride, water, emulsifier (e.g., alkyl aryl or alkyl sulfonate), and monomer-soluble initiator (e.g., lauroy peroxide) is homogenized in a separate vessel and pumped into the reactor. The mixture is then heated with agitation to the polymerization temperature. A stable emulsion with particle size of 0.1-5 μm is formed. This is in contrast to suspension or bulk polymerization which produces much larger particles (100 μm and above). Microsuspension polymerization dif- fers from emulsion polymerization in the type of initiator used (oil or water soluble, respectively) and an additional operation used in microsuspension (homogenization), which is required to obtain regular, small particles. In the case of emulsion polymerization, reaction begins in a water phase taking advantage of the fact that vinyl chloride is slightly soluble in water. Typical initiators of emulsion polymerization are ammonium and potassium persulfates and hydrogen perox- ide. Polymerization continues in the monomer phase because polymer is not soluble in either monomer or water but monomer is very soluble in polymer (30 wt% monomer can be absorbed by polymer). Good control of particle size due to the use of proper mixing and effective emulsifier permits production of PVC with very narrow particle size distribution. Frequently, emulsion polymer is mixed in a production plant with microsuspension poly- mer which has a much broader particle size distribution. By such premixing, an adequate proportion of various particle sizes can be achieved in order to obtain a maximum packing density of polymer which reduces plasticizer demand (and viscosity) of PVC plastisols. In the emulsion polymerization, the type and the quantity of emulsifier is the most crucial feature of the process which impacts particle size distribution of the resultant resin. Usually anionic emulsifiers are used and their concentrations are high compared with rela- tively pure suspension polymerizates. Unlike in other types of polymerization, emulsifier may not be added at the beginning but is metered throughout the process to control the rate of initiation and the size of particles (if more surfactant is added, more new growth sites are formed and particles become smaller). The rates of polymerization can be increased by the addition of reducing agents which increase the rate of initiation and help in reducing polymerization temperature. Particle size distribution of emulsion polymerization can be further controlled by the so-called seeded emulsion polymerization. By choice of size and number of seed latex, the number of growth sites (particles) is controlled, giving superior control over particle size Table 2.4. Typical formulations for microsuspension and emulsion polymerizations.6,10 Concentration, parts microsuspension emulsion Vinyl chloride 100 100 Water 137 110-180 Oils soluble initiator 0.2 - Water-soluble initiator - 1-1.2 Emulsifier 1 1-2 2.3 Polymerization technology 35 distribution. If combination of large and small particles is required, this is usually achieved by the use of a blending tank in which latices of different particle sizes are mixed prior to drying. The polymerization reactor is protected against deposition of particles either by com- position (glass lining or use of stainless steel reactors) or by a protective coating similar to suspension polymerization. Conversion of monomer for both microsuspension and emul- sion polymerization is in the range of 80-90%. In the next step, reaction mass must be stripped of vinyl chloride. This is done by a combination of vacuum and temperature, most frequently in a reaction vessel but also in special tanks, columns, or thin film evapo- rators. If the temperature of the process is higher than the glass transition temperature of PVC (82oC), polymer will begin to degrade. A slurry stripped to at least 8.5 ppm of vinyl chloride is transferred to dryer. Slurry is pumped to a spray dryer where it is atomized into fine droplets falling into hot air. The selection of drying temperature (inlet and outlet temperatures) is important for quality since PVC without thermal stabilizers can be degraded and also agglomerated (if temperature is close to the glass transition temperature). The resin agglomeration also depends on atomizer tip speed. Atomizers are either equipped with a spinning disc or noz- zle.6 The performance of atomizer controls the size of the latex droplet and combination of inlet and outlet temperature and atomizer tip speed controls percentage of agglomerates. Higher atomizer tip speed produces a higher percentage of coarse particles. Inlet tempera- ture increase causes an increase in the amount of agglomerates because more particles become fused. Similar effect has an increase in outlet temperature. Fine particles of paste resin tend to lump and cake, and for this reason bag shipment is more frequently used than bulk transportation. The above description of the process shows that there are several points where qual- ity of polymer may be affected so as to influence its future thermal and environmental sta- bility. These include: • effect of initiator type and concentration • effect of polymerization temperature on molecular weight and polydispersity • total conversion of monomer (defects types and their concentrations) • type and quantity of emulsifier • size of polymerization reactor (ease of cooling which affects efficiency and molecular weight and its distribution) • reactor wall coating and rinsing efficiency (concentration of particles, which cause fish eye defects) • stripping temperature and duration (affect color and stability of resin) • inlet and outlet temperatures in spray dryer • atomizer tip speed (influences time of exposure to elevated temperature). Some of these influences (formation of defects) were already discussed and others are discussed in the next chapter. 36 PVC Manufacture Technology 2.3.3 BULK Developed by Saint Gobain, later Rhone-Poulenc, and now Arkema (Lacovyl), the bulk (or mass) polymerization is the simplest process which uses only vinyl chloride monomer and initiator.6 It does not require preparation of solutions, water, drying, separation, and waste water treatment. It is very efficient process, which, conducted in a small reactor, yields a large production output. It also results in the purest polymer because of simple polymerization formulation. There are also many disadvantages, such as • long heating required to strip residual monomer • difficulties in cleaning off-gases (steam and/or nitrogen) from residual monomer • problems with efficient wall coating and thus batch process is required • production of fine particles which are difficult to handle • complicated, two-stage process. The last disadvantage is caused by the fact that during polymerization the reaction mixture changes from liquid to solid and it needs different conditions of mixing. The first commercial process in 1939 used a single reactor but sine 1962, two reactors are used: pre-polymerizer and post-polymerizer. In the pre-polymerizer, reaction is conducted up to 10% conversion, and in the post-polymerizer, up to 80% conversion of monomer. Since the first stage process is more efficient (reaction takes only 20 min), one pre-polymerizer is serving several post-polymerizers (initially horizontal but from 1978, vertical reactors).6 Reaction temperature at the first stage is higher (usually 60-70oC) than in the second stage (usually 50-60oC). Initiators for the first and the second stage are different, considering temperature difference. The first stage produces seed particles and therefore controls parti- cle morphology. The higher reaction temperature increases aggregation. Aggregation is also controlled by the type and quantity of initiator added in the first stage. Also, agitation speed helps to control primary particle size. The temperature of the second stage is selected to control molecular weight of the product. Second stage initiators have longer half-life time. Usually a combination of 2-3 initiators is used to obtain constant tempera- ture of reaction. Material becomes powdery at about 20-25% conversion.6 Final conver- sion of monomer is very important for porosity of particles. A high conversion rate will cause densification of particles and reduction of porosity. The above description of the process shows that there are several points where qual- ity of polymer may be affected so as to influence its future thermal and environmental sta- bility. These include: • effect of initiator type and its concentration in each stage of polymerization • effect of polymerization temperature on molecular weight and particle size • total conversion of monomer (defects types and their concentrations) • reactor wall coating and cleaning efficiency (concentration of particles, which cause fish eye defects) • stripping temperature and duration (affects color and stability of resin). Some of these influences (formation of defects) were already discussed and others are discussed in the next chapter. 2.4 Polymerization conditions and PVC properties 39 viscosity of reaction medium produces smaller particles.16 Oxygen is capable of reacting with initiator primary radicals and oligomers in the aqueous phase to produce vinyl polyperoxides.17 The vinyl polyperoxides can decompose into radicals, which are capable of initiating new polymer chains. The average particle size exhibited a U- shaped behavior with respect to the initial oxygen concentration. This behavior is explained by the combined role of vinyl polyperoxides as radical generators and ionic strength promoters (Figure 2.5).17 Processes following polymerization also affect particle size, as seen from the example of drying emulsion polymerizate (Figure 2.6).6 Not only atomizer speed but especially inlet and outlet air temperatures affect agglomeration and thus particle sizes. In addition, inlet and outlet temperatures cause initial degradation of PVC. Pore diameter (Figure 2.7)16 and porosity (Figure 2.8)18 of particles decrease with conversion and/or polymerization time increasing. Similar is the case of specific surface area (Figure 2.9), which also decreases with conversion.14 This is understandable consid- ering that monomer diffuses into polymer grains and fills the available free volume. Figure 2.5. Effect of the initial oxygen concentration on the dimensionless average particle diameter. [Data from Kiparissides, C.; Achilias, D. S.; Frantzikinakis, C. E., Ind. Eng. Chem. Res., 41, 3097-3109, 2002.] Figure 2.6. PVC agglomeration vs. atomizer tip speed in spray drying of emulsion polymer. [Data from Saeki, Y.; Emura, T., Prog. Polym. Sci., 27, 2055-2131, 2002.] Figure 2.7. Mean pore diameter vs. conversion. [Data from Cebollada, A. F.; Schmidt, M. J.; Farber, J. N.; Capiati, N. J.; Valles, E. M., J. Appl. Polym. Sci., 37, 145-166, 1989.] 40 PVC Manufacture Technology Porosity of suspension polymer depends also on composition of reaction mixture. Poly(vinyl alcohol) is frequently used as protective colloid. Poly(vinyl alcohol) is obtained by hydrolysis of poly(vinyl acetate). The degree of hydrolysis, in the range from 45 to 81.5%, affects porosity of PVC particles, as can be seen from Figure 2.10.2 Polymerization temperature affects glass transition temperature of polymer which is inversely proportional to polymerization temperature (Figure 2.11).19 Figure 2.12 shows that glass transition temperature of polymer affects its processing properties.20 The higher Figure 2.8. Porosity vs. polymerization time. [Data from Scherrenberg, R. L.; Reynaers, H.; Gondard, C.; Booij, M., J. Polym. Sci., Part B: Polym. Phys., 32, 1, 99-109, 1994.] Figure 2.9. Specific surface area of PVC grains vs. conversion. [Data from Zhao, J.-S.; Wang, X.-Q.; Fan, C.