University of Groningen Effect of the modification of the polymer-rich phase composition on the formation of structural defects in radical suspension PVC Purmova, Jindra IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Publication date: 2007 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Purmova, J. (2007). Effect of the modification of the polymer-rich phase composition on the formation of structural defects in radical suspension PVC. s.n. Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date: 13-03-2022
Transcript
University of Groningen
Effect of the modification of the polymer-rich phase composition on the formation of structuraldefects in radical suspension PVCPurmova, Jindra
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.
Document VersionPublisher's PDF, also known as Version of record
Publication date:2007
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):Purmova, J. (2007). Effect of the modification of the polymer-rich phase composition on the formation ofstructural defects in radical suspension PVC. s.n.
CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license.More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne-amendment.
Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.
Effect of the addition of a solvent for PVC on the formation of structural defects
Purmová, J.; Pauwels, K.F.D.; Van Zoelen, W.; Vorenkamp, E.J.; Schouten, A.J. ABSTRACT: The effect of the addition of a good solvent for PVC on the frequency of the side reactions occurring during the suspension radical polymerization of vinyl chloride was studied. Ortho-dichlorobenzene was chosen in this study because of the limited number of good solvents for PVC and the requirement of low miscibility with water. This resulted in less hindrance of the reactions susceptible to diffusion control (such as propagation, transfer to monomer, and intermolecular hydrogen abstractions) even at high polymer concentrations. The intramolecular transfer reactions are not very dependent on the density of the polymerization medium. Their frequency was therefore not affected by the dilution of the polymer-rich phase, whilst the polymerization was less diffusion controlled. The logical consequence was a strong decrease in the content of butyl branches – structural defects having origin in the intramolecular transfer reactions. Easier transfer to monomer, however, cancelled out the decrease in intramolecular formation of the intermediate radicals of internal double bonds. Only a slight decrease was registered for these defects. The same applies for long-chain branches because of the easier intermolecular transfer. The positive changes in the labile chlorine content were reflected in the thermal degradation behavior of the polymers. The elimination of the first HCl molecules was delayed and less conjugated double bonds sequences were formed.
CHAPTER 5
110
Introduction The effect of an earlier and more intensive precipitation of the polymer chains, and longer existence of the monomer-rich phase on the number of defects was described in the previous chapter. An increase in the concentration of all kinds of defects was observed, suggesting the key role of the polymerization, which takes place in the polymer-rich phase, in their formation. Although premature precipitation of the PVC chains produced more branching and unsaturations, polymerizations taking place in a homogeneous medium like solution polymerization are known to give rise to higher concentrations of defects (especially butyl branches) than radical suspension polymerization.1 This can be attributed to the higher mobility of the polymer chains. When the polymer is in solution the local concentration of the polymer segments is lower than in bulk. Intermolecular transfers are therefore less probable. That leads to higher availability of polymer chains to backbiting especially at high monomer conversions, when the monomer is scarce and propagation considerably slowed down. However, studying the monomer conversion dependence of the formation of the various types of defect structures, an enormous increase in the formation of both branched and internal unsaturated structures2 above 85% monomer conversion was found. This was reflected by a sudden decrease in stability of the resulting PVC polymer. It can be deduced that above this so-called “threshold conversion” the polymer-rich phase becomes extremely dense, making the diffusion coefficient of the monomer much lower than in the monomer-rich phase.3-5 This causes substantial reductions in the propagation rate,3,5 thereby allowing the chain transfer processes to compete more effectively. The addition of a solvent for both PVC and VCM to the polymerization system makes the PVC in the polymer-rich phase more swollen. The monomer in the polymer-rich phase is diluted due to the addition of the solvent; however, the total amount does not change.6 The polymer-rich phase is expected to remain swollen even at very high monomer conversions, where usually its Tg would almost be reached. In the presence of small amounts of the solvent, the polymerization process takes place in a two-phase system until the critical conversion, when the liquid monomer-rich phase disappears and the reaction continues only in the polymer-rich phase swollen in monomer. A solvent, miscible with VCM and insoluble in the aqueous suspension medium, ortho-dichlorobenzene (o-DCB) was selected. Aromatic functional groups are able to exhibit non-covalent interactions with the atoms of PVC chains, which is one of the basic principles of action of commonly used plasticizers.7,8 By using an aromatic solvent an attempt was made to explore the potential of these functional groups to protect the polymer chain from the side reactions occurring during the polymerization. In order to investigate the importance of the interaction of PVC chains with chlorine atoms in o-
Effect of the addition of a solvent for PVC on the formation of structural defects
111
DCB, polymerizations were also carried out in the presence of chlorobenzene − aromatic solvent carrying only one chlorine. The effect of the moderate dilution of the polymer-rich phase and the presence of functional groups able to interact with PVC, on the kinetics of the side reactions is examined in the following text. A new explanation will be added to account for the contradictory results of an increasing amount of methyl branches along with a decreasing number of chloroallylic end groups on conversion. Experimental The polymers were produced by suspension polymerization in a 1L steel autoclave as described in Chapter 2, with the modifications mentioned in the Chapter 4. Instead of non-solvents, different quantities of solvents (0.5 – 4.4 mol % of o-DCB or 1.1 mol% of chlorobenzene with respect to monomer) were added to the reaction mixture. The solubility of o-DCB in water is as low as 0.16 g L-1 at 25 °C9 and in the case of chlorobenzene 0.50 g L-1.10 The water solubility might be slightly higher at the polymerization temperature (57.5 °C), but it can be assumed that the amounts of both solvents in the monomer droplets are not substantially influenced. Moreover, the vapor pressure at this temperature is also low (approximately 1.81 and 16 mbar for o-DCB and chlorobenzene respectively)11, therefore we expect that almost all the added solvent is mixed homogenously with VCM. The reactions were stopped after a 3 bar pressure drop in the reactor. For the “standard” reaction, this correlates with 87% conversion. The conversions of all reactions mentioned in this Chapter are presented in Table 5-1. The resins were characterized by the analytical methods comprehensively described in the experimental section of Chapters 2 and 4. Average particle size was roughly determined from light microscopy pictures. Table 5-1. Characteristics of PVC samples studied in this chapter.
