European Polymer Journal 45 (2009) 3389–3398

Contents lists available at ScienceDirect

European Polymer Journal

journal homepage: www.elsevier .com/locate /europol j

Influence of physical–chemical interactions on the thermal stabilityand surface properties of poly(vinyl chloride)-b-poly(hydroxypropylacrylate)-b-poly(vinyl chloride) block copolymers

Nuno Rocha a,c,*, J.A.F. Gamelas a, Pedro M. Gonçalves b, M.H. Gil a, J.T. Guthrie c

a CIEPQPF, Chemical Engineering Department, University of Coimbra, 3030-790 Coimbra, Portugalb Cires S.A. – Companhia Industrial de Resinas Sintéticas, Apartado 20, Samoqueiro – Avanca, 3864-752 Estarreja, Portugalc Department of Colour Science, School of Chemistry, University of Leeds, Leeds LS2 9JT, United Kingdom

a r t i c l e i n f o

Article history:Received 14 August 2009Received in revised form 17 September 2009Accepted 30 September 2009Available online 7 October 2009

Keywords:Poly(vinyl chloride)Living radical polymerisationBlock copolymersThermal degradation kineticsInverse gas chromatographyInteraction parameters

0014-3057/$ - see front matter � 2009 Elsevier Ltddoi:10.1016/j.eurpolymj.2009.09.020

* Corresponding author. Tel.: +351 917053571; faE-mail address: nunopx@eq.uc.pt (N. Rocha).

a b s t r a c t

The synthesis of poly(vinyl chloride) (PVC) homopolymers and poly(vinyl chloride)-b-poly(hydroxypropyl acrylate)-b-poly(vinyl chloride) (PVC-b-PHPA-b-PVC) block copoly-mers via a single electron - degenerative transfer mediated living radical polymerisationwas carried out on a pilot scale in industrial facilities. The thermal stability of the productswas assessed conductimetrically. The block copolymers, that contained a low content ofPHPA (below 12 wt.%), showed thermal stability that was approximately three timesgreater than that of conventional PVC. Inverse gas chromatography study of the copoly-mers surface showed that there was a decrease in the dispersive component and greaterLewis acidity and basicity constants were observed relative to those of PVC. The thermalstabilisation of PVC when in the presence of PHPA is explained by the interactions betweenits functional groups and the structures formed during the thermal degradation. The ther-mal stability and the surface properties of PVC-b-PHPA-b-PVC were strongly dependent onthe molecular weight of the block copolymer. Lewis acid–base interaction parameters weredetermined and are interpreted as evidence of the PVC-b-PHPA-b-PVC compatibilisingfunction in PVC-wood flour composites.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Some PVC copolymers are long-established commercialproducts. The oldest is the vinyl chloride/vinyl acetate (VC/VAc) random copolymer. However, grafted copolymers,such as with ethylene/vinyl acetate (EVA), are also of com-mercial interest [1]. The chief properties that result fromthe presence of a significant proportion of a co-monomerin the vinyl chloride polymer chain are normally similar,but more permanent, to those of additives. The propertiesof these products include improvements to the weather-ability, impact resistance, viscosity control, thermal stabil-ity and adhesion to non-plastic surfaces [2].

. All rights reserved.

x:+351 239 798 703.

The discovery of living radical polymerisation (LRP)techniques opened the way for the preparation of well-de-fined block copolymers by controlling the radical polymer-isation reaction [3]. Since more than 80% of PVC isproduced by suspension polymerisation, the synthesis ofPVC-based copolymers, through LRP in an aqueous envi-ronment, is of great interest [4]. Single electron transfer/degenerative chain transfer mediated LRP (SET–DTLRP)[5–8] developed methods have been used in the synthesisof a-x-di(iodo)PVC and a-x-di(iodo)polyacrylates that canbe used either for the subsequent functionalisation of theirchain ends or as macroinitiators for the synthesis of ABAblock copolymers [9,10].

The ability to synthesise PVC block copolymers thatcontain acrylates, using SET–DTLRP methods, is extremelyimportant. It provides a basis, for the preparation of mate-rials with different morphologies that can be adapted to

3390 N. Rocha et al. / European Polymer Journal 45 (2009) 3389–3398

different chemical environments, and with differentmechanical properties. A conventional experimental canbe used as can an environmental friendly reaction med-ium. These points make the process attractive from theindustrial standpoint [11]. The SET–DTLRP technique hasbeen used to synthesise poly(vinyl chloride)-b-poly-(hydroxypropyl acrylate)-b-poly(vinyl chloride) (PVC-b-PHPA-b-PVC) on a pilot scale, using common industrialfacilities [12].