-G., Polymer, 32, 14, 2674-2679, 1991.] Figure 2.10. Effect of the degree of hydrolysis of PVAc on grain porosity. [Data from Zimmermann, H., J. Vinyl Additive Technol., 2, 4, 287-294, 1996.] Figure 2.11. Glass transition of PVC vs. its polymer- ization temperature. [Data from Daniels, C. A., J. Vinyl Technol., 1, 4, 212-217, 1979.] 2.4 Polymerization conditions and PVC properties 41 the glass transition temperature of polymer, the higher the gel temperature of plastisol. Polymers with higher glass transition temperatures require more heat to be processed. The above influences are not surprising, considering that molecular weight of poly- mer depends on conversion and polymerization time (Figures 2.13-2.15). Increase in polymerization time under particular conditions increases monomer conversion (Figure 2.13).18 Increase in polymerization temperatures causes an almost linear decrease in the Figure 2.12. Relationship between gel temperature of plastisol and PVC glass transition temperature. [Adapted, by permission from Nakajima, N.; Yavorni- tzky, C. M.; Roche, E. J.; Harrell, E. R., J. Appl. Polym. Sci., 32, 1986.] Figure 2.13. Vinyl chloride conversion vs. polymeriza- tion time. [Data from Scherrenberg, R. L.; Reynaers, H.; Gondard, C.; Booij, M., J. Polym. Sci., Part B: Polym. Phys., 32, 1, 99-109, 1994.] Figure 2.14. PVC molecular weight vs. polymeriza- tion time. [Data from Scherrenberg, R. L.; Reynaers, H.; Gondard, C.; Booij, M., J. Polym. Sci., Part B: Polym. Phys., 32, 1, 99-109, 1994.] Figure 2.15. PVC Mw vs. monomer conversion. [Data from Bao, Y. Z.; Brooks, B. W., J. Appl. Polym. Sci., 85, 1544-1552, 2002.] 44 PVC Manufacture Technology amount of thermal energy required for processing by which they affect thermal stability of polymer. The direct studies on polym- erization temperature and conver- sion on thermal stability of PVC show that both have a strong effect. Increased polymerization temperature increases dehydro- chlorination rate (Figures 2.22 and 2.23). In addition, after initial stages of conversion further polymerization decreases poly- mer stability (Figure 2.23) because of formation various defects as it was discussed in Chapter 1. The data which were ana- lyzed in this section show that PVC stability is influenced by conditions of polymerization. These influences together with the structural defects in PVC chain make behavior of PVC dehydrochlorination rate very complex. 2.5 OPTIMAL OPERATION OF INDUSTRIAL PVC DRYER Several publications are now available on optimal operation of fluidized bed dryer for PVC.27-29 Figure 2.24 shows a schematic diagram of fluidized bed dryer for PVC.27 After polymerization, a mixture containing 30% PVC and 70% water is stripped of some water in a centrifuge, the remaining 20-30% water is removed in dryer shown in Figure 2.24. Temperature and outlet relative humidity are controlled in the dryer.27 Proper control of temperature difference in the dryer allows to save 16% energy required for drying and to increase process output by 22%.28 Numerical model was developed for control of drying operation.29 Model further enhances economical operation of the dryer.29 REFERENCES 1 Endo, K., Prog. Polym. Sci., 27, 2021-2054, 2002. 2 Zimmermann, H., J. Vinyl Additive Technol., 2, 4, 287-294, 1996. 3 Xie, T. Y.; Hamielec, A. E.; Rogestedt, M.; Hjertberg, T., Polymer, 35, 7, 1526-34, 1994. 4 Yamamoto, K.; Maehala, T.; Mitani, K.; Mizutani, Y., J. Appl. Polym. Sci., 51, 4, 755-9, 1994. 5 De Roo, T.; Wieme, J.; Heynderickx, G. J.; Marin, G. B., Polymer, 46, 19, 8340-8354, 2005. 6 Saeki, Y.; Emura, T., Prog. Polym. Sci., 27, 2055-2131, 2002. 7 De Faria, J. M.; Machado, F.; Lima, E. L.; Pinto, J. C., 10th International Symposium on Process Systems Engineering - PSE2009. 8 Etesami, N.; Esfahany, M. N.; Bagheri, R., J. Appl. Polym. Sci., 117, 2506-14, 2010. 9 Tacidelli, A. R.; Alves J. J. N.; Vasconcelos L. G. S.; Brito, R. P., Chem. Eng. Process., 48, 485-925, 2009. 10 Matthews, G., PVC: Production, Properties, and Uses, The Institute of Materials, London, 1996. 11 Percec, V.; Popov, A. V., World Patent, WO2003002621, University of Pennsylvania, Jan. 9, 2003. 12 Simoes, P. N.; Coelho, J. F. J.; Goncalves, P. M. F. O.; Gil, M. H., Eur. Polym. J., 45, 1949-59, 2009. Figure 2.24. Schematic diagram of the continuous PVC fluidized bed drying process. [Adapted, by permission, from de Araujo, A. C B.; Neto, J. J. N.; Shang, H, 10th International Symposium on Pro- cess Systems Engineering - PSE2009. References 45 13 Coelho, J. F. J.; Fonseca, A. C.; Goncalves, P. M. F. O.; Popov, A. V.; Gil, M. H., Polymer, 52, 2998-10, 2011. 14 Zhao, J.-S.; Wang, X.-Q.; Fan, C.-G., Polymer, 32, 14, 2674-2679, 1991. 15 Tornell, B.; Uustalu, J. M.; Jonsson, B., Coll. Polym. Sci., 264, 439-444, 1986. 16 Cebollada, A. F.; Schmidt, M. J.; Farber, J. N.; Capiati, N. J.; Valles, E. M., J. Appl. Polym. Sci., 37, 145-166, 1989. 17 Kiparissides, C.; Achilias, D. S.; Frantzikinakis, C. E., Ind. Eng. Chem. Res., 41, 3097-3109, 2002. 18 Scherrenberg, R. L.; Reynaers, H.; Gondard, C.; Booij, M., J. Polym. Sci., Part B: Polym. Phys., 32, 1, 99-109, 1994. 19 Daniels, C. A., J. Vinyl Technol., 1, 4, 212-217, 1979. 20 Nakajima, N.; Yavornitzky, C. M.; Roche, E. J.; Harrell, E. R., J. Appl. Polym. Sci., 32, 1986. 21 Bao, Y. Z.; Brooks, B. W., J. Appl. Polym. Sci., 85, 1544-1552, 2002. 22 Starnes, W. H., Prog. Polym. Sci., 27, 2133-2170, 2002. 23 Rogestedt, M.; Hjertberg, T., Macromolecules, 26, 1, 60-4, 1993. 24 Tripathi, A.; Tripathi, A. K.; Pillai, P. K. C., J. Mater. Sci., 25, 1947-1951, 1990. 25 Xie, T. Y.; Hamielec, A. E.; Wood, P. E.; Woods, D. R.; Chiantore, O., Polymer, 32, 9, 1696-1702, 1991. 26 Purmova, J.; Pauwels, K. F. D.; Van Zoelen, W.; Vorenkamp, E. J.; Schouten, A. J.; Coote, M. L., Macromolecules, 38, 15, 6352-6366, 2005. 27 de Araujo, A. C B.; Neto, J. J. N.; Shang, H, 10th International Symposium on Process Systems Engineering - PSE2009. 28 de Araujo, A. C. B.; Vasconcelos, L. G. S.; Alves, J. J. N.; Shang, H, Drying Technol., 29, 19-34, 2011. 29 Tacidelli, A. R.; Tavernard, A.; Neto, P.; Brito, R. P.; de Araujo, A. C. B.; Vasconcelos, L. G. S.; Alves, J. J. N., Chem. Eng. Technol., 35, 12, 2107-19, 2012. 3 PVC Morphology This is an important subject for understanding the processing behavior of PVC. Unfortu- nately most monographic sources either omit this discussion or restrict it to a description of PVC grain morphology. Morphology of a polymer is determined by the size of molecule (molecular weight or molecular weight between crosslinking), structure of chain segments (cis-trans isomer- ization, tacticity, and conformation), chain spatial distribution and interaction (entangle- ments, spacing, and crystalline structure), and structure of particles (size, primary particles, agglomerates, skins, pores, etc.). This also will be the sequence of our discussion here, which will be finalized by a short discussion on how morphology influences PVC processing, product structure and properties, and its durability. Few authors have commented on this subject, especially because of the difficulties in either measuring techniques or quantum mechanical calculations. PVC, being a polymer of a low degree of crystallinity, cannot be studied effectively by the usual methods applied to crystalline materials. The major difficulty is in interpretation of results caused by the fact that the morphology is a descriptive feature, not a parameter, because it is influenced by so many variables. In spite of all these difficulties and multidisciplinary character, the available research nevertheless has given us some interesting results, which are com- mented on below. 3.1 MOLECULAR WEIGHT OF POLYMER (CHAIN LENGTH) In an ideal case, we would like to have a straightforward correlation between either the molecular weight of polymer or the distribution of molecular weight and its thermal stabil- ity and other typical properties of PVC. In the early studies, such research was conducted,1 resulting in more confusion than clear explanation of the effect of molecular weight on PVC thermal stability. Table 3.1 shows some of the data.1 Table 3.1. The number-average molecular weight versus PVC dehydrochlorination rate. [Data from Crosato-Arlandi, A.; G. Palma, G.; Peggion, E.; Talamini, G., J. Appl. Polym. Sci., 8, 747, 1964.] Mn Polymerization temp., K Dehydrochlorination rate, 106 dx/dt, s-1 55,000 298 1.66 95,000 273 1.78 142,000 253 1.56 136,000 233 1.61 50 PVC Morphology where: critical end-to-end distance for entanglements Mc critical molecular weight mean-square end-to-end distance for polymer M molecular weight of polymer. The studies of entanglements compared end-to-end distance for entanglements with chain diameter for many polymers. The chain diameter can be obtained from the following equation:8 [3.3] where: S cross-sectional area of polymer obtained from crystallographic data. Critical end-to-end distance increases with increasing molecular diameter. Thicker chains form longer entanglements, most likely because rigidity of chain increases with molecular diameter. Out of 36 polymers tested, PVC has relatively short Mc (6,250; Mc for tested polymers varies from 2,500 to 114,000; and relatively small chain diameter of 0.52 nm; chain diameters for the tested polymers ranged from 0.42 to 11.6 nm).8 Chapter 1 characterizes changes in the composition and structure of PVC at different rates of conversion and molecular weights. In the beginning of this chapter, an example is given (Table 3.1) which shows that molecular weight is not the only parameter controlling its thermal stability. From Chapter 1 it is known that polymerization conditions affect con- centration of labile groups. The effect of these groups on PVC thermal stability will be discussed in depth in Chapter 4. Many other processing results and parameters are affected by molecular weight and its distribution. Figures 2.18 and 2.19 show that melt viscosity is affected by molecular weight and its distribution. Figures 2.20 and 2.21 show the effect of molecular weight and its distribution on fusion time. Gelation of plastisols9 and sheets manufactured by roll milling10 was also enhanced by lower molecular weight PVC. In industrial practice, a technological compromise is the most usual case. Impact resistance, weather stability, and cost were the factors analyzed to develop the process of production of extruded PVC profiles.11 Impact modification was thought to be affected by the type and the concentration of impact modifier and the molecular weight of PVC resin. Cost was mostly determined by the production rate and the amount of impact modifier. Production rate can be optimized by reducing molecular weight of resin but reduction of molecular weight causes decrease in impact strength. These two effects impact the choice of PVC resin to favor medium molecular weight. With such resin it is possible to maintain required impact strength and produce profiles with increased rate and low concentration of impact modifier. Rc R2 0 D S1 2⁄= 3.2 Configuration and conformation 51 3.2 CONFIGURATION AND CONFORMATION The first-order structure of the polymer chain is determined by mers arrangement, config- uration and conformation. The mers arrangement includes head-to-head and head-to-tail structures that have already been discussed. By configuration one should understand the spatial arrangement of atoms, and groups formed by them, in reference to a particular point of a chain-like center of asymmetry or chain backbone. Structural isomerism is the simplest form of configuration. All structural isomers have the same number of atoms (the same empirical formula) but physical properties of isomers are different. Hydrocarbon having 30 atoms may theoretically form 4,111,846,763 isomers. Polymers have a much larger number of monomeric units, therefore, potentially, they