Results and Discussion Addition of any kind of solvent produced PVC containing the same types of defect structures as standard PVC made in the laboratory of Polymer Chemistry at University of Groningen (see Chapter 2). However, variations in the concentration of these defects were observed. The relation with the type and amount of solvent is further discussed. Polymerizations with additions of solvents exhibited somewhat prolonged reaction times. However, the reactions were stopped at the same moment along the reaction profile (pressure drop of 3 bar) and therefore exhibited approximately the same conversion as without the addition of solvent (Table 1). A noticeable difference was the particle size, which increased greatly. With additions of large amounts of solvents, eventually no separate particles were formed anymore leading to uncontrolled reaction paths. Moreover, the appearance of the Trommsdorff effect, as measured from the heat production in the reactor, changed considerably. Although this effect dominates the whole polymerization, especially at the end of the reaction with conversions exceeding 75%, an accumulation of heat production was normally observed.12 Addition of o-DCB diminished this effect and ultimately made it disappear. These observations indicate that the addition of small amounts of solvents greatly influences the physical-chemical conditions for the suspension polymerization.12 In this Chapter, we will focus on the content of defects measured in samples of polymerizations with different but small amounts of solvents. Again, we have to consider possible effects on the bulk of the monomer droplet and possible effects on the interfacial layer or skin of the particle. One problem with the solvents used in this study is their high boiling point, which makes the removal of the solvents from the particles after reaction, without disturbing the particle morphology, quite troublesome. All these observations lead to the conclusion that these solvents cannot be seen only as diluters but effects on the interfacial processes have to be considered as well. Particle size. It is clear from microscopic pictures (Figure 5-1) that the particle size increased from about 133 μm to 446 μm upon addition of up to 4.4 mol % (10.3 weight %) o-DCB, see Table 5-2. Table 5-2. Average particle size, measured manually from the pictures shown in Figure 5-1
Amount of o-DCB added (mol %) 0 0.7 1.1 1.3 4.4
Size in μm 133 304 257 435 446
The morphology has changed somewhat too, the particles resemble aggregates, and smaller particles in the aggregate can be identified. The consequence of this is that diffusion pathways from the water phase to the particle interior might be enlarged greatly and that all reactions involving a rate that could be
Effect of the addition of a solvent for PVC on the formation of structural defects
113
diffusion limited might be extra hindered. Addition of supplementary protecting colloid would probably be necessary to overcome this phenomenon, but this was not so far investigated. The diffusion limitation became obvious at higher o-DCB content when reaction profiles exhibited a greatly enhanced reaction time needed to obtain a vapor pressure decay of 3 bars, the normal point of stopping the reaction. These concentrations, however, are not considered in this Chapter.
(a) (b)
(c) (d)
(e) Figure 5-1. Influence of the addition of o-DCB on the particle size of PVC with approximately 87% conversion prepared at 57.5 °C. a) 0 mol %, b) 0.7 mol%, c) 1.1mol%, d) 1.3mol %, e) 4.4 mol% o-DCB.
CHAPTER 5
114
Molecular weight. The molecular weight of the PVC was measured with GPC and the
results are presented in Figure 5-2 in terms of nM , wM and wM / nM . Obviously, wM
decreases, whereas nM remains constant. This means that overall the polymerization kinetics hardly changes. The molecular weight of PVC after addition of 1.3 mol percentage of chlorobenzene followed the same trend (Figure 5-2). The results indicate that these solvents do not behave as chain transfer agents like the aromatic non-solvents investigated in Chapter 4.
25
40
55
70
85
100
0 1 2 3 4 5Solvent [mol %]
MW
[kg
mol
-1]
0.0
1.0
2.0
3.0
4.0
5.0
nM
wM
Figure 5-2. Development of the molecular weight with increasing amount of o-DCB. ▲ nM ; � wM ;
wM / nM ; white labels correspond to the polymers made with the addition of chlorobenzene. This is, however, only indirect evidence. The proton chemical shifts of the functional groups introduced into the PVC molecule by the chain transfer coincide with the signals of the unsaturations.13,14 The capacity of a chain transfer agent to decrease the content of internal unsaturations was proved by examining PVC made with the addition of CBr4 (see Chapter 3 for more details). Assuming the chain transfer actually takes place, chain transfer constants to o-DCB (50 x10-4) and chlorobenzene (65 x10-4) were calculated with the help of the well-known Mayo equation15 (See Chapter 4 for more details). This is an indication of a low chain transfer constant capacity for both solvents and the result of the competition among the side reactions and chain transfer to the additive is not therefore very evident. The addition of an increasing amount of both solvents led to the gradually decreasing polymerization rate. Taking into account the weak effect of the solvent on the molecular weight (Figure 5-2) this is probably mainly the result of the deceleration of the reaction due to the decreasing monomer concentration. In addition, a less marked gel effect can also play a role in the elongation of the polymerization time.