PHPA has been used as a co-monomer in the hydropho-bic/hydrophilic control of copolymeric systems [13,14] andin the creation of hydrogel systems [15]. Hydroxylpropylacrylate (HPA)-related monomers have been used in thesynthesis of amphiphilic block copolymers via other LRPmethods [16,17]. The inferior mechanical performance ofPVC and wood flour composites [18] can be improved byadding small amounts of PVC-b-PHPA-b-PVC to the com-posite formulations [12]. PVC-b-PHPA-b-PVC enables cou-pling between hydrophobic PVC and hydrophilic woodflour surfaces, due to creation of strong intermolecularforces.

The thermal degradation of PVC involves the sequentialloss of hydrogen chloride molecules, accompanied by thegeneration of conjugated polyene sequences [19,20]. Stud-ies in blends of PVC with a range of poly(methyl methacry-late)s (PMMAs) have shown a dependency of the PVCthermal stability on the methacrylate chemical structure[21]. Studies in PVC/EVA blends have related the availabil-ity of acetate groups to the capture of protons releasedduring the dehydrochlorination, preventing the auto-cata-lytic effect [22].

Inverse gas chromatography (IGC) is a method that canbe used to examine the surface characteristics of solids[23]. IGC has been of great interest to the evaluation ofinteraction parameters for the different components inPVC formulations [24,25] and the surface energy and Lewisacid–base behaviour of PVC-based materials [26–28]. Thetensile strength of cellulosic fibre composites that con-tained PVC has been correlated with the acid–base param-eters, determined by IGC, of each main component of thecomposite [29].

The addition of low contents (3.5 wt.%) of PVC-b-PHPA-b-PVC, rich in PHPA (� 90 wt.%), can enhance the perfor-mance properties of PVC and wood flour composites [12].This work concerns a study of the influence of low contentsof PHPA, in the block copolymer, on the thermal stability ofthe copolymers and on its surface properties. PVC-b-PHPA-b-PVC copolymers, rich in PVC, are attractive for use in tra-ditional PVC manufacturing and processing technology,especially when the PHPA content is low and yet sufficientto create the desired physical properties.

2. Experimental

2.1. Materials

Different PVC suspension polymerisation grades wereprovided by Cires, S.A. (Estarreja, Portugal). These samplesare identified as S (meaning suspension) followed by thedegree of polymerisation: 700, 950, or 1200. Pine wood

flour, Fibreton�40, was supplied by WTL International(Bosley, United Kingdom).

Hydroxypropyl acrylate (mixture of isomers, 95%), so-dium dithionite ðNa2S2O4Þ (85%), sodium bicarbonateðNaHCO3Þ (99+%), and iodoform ðCHI3Þ (99%) were pur-chased from Sigma–Aldrich and used as received.Hydroxypropyl methylcellulose (MF50) (1.86 wt.% aque-ous solution) and the partially hydrolysed poly(vinyl alco-hol) (PVA) (3 wt.% aqueous solution) were provided byCires, S.A., Portugal. Vinyl chloride (VC) was purchasedfrom ShinEtsu, Japan.

2.2. Synthesis of PVC-b-PHPA-b-PVC

The block copolymer was prepared in a two steps reac-tion. The first step was the synthesis of PHPA. The reactorwas charged with the compounds in the respectiveamounts that are presented in Table 1. For each batch,the recipe was adjusted to the desired degree of polymer-isation of PHPA (DP1). The recipe presented in Table 1 isdesigned to give synthesis of a macroinitator with degreeof polymerisation (DP) of 23, considering that the HPA con-version is 90% [30], in a copolymer containing 6.7 wt.% ofPHPA. The volume of HPA that was charged into the reactordepended also on the fraction of PHPA that was required inthe final copolymer, since a second charge was to be added,i.e. the amount depended on the volume with which thereactor was charged at the second stage. The reactionwas undertaken in a inert atmosphere, nitrogen (5 bar),over 6 h, at 25 �C, with an agitation speed of 375 rpm.