may form an infinite num- ber of structural isomers. In practice, the amount of these isomers is restricted by a mech- anism of polymerization which favors a preferred arrangement, but still a polymer is a mixture of a large number of structural isomers. Chain branching is one of the resultant structures of structural isomerism. This subject was extensively investigated in Chapter 1 and implications of structural isomerism on PVC degradation are discussed in the next six chapters in reference to their influence on thermal, UV, irradiance, high energy, mechani- cal, and chemical degradations. Geometrical isomerism is another subset (or type) of configuration. It occurs when double bonds are present in a molecule. It is popularly known as cis-trans isomerism. Cis- trans isomerism has not been studied in PVC, either experimentally or by molecular cal- culations; therefore, in order to explain this phenomenon, we may search data for related substances. We have two possibilities: either to consider the low molecular models having a similar number of conjugated double bonds, such as, for example, carotenoids (but in this case we shall expect the influence of an ionon ring on charge distribution and on con- formational changes); or we may consider a polymer having in its backbone a conjugated double bonds, and here polyacetylene serves as a good example. Based on the results of polyacetylene studies,12-20 we are able to say more about the polyene structure. Polyacety- lene exists in four configurational forms that are well characterized by the following struc- tures: 52 PVC Morphology If the polymer is obtained at considerably low temperature (e.g., 195K), it yields pri- marily cis-(CH)x, which is, however, a thermodynamically unstable form that is subse- quently isomerized into all-trans polyacetylene. The polymer initially obtained by low temperature polymerization exists in the cis-transoid conformation. The isomerization process can occur due to thermal treatment, even at low temperatures; for example, it was discovered that isomerization occurs during IR spectrum measurement, at which the tem- perature was estimated to be around 313K. The activation energy of the isomerization pro- cess is initially very low (67 kJ/mol). As the isomerization process continues, the activation energy increases to 117 kJ/mol. It is thought that the first step in isomerization is defect-induced.18 Considering structural changes, gradual isomerization occurs by rota- tion around single bonds as shown below: cis-transoid trans-cisoid trans-transoid cis-cisoid 3.2 Configuration and conformation 55 Polymer tacticity exhibits three basic structures: • isotactic, in which all the groups are on the same side of a chain (in PVC, all the Cl atoms); • atactic, in which these groups are randomly and irregularly distributed; • syndiotactic, in which groups, e.g., Cl, are regularly distributed on both sides of the chain or its segment in alternative arrangement. Isotactic and syndiotactic segments can be expressed by the following structures: Configurational content can be studied by high resolution NMR, IR and Raman spectroscopies. NMR spectroscopy was used by Sorvik35 who distinguished six bands assigned to syndio-, iso-tactic and combined sequences. The chief drawback of this method is that it can be used only for polymers in solution, while highly syndiotactic poly- mers are insoluble; moreover, they may change their initial properties in solution and under conditions of dissolution. IR spectroscopy is the most frequently applied method for tacticity evaluation. In this method, by comparing the intensities of two bands, one can obtain an index of tactic- ity. Most frequently used ratios are A615/A69036 and A1428/A1434.37 The sample prepa- ration method plays an important role here, as can be seen from data in Table 3.3. There is an evident correlation between both sets of data, but the values differ con- siderably. From this and other studies,38 it is evident that the tacticity index increases when polymerization temperature decreases. Martinez39 determined the PVC dehydrochlorina- tion rate for samples of varying tacticity index obtained by fractionation of polymers of varying polymerization temperature. Figure 3.1 shows the relationship of both values. Martinez39 explains the increase of dehydrochlorination rate, which is parallel to the increase in syndiotacticity, by easier propagation of polyenes along syndiotactic sequences as compared with their propagation along atactic sequences. His explanation is based on Table 3.3. Tacticity index measured by IR in KBr pellets and cast film. (Data from Ref. 36.) Polymerization temperature, K Tacticity index KBr disc cast film 318 1. 25 1. 75 258 1. 50 2.20 233 1. 62 2.40 213 1. 60 2.55 183 2.28 3.30 CH2 CH2H2C CH2 C C C Cl Cl ClH H H CH2 CH2H2C CH2 H C C C H ClCl H Cl isotactic syndiotactic 56 PVC Morphology studies of chain scissions (by ozonolysis) for samples degraded to 0.1 and 0.3% con- version. The difference in the number of chain scissions for a sample with the lowest tacticity index (1.2) is very high, while there is no considerable change in the num- ber of chain scissions for two other samples of tacticity index 1.5 and 1.62. Unfortu- nately, the data as presented cannot be used to estimate if initiation can be correlated with tacticity index because the authors39 did not give the number of scissions for vir- gin polymer. Instead, they compared only the molecular weight of initial polymer with the molecular weight of polymer which was degraded and then subjected to ozonolysis. Therefore, the final result includes the sum of double bonds initially present in the polymer and those newly formed. In further studies by the same research group,40 we can find data showing that the initiation begins more easily from atactic sites than it does at syndiotactic ones. In sum- mary, the syndiotactic segments are less vulnerable to degradation initiation but they pro- mote polyene propagation after it begins. More recent studies41 show that the tacticity of macromolecules has no significant influence on the initial stage of the thermal degradation of PVC (or its initiation). This subject will be further discussed below as a part of analysis of the effect of conformation on various aspects of PVC behavior. Polymer tacticity influences other properties of polymer. It has a major influence on crystallinity. Isotactic and syndiotactic fragments can become crystalline, but the atactic ones are usually completely amorphous.42 The crystallinity occurs due to the folding and entanglements of the long molecular chains with bulky side groups. The bulkier and lon- ger the side groups, the more the entanglements and folds, the higher the crystallinity. Through crystallinity, tacticity affects many other properties of PVC as discussed in a sep- arate section of this chapter. The tacticity has a significant influence on the microstructure caused by steric hin- drance. Isotacticity induces a TGTG conformation (G and T refer to “gauche” and “trans” conformations of chain). This is responsible for a helical shape of the chain. Syndiotactic- ity induces a TTTT conformation which gives an origin to a “zigzag” shape of the chain More explanation on the subject of conformation follows).43 Syndiotactic segments bring chains closer together due to increased hydrogen bonding. Thus, syndiotacticity lowers free volume and mobility of chain more extensively than isotacticity. It is suggested that tacticity causes antiplasticization. Antiplasticization causes an increase in modulus and in tensile strength and a progressive disappearance of β-transition which is opposite to the expected action of plasticizer.44 According to morphological stud- Figure 3.1. PVC dehydrochlorination rate at 453K up to 0.3% conversion vs. tacticity index of polymer. [Data from Martinez, G.; Millan, J., J. Macromol. Sci.-Chem., A12, 489, 1978.] 3.2 Configuration and conformation 57 ies the antiplasticization originates from the occurrence of strong and specific interactions between structures associated with long isotactic sequences and the plasticizer, thereby hindering the motions at the latter structures.44 The last two cases base their conclusions on studies of PVC conformation. So far, discussed configuration refers to the order that is determined by chemical bonds. The con- figuration of a polymer cannot be altered unless chemical bonds are broken and reformed. Conformation refers to the order that arises from the rotation of molecules about the single bonds. Conformation is thus a spatial arrangement of substituent groups that are free to assume different positions in space without breaking any bonds, because of the freedom of bond rotation. Several typical notations are used to describe conformation of polymer chain. Two adjacent structural units form a diad. Meso diad consists of two identically oriented units, and racemo diad of two opposite oriented units: Synonyms and their notations are given below Fischer projections showing structures of these diads. Triads contain three adjacent structural units. Isotactic triad has notation mm or GG and syndiotactic triad has notation rr or TT. Longer microconformers are tetrads, pentads, etc. Thermodynamical stability of macroconformer depends on its structure. The more stable the conformer, the lower Gibbs’ free energy. Isotactic and syndiotactic con- formers have the lowest Gibbs’ free energy and that is why they are thermodynamically most stable. At the same time, each conformer has freedom of assuming different position without breaking any bonds under influence of thermal energy by rotating around single bonds. Figure 3.2 shows that with increasing concentration of syndiotactic triads, the stabil- ity of PVC increases. The concentration of isotactic triads decreases with increasing con- centration of syndiotactic triads.45 This figure shows one fundamental observation from conformation studies which, with some exceptions,41 is the generally accepted common view. The activation energy for isotactic addition is higher than that for syndiotactic addi- tion by approximately 500–600 cal/mol. Thus, the content of syndiotacticity of the PVC obtained from radical polymerization decreases with increasing reaction temperature.46 Cl H H H H H H Cl Cl Cl H H H H H H racemic r trans T syndio- meso m gauche G iso-tactic 60 PVC Morphology 3.4 CHAIN THICKNESS Privalko,54 based on the following equation: [3.4] where: A macromolecular cross-sectional area (A = a2), a chain thickness, vC crystalline specific volume, M molecular weight of monomer unit, d identity period, m0 = M/n, n number of main chain bonds in mer, d0 = d/n, NA Avogadro's number. has calculated chain cross-sectional areas for different polymers. For vinyl polymers this equation can be simplified as follows: A = 1.41m0 [3.5] From these calculations, we have a value for PVC of 2.718 nm, and for polyacety- lene, 1.828 nm. These values may not correspond to values obtained from either crystallo- graphic measurement or conformational calculations. From conformational calculations Conte51 has arrived at a lamellar thickness of 6.0 nm for PVC which seems to be in the range of the above thickness. From the small angle diffraction, the X-ray long spacing for PVC also equals 6.16 nm.55 For cis-polyacetylenes the lamellar thickness is estimated to be in the range of 5-10 nm, while for all-trans conformation it should be as estimated by Privalko.54 3.5 ENTANGLEMENTS Entanglements result from the interpenetra- tion of random-coil chains and are impor- tant in prediction of synthesis of PVC, its rheology, fusion, crazing, UV and thermal stability, shape memory, and fracture prop- erties. Figure 3.4 shows an entangled linear chain network. An entanglement network in an isotropic concentrated melt can be explored by counting the number of bridges and chains intersecting an arbitrary plane, as shown in Figure 3.4. A bridge is a seg- ment of chain which crosses the plane three times (bold section of chain in Figure 3.4). It is sufficiently long to complete one circu- lar loop through the plane. The bridge is A vCM dNA⁄ m0vC d0NA⁄= = Figure 3.4. Entanglements in polymer melt. [Adapted by permission, from Wool, R. P., Macromolecules, 26, 7, 1564-9, 1993.] 