The decrease in wM will be discussed in relation to the change of number of branches.
Effect of the addition of a solvent for PVC on the formation of structural defects
115
Unsaturated structures. Polymerizations were carried out up to the same conversions (around 87 %) with different but small amounts of o-DCB. The effects of these additions on the defects are shown in Figures 5-3 to 5-13. It appeared possible to add more o-DCB before unacceptable particle agglomeration prevents the normal reaction procedure. However, from a technical point of view, additions of high amounts of solvents are not attractive and therefore we limited our further investigations to 4.4 mol %. Figure 5-3 shows a 25% increase in the number of chloroallylic end groups per chain in comparison to polymer of similar conversion without the addition of a solvent. Moreover, the total effect already appears after addition of only 0.7mol % and remains constant upon addition of more solvent. This is in great contrast to the effect of addition of non-solvents (Figure 4-1). The discussion of the possible causes of this effect should also include the effect on the content of methyl braches, as they both originate from the same species, the 1-2 chlorine-shifted polymer radical after a head-to-head addition:
CH2
CHCH2
CHCH
CH2
Cl
ClCl
This species is the central starting point for both defects and from the known mechanisms it has the capacity to react with monomer, either to continue polymerization (methyl branch) or to transfer a chlorine atom and end up as a chloroallylic end group (Scheme 1-3). However, it should be clear that upon decreasing monomer concentration (dilution effect) both reactions should be suppressed, so, a simple explanation based on dilution would not apply here. In Chapter 2 we considered the possibility of copolymerization of the formed chloroallylic end group with a normal polymer chain during further polymerization. Based on theoretical calculations this was not such an unlikely process as it appears. Moreover, if diffusion of monomer to this reactive chain end would be rate limiting again both reactions should be affected to the same relative extent. Clearly, this is not the case here, as it was not observed in the experiments described in Chapter 2. Further on in the text we will come back to this effect. Figure 5-4 displays the number of chlorovinyl end groups as a function of the added amount of o-DCB. In this case, the content of this defect increases linearly with the added amount, which can be explained most easily by an increasing contribution of normal bimolecular termination via disproportionation of the growing chains, the process that is responsible for this end group. In fact, the Trommsdorff effect decreases here due to dilution and consequent lowering of the viscosity. This corresponds nicely with the observation from the reaction profiles that the heat produced during the reaction does not exhibit a maximum during the later stages of the reaction upon o-DCB addition.16
CHAPTER 5
116
The internal double bonds and the methyl, butyl and long branches (and total branches accordingly), however, clearly show a leveling off effect upon addition of more solvent (Figures 5-4 to Figure 5-7). This is not easily understood using only the common theories about the origins of these effects. To obtain more information about this effect we analyzed a series of polymers produced with the addition of 1.3 mol % o-DCB polymerized to different conversions. The results are shown in Figure 5-8 and Figure 5-9 (Section Impact on different stages of the polymerization). It is clearly seen from this set of experiments that the observed decrease in internal double bonds and total branches, as well as the increase in the chloroallylic end groups, is found for all conversions. This indicates that the changing physical-chemical conditions, resulting from adding a solvent, do not interfere with reactions subjected to monomer diffusion limitation. If this were the case, the effect should not be present at low conversions, where diffusion limitation hardly plays a role. The results from Figure 5-8, however, clearly indicate that major changes are introduced just by adding small amounts of solvents. During polymerization, only the monomer is consumed and thus the concentration of VCM decreases in both phases during the course of the reaction. The quality of the solvent mixture for PVC therefore increases. This should lead to an expansion of the polymer coils, and consequently to an increase of the total volume fraction of the polymer rich phase in the monomer droplet, at the expense of the polymer-lean phase. We have to assume that because of the small amounts of solvent used this effect will also be small. Therefore, major viscosity changes upon addition of the solvents are not expected. Only at the very end of the polymerization, one might expect a relatively high impact from the presence of o-DCB in the polymer matrix, decreasing the Tg of the polymerizing mixture and possibly enhancing segmental diffusion. So far, we cannot refer to any described mechanism to explain our results. A further analysis of the reaction was therefore carried out.
0.3
0.4
0.5
0.6
0 1 2 3 4 5Solvent [mol %]
Chl
oroa
llylic
end
gro
ups/
chai
n
Figure 5-3. Dependence of the number of chloroallylic end groups on the percentage of solvent added during the polymerization. ▲ o-dichlorobenzene; r chlorobenzene.