Then, an aliquot sample was taken. The reactor wasopened and charged with the recipe presented in Table 1(PVC-b-PHPA-b-PVC stage). The theoretical design, givenas an example above, is met when the added DP of VC is672 (DP2). VC was charged after the other compoundsand a vacuum was achieved inside the reactor. The reac-tion was undertaken over 20 h, at 42 �C, and with an agita-tion speed of 750 rpm.

This last stage made it possible for the completion ofthe synthesis of the copolymer PVC-b-PHPA-b-PVC to berealised. The product was left settling for two days. Thewater was removed and the product dried in an oven at70 �C for 48 h. Then the product was rinsed with distilledwater, until the washing water had shown no variation inits conductivity. Note that, reaction residues (catalyst andbuffer) are ionic. Thus, their presence in the washing watercan be detected by conductimetry.

2.3. Standard characterisation of synthesised materials

2.3.1. Fourier transform infrared spectroscopyFourier transform infrared spectroscopy (FTIR) spectra

were obtained with a Spectrum-One FT-IR Spectrometer(Perkin-Elmer) using an attenuated reflectance accessory,containing a diamond crystal. The products were analysedas prepared and the number of scans was 100.

2.3.2. K-value determinationSamples of 0:500� 0:0005 g of the product were

weighed and added to 100 mL volumetric flasks. The flaskswere filled with 80 mL of cyclohexanone and carefully stir-

Table 1Typical recipe for the synthesis of PVC-b-PHPA-b-PVC in a 6.5 L reactor.

Stage Role Compound Ratio Mole (mol) Charge

PHPA Monomer HPA 23 (DP1) 0.676 0.10 LInitiator CHI3 1 mol/(mol HPA/DP1) 0.0293 11.55 gCatalyst Na2S2O4 2 mol/(mol HPA/DP1) 0.0587 10.21 gBuffer NaHCO3 4 mol/(mol HPA/DP1) 0.117 9.86 gSuspension agent MF50 420 ppm (of HPA) 1.99 gSuspension agent PVA 980 ppm (of HPA) 2.87 gEnvironment Distilled water 3 L/(L HPA) 0.30 L

PVC-b-PHPA-b-PVC Monomer VC 672 (DP2) 19.7 1.25 LCatalyst Na2S2O4 8 mol/(mol VC/DP2) 0.235 40.86 gBuffer NaHCO3 1.45 mol/(mol VC/DP2) 0.0425 3.57 gSuspension agent MF50 1680 ppm (of VC) 111.28 gSuspension agent PVA 3920 ppm (of VC) 160.98 gEnvironment Distilled water 2 L/(L VC) 2.50 L

N. Rocha et al. / European Polymer Journal 45 (2009) 3389–3398 3391

red in a temperature controlled, glycerine-containing heat-ing bath, at 25� 1 �C. After complete dissolution of thesample, filtration through porous plaque G-1 was under-taken. Using an AVS 50 viscometer, the time constant forthe solvent and the sample were determined. To validatethe sample measurements, three determinations were car-ried out for each sample, considering that the maximumadmissible difference is 0.25%. This method is in accor-dance with the standard procedure ISO 1628–2:1998.

2.3.3. Size exclusion chromatographySize exclusion chromatography was undertaken using

three detectors (TriSEC). The chromatography parametersof the samples were determined using a HPSEC unit (HighPerformace Size Exclusion Chromatography); Viscotek(Dual detector 270, Viscotek, Houston) with differentialviscometry (DV); right-angle laser light scattering (RALLS,Viscotek) and refractive index (RI) (Knauer K-2301) detec-tion. The column set consisted of a PL 10-lL guard columnð50� 7:5 mm2Þ followed by two MIXED-B PL columnsð300� 7:5 mm2; 10 lmÞ. The tests were done at 30 �Cusing an Elder CH-150 heater. Before injection (100 lL),the samples were filtered through a PTFE membrane with0.2 lm pores.

2.3.4. Elemental analysisThe determination of carbon, hydrogen, nitrogen, sulfur,

and oxygen contents was undertaken in a flash combustionusing a Thermo Flash EA 1112 series.