3.5 Entanglements 61 capable of transmitting forces across the plane in the melt for a time dependent on the relaxation of this chain segment.56 The number of chain segment crossings per unit area is independent of molecular weight, but the number of chains intersecting the plane decreases with increasing molecular weight. Thus, by varying the molecular weight, we can reach a state where the number of bridges is comparable to the number of chains. The number of bridges per chain is given by the following equation:56 [3.6] where: pc number of bridges, p number of chain segment crossing per unit area, n number of chains intersecting the plane. Critical molecular weight of entanglement is given by the following equation for vinyl polymers:56 [3.7] where: Mc critical entanglement molecular weight, = characteristic ratio, M0 monomer molecular weight. Critical molecular weight of an entanglement determined experimentally for PVC was 11,000 which compares with 10,700 obtained from modelling.56 There is still another parameter used to quantify length of entanglements. It is the critical end-to-end distance for entanglements given by the following equation:57 [3.8] where: the mean-square end-to-end distance for polymer having molecular weight M in the theta solvent Using this equation we can calculate critical end-to-end distance for PVC to be 8.4 nm and chain thickness of 0.52 nm. During polymerization, at high conversions, monomer concentration decreases with conversion, the number of chain entanglement points increases rapidly and the bimolecu- lar-termination rate constant for polymer radicals falls dramatically.58 This causes forma- tion of an increased number of short and long branches and thus structural defects which increase degradation of PVC. More numerous long branches give greater probability of chain entanglement during processing.59 UV degradation rate is also affected by changes in entanglements. When, in the course of degradation process, the molecular weight becomes lower than a critical value Mc, the entanglement network is destroyed, plastic deformation cannot occur and the pc 1 2 -- p n -- 1–= Mc 30C∞M0≈ C∞ Rc Mc 1 2⁄ R2 0 M⁄( ) 1 2⁄ = R 0 62 PVC Morphology toughness decreases sharply by two or three orders of magnitude.60 Cracks propagate rap- idly in the oxidized layer and material is considered severely degraded. The degree of polymer fusion strongly affects entanglements.61 An increase in the degree of fusion causes a higher density of entanglements and secondary crystalline junc- tions. Fusion is a function of temperature, shear, and time. A proper fusion level is neces- sary for good mechanical properties. The fusion process involves melting and reforming of crystallites and molecular entanglement by diffusion.62 Increased entanglements during processing also apply to polymeric additives such as process aids.63 These entanglements with process aids strongly increase the mechanical properties of PVC. Higher elongation at break, higher elasticity, and higher melt strength result from increased entanglements with process aid.63 Polymers having a high chain entanglement density tend to deform by shear yielding (polycarbonate, polyamide, polyester, polyethylene, and polypropylene are examples). Polymers having a low chain entanglement density tend to deform by crazing (examples are polystyrene, poly(methyl methacrylate) and poly(styrene-co-acrylonitrile)). Poly(vinyl chloride) has an intermediate chain entanglement density. Shear yielding is the main toughening mechanism for PVC when toughened by rubber.64 Crazing is the deformation mechanism when the entanglement density of the polymer is below a certain critical value. At higher entanglement densities both crazing and shear yielding take place.65 Entangle- ments strongly influence the dynamic properties and the glassy state properties such as viscosity and toughness.65 The motion of polymer chains is dominated by the presence of direct polymer chain- polymer chain interactions or entanglements in nondilute solutions.66 The increase in spe- cific viscosity is associated with the more numerous entanglements and stronger interac- tions between the polymer chains which leads to a more extensive energy dissipation. The density of entanglement coupling is independent of the solvent nature at higher concentra- tions.66 Shape memory results from crystalline structure, glassy state, entanglements, and/or crosslinking.67 Because PVC has low crystallinity, its shape memory is mostly controlled by actual entanglement density (for example, if entanglement density decreases during exposure to UV also its shape memory fails). To increase shape memory in PVC, cross- linking is used.67 3.6 CRYSTALLINE STRUCTURE Structural studies are usually associated with results obtained by X-ray diffraction. Natta and Corradini68 have measured the unit-cell dimensions as follows: a=1.06 nm, b=0.54 nm and c=0.51 nm. The space group is Pcam, Z=4 and the unit cell is orthorhombic. These data have been confirmed later by Wilkes.55 Each cell contains two PVC chains arranged in planar zig-zag conformation, with the c direction corresponding to two chemical repeat units. Since lengths of syndiotactic sequences are on average no more than 5–6 repeat units, crystallites in the c direction are very thin.69 The amorphous trace for PVC has two broad diffraction peaks. The first one at d=0.36 nm corresponds to van der Waals spacings between groups of atoms. The peak corresponding to d=0.50 nm has been attributed to 3.6 Crystalline structure 65 temperature. It has to be noted that there is no direct knowledge of morphological struc- ture.81 Crystallites are very small and they have an average distance between crystalline regions of 10 nm. In order to form a network, crystallites need to be connected together by tie molecules. According to Summers et al.84 PVC having K-value close to 50 has the average radius of gyration close to 10 nm. For K-values of 58 and 68 they estimated the radius of gyration to be approximately 13 nm and 20 nm, respectively.85 There is still another explanation of the connectivity between crystallites by the fiberlike model which describes the molecular structure of PVC aggregates.86 According to this model, fibrous crystals arise from the crys- tallization of rigid or semirigid chains and there are no fold chains in the crystals. The fiberlike morphology has also been experimentally observed in the PVC gels in good and bad solvents. Size of crystallites seems to depend on conditions of polymerization and sample preparation. It varies in a quite narrow range of 0.7 to 15 nm with spacing between crystal- lites from 0.5 to 20 nm. There is almost full agreement that PVC does not form spherulites because of the small size crystals. But there is an exception in one paper according to which small spherulites are formed (see Figure 3.7).87 This result of research should be treated as a specific case because research material was obtained under specific conditions of catalyzed nucleation on special crystallization cores.87 Figure 3.5. The fringed micelle model of crystalliza- tion. Figure 3.6. Lamellar structure of crystallite. Figure 3.7. PVC spherulite. Adapted, by permission, from Dobreva, A.; Gantcheva, T.; Gut- zow, I., Cryst. Res. Technol., 27, 6, 743-748, 1992. 66 PVC Morphology The theoretical melting point of a per- fectly syndiotactic PVC is estimated to be about 400oC, but such a polymer has not been synthesized.46 The melting points of PVC synthesized by radical polymerization are 102–230oC because of small size crys- tallites and imperfections in the ordered structure. Temperature and orientation are the most influential factors which impact sec- ondary crystallization (the number of crys- tallites, their structure and size, and their spatial distribution and interaction). Figures 3.8 and 3.9 show effect of annealing tem- perature and draw ratio on order factor.88 It is well established that drawing increases tensile strength and elongation of mate- rial.88-91 The presence of crystallites is very important for product performance because they serve as physical crosslinking points which reinforce material and improve its viscoelastic properties. It is therefore very important to notice that the mechanism peculiar to PVC allows for melting of crystallites and their formation on cooling. Melting crystallites improve their processability, and their recovery improves performance characteristics of PVC. Heat of crystallite melting is 74 kJ kg-1.73 Figure 3.10 shows effect of processing temperature on fusion enthalpies of primary and secondary crystallites.92 The gelation Figure 3.10. Effect of processing temperature on fusion enthalpy of primary and secondary crystallites. [Data from Li, T.; Qi, K., J. Appl. Polym. Sci., 63, 13, 1747- 1754, 1997.] Figure 3.8. Effect of annealing temperature on the order factor in PVC. [Data from Liu, Z.; Gilbert, M., Polymer, 28, 1303-1308, 1987.] Figure 3.9. Effect of drawing on the order factor in PVC. [Data from Liu, Z.; Gilbert, M., Polymer, 28, 1303-1308, 1987.] 3.6 Crystalline structure 67 degree was estimated based on the melting enthalpies of primary and secondary crystal- lites according to the following formula:93 [3.10] where: G gelation degree of PVC as estimated by DSC Hp melting enthalpy of primary crystallites Hs melting enthalpy of secondary crystallites Several other factors were found to affect crystallinity. Milling is one such factor, although it is not clear whether crystallinity is increased or decreased by milling. In one study,88 both crystallinity and crystallite size were increased when PVC mixed with ther- mal stabilizers was milled at 165oC.88 Authors attributed this growth to combination of shear effects and intergrain crystallization. In another study,82 the effect of vibromilling and jet milling on gelation was investigated. Apparent density and crystallinity of PVC measured by infrared were decreased during milling. The effect of plasticizer on crystallinity is not fully understood. In one study,94 addi- tion of DOP did not affect crystallinity up to 50% concentration. In another study,77 the effect was dependent on compatibility of plasticizer (more compatible plasticizers decrease crystallinity more extensively). It seems that if there is such effect it is negligible because plasticizer cannot easily penetrate crystalline domains. UV degradation was found to increase crystallinity as measured by infrared.95 It is suggested that crystallinity increased because of degradation of the amorphous phase. Crystallinity affects many properties of PVC. We have already discussed the effect on density, antiplasticization, fusion, tensile, and elongation. Crystallinity also affects dehydrochlorination during UV exposure.96 Since degradation occurs at temperatures below the glass transition temperature, crystalline domains remain unaffected and degra- dation occurs in the amorphous phase. The physical aging of PVC is affected by the degree of crystallinity.97 The effect is not only restricted to the overall percentage of crystallinity, but also to the entire structure of the material. It seems that the quality of crystallites influences the structure of the amor- phous phase (i.e., the extent of disturbed regions, as well as the segmental mobility of undisturbed regions) which is responsible for the aging phenomena.97 G Hp Hp Hs+ ------------------ 100%×= 70 PVC Morphology mer, which is unique feature of PVC polymerization (most poly- mers are readily soluble in their monomers).102 Figure 3.15 shows morphol- ogy of suspension PVC at its final stage. The PVC grains obtained by suspension polymerization exhibit a hierarchical morphology. They are constituted from agglomerates trapped inside a skin made of hydrosoluble polymers. The agglomerates are made out of primary particles. Inside the pri- mary particles, domains are visi- ble.103 This peculiar morphology of PVC was described by Geil104 over 30 years ago and since that time this nomenclature is consis- tently used to describe properties or features of suspension PVC morphology. Figure 3.16 shows a sche- matic diagram of suspension PVC grain.105 It compares very well with features of PVC micrographs in Figure 3.15. The details of morphology are discussed in Table 3.4 which includes modifications made by Allsoop106 to improve the original Gail’s nomenclature. Figure 3.16. Model of PVC grain. [Adapted, by per- mission, from Saeki, Y., Kagaku Keizai, 47, 11, 78–86, 2000.] Figure 3.15. Morphology of PVC grains. [Adapted, by permission from Diego, B.; David, L.; Girard-Reydet, E.; Lucas, J.