Effect of the addition of a solvent for PVC on the formation of structural defects
Figure 5-4. The number of chlorovinyl end groups (–CHClCH=CHCl) per chain and the solvent percentage. ¿ o-DCB; ¯ chlorobenzene. Addition of o-DCB caused a slight decrease (10%) in the concentration of internal unsaturations per chain (Figure 5-5). The effect of the addition of chlorobenzene was very similar (Figure 5-5). This is probably the effect of the decrease in total surface area (because of the increase in particle size, see Figure 5-1) leading to formation of less structural-defect-rich material on the water polymer interface.
0.20
0.25
0.30
0.35
0.40
0 1 2 3 4 5Solvent [mol %]
Inte
rnal
dou
ble
bond
s/ch
ain
Figure 5-5. Dependence of the number of internal unsaturations per chain on the solvent percentage. � o-DCB; � chlorobenzene. Internal double bonds are formed by either inter- or intramolecular hydrogen abstraction followed by a chlorine abstraction, which occurs exclusively by transfer to monomer.17 The intramolecular process was shown to be predominant (Chapter 3, Table 3-3, and Scheme 3-6). The decrease in monomer concentration in both phases caused by dilution with o-DCB led to a decrease in transfer to monomer and subsequently to a decrease in
CHAPTER 5
118
the content of defects involving this reaction such as internal unsaturations (See Chapter 1, Scheme 1-3, 1-4). Branches. The total number of branches decreased about 9% with the addition of solvent (Figure 5-6) followed by an initial decay reaching a plateau of seven
branches/1000 VC units after 1.3 mol % of additive. The values of wM determined by GPC are known to be affected by the presence of branching, producing a high molar mass tail.18 Decreased short chain branch content is probably therefore the cause of the
decreased value of wM . The number of all types of branches per 1000 VC units exhibited the same decreasing tendency with the addition of both solvents (Figure 5-7), as well as the total number of branches (Figure 5-6). Some noticeable differences were observed. The strongest relative decrease of 25% is shown by the butyl branches, followed by the number of ethyl, long and methyl branches (all approximately 13%).
6.0
6.5
7.0
7.5
8.0
0 1 2 3 4 5Solvent [mol %]
Tota
l bra
nche
s/10
00VC
Figure 5-6. Relation between the total number of branches per 1000 VC units and the amount of solvent added during the polymerization. ¢ o-DCB; £ chlorobenzene. In contrast to the couple methyl branch/chloroallylic end group, butyl branches and internal allylic groups exhibit the same trend. Apparently both backbiting reactions:
CH
CClCl Cl Cl
Cl
Cl
H
CHH
Cl
Cl
ClClCl
HCH
Cl
Cl
ClClClCl
CH
CHCl
Cl
ClClClCl
H
VCM
VCM
Butyl branch
Internal unsaturation
Effect of the addition of a solvent for PVC on the formation of structural defects
119
are suppressed in the same way, or alternative reactions are favored. Because these backbiting reactions are supposed to be unimolecular, monomer concentrations and macro viscosity are not important parameters. Nonetheless, the following reaction with monomer might be kinetically influenced by diffusion control or monomer concentration. The relatively small amounts of solvent added, however, can never explain such a dramatic change of defect contents. Moreover, the result that the effect is already present at 0.7 mol% contradicts such an explanation. Another possibility would be the specific interaction with the growing polymer chain end. The molecules of o-DCB would have to interact preferentially with the monomer units near to the chain end of the growing macroradical to achieve such an effect. So far, we do not have any definitive proof of that, nor are we aware of any comparable phenomena in the literature. The capacity of the hydrogen abstractions from a dead polymer, leading to formation of the long branches, to compete with the propagation decreases only slightly with the addition of solvent. This might be explained by chain transfer to the additive, mentioned in the section about unsaturations. The formation of methyl and ethyl branches by the generally accepted mechanism19,20 involving chlorine shift after head-to-head addition (Scheme 1-3) should not be affected by either decrease in monomer concentration or diminishing of the Trommsdorff effect. There is a possibility, mentioned in Chapter 2, that some of the methyl branches might be formed by combination of Cl shifted head-to-head radicals with growing polymer chains. From the diminishing of the heat peak associated with the Trommsdorff effect, it can be deduced that the presence of o-DCB in the reaction mixture favors bimolecular terminations. This is also reflected in the trend of chlorovinyl end groups (Figure 5-4). The apparently contradictory decrease in the number of methyl branches can be attributed to the increase in disproportionations paired with a decrease in combinations. Addition of chlorinated aromatic compounds probably changes the polarity of the polymerization medium enough to change the composition of bimolecular terminations. Disproportionation of a 1-2 Cl shifted head-to-head radical would render a chlorovinyl end group and saturated chain end (Scheme 4-1). Another possibility includes the formation of saturated chain end and the chloroallylic-like end group (Scheme 4-1).