2.4. Dehydrochlorination of samples

Dehydrochlorination of the samples was determined ina Metrohm 763 PVC Thermomat, an instrument for theautomatic determination of the thermal stability of PVCand related polymers. The samples were used as prepared,in powder form without additives. 0:5� 0:02 g of eachsample were placed in the reaction vessel and put in thethermomat, when the temperature had stabilised at170 �C. A nitrogen gas flow of 6.9 mL/min was used totransfer the evolved HCl in the reaction vessel to a measur-ing vessel filled with 60� 0:02 mL deionised water. Theconductivity of the solution in the measuring vessel, versustime, was measured using a conductimeter and recordedby a computer. Each sample was tested at least 3 times.

2.5. Inverse gas chromatography (IGC)

A DANI GC 1000 digital pressure control (DPC) gas chro-matograph, equipped with a hydrogen flame ionizationdetector (FID), was used for IGC data collection. Stainlesssteel columns, 0.5 m long and 0.4 mm ID, were degreased,washed and dried before packing. Depending on the den-sity, 2.5–5.0 g of sample were packed into the GC columnsusing a vacuum pump. The columns were shaped in asmooth ‘‘U” shape to fit the detector/injector geometry ofthe instrument. The packed columns were conditionedovernight at 50 �C, under a helium flow, before any mea-surements were made. This procedure was used to removeany volatiles, including water molecules, that might havebeen adsorbed on the stationary phase surface and that,consequently, could affect the retention of the probemolecules.

Experiments were carried out at a column temperatureof 50 �C. The injector and detector were kept at 180 and200 �C, respectively. Helium was used as carrier gas andits flow was selected to ensure that neither absorptionnor diffusion of the probes would occur inside the columnstationary phase. Small quantities of probe vapour (<1 lL)were injected into the carrier gas flow to ensure that theexperiments took place at infinite dilution. The probes,methane, n-hexane (C6), n-heptane (C7), n-octane (C8),n-nonane (C9), n-decane (C10), trichloromethane (TCM),tetrahydrofuran (THF) and ethyl acetate (ETA), were ofchromatographic grade and were used as received (Sig-ma–Aldrich). The retention times were the average of atleast three injections and were determined with the Con-der and Young method. This method provides a goodapproximation of the retention time, taken as the masscentre of the peak for asymmetrical peaks [31].

3. Results and discussion

3.1. Composition of the SET–DTLRP synthesised products

Fig. 1 shows the FTIR spectra of a conventional PVCsample (S950) and of an example of one of the synthesisedblock copolymers (L539-A). The functional groups of bothsegments (PVC and PHPA) are shown in the FTIR spectrumof the block copolymer. The presence of PHPA is confirmed

Fig. 1. FTIR spectra of the L539-A and S950.

3392 N. Rocha et al. / European Polymer Journal 45 (2009) 3389–3398

by the broad band at the 3600� 3100 cm�1 region, due tothe hydroxyl (OAH) stretching, and by the band at1720 cm�1, assigned to the carbonyl (C@O) stretching inthe ester functionality (ACH2ACOAOAR). The PVC seg-ments are represented by the stretching of aliphatic car-bon–chloride bonds (CACl) at 616 cm�1.

The various products that were synthesised in thiswork, either PVC homopolymers (L707, L901, and L920)or PVC-b-PHPA-b-PVC block copolymers (L210-A, L405-A,L539-A, L741-A, L759-A, L781-A, and L1075-A), are listedin Table 2. Table 2 includes the molecular weight, the poly-dispersity index (PDI), and the content of PHPA of each ofthe synthesised products. The molecular weight that ispresented is the one given by the K-value determination,since the processing conditions of PVC materials are gener-ally set on the basis of its K-value [1]. The PDI was deter-mined by the TriSEC technique and the PHPA contentwas determined by elemental analysis.

3.2. Thermal stability of PVC homopolymers and PVC-b-PHPA-b-PVC

3.2.1. PVC dehydrochlorination kineticsThe dehydrochlorination of PVC is initiated mainly by a

tertiary chlorine atom and from chloroallylic labile struc-tures, that arise as structural defects of PVC, created duringits polymerisation [20]. Since HCl is the only volatile prod-

Table 2Molecular weight, polydispersity index and PHPA content of each synthes-ised product.

Product Mn (kg/mol) PDI PHPA (wt.%)

L707 52.6 2.3 0.0L901 65.3 2.0 0.0L920 66.7 1.6 0.0L210-A 15.2 1.3 24:2� 3:3L405-A 30.1 3.3 5:1� 0:4L539-A 40.1 2.2 6:5� 0:5L741-A 53.7 1.6 4:7� 0:9L759-A 55.0 2.3 12:3� 1:4L781-A 58.1 2.1 12:3� 1:0L1075-A 77.9 1.7 10:0� 2:6

uct of PVC decomposition below 220 �C [32], monitoringthe amount of this acid that is evolved during the PVC ther-mal degradation allows determination of the rate of dehy-drochlorination of the PVC.