-M.; Denizart, O., Polym. Intern., 53, 5, 515-522, 2004.] 3.7 Grain morphology 71 Table 3.4 omits one essential compo- nent of suspension PVC structure, which is skin. Skin is formed on the surface of the PVC grains because the surface of the VC droplets is in contact with an aqueous phase and absorbs dispersant dissolved in water at the initial stage of polymerization and graft copolymerization with vinyl chloride and dispersant occurs.107 Daniels mentions a skin-like membrane, thought to originate from the dispersant used in the polymeriza- tion.108 Cuthbertson et al.109 point out that poly(vinyl alcohol) is added to the polymer- ization recipe to prevent coalescence of PVC particles swollen by monomer, and that it functions by coating the particles with a skin. This explanation does not explain presence of skin which engulfs the entire grain as seen in Figure 3.15. Nakajima and Harrell110 offer the opinion that the suspending agent, to which PVC microdomains are attached, forms a skin layer on the grain. Inside of the skin layer are the subparticles and interconnected pores.110 Shah and Poledna111 con- clude that since conventional suspension polymerized PVC has a pericellular membrane (skin), the skin can impede the absorption of plasticizer. Summers102 concludes that early in the polymerization, particles of PVC deposit onto the membrane from both the mono- mer and the water sides forming a skin 0.5-5 μm thick that can be observed on grains after polymerization. There is no question about the presence of such membranes (skins) in many suspen- sion grades since we can see them on SEM photomicrographs (Figure 3.17).112 Based on the above observations, we may conclude that all information points in the same direction: that the skin is formed due to the presence of suspending agent; but it is also clear that we do not know the composition of the skin (whether it is PVC, VC-VAc copolymer, or Table 3.4. Suspension PVC nomenclature. (Based on refs. 104-106) Structural element Size range, μm Description Grain 50-250 Visible constituent of free flowing powder Agglomerate 1-10 Formed by coalescence of primary particles as shown in Figure 3.13 Primary particle 0.5-1 Grown droplet of monomer Domain 0.1-0.2 Formation of inside of primary particle visible only at conver- sions below 2% Microdomain 0.01-0.02 Smallest visible particle made out of approximately 50 chains. It is not visible at conversions higher than 2% Figure 3.17. Membrane (skin) surrounding PVC parti- cle. 480x. [Adapted, by permission, from Cebollada, A. F.; Schmidt, M. J.; Farber, J. N.; Capiati, N. J.; Valles, E. M., J. Appl. Polym. Sci., 37, 145-166, 1989.] 72 PVC Morphology PVAc). It is more likely one of the last two, since skin affects fusion and plasticizer penetration. It should be pointed out that thermal treatment (e.g., drying) also affects PVC morphology. Figure 3.18 shows that the surface of grains was melted.113 Figure 3.19 shows clear particles which were found to cause “fish-eyes”.114 “Fish-eyes” are solid, unmelted looking small spots which may be formed due to various reasons (in the case of PVC, large PVC particles, less porous and/or usu- ally clear particles of all PVC, and contaminations in PVC, which do not melt under test conditions).114 When suspension PVC was dissolved in tetrahydro- furan and dropwise added to methanol to precipitate PVC, its morphology was drastically changed.115 The original grain structure and skin were destroyed.115 Resulting polymer required more time to melt because PVC grains caused self-heating which increased tem- perature faster on mixing.115 At the same time, precipitated polymer had much higher absorption of plasticizer due to its substantially smaller bulk density and more developed structure.115 Due to very complex morphology, it is very difficult to determine particle size distri- bution of suspension PVC. Comparative studies show that application of image analysis in determination of particle size gives a lower error compared with other methods.116-117 Figure 3.18. Suspension PVC particle (A) and particles exposed to 200oC (B). [Adapted, by permission, from Rabek, J. F.; Lucki, J.; Kereszti, H.; Hjertberg, T.; Jun, Q. B., J. Appl. Polym. Sci., 39, 1569-1586, 1990.] Figure 3.19. Photograph of a clear particle responsible for “fish-eye” formation. [Adapted, by permis- sion, from Arslanalp, C.; Erbay, E.; Biligic, T.; Savasci, O. T., Angew. Makromol. Chem., 193, 3321, 99-112, 1991.] References 75 particles were formed with smooth, even grain surface.120 On the right, almost spherical particles were produced with some agglomeration.120 3.7.2 EFFECT OF MORPHOLOGY ON DEGRADATION Trapping of HCl is potentially the most important influence on PVC degradation. Cuthb- ertson et al.109 studied the effect of the concentration of suspending agent (PVAc) on dehy- drochlorination. They concluded that the thicker skin of the 1000 ppm PVAc polymerization traps more hydrogen chloride liberated during the annealing process and this causes autocatalytic processes during the following stages of degradation. The pres- ence of hydrogen chloride leads to longer polyene sequences and this is the most plausible explanation for the differences observed with the two PVAc levels. Kip et al.31 argued that the polyene sequence distribution might not be the same for the skin and core of the PVC grains. Braun et al.121 believe that sample morphology may affect dehydrochlorination rate. Formation of film by pressing affected the formation of polyenes as compared with free powders, which was attributed to changes in morphology and restriction of HCl removal.122 These few contributions do not carry sufficient experimental material for direct proof of influence of morphology on dehydrochlorination. It is due to the fact that it is difficult to design an experiment which will separate this influence from many other influences which cause dehydrochlorination. At the same time, there should be no doubt that HCl dif- fusion process from porous grains of bulk and, especially, suspension polymer meets many obstacles due to the tortuosity of its pathways. This has to be factored in modelling processes of degradation and stabilization. In addition to morphology created by polymerization, we may also expect that struc- tural damage and formation of skin on the surface of particles during their drying will influence dehydrochlorination rates due to the formation of defects and lower diffusion rate of HCl, respectively. We may also consider that the heat required for melting crystallites increases the duration of fusion processes. Also, any other obstacle to fusion, such as, for example, skin on the surface of particles or coating by residual emulsifier, will affect exposure to thermal treatment and it will affect the extent of thermal degradation. Finally, we should also consider that complex morphological structures affect distri- bution of stabilizers because they do not act in a homogeneous manner and decrease ther- mal stability of PVC. REFERENCES 1 Crosato-Arlandi, A.; G. Palma, G.; Peggion, E.; Talamini, G., J. Appl. Polym. Sci., 8, 747, 1964. 2 Hamielec, A. E.; Gomez-Vaillard, R.; Marten, F. L., J. Macromol. Sci.-Chem., A17, 1005, 1982. 3 Graessley, W. W.; Uy, W. C.; Ghandi, A., Ind. Eng. Chem., Fundam., 8, 696, 1969. 4 Pang, S.; Rudin, A., J. Appl. Polym. Sci., 49, 7, 1189-96, 1993. 5 Pepperl, G., J. Vinyl Additive Technol., 6, 2, 88-92, 2000. 6 Pepperl, G., J. Vinyl Additive Technol., 8, 3, 209-213, 2002. 7 Flores, R.; Perez, J.; Cassagnau, P.; Michel, A.; Cavaille, J. Y., Polymer, 35, 13, 2800-7, 1994. 8 Zang, Y.-H.; Carreau, P. J., J. Appl. Polym. Sci., 42, 1965-1968, 1991. 9 Marcilla, A.; Garcia, J. C.; Beltran, M., Eur. Polym. J., 33, 5, 753-759, 1997. 10 Fujiyama, M.; Kondou, M., J. Appl. Polym. Sci., 92, 3, 1915-1938, 2004. 76 PVC Morphology 11 Tseng, A. A.; Borrosky, M. A.; O'Connor, S. M.; Knotts, J. J., Adv. Polym. Technol., 10, 3, 205-218, 1990. 12 Terlemezyan, L.; Mihailov, M., Makromol. Chem., Rapid Commun., 3, 613, 1982. 13 Yamabe, T.; Akagi, K.; Shirakawa, H.; Ohzeki, K.; Fukui, K, Chem. Ser., 17, 157, 1981. 14 Bernier, P.; Schue, F.; Sledz, J.; Rolland, M.; Giral, L.; Chem. Scr., 17, 151, 1981. 15 Yamabe, T.; Akagi, K.; Ohzeki, K.; Fukui, K., J. Phys. Chem. Solids, 43, 577, 1982. 16 Aldissi, M.; Rolland, M.; Scue, F., Phys. Stat. Sol., 69, 733, 1982. 17 Montaner, A.; Galtier, M.; Benoit, C.; Aldissi, M., Solid State Commun., 39, 99, 1981. 18 Kuzmany, H.; Imhoff, E. A.; Fitchen, D. B.; Sarhangi, A., Mol. Cryst. Liq. Cryst., 77, 197, 1981. 19 Yamabe, T.; Matsui, T.; Akagi, K.; Ohzeki, K.; Shirakawa, H.; Mol. Cryst. Liq. Cryst., 83, 125, 1982. 20 Simonescu, C. I.; Percec, V., Prog. Polym. Sci., 8, 133, 1982. 21 Tran, V. H.; Garrigues, C.; Nguyen, T. P.; Molinie, P., Polym. Deg. Stab., 42, 2, 189-203, 1993. 22 Tran, V. H.; Guyot, A.; Nguyen, T. P.; Molinie, P., Polym. Deg. Stab., 49, 3, 331-7, 1995. 23 Tran, V. H.; Guyot, A.; Nguyen, T. P.; Molinie, P., Polym. Deg. Stab., 37, 209-216, 1992. 24 Tran, V. H.; Nguyen, T. P.; Molinie, P., Polym. Deg. Stab., 44, 2, 151-61, 1994. 25 Tran, V. H.; Nguyen, T. P.; Molinie, P., Polym. Deg. Stab., 53, 3, 279-288, 1996. 26 Chatterjee, N.; Basu, S.; Palit, S. K.; Maiti, M. M., J. Polym. Sci., Part A: Polym. Chem., 32, 7, 1225-36, 1994. 27 Fisch, M. H.; Bacaloglu, R., J. Vinyl Additive Technol., 5, 4, 205-217, 1999. 28 Fisch, M. H.; Bacaloglu, R.; Biesiada, K.; Brecker, L. R., J. Vinyl Additive Technol., 5, 1, 45-51, 1999. 29 Karpfen, A.; Hoel1er, R., Solid State Commun., 37, 179, 1981. 30 Wypych, J., Polyvinyl chloride degradation; Polymer Science Library 3; Jenkins, A. D., Ed., Elsevier, Amsterdam, 59-159, 1985. 31 Kip, B. J.; van Aaken, S. M.; Meier, R. J.; Williams, K. P. J.; Gerrard, D. L., Macromolecules, 25, 4290-4296, 1992. 32 Lemaire, J.; Siampiringue, N.; Chaigneau, R.; Delprat, P.; Parmeland, G.; Dabin, P.; Spriet, C., J. Vinyl Additive Technol., 6, 2, 69-79, 2000. 33 Kershaw, J. R.; Edmond, L.; Rizzardo, E., Polymer, 30, 360-363, 1989. 34 Lemaire, J.; Siampiringue, N.; Girod, N.; Delprat, P.; Parmeland, G.; MacDonald, N.; Spriet, C., J. Vinyl Additive Technol., 9, 2, 54-60, 2003. 35 Sorvik, E. M., J. Appl. Polym. Sci., 21, 2769, 1977. 36 Carrega, M. E., Pure Appl. Chem., 49, 569, 1977. 37 Pham, Q. T.; Millan, J.; Madruga, E. L.; Makromol. Chem., 175, 945, 1974. 38 Millan, J.; Mijangos, C.; Martinez, J.; Berticat, B.; Chauchard, J., Eur. Polym. J., 15, 615, 1979. 39 Martinez, G.; Millan, J., J. Macromol. Sci.