CHAPTER 5
120
CH2 CH
C HCH2
Cl
Cl
CH2
C HCl
Disproportionation
CH
CHCl
CH2 CH
CH2
CH2
Cl
Cl
CH2 CH
C HCH2
Cl
Cl
CH2
C HCl
Disproportionation
CH2
CH2
Cl
CH2 C
CH
CH2
Cl
Cl
or
Scheme 5-1. Termination of 1-2 Cl shifted head-to-head radical by disproportionation
3.0
3.5
4.0
4.5
5.0
0.0
0.5
1.0
1.5
2.0
0 2 4 6 8 10 12
o -DCB [%]
Def
ects
/100
0 VC
0.0
0.5
1.0
1.5
2.0
0 1 2 3 4 5Solvent [mol %]
Bra
nche
s/10
00 V
C
3.0
3.5
4.0
4.5
5.0
Figure 5-7. Correlation of the number of different types of branches per 1000 VC units with percentage of o-DCB added during the polymerization. X Methyl branches, � Butyl branches, ▲ Long branches, u Ethyl branches. Grey points represent defects in PVC made in the presence of chlorobenzene. Impact on different stages of the polymerization. In order to investigate the mechanism of the influence of o-DCB on the polymerization in more detail, samples of lower and higher conversions were prepared in the presence of 1.3 mol % of o-DCB. The number of chloroallylic end groups was higher in PVC made in the presence of o-DCB at all conversions and had the same tendency to decrease with conversion as in the case of standard PVC (Figure 5-8a). The internal unsaturations were lower for all conversions as were the total number of branches. Moreover, the tendencies for both defect classes as a function of conversion were the same as the polymer without the solvents. This clearly demonstrates that the
Effect of the addition of a solvent for PVC on the formation of structural defects
121
effect of the addition is not limited to the high conversion situation or has anything to do with neither diffusion limitation nor dilution effects. As revealed in Figures 5-8b and c the number of both defects is lowered over the whole conversion range after the addition of solvent, more or less to the same extent. However, the decrease is slightly less noticeable at conversions >87% (Figure 5-8b,c).
0.2
0.3
0.4
0.5
0.6
0.7
10 20 30 40 50 60 70 80 90 100Conversion [%]
Chl
oroa
llylic
end
gro
ups/
chai
n
0.2
0.3
0.4
0.5
0.6
10 20 30 40 50 60 70 80 90 100Conversion [%]
Inte
rnal
dou
ble
bond
s/ch
ain
(a) (b)
6.0
6.5
7.0
7.5
8.0
8.5
9.0
9.5
10.0
10 20 30 40 50 60 70 80 90 100Conversion [%]
Tota
l bra
nche
s/10
00 V
C
(c) Figure 5-8. Comparison of the number of defects present in PVC of different conversions made with or without addition of 1.3 mol % of solvent. a) Chloroallylic end groups/chain, b) internal double bonds/chain, c) total number of branches/1000 VC. Black labels: no solvent; white labels: o-DCB; grey labels: chlorobenzene. Analyzing the different types of branches separately, the following trends were observed after the addition of both solvents. The number of methyl branches was lower even at a conversion of about 12% and followed the same tendency to increase with monomer conversion as in case of polymerization without solvent (Figure 5-9a). Ethyl branches displayed only a very slight decrease at conversions >87% (5-9b), although within experimental error this cannot be stated definitively. The relative decrease in the number of these branches at high conversions was approximately the same (~15%). The number of long branches did not present any visible changes (Figure 5-9c). Butyl branches were formed to a lesser extent when a solvent was added to the polymerization
CHAPTER 5
122
system. This effect was not very pronounced at the beginning of polymerization, but was quite large at high conversions, achieving the relative decrease of 30% (Figure 5-9b). Both solvents apparently inhibit preferentially the intramolecular transfer reactions, which lead to the formation of internal double bonds and butyl branches (Scheme 1-2), and start to act at higher concentration of the polymer chains.
3.0
3.5
4.0
4.5
5.0
5.5
6.0
10 20 30 40 50 60 70 80 90 100Conversion [%]
Met
hyl b
ranc
hes/
1000
VC
0.2
0.6
1.0
1.4
1.8
2.2
10 20 30 40 50 60 70 80 90 100Conversion [%]
Bra
nche
s/10
00 V
C
(a) (b)
0.2
0.4
0.6
0.8
1.0
1.2
1.4
10 20 30 40 50 60 70 80 90 100Conversion [%]
Long
bra
nche
s/10
00 V
C
(c) Figure 5-9. The number of different types of branches present in PVC of different conversions made with or without addition of solvent. a) Methyl, b) � butyl and ¿ ethyl and c) long branches. Black labels: no additive; white labels: 1.3 mol % o-DCB. The presence of the solvent probably also influences the situation on the surface of the monomer droplets, clearly observed from the particle size differences shown in Figure 5-1. As a result of the analysis of the distribution of different structural defects in the fractions of various molecular weights, it was proposed that some part of the low molecular weight material is formed on the surface of the PVC particles (see Chapter 3). The properties of the polymer formed on the water-polymer boundary should be similar to those of the PVC obtained at subsaturation conditions.21,22 A substantial increase in the number of long, butyl and ethyl branches and decrease in the number of methyl branches was observed for polymers made at lower monomer pressure.23 At conditions when the monomer-rich phase still exists and/or the polymer-rich phase is sufficiently swollen in
Effect of the addition of a solvent for PVC on the formation of structural defects
123
monomer these chains migrate back to the interior of the PVC particle and continue their growth (see also Chapter 3 for a more detailed discussion about this phenomenon). This migration becomes slower and eventually stops with the decreasing amount of monomer in the polymer-rich phase. In the presence of a good solvent for PVC, the polymer matrix stays swollen even at high conversions, because the solvent is still present when VCM is almost depleted. It is therefore tempting to speculate that the polymer chains, formed at the water interface, will be able to re-enter the polymer-rich phase even at higher conversions than normally. This together with the decrease in total surface area caused by the increase in the mean particle size would lead to the formation of less “subsaturation like” low molecular weight material, which might also contribute to the lower number of branches. However, it is a little surprising that the content of the long branches remains the same. This is in contradiction against such an explanation. Interactions with solvent molecules. The protection of the polymer chain against hydrogen abstractions