Using the PVC Thermomat, PVC materials were heatedin the presence of nitrogen gas. The conductivity of the ori-ginal distilled water, in a collection vessel, changes withtime, due to the dissolution of evolved HCl. The time atwhich the conductivity K begins to increase is the induc-tion time and the period when the conductivity reaches50 lS/cm, since the induction, is the stability time. Fig. 2shows the variation in the conductivity of the water, insidethe collection vessel, with time, in a Thermomat test, forthree samples of S950. A total of fourteen S950 test sam-ples gave an induction time of 309� 84 s and a stabilitytime of 1913� 66 s. PVC degradation has two regions oflinear increase of dehydrochlorination. The first is attrib-uted to the initiation mechanism and the second to thepropagation mechanism [33].

The conductivity of an electrolyte solution, K, dependson the charge of each specie, z, and on the mobility, U(7:913� 10�8 and 36:35� 10�8 m2=ðV � sÞ respectively forCl� and Hþ ions at 25 �C in a HCl aqueous solution [34]).It also depends on the concentration of ions in the solution,C, and on the Faraday constant, F (96485 C/mol), given byEq. 1. Considering that in the dehydrochlorination process,the protons Hþ and the anions Cl� are the only speciesbeing dissolved in the collection vessel, at the same rate,the concentration of dehyclorinated sites hHCl (expressedin lmol of HCl per gram of sample) can be given by Eq.2. Here, Vv is the volume of distilled water in the collectionvessel and ms is the mass of sample submitted to theThermomat test.

K ¼ F �X

i

jZij � Ci � Ui ð1Þ

hHCl ¼K

F � UCl� þ UHþð Þ �Vv

msð2Þ

From the Thermomat data obtained, it is possible torepresent the rate of dehydrochlorination as a function oftime, as shown in Fig. 3, demonstrating the two zero orderdehydrochlorination rate regions [33,35], and the transi-tion between these regions.

Fig. 2. Water conductivity in the collection vessel for S950 samples inThermomat test.

Fig. 3. Rate of dehydrochlorination of sample S950.

Fig. 5. Dependency of the stability time with the degree of polymerisa-tion of samples, SPVC, LPVC, and PVC-b-PHPA-b-PVC.

N. Rocha et al. / European Polymer Journal 45 (2009) 3389–3398 3393

Eq. 3 descibes the model for PVC dehydrochlorination[36]. Here, k is the rate constant of dehydrochlorinationand h0 is the concentration of possible sites for dehydro-chlorination (16.0 mmol/gPVC). The values obtained forthe rate constant k, of sample S950, were 3:2� 10�7 and3:7� 10�5 min�1, respectively, for the zero order dehydro-chlorination process, at the initiation and at the propaga-tion stages.

dhHCl

dt¼ k � h0 ð3Þ

3.2.2. Dehydrochlorination kinetics of PVC homopolymersand PVC-b-PHPA-b-PVC copolymers

Fig. 4 shows the Thermomat data that were obtained forthe PVC homopolymers and for the PVC-b-PHPA-b-PVCblock copolymers. Fig. 5 shows the dependency of the sta-bility time on the degree of polymerisation of the differentsamples, predicted from the K-value analysis. The SPVCproducts are different commercial PVC samples, polymer-ised via conventional suspension polymerisation. The LPVCproducts are the homopolymer samples and PVC-b-PHPA-b-PVC the block copolymer samples, that were synthesisedvia LRP technique.

Comparing the SPVC samples (S700, S950, and S1200),the S1200 sample shows an increase in its thermal stabilitythat is related to the lower polymerisation temperature

Fig. 4. Thermomat profiles obtained for the PVC homopolymers and forthe PVC-b-PHPA-b-PVC block copolymers.

that was used [20]. However, the thermal stability timevariation with the molecular weight for SPVC is much lessthan that for the LRP products.

The stability time of the LRP products, either as thehomopolymer or as the copolymer, is directly related tothe molecular weight of the sample. The stability time ofthe LPVC products is much less than that of the PVC-b-PHPA-b-PVC block copolymers of similar molecularweight.