-Chem., A12, 489, 1978. 40 Millan, J.; Martinez, G; Mijangos, C, J. Polym. Sci., Polym. Chem. Ed., 18, 505, 1980. 41 Troitskii, B. B.; Yakhnov, A. S.; Novikova, M. A.; Ganyukhina, T. G.; Denisova, V. N., Eur. Polym. J., 33, 4, 505-511, 1997. 42 Tseng, A. A.; Borrosky, M. A.; O'Connor, S. M.; Knotts, J. J., Adv. Polym. Technol., 10, 3, 205-218, 1990. 43 Vella, N.; Toureille, A.; Guarrotxena, N.; Millan, J. L., IEEE Trans. Dielectrics Electrical Insulation, 7, 3, 329-333, 2000. 44 Tiemblo, P.; Martinez, G.; Gomez-Elvira, J. M.; Millan, J., Polym. Bull., 32, 3, 353-9, 1994. 45 Castillo, F.; Martinez, G.; Sastre, R.; Millan, J.; Bellenger, V.; Gupta, B. D.; Verdu, J., Polym. Deg. Stab., 27, 1-11, 1990. 46 Endo, K., Prog. Polym. Sci., 27, 2021-2054, 2002. 47 Martinez, G.; Mijangos, C.; Millan, J., J. Appl. Polym. Sci., 28, 33, 1983. 48 Simon, P.; Cernay, P.; Valko, L., Eur. Polym. J., 25, 3, 245-249, 1989. 49 Behnisch, J.; Zimmermann, H., Makromol. Chem., 190, 2347-2352, 1989. 50 Conte, C.; D'Ilario, L.; Pavel, N. V.; Giglio, E., J. Polym. Sci., Polym. Phys. Ed., 17, 753, 1979. 51 Nakajima, A; Hayashi, S., Kolloid Z. Z. Polym., 229, 12, 1969. 52 Warrier, A. V. R.; Krimm, S., Macromolecules, 3, 709, 1970. 53 Braun, D., Pure Appl. Chem., 53, 549, 1981. 54 Privalko, V. P., Macromolecules, 13, 370, 1980. 55 Wilkes, C. E.; Folt, V. L.; Krimm, S., Macromolecules, 6, 235, 1973. 56 Wool, R. P., Macromolecules, 26, 7, 1564-9, 1993. 57 Zang, Y.-H.; Carreau, P. J., J. Appl. Polym. Sci., 42, 1965-1968, 1991. 58 Xie, T. Y.; Hamielec, A. E.; Wood, P. E.; Woods, D. R.; Chiantore, O., Polymer, 32, 9, 1696-1702, 1991. 59 Dashora, P.; Gupta, G., Polymer, 37, 2, 231-4, 1996. 60 Anton-Prinet, C.; Mur, G.; Gay, M.; Audouin, L.; Verdu, J., J. Mater. Sci., 34, 2, 379-384, 1999. References 77 61 Covas, J. A.; Gilbert, M., Polym. Eng. Sci., 32, 11, 743-750, 1992. 62 Zdilar, A. M.; Tanner, R. I., J. Rheology, 38, 4, 909-20, 1994. 63 Disson, J.-P; Girois, S., J. Vinyl Additive Technol., 9, 4, 177-187, 2003. 64 Liu, Z. H.; Zhang, X. D.; Zhu, X. G.; Qi, Z. N.; Wang, F. S.; Li, R. K. Y.; Choy, C. L., Polymer, 39, 21, 5027-5033, 1998. 65 Zhou, C.; Bao, X. Y.; Tan, Z. Y.; Sun, S. L.; Ao, Y. H.; Yang, H. D.; Zhang, H. X., J. Polym. Sci., Part B: Polym. Phys., 44, 4, 687-695, 2005. 66 Wildman, G. C.; Hsu, S. J. R., J. Appl. Polym. Sci., 30, 4385-4408, 1985. 67 Jeong, H. M; Song, J. H.; Lee, S. Y.; Kim, B. K., J. Mater. Sci., 36, 22, 5457-5463, 2001. 68 Natta, G.; Bassi, I. W.; Corradini, P.; Atti Accad. Naz. Lincei Cl. Sci. Fis. Mat. Nat. Rend., 31, 17, 1961. 69 Gilbert, M.; Ross, D. H.; Kim, H. C., J. Appl. Polym. Sci., 104, 1, 528-535, 2007. 70 Guerrero, S. J.; Meader, D.; Keller, A., J. Macromol. Sci.-Phys., B20, 185, 1981. 71 Bartczak, T. J.; Galdecki, Z.; Wypych, A.; Wypych, J., unpublished studies. 72 Guerrero, S. J.; Veloso, H.; Radon, E., Polymer, 31, 1615-1622, 1990. 73 Komoroski, R. A.; Lehr, M. H.; Goldstein, J. H.; Long, R. C., Macromolecules, 25, 3381-3384, 1992. 74 Chartoff, R. P.; Lo, T. S. K.; Harrell, E. R.; Roe, R. J., J. Macromol. Sci.-Phys., B20, 287, 1981. 75 Ulutan, S., J. Appl. Polym. Sci., 90, 14, 3994-3999, 2003. 76 Tran, V. H.; Nguyen, T. P.; Molinie, P., Polym. Deg. Stab., 53, 3, 279-288, 1996. 77 Beltran, M.; Garcia, J. C.; Marcilla, A., Eur. Polym. J., 33, 4, 453-462, 1997. 78 Chen, R.; Tyler, D. R., Macromolecules, 37, 5430-5436, 2004. 79 Voyiatzis, G. A.; Andrikopoulos, K. S.; Papatheodorou, G. N.; Kamitsos, E. I.; Chryssikos, G. D.; Kapoutsis, J. A.; Anastasiadis, S. H.; Fytas, G., Macromolecules, 33, 15, 5613-5623, 2000. 80 Gilbert, M.; Liu, Z.; Hitt, D. J., Polym. Eng. Sci., 37, 11, 1858-1864, 1997. 81 Flores, R.; Perez, J.; Cassagnau, P.; Michel, A.; Cavaille, J. Y., Polymer, 35, 13, 2800-7, 1994. 82 Xu, X.; Guo, S.; Wang, Z., J. Appl. Polym. Sci., 64, 12, 2273-2281, 1997. 83 Fiaz, M.; Gilbert, M., Adv. Polym. Technol., 17, 1, 37-51, 1998. 84 Summers, J. W.; Rabinovitch, E. B.; Booth, P. C., ANTEC Tech. Papers, 31, 1027, 1985. 85 Hinrichsen, E. L.; Thorsteinsen, P., J. Vinyl Additive Technol., 2, 1, 18-29, 1996. 86 Li, L.; Uchida, H.; Aoki, Y.; Yao, M. L., Macromolecules, 30, 25, 7842-7848, 1997. 87 Dobreva, A.; Gantcheva, T.; Gutzow, I., Cryst. Res. Technol., 27, 6, 743-748, 1992. 88 Liu, Z.; Gilbert, M., Polymer, 28, 1303-1308, 1987. 89 Fras, I.; Boudeulle, M.; Cassagnau, P.; Michel, A., Polymer, 39, 20, 4773-4783, 1998. 90 Tseng, A. A.; Borrosky, M. A.; O'Connor, S. M.; Knotts, J. J., Adv. Polym. Technol., 10, 3, 205-218, 1990. 91 Tsitsilianis, C., J. Appl. Polym. Sci., 43, 835-838, 1991. 92 Li, T.; Qi, K., J. Appl. Polym. Sci., 63, 13, 1747-1754, 1997. 93 Piszczek, K.; Sterzynski, T.; Lukaszewicz, E., Polym. Testing, 22, 115-119, 2003. 94 Zorn, R.; Monkenbusch, M.; Richter, D.; Alegria, A.; Colmenero, J.; Farago, B., J. Chem. Phys., 125, 15, 154904/1-154904/13, 2006. 95 Denizligil, S.; Schnabel, W., Angew. Makromol. Chem., 229, 73-92, 1995. 96 Behnisch, J.; Zimmermann, H., Makromol. Chem., 190, 2347-2352, 1989. 97 Tsitsilianis, C.; Tsapatsis, M.; Economu, C., Polymer, 60, 1861-1866, 1989. 98 Allsopp, M. W., Pure Appl. Chem., 53, 449–65, 1981. 99 Saeki, Y.; Emura, T., Prog. Polym. Sci., 27, 2055-2131, 2002. 100 Zimmermann, H., J. Vinyl Additive Technol., 2, 287–94, 1996. 101 Zerfa, M.; Brooks, B. W., J. Appl. Polym. Sci., 65, 1, 127-134, 1997. 102 Summers, J. W., J. Vinyl Additive Technol., 3, 2, 130-139, 1997. 103 Diego, B.; David, L.; Girard-Reydet, E.; Lucas, J.-M.; Denizart, O., Polym. Intern., 53, 5, 515-522, 2004. 104 Geil, P. H., J. Macromol. Sci., Phys., B14, 171, 1997. 105 Saeki, Y., Kagaku Keizai, 47, 11, 78–86, 2000. 106 Allsopp, M. W., Pure Appl. Chem., 53, 449–65, 1981. 107 Zhao, J.-S.; Wang, X.-Q.; Fan, C.-G., Polymer, 32, 14, 2674-2679, 1991. 108 Daniels, C. A., J. Vinyl Technol., 1, 4, 212-217, 1979. 109 Cuthbertson, M. J.; Bowley, H. J.; Gerrard, D. L.; Maddams, W. F.; Shapiro, J. S., Makromol. Chem., 188, 2801-2809, 1987. 110 Nakajima, N.; Harrell, E. R., J. Colloid Interface Sci., 238, 1, 105-115, 2001. 111 Shah, A. C.; Poledna, D. J., J. Vinyl Additive Technol., 8, 3, 214-221, 2002. 112 Cebollada, A. F.; Schmidt, M. J.; Farber, J. N.; Capiati, N. J.; Valles, E. M., J. Appl. Polym. Sci., 37, 145-166, 1989. 80 Principles of Thermal Degradation much lower than that of secondary chlo- rine.7 Also, bond energy of C-Cl attached to the tertiary carbon atom is lower by 10 kcal/ mol compared with chlorines attached to the secondary carbon atoms.8 The synperiplanar four-center transi- tion state is forbidden by Woodward-Hoff- mann rules. Consequently, in order for it to occur, a non-simultaneous mechanism requiring strong polarization of the C-Cl bond must prevail. Experimental data and theoretical calculations support the pre- dicted polarization. At the same time, this slows down reaction rate.4 Highly polar sol- vents ionize tertiary chloroalkanes, chang- ing the mechanism of elimination to a process which requires lower activation energy.4 The effect of tertiary carbon atoms on the initiation of dehydrochlorination was con- firmed by experiment in which tertiary chlorines were eliminated by the reaction with tri- alkyl aluminum, as follows:9 PVC reacted with trimethyl aluminum had a reduced rate of dehydrochlorination.9 Figure 4.1 shows the effect of concentration of tertiary chlorines in PVC on its dehy- drochlorination rate.2 Using linear regression, the following equation was obtained from the data in Figure 4.1:2 [4.1] where: dHCl/dt dehydrochlorination rate, [Cl]T tertiary chlorines/1000 VMC. This data shows that there is little doubt that tertiary chlorine atoms are capable of initiation of PVC dehydrochlorination. Bacaloglu et al.3 pointed out that initial rate of decomposition of PVC is higher because of the presence of tertiary chlorines and it later slows down when they are exhausted. This seems to coincide with observations from PVC degradation testing during which initial emission of HCl is indeed more extensive and it C Cl CH2 3 + Me3Al 3 C CH2 Me + AlCl3 dHCl dt ------------- 3.3 12.4 Cl[ ]T 10 3– min 1–+= Figure 4.1. Effect of tertiary chlorines on PVC dehy- drochlorination rate. [Adapted, by permission, from Xie, T. Y.; Hamielec, A. E.; Rogestedt, M.; Hjertberg, T., Polymer, 35, 7, 1526-34, 1994.] 4.1 The reasons for polymer instability 81 eventually slows down to produce a constant degradation rate during the phase of polyene extension. This initial increase may also be caused by the fact that HCl is removed from the sample by diffusion in the beginning but later, when its concentration rises, the sample becomes porous, which suddenly increases HCl output. 4.1.1.3 Unstaturations Unsaturations alone do not increase the rate of initiation but there is common agreement that internal double bonds, due to the presence of chloroallyl groups, may serve as active initiation centers of dehydrochlorination. Chlorallyl groups and tertiary chlorines are the most important defects which initiate thermal degradation.1 Unsaturations at the chain ends have no effect on polymer thermal stability.10-11 Minsker et al.10 claimed that chloroallyl groups rapidly oxidize to ketochloroallyl groups and that they contribute to chain reaction initiation, but this has not been accepted and it is not considered valid. Chloroallylic bond between chlorine and carbon has energy of 58 kcal/mol compared with the same bond in tertiary chlorine having energy of 67 kcal/mol.8 The constant of dehydrochlorination is 1.17 x 10-3 for allylic chlorine and 1.75 x 10-3 for tertiary chlorine.5 This shows that chloroallyl groups are better initiators of thermal degradation but at the same time they are less numerous. Figure 4.2 shows that there is no correlation between chloroallyl concentration and the rate of thermal degradation. This indicates that there are more influential factors (e.g., concentration of tertiary chlorines, see Figure 4.1) which affect degradation rate. Figure 4.3 shows otherwise, which may suggest that the influence of allylic groups is determined by the conditions of synthesis, especially taking into con- sideration that the formation of internal double bonds is independent of conversion, whereas concentration of tertiary chlorines rapidly increases above 70% conversion.2 This Figure 4.2. Effect of concentration of internal double bonds on dehydrochlorination rate. [Adapted, by per- mission, from Xie, T. Y.; Hamielec, A. E.; Rogestedt, M.; Hjertberg, T., Polymer, 35, 7, 1526-34, 1994.] Figure 4.3. Effect of concentration of chloroallyl groups on dehydrochlorination rate. [A − data from Dean, L.; Dafei, Z.; Deren, Z., Polym. Deg. Stab., 22, 31-41, 1988; B − data from Xu, P.; Zhou, D.; Zhao, D., Eur. Polym. J., 25, 6, 575-579, 1989.] 