by reversible non-covalent interactions with solvent molecules might also play a role in the decrease of concentration of the internal unsaturations. As was described in previous sections, the addition of o-DCB preferentially reduces the concentration of defects formed by intramolecular processes. Chlorine atoms present in o-DCB molecules could form weak hydrogen bonds with the methynic protons of PVC. The existence of similar weak hydrogen bonds has already been reported in the literature.24,25However, it is difficult to imagine that, in that case, growing chain ends would be the preferred site of interaction, because on a base molar scale we added only 4.4 mol % at maximum. In order to investigate the relative importance of these interactions, a solvent containing only one chlorine atom was added. A monosubstituted solvent would not be able to create the same structure as the disubstituted molecule and its effect on the defect structures should be weaker. The Figures 5-1 to 5-7 show that the effect of monosubstituted solvent (chlorobenzene) was basically the same as that of disubstituted solvent (o-DCB). This result makes the above-mentioned explanation unlikely. The absence of any substantial changes in the number of long branches also excludes the protection of polymer chains from hydrogen abstractions by means of CH-π interactions among solvent and PVC molecules. Distribution of defects in fractions of different molecular weight. The influence of the addition of an additive on the chain length dependence of unsaturated and branched structures was also studied for polymers made in the presence of solvent (in this case 4.4 mol% of o-DCB).
CHAPTER 5
124
A decrease in the concentration of the chloroallylic end groups (per chain) with increasing molecular weight was registered (Figure 5-10). Compared to the polymer prepared without solvent, the decrease is much more pronounced. It is clear that the extra chloroallylic end groups are found preferentially in the low molecular weight components. This is caused by more frequent transfer to monomer (necessary for the formation of chloroallylic end groups), but the reason why is not seen so easily. One possibility is the enhanced contribution of interfacial effects.
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 10 20 30 40 50 60 70 80 90 100
Chl
oroa
llylic
end
gro
ups/
chai
n
]molkg[M 1n −
Figure 5-10. Dependence of the number of chloroallylic end groups per chain on nM for fractions of different PVC samples made with addition of o-DCB. 87.2% conversion; n 4.5 mol % o-DCB. Summary of the most pronounced features appearing upon addition of o-DCB:
- Decreasing Trommsdorff effect - Continuously increasing unsaturated chlorovinyl end groups
- Decreasing wM with constant nM - Decreasing content of methyl branches even with small amounts of o-DCB - Increasing number of chloroallylic end groups even with small amounts of o-DCB - Increasing number of chloroallylic end groups, especially in the low molecular
weight components - Decreasing number of butyl branches already even with small amounts of o-DCB - Decreasing total number of branches /1000VC even with small amounts of o-
DCB - Decreasing number of internal double bonds per chain even with small amounts
of o-DCB - Effects are independent of the conversion - Long branches are not affected
Effect of the addition of a solvent for PVC on the formation of structural defects
125
The disappearance of the extra heat development in our reactor system resulting from an enhancement of the Trommsdorff effect was not achieved after addition of only a small amount of o-DCB, but the effect gradually diminished upon addition of more solvent. This could be interpreted in terms of enhanced bimolecular termination reactions, especially of those small chains, which have degrees of polymerization below the critical chain length for entanglements. It is now well known that these chains exhibit a completely different diffusion behavior than long chains, where reptation behavior is accepted as the sole mode of diffusion.26 The effect of the solvent would be to increase swelling of the entanglement network, increase the average chain length between entanglements, and consequently opening the possibility for longer chains to terminate bimolecularly. This is in agreement with the continuously increasing content of chlorovinyl end groups, which originate from bimolecular termination by disproportionation. It may be speculated that the features exhibited by the molecular weights also correspond to this effect, although the changing kinetics of all reactions will have an influence on the molecular weights and distributions and therefore the overall effect is difficult to predict. Certainly, the changes found do not in contradict this theory. All the effects produced in the defects after addition of only a small amount of o-DCB and which do not change anymore upon addition of more solvent, might be correlated with the changing particle diameter as shown in Figure 5-1. It is clear that the addition of a small amount of solvent already induces a big change, and that this change is more or less constant under further addition of solvent. This effect was not expected, and remains to be explained by further study of the polymerization with additional or alternative protecting colloids. For the time being, we have to use the phenomenon described above in the explanation of the results. Particle size is already determined in the very early beginning of the suspension polymerization. This corresponds nicely with our observation that all the changes in defect concentrations are already present at low conversions. The consequence of this particle agglomeration is that we create smaller interface with water. On the other hand, longer pathways are formed for monomer to enter into the inner parts of the particles. Diffusion limitation of reactions are therefore expected to be affected at the point in the reaction where monomer from the water phase and the gas phase above the solution has to be transported to the particle, either to the very inside or to the interfacial layer. Another important new viewpoint related to the interfacial layer was introduced in Chapter 3, where we concluded that some chains might grow for some time in the interfacial layer or even in the water phase. Depending on the viscosity of the bulk of the particle, these chains might then diffuse into the bulk of the particle. It is expected that, in the present case, the addition of o-DCB greatly influences the bulk viscosity at later stages of the reaction. The overall effect will be that the contribution of chains formed under these special conditions will be lower, especially at higher conversion. This explanation