The stability time of PVC-b-PHPA-b-PVC block copoly-mers increases dramatically with their molecular weight.For the higher molecular weight copolymers, the stabilitytime is between 3 and 4 times that of the conventionalPVC products. Only for the very low molecular weightcopolymers, is the thermal stability inferior to that of theconventional PVC products.

Comparing samples that have different characteristics,such as conventional PVC (S950), the LRP PVC homopoly-mer (L901), and the LRP copolymer PVC-b-PHPA-b-PVC(L781-A), but have similar molecular weights, one can con-clude that the copolymers have greater thermal stabilitythan any of the PVC homopolymers.

If the criterion for stability time is amended, accordingto the PHPA content of L781-A (12.0 wt.%), then the stabil-ity time should be amended to a variation in the conductiv-ity of 44.0 lS/cm (see Eq. 1). This would lead to a stabilitytime of 4980� 130 s that, in comparison with that of5340� 140 s, obtained considering a criterion of 50 lS/cm, gives to an error of 6.7%. The obtained amendmentfor the block copolymers, based on the PHPA content ofthe copolymers, are much less than the differences in sta-bility time due to the molecular weight influence. More-over, the gain obtained in the thermal stability is muchgreater than that obtained with PVC homopolymers withmolecular weights that are typically used in products de-signed for applications that require the product to be rigid,S700 and S950.

Table 3 shows the values for the initiation rate con-stants and for the propagation rate constants obtainedfor the different samples created in this study. The closevalues of the standard deviations and the initiation rateconstants makes it extremely difficult to distinguish

Table 3Initiation and propagation rate constants for different samples.

Group Batch k=ð10�7 �min�1Þ

Initiation Propagation

SPVC S700 4:25� 0:23 306� 6S950 3:22� 1:57 368� 17S1200 6:10� 1:96 251� 6

LPVC L707 6:73� 2:48 615� 44L901 7:23� 0:04 341� 27L920 7:32� 1:05 351� 29

Block copolymers L210-A 15:07� 4:19 221� 34L405-A 12:99� 5:36 310� 34L539-A 3:27� 0:95 237� 27L741-A 2:34� 1:46 181� 23L759-A 3:45� 0:95 151� 19L781-A 2:87� 1:68 148� 3L1075-A 1:91� 1:73 129� 16

Table 4Properties of probes used in the calculation of the surface properties by IGC.

Probe A ðÅ2Þ cdl ðmJ �m�2Þ AN� (kJ/mol) DN (kJ/mol)

C6 51.5 18.4C7 57.0 20.3C8 63.0 21.3C9 69.0 22.7C10 75.0 23.4THF 45.0 22.5 2.09 83.7TCM 44.0 25.9 22.6 0ETA 48.0 19.6 6.3 71.6

3394 N. Rocha et al. / European Polymer Journal 45 (2009) 3389–3398

between the initiation mechanisms of the samples. How-ever, those copolymers with a very low molecular weightseem to have a much greater initiation dehydrochlorina-tion rate. The PVC homopolymers that were synthesisedvia LRP seem to have a slightly greater initiation rate con-stant than does conventional PVC. The propagation rateconstants show better conformity with the stability timevalues. The copolymers have a propagation rate constantthat is lower than that of the PVC homopolymers, even ifa correction is applied for the PHPA content (see Table 2),decreasing with the molecular weight. These resultsstrongly suggest the operation of a stabilising effect ofPHPA groups on the dehydrochlorination process.

This increased thermal stability of the copolymers canbe explained by consideration of the interaction of thechemical functionalities. These interactions were intro-duced by the PHPA segments and influenced by blockcopolymer molecular weight design. The interactions tookplace with the PVC molecular segments, during the dehy-drochlorination process.

3.3. Surface properties determined by IGC

IGC analysis was undertaken on selected samples (S950,L405-A, L539-A, and L1075-A), in an attempt towards gain-ing a greater understanding of the chemical interactionsand the surface properties, that resulted from the copoly-mer design, in the PVC thermal stabilisation process.

IGC has been used to investigate the susceptibility ofmaterials towards dispersive and acid–base intermolecularinteractions, due to its ability to be used in monitoring sur-face adsorption on polymer surfaces [37]. Estimation of thedispersive component of the surface energy is determinedaccording to the level of adsorption of apolar probes (n-al-kanes), whereas the specific interaction capacity (non-dis-persive forces) of a solid surface is obtained by comparisonof the adsorption of apolar and polar probes [38].