82 Principles of Thermal Degradation permits combinations with exaggerated influence of chloroallyl groups such as presented in Figure 4.3, but the stability of the general purpose, commercial PVC, is more influ- enced by tertiary chlorine than by the chloroallyl group. It should be noted that analytical methods and methods of defect calculation were different in all three cases reported. Fisch and Bacaloglu,14 based on studies of low molecular weight models, proposed a six atom transition state which lowers energy of dehydrochlorination of allylic chlorine: Unlike the four-member state which also explains the mechanism of dehydrochlori- nation, this state is not forbidden by Woodward-Hoffmann rule and reaction occurs faster. Geometric isomerism (see Chapter 3) is important in explanation of reaction mecha- nism and rate. Chemical structures below show explicitly why cis-configuration is reac- tive (H and Cl are in proximity) and why trans-configuration causes interruption of the dehydrochlorination chain of reactions:3 Cis-allylic chlorine is eliminated much faster15 (about 50-200 times faster than trans).4 Bacaloglu et al.16 concludes that the chain of dehydrochlorination stops when it meets trans configuration on its way. Trans form has higher activation enthalpy and it is not considered to be an initiator of dehydrochlorination.3 It is also possible that trans-con- figuration isomerizes in solution to cis-configuration.4 It should be pointed out that chloroallylic group is also very essential during propa- gation of dehydrochlorination and cis-trans isomerism may be a possible explanation of growth termination of polyene chains. When polymer is exposed to the thermal processing as in the case of drying or repro- cessing of waste polymer, large number of conjugated double bonds are formed in such polymer. This exposure and formation of conjugated double bonds dramatically decreases thermal stability as was experimentally confirmed.17 When polymer lost 0.4 mol% of HCl due to the exposure for 30 min at 200oC, its onset decomposition temperature was 50- 60oC lower than that of virgin resin.17 C H C H C H C Cl C H C H CC Cl C H C H CC H HCl H ClClH trans stablecis reactive 4.1 The reasons for polymer instability 85 There are almost no data comparing different initiators in terms of their effect on thermal stability. One mechanism is proposed according to which initiator rests residing at chain ends may dissociate:40 the radical formed becomes mobile and it can readily abstract hydrogen causing formation of macroradical and thus increasing degradation rate. In recapitulation, free initiator molecules and initiator rests can influence dehydro- chlorination rate and its initiation but their performance and concentration are easily con- trolled, and most likely the initiators do not pose problems under normal conditions of PVC processing. 4.1.2.2 Transfer agent rests Monomer is commonly used in practice as a chain transfer agent and does not influence PVC thermal degradation because it is a regular building block. In some cases,41-43 chain- transfer agents are added to polymerization mixture mainly to control (reduce) molecular weight. Depending on their chemical composition, they may have no negative effect on dehydrochlorination (mercaptoethanol),41 slightly increase degradation rate (dodecylmer- captan),42 or rapidly increase degradation (formaldehyde).43 Considering that these are mostly experimental products, a user of commercial products is typically not affected by their presence. 4.1.2.3 Polymerization additives Suspension agents and emulsifiers are common additives found in polymerization recipes. Emulsifiers are known to modify flow properties,44 affect viscosity changes over time,45 affect results of K-number measurements (lower K-value by about 2 units),46 and improve antistatic properties.47 Suspending agents influence porosity and become part of the peri- cellular membrane, a skin encompassing particles of suspension PVC.48 There is no data or information on direct involvement of these additives in support- ing degradative changes. It could be predicted that, since viscosity is influenced by tem- perature, any influence on viscosity will be corrected by temperature change; therefore these additives can be linked to indirect involvement in degradative processes. Also, for- mation of skin is related to problems with HCl diffusion as well as higher energy require- ment to homogenize materials (fusion or gelation) before forming. This is also an example of indirect influence on degradative processes. 4.1.3 METAL DERIVATIVES 4.1.3.1 Metal chlorides Metal chlorides discussed here are mostly products of reaction of thermal stabilizers with HCl. They remain in the material and they may influence the material’s stability. Calcium and barium chlorides Calcium and barium are the next most electropositive metals after the alkali metals. Like the alkali metals, they form stable salts of which anions are the chief points of interest. The cations in salts have little deforming power, and hence they stabilize anions. RCH2CHClCH2CHCl R + CH2CHClCH2CHCl 86 Principles of Thermal Degradation Figure 4.4. shows the effect of calcium stearate concentration on the induction period of PVC degradation.49 The induction period is determined experimentally as the begin- ning of emission of HCl from PVC degraded under the nitrogen flow. Induction period is calculated from the equation derived by this author:50 [4.2] where: n induction period, PA HCl absorption capacity of stabilizer (determined by conductometric titration),50 v0 dehydrochlorination rate of PVC without thermal stabilizer. Both experimental and calculated val- ues give linear relationships but calculated results predict that degradation should have occurred faster. Equation 4.2 accounts for reaction of stabilizer but does not consider that unreacted HCl autocatalyzes PVC dehydrochlorination which is accounted for in experimental results. Figure 4.4 also shows that CaCl2 (and none of intermediate products of reaction with HCl which could be formed) does not catalyze PVC degrada- tion. Other metal chlorides that are thought not to influence the PVC dehydrochlorina- tion rate are Na, K, Ba, Sr, Mg, Pb.51-54 More recent publications either mention that CaCl2 does not accelerate degrada- tion,55 or that calcium is more ionic and therefore does not increase degradation rate of PVC,56 or that calcium chloride is not Lewis acid which explains why it does not catalyze dehydrochlorination.57 The above reported research does not answer questions regarding the mechanism of reaction with HCl, which can be schematically written as follows: This scheme of reactions generates one essential question regarding the sequence of events. There are two possible scenarios: n PA v0 -------= Ca(OOR)2 + HCl CaOORCl + HCl ROO = fatty acid rest CaOORCl + ROOH CaCl2 + ROOH Figure 4.4. Induction period of PVC degradation vs. concentration of calcium stearate. [Data from Wypych, G., Selected properties of PVC plastisols, Wroclaw Technical University Press, 1977.] 4.1 The reasons for polymer instability 87 • molecules of calcium stabilizer (calcium stearate or salt of any other fatty acid) react first with one molecule of HCl forming partially reacted stabilizer, CaOORCl • molecules of calcium stabilizer react with two molecules of HCl forming final chloride, CaCl2 Several studies were performed by this author,58 including elemental analysis, IR spectroscopy, and x-ray analysis of samples of barium and calcium stearates reacted in moisture-free acetone with 50% of stoichiometric amount of HCl required to completely react with stabilizer. Elemental analysis of the product of reaction showed that its concen- tration of carbon, hydrogen, oxygen, chlorine, and calcium was very similar to concentra- tion of these elements in CaClSt (BaClSt) and very different than concentration of these elements in calcium (or barium) stearates.58 IR showed an additional peak around 1660 cm-1 which could not be attributed to stearic acid and was not present in the initial stear- ates.58 By x-ray analysis it was possible to directly monitor concentration of the initial cal- cium stearate which was still present after addition of 35.6% stoichiometric quantity of HCl. Only after 81.3% of stoichiometric HCl was added, the peak for calcium stearate was not found.58 These results show that: • partially reacted calcium or barium stabilizers (CaOORCl or BaOORCl) are sta- ble and can exist in this form without converting into final chlorides, • considering that PVC dehydrochlorination rate is not increased by the presence of these stabilizers (Figure 4.4), it is safe to conclude that these partially reacted with HCl products do not catalyze PVC dehydrochlorination. The results of the above studies were neither confirmed nor rejected by more recent research. We may summarize properties of calcium, barium and the like, as follows: • metal chlorides and partially reacted with HCl products do not catalyze dehydro- chlorination, • by reacting with HCl, these stabilizers eliminate its catalytic effect, • there is no emission of HCl until they are not exhausted, • products partially reacted with one HCl molecule are stable in the PVC environ- ment and they undergo further reaction with HCl only after the first step of the above reactions was completed. Cadmium and zinc chlorides Figure 4.5 shows data obtained in a similar way as in Figure 4.4 but for zinc stearate. There is substantial difference between the two figures. In the case of calcium stearate experimental data form a linear relationship and their values are slightly lower than the values predicted from the equation. In Figure 4.5 experimental data follow a non-linear relationship and their values are lower than predicted from equation 4.2. There may only be one reason for this difference, namely that products of zinc stearate reaction with HCl have a catalytic influence on PVC dehydrochlorination rate. From the character of the experimental curve we may also conclude that this catalytic activity intensifies when con- centration of stabilizer increases. Numerous research works were published on the effect of reaction products of zinc and cadmium carboxylates on the PVC dehydrochlorination rate. Czako59 found that reac- 90 Principles of Thermal Degradation zinc products of reaction. Gerrard et al.86 were not able to repeat the experiment by Owen and Msayib.84 They obtained similar spectra but of substantially different character. Vymazal put forward a concept which is explained by the following chemical reac- tion scheme:87 Cadmium is known to form complexes having a coordination number of 6. The com- plex formed dissociates on a polymer side with subtraction of HCl molecule which remains associated with cadmium stabilizer. This model of reaction has substantial weak- ness: • it assumes randomness of initiation, • it implicates stabilizer in subtraction of HCl rather than its expected role of bring- ing stability to polymer, • it bypasses a concept of defect structures which are considered responsible for initiation of dehydrochlorination. In summary, the right principle was wrongly applied. It was postulated by this author in the previous edition of this book88 that metal carboxylate has molecular symmetry which stabilizes formation of complexes as proposed by Vymazal but this may be affected by polar solvents affecting stability of partially reacted with HCl carboxylate. In the case of zinc carboxylate it is less relevant, considering that it is not very stable in PVC compo- sition (see the above discussion). Exchange of chlorine between different stabilizing forms is the most influential reac- tion for PVC stabilization and its mechanisms:49,50,57,58,84,89-92 These two reactions and many other combinations of different reaction products or metals (e.g., Cd and Ba) constitute this mechanism aiming at lowering catalytic activity of metal salts. The principle is to keep a low concentration of chloride of metal which is capable of forming Lewis acids. The concentration of the zinc compound in the stabilizer form must be sufficient for the substitution of active chlorine atoms at a proper speed, but must not overcome the capacity of calcium stearate to convert zinc chloride into zinc stearate. This reaction forms one of the principles of stabilization and it will be further dis- cussed in Chapter 11 in the section explaining mechanisms of PVC thermal stabilization. Cl CH2CHCH2CH Cl O O O O C C RR CH2CHCH2CH Cl Cd Cl ZnOORCl + CaOORCl Zn(OOR)2 + CaCl2 ZnCl2 + Ca(OOR)2 Zn(OOR)2 + CaCl2 4.1 The reasons for polymer instability 91 Tin chlorides Tin stabilizers react with HCl according to the following sequence:93 Four chlorine-containing compounds can potentially be formed. These compounds differ in Lewis acidity. SnCl4 and RSnCl3 are strong Lewis acids and they affect dehydro- chlorination rate similar