CHAPTER 5
126
can then also be used for the last five effects summarized in the beginning of this section. When less growing chains are trapped in the interfacial layer, where the monomer concentration is lower, the growth processes take place in surroundings where more monomer is present. Accordingly, the increase in the chloroallylic end groups (especially in the low molecular weight part) can be explained in the same way, because this process needs monomer and the monomer concentration in the interior will be higher than in the interfacial layer. Unimolecular backbiting reactions will then be suppressed and propagation will be favored. Thermal stability. The presence of o-DCB in the reaction mixture during the polymerization delayed the beginning of the dehydrochlorination process (Figure 5-11a). Once the first eliminations of HCl started, these polymers degraded as fast as regular PVC (Figure 5-11b). Internal unsaturations and all kinds of branches exhibited clear decrease along the whole range of the solvent amounts (Figure 5-5, 5-6 and 5-7). Therefore, the lower susceptibility to the degradation is probably the effect of the changes in the content of allylic and tertiary chlorine containing defects. The addition of the solvent during the polymerization also affected the morphology of the resulting particles. The mean pore diameter was larger and the overall porosity was lower.12 As was already mentioned in Chapter 4, these morphological characteristics could bias the results of the standard powder dhc-test. Indeed, when the dehydrochlorination was measured in solution, a much smaller increase in the initiation time was observed whilst the dhc-rate kept its tendency (Figure 5-11).
5.0
6.0
7.0
8.0
9.0
10.0
11.0
0 1 2 3 4 5 o -DCB [mol % ]
t i [m
in]
3.0
4.0
5.0
6.0
7.0
0 1 2 3 4 5 o -DCB [mol % ]
dhc-
rate
x10
00 %
/min
(a) (b) Figure 5-11. Development of the a) induction time of dhc (ti) and b) dhc-rate with the amount of o-DCB added during the polymerization. Grey labels: dhc-test in bulk; black labels: dhc-test in solution. Lower content of defects had a negligible effect on the average length of the polyene sequences (Figure 5-12a). On the other hand, concentration of polyenes formed during the degradation presented a clear decrease (Figure 5-12b). Residual o-DCB (5 wt % measured by TGA) might work as a sort of stabilizer inhibiting the initiation of the
Effect of the addition of a solvent for PVC on the formation of structural defects
127
unzipping. However, Bengough and Sharpe27,28 observed that o-DCB, when used as a solvent for PVC during the dehydrochlorination test performed in solution, did not play an active role in the dehydrochlorination. The decrease in the average concentration of conjugated double bonds can be therefore exclusively attributed to the less frequent initiation of the polyene sequences due to the lower number of labile sites.
0.0
1.0
2.0
3.0
4.0
5.0
6.0
0 10 20 30 40 50 60 70
Degradation time [min]
Ave
rage
pol
yene
leng
th
0.0
0.4
0.8
1.2
1.6
0 10 20 30 40 50 60 70 80 90
Degradation time [min]
Poly
enes
/100
0 VC
(a) (b) Figure 5-12. Comparison of (a) average polyene length and (b) concentration in PVC samples of approximately 87 % monomer conversion made with and without the addition of o-DCB. −¿− 0, --¯-- ~4.4 mol% o-DCB. Conclusions The effect of the presence of aromatic solvents for PVC in the polymerization mixture on the course of the side reactions and thermal stability of the resulting polymer was studied. The number of all unsaturated and branched defects decreased. The only exception was the increase in chloroallylic and chlorovinyl end groups. Addition of moderate amounts of solvents greatly influenced the physical-chemical reaction conditions, leading to prolonged polymerization times and larger particle size. The Trommsdorff effect observed in standard VCM polymerization was also gradually diminished with increasing amount of solvent added. On the other hand, the molecular weight was hardly affected. These results can be explained by enhanced bimolecular termination reactions, especially of chains shorter than the critical chain length for entanglements. A more swollen entanglement network (caused by the presence of the solvent) also enabled longer chains to terminate bimolecularly. The consequence was a continuously increasing content of end groups formed by termination by disproportionation (chlorovinyl end groups). The most striking decrease was observed for butyl branches (formed by intramolecular hydrogen abstraction) and the lowest for long branches (formed after intermolecular hydrogen abstractions). This would suggest selective CH-π interactions among the solvent and monomer units near to the chain end carrying the radical. Nevertheless, chlorobenzene caused exactly the same effect on the number of structural defects, which
CHAPTER 5
128
practically rules out the influence of such non-covalent interactions. A more likely explanation is diminished formation of “subsaturation-like” PVC on the water-polymer interface due to the increase in particle size. Relatively facile access of these chains back into the polymer-rich phase, which is swollen by the solvent for PVC, might also play a role. The decrease in the number of methyl branches does not fit to this picture. However, a plausible explanation can be given by changes in the polarity of the polymerization medium, which in turn modify the course of bimolecular terminations in favor of disproportionation. Combination of a 1-2 Cl shifted head-to-head radical with growing macroradical renders a methyl branch, whilst disproportionation results in different types of unsaturated end groups. The positive changes in the content of structural defects were reflected in the thermal degradation behavior of the polymers. The elimination of the first HCl molecules was delayed and less conjugated double bonds sequences were formed. References 1. Darricades-Llauro, M.-F.; Bensemra, N.; Guyot, A.; Pétiaud, R. Makromol. Chem. , Macromol. Symp.