to zinc and cadmium chloride.22 Similar to zinc chloride, they cause catastrophic failure when they accumulate in sufficient concentration.94 RSnCl3 was even found to form green-blue compounds with polyenes in the same way as zinc chlo- ride.86 Lewis acidity also depends on the character of group R. For example, diphenyltin dichloride has higher Lewis acidity than dioctyltin dichloride because the phenyl group has electron withdrawing capability.82 Dichlorides (R2SnCl2) should have theoretically lower stabilizing activity than monochlorides (RSnCl3) from the point of view of Lewis acidity but they are better stabilizers because of formation of stabilizing coordination complexes as follows:82 Chlorine exchange reaction is also central to the performance of tin stabilizers and their mechanism of action:22,95-97 These and many other combinations of chlorine exchange reactions occur, which help to reduce catalytic activity of products of reaction with HCl. R2SnY2 HCl R2SnClY HCl R2SnCl2 HCl RSnCl3 HCl SnCl4 Cl Cl Sn Cl Cl Sn Cl Cl Sn R2SnY2 + R2SnCl2 2RSnY3 + RSnCl3 2R2SnClY 3RSnClY2 92 Principles of Thermal Degradation Also, exchange reaction with calcium stabilizers occurs:97 Aluminum chloride Catalytic dehydrochlorination of PVC occurs in the presence of aluminum chloride (1:1 molar ratio of PVC and AlCl3), leading to completely dehydrochlorinated PVC − poly- acetylene.98 It should be noted that crosslinking occurs during this process. PVC heated with smaller concentrations of AlCl3 (0.05 to 4%) dehydrochlorinates, forming shorter polyenes than without the presence of AlCl3.99 4.1.3.2 Copper and its oxide Studies were conducted to understand the processes of incineration of wire and cable which contain, in addition to PVC, copper, its oxide, and then chloride as a product of reaction with HCl.100 Thermal degradation of PVC in the presence of copper and its com- pounds occurs at higher temperatures with major changes in the concentrations of the components of the gaseous products. These changes are caused by extensive crosslinking in the presence of copper and its oxide. Also, yields of aromatic hydrocarbons, such as benzene, are substantially decreased and yields of aliphatics are increased. Total chlorine emissions are reduced by 40% in air and 20% in nitrogen and benzene emissions are reduced by at least 90% in air and nitrogen.100 4.1.4 HYDROGEN CHLORIDE In the early works on PVC dehydrochlorination, Arlman101 ruled out the possibility of an autocatalytic effect of HCl on PVC degradation kinetics. Around the same time, other sci- entists came to the conclusion that only in an inert atmosphere does HCl not affect the dehydrochlorination process.102-105 These early studies developed a lot of interest in HCl catalysis, and from the early sixties, after work done by Rieche,106 it become evident that HCl plays an important role in the PVC degradation process. During this period an impressive account of HCl’s autocatalytic effect was collected.49,107-124 It is interesting to note a common agreement considering that it is not possible to obtain direct evidence that HCl actually causes increase in the dehydrochlorination rate because HCl is always pres- ent in unknown concentration in the reaction environment. Endo and Emori125 conducted an interesting experiment to determine influence of HCl on dehydrochlorination rate. Their dehydrochlorination experiments were conducted under a nitrogen blanket and in water heated under high pressure (19.6 MPa). In water, HCl was quickly dissolving and therefore it was removed faster from contact with PVC. Figure 4.7 shows that dehydrochlorination in water occurs with a reduced rate. Figure 4.8 shows another experiment which compares dehydrochlorination rates with and without continuos removal of HCl. Degradation rate is faster if HCl is not removed. In addition, HCl was added in two different concentrations, causing an increase of degradation rate proportional to its concentration.126 Other researchers conducted simi- R2SnCl2 + Ca O O CaCl2 + O O R2Sn 4.1 The reasons for polymer instability 95 mechanism of chain extension and equation given by Troitskii, the mechanism of catalysis becomes clear and consistent with experimental data: There should be very little doubt about this mechanism because it agrees with all underlining principles of PVC degradation which have been established by most research- ers so far. These mechanisms are by no means recent discoveries. Ionic mechanism was proposed by van der Ven113 in 1969, four-center HCl catalysis by Hjertberg and Sorvic119 in 1978 and a mechanism similar to cis geometry by Morikawa134 in 1968. These mecha- nisms were reported in the first edition of this book more than 30 years ago. It is important to mention it because it shows that many research works scrutinized these concepts and must have found them sound since they are still considered as valid. The next question regards the influence of HCl from the point of view of physics. HCl has molar volume of 0.838 l/mol, molecular diameter of 0.322 nm, and an apparent activation energy of diffusion of 62.8 kJ/mol.136 It is a small molecule easy to diffuse in any environment which does not preclude the fact that diffusion time depends on environ- ment and tortuosity of pathways. Many authors claim “efficient” removal of HCl.132 But with the most “efficient” removal there is always some time (no matter how small) which permits HCl to interact with sequences of chain according to the above equation. We know that metal soap stabilizers are very reactive and they are homogeneously dispersed in polymer, and this author has found that stabilizer presence reduces the catalytic effect of hydrogen chloride (see Figure 4.4).49 Real samples have three dimensions (for this purpose their thickness is the most important measure). Minsker124 considered influence of sample size on the results of determination of dehydrochlorination rate and used samples of a thickness less than the critical one calculated from the following equation: [4.3] where: D coefficient of diffusion, k3 effective constant of reaction catalyzed by HCl. The problem is even more complicated than Minsker predicted.124 If one combines his equation with Arrhenius' equation for constant k3 and Gillands' equation for the diffu- sion coefficient, one obtains the following relationship after some parameters of the equa- tions have been simplified: [4.4] where: T temperature, E,R constants. Cl H H Cl + 2HCl lcr π D k3⁄( ) 0.5= l CT0.75eE 2RT⁄= 96 Principles of Thermal Degradation We can see, first of all, that the function is not complete. It has a minimum at T=2E/3R; therefore, a sample of thickness of lcr would be affected by catalysis less than a sample of any other thickness, which does not offer stable conditions for measurement. Early studies137-139 revealed that when the electrical conductivity of samples is mea- sured, the grain structure of PVC in the sample affects readings. Studies on PVC morphology140-142 have demonstrated that PVC grain is built up from smaller particles causing formation of a free space inside the grains (see more in Chapter 3). Also, in sus- pension PVC the grain is surrounded by a sack-like skin, which contains microparticles and submicroparticles. One can easily imagine tortuosity of pathways complicating diffu- sion of HCl to the outside of the sample. These structures are completely or partially destroyed during processing, therefore diffusion constantly changes. Different structures provide conditions for intermediate trap- ping of the HCl evolved. These problems remain unresolved both theoretically and exper- imentally and we do not know the magnitude of influence of HCl on dehydrochlorination rate because this must be related to the time of contact, which cannot be easily predicted in these complex geometries and influence of composition on the HCl diffusion rate. Influence of HCl on polyene length was also a subject of studies. Owen121 assumed that HCl catalysis may influence the distribution of polyene sequences during degradation and that it may also affect photochemical crosslinking reactions of polyenic cations. Braun143 has done spectrophotometric studies in order to compare the length of polyene sequences when PVC samples in the form of powder and film were degraded. In the case of PVC powder, he found that the higher the temperature, the shorter the polyenes, whereas PVC in the form of a film exhibited just the opposite relationship. Thus, the form of a sample may influence the result, and finally, the conclusions. There is now general agreement that HCl presence increases polyene length.3,144,145 There are a few additional observations regarding this topic. Chlorinated PVC is more stable because double bonds were saturated with chlorine during the chlorination process.146 Some authors believe that polyenes can be shortened by HCl readdition into polyene sequence.147 Entropy of PVC dehydrochlorination reaction is lower than entropy of model compounds because of the influence of HCl autocatalysis.3 It was calculated that it will take 2 billion years for PVC to lose 1% chlorine at 40oC.148 4.1.5 IMPURITIES Many research papers mention impurities but do not identify their composition and the way in which they may interfere with thermal degradation. Some impurities were found to be introduced with additives, such as fillers,149,150 plasticizers,151 and solvents (most likely in the case of sample preparation but sometimes they are introduced with additives).152 There is no information in these papers on how these impurities may influence thermal degradation. It was pointed out that polymerization impurities do not seem to interfere with thermal stability of polymer because it does not change after reprecipitation.36 The only documented dangers of influencing thermal stability of PVC are reported in works on recycling.31,153-155 The most frequent impurities include polymers (e.g., PET and PE),31 paint and ink residues, glue residues, metal traces, and contact media.154,155 This 4.1 The reasons for polymer instability 97 type of contamination poses danger to poly- mer stability as well as to retention of mechanical properties and esthetic values. 4.1.6 SHEAR Shear increases temperature of melt (Figure 4.9).156 It is usually difficult to the separate effect of shear and heating on temperature but capillary rheometry permits their sepa- ration.16 Figure 4.10 shows rate constants of stabilizer reaction with PVC degradation product as a function of shear stress and temperature.16 Increase in temperature and shear stress increases stabilizer consump- tion (and dehydrochlorination rate).16 This information shows that stationary testing of thermal degradation does not include all influences and a correction must be intro- duced to account for the effect of shear on temperature increase. It was demonstrated (Figure 4.11) that jet milling affects molecular weight of PVC.157 Also its molecular weight distribution is affected. In injection molding process, rate of material flow and its distance from the wall influence temperature.158 If conditions of flow are too severe, temperature rises above crit- ical for thermal stability of material and injection molded parts show shear burning.158,159 Figure 4.11. Number of jet milling cycles vs. molecular weight of PVC. [Data from Xu, X.; Guo, S.; Wang, Z., J. Appl. Polym. Sci., 64, 12, 2273-2281, 1997.] Figure 4.9. Screw speed vs. extrusion temperatures. [Data from Covas, J. A.; Gilbert, M., Polym. Eng. Sci., 32, 11, 743-750, 1992.] Figure 4.10. Rate constant of stabilizer reaction vs.shear stress for PVC siding capstock compound at two different temperatures. [Data from Bacaloglu, R.; Fisch, M.; Steven, U.; Bacaloglu, I.; Krainer, E., J. Vinyl Additive Technol., 8, 3, 180-193, 2002.]