1989, 29, 171-184. 2. Purmova, J.; Pauwels, K. F. D.; Van Zoelen, W.; Vorenkamp, J. E.; Schouten, A. J.; Coote, M. L.
Macromolecules 2005, 38, 6352-6366. 3. Xie, T. Y.; Hamielec, A. E.; Wood, P. E.; Woods, D. R. Polymer 1991, 32, 537-557. 4. Mita, I.; Horie, K. J. Macromol. Sci.-Reviews in Macromol. Chem. Phys. 1987, C27, 91-169. 5. De Roo, T.; Heynderickx, G. J.; Marin, G. B. Macromol. Symp. 2004, 206, 215-228. 6. Olaj, O. F.; Breitenbach, J. W.; Reif, H.; Parth, K. J. Angew. Chem. 1971, 10, 351 7. Sears, J. K.; Darby, J. R. Technology of Plasticizers, John Wiley & Sons: New York., 1982 8. Wypych, G. Handbook of Plasticizers, Handbook of Plasticizers ed.; Chemtec Publishing/William
Andrew Publishing: Toronto, 2004. 9. Banerjee, S.; Yalkowsky, S. H.; Valvani, S. C. Environmental Science & Technology 1980, 14, 1227-1229. 10. Horvath, A. L. Halogenated Hydrocarbons. Solubility-Miscibility with Water, Marcel Dekker, Inc.: New York,
1982. 11. Daubert, T. E.; Danner, R. P. Data compilation tables of properties of pure compounds, Design Institute for
Physical Property Data, American Institute of Chemical Engineers: New York, 1985. 12. Pauwels, K. F. D. New aspects of the supension polymerization of vinyl chloride in relation to the low thermal
stability of poly(vinyl chloride) , Dissertation: University of Groningen, 2004, p99-119. http://dissertations.ub.rug.nl/faculties/science/2004/k.f.d.pauwels/
13. ACD/CNMR Predictor, 3.50; Advanced Chemistry Development Inc.: Toronto, 1998. 14. Chem Draw® Ultra, 8.0.3; CambridgeSoft Corporation: Cambridge, MA,USA, 2003. 15. Gregg, R. A.; Mayo, F. R. J. Am. Chem. Soc. 1948, 70, 2373-2378. 16. Pauwels, K. F. D. New aspects of the suspension polymerization of vinyl chloride in relation to the low thermal
stability of poly(vinyl chloride) , Dissertation: University of Groningen, 2004, p 145. http://dissertations.ub.rug.nl/faculties/science/2004/k.f.d.pauwels/
17. Starnes, W. H., Jr.; Chung, H.; Wojciechowski, B. J.; Skillicorn, D. E.; Benedikt, G. M. ACS, Adv. Chem. Ser. 1996, 249, 3-18.
18. Podzimek, S.; Vlcek, T. J. Appl. Polym. Sci. 2001, 82, 454-460. 19. Rigo, A.; Palma, G.; Talamini, G. Makromol. Chem. 1972, 153, 219-228. 20. Starnes, W. H., Jr.; Wojciechowski, B. J. Makromol. Chem. , Macromol. Symp. 1993, 70/71, 1-11. 21. Hjertberg, T.; Sörvik, E. M. J. Polym. Sci. , Polym. Chem. Ed. 1978, 16, 645-658. 22. Sörvik, E. M.; Hjertberg, T. J. Macromol. Sci. , Part A, Chem. 1977, A11, 1349-1378.
Effect of the addition of a solvent for PVC on the formation of structural defects
129
23. Hjertberg, T.; Sörvik, E. M. ACS Symp. Ser. 1985, 280, 259-284. 24. Desiraju, G. T.; Steiner, T. The Weak Hydrogen Bond: Applications to Structural Chemistry and
Biology, Oxford Univ. Press: 1999. 25. Steiner, T. Acta Crystall. B-Stru. 1998, B54, 456-463. 26. De Kock, J. B. L.; Van Herk, A. M.; German, A. L. J Macromol Sci-Polym Rev 2001, C41, 199-252. 27. Bengough, W. I.; Sharpe, H. M. Makromol. Chem. 1963, 66, 31-44. 28. Bengough, W. I.; Varma, I. K. Eur. Polym. J. 1966, 2, 49-59.