The physical properties of the probes used in this workare listed in Table 4 [39–41]. Here, A is the surface area ofthe probe, cd

l is the dispersive component of the surface en-ergy of the probe, and AN� and DN are, respectively, theGuttman’s electron acceptor and electron donor numbersof the polar probes.

Table 5 presents details of the operating conditions ofthe IGC experiments for each sample. Also presented aredetails of the retention times, obtained by application ofthe Conder and Young method [31] for each probe, in eachtested sample (S950, L405-A, L539-A, L1075-A).

The net retention volume Vn, which is the volume of in-ert carrier gas that is necessary to push the probe moleculethrough the chromatographic column containing the solidsample, depends on the sample–probe interactions and canbe calculated from IGC data using Eq. 4. Here, tr is theretention time of the injected probe through the column,t0 is the retention time of the non-interacting probe (meth-ane), F is the flow rate of the inert carrier gas (measuredwith a digital flow meter), and J is the James–Martin com-pression correction factor. J correction term is determinedby Eq. 5. Here, P1 is equal to Pa (atmospheric pressure) plusthe pressure drop in the column [41].

Vn ¼ðtr � t0Þ � F � J ð4Þ

J ¼32

1� P1Pa

� �2

1� P1Pa

� �3 ð5Þ

3.3.1. Determination of the dispersive component of thesurface energy

The net retention volume Vn can be related to the dis-persive components of the interacting solid and probe,respectively, cd

s and cdl , by Eq. 6 [42]. Here, NA is the Avoga-

dro number, R is the gas constant, T is the column absolutetemperature. The constant C is dependent on the chosenreference state [43].

R � T � lnðVnÞ ¼ NA � A � 2ffiffiffiffiffiffiffiffiffiffiffiffifficd

l � cds

qþ C ð6Þ

According to Eq. 6, it is possible to estimate the disper-sive component of the surface of each sample, from theslope of the linear fit of R � T � lnðVnÞ as a function ofNA � A � 2ðcd

l Þ1=2, using the IGC data obtained with the apolar

probes. The results are shown in Table 6. The dispersivecomponent of the surface energy represents the potentialof materials to undergo London dispersion interactions. Atrend towards a decrease in the apolarity in PVC materialsthat had been combined with polar materials has been re-ported [44,45]. In PVC-b-PHPA-b-PVC, the PHPA segmentscontribute greatly to the non-dispersive interactions sincethey contain basic groups (ester functionality) and acidicgroups (hydroxyl functionality) [42]. PVC is moderatelypolar and the carbon–chlorine bond has a special attractionto the carbonyl bond [46]. The presence of carbonyl groups

Table 5IGC data for the different probes in S950, L405-A, L539-A, and L1075-A.

Parameter/Probe S950 L405-A L539-A L1075-A

P1 (bar) 1.25 1.60 1.60 1.70F �C (mL/min) 19.0 22.4 22.3 19.4Retention time (min) Methane 0:187� 0:001 0:262� 0:003 0:291� 0:002 0:323� 0:003

C6 0:292� 0:003 0:350� 0:007 0:372� 0:001 0:408� 0:005C7 0:428� 0:003 0:482� 0:004 0:481� 0:002 0:515� 0:002C8 0:765� 0:006 0:706� 0:004 0:735� 0:001 0:769� 0:002C9 1:606� 0:013 1:280� 0:007 1:319� 0:005 1:332� 0:005C10 3:756� 0:024 2:646� 0:011 2:747� 0:012 2:715� 0:013TCM 0:748� 0:018 1:017� 0:012 1:041� 0:018 1:888� 0:046THF 1:103� 0:013 1:331� 0:025 1:279� 0:029 3:177� 0:031ETA 1:065� 0:023 1:311� 0:004 1:978� 0:081 3:754� 0:179

Table 6Values of cd

s for the different samples.

Sample cds ðmJ=m2Þ Correlation coefficient Decay (%)

S950 30.7 0.9997 –L405-A 26.0 0.9994 15.3L539-A 28.5 0.9998 7.3L1075-A 27.4 0.9998 10.7

N. Rocha et al. / European Polymer Journal 45 (2009) 3389–3398 3395

(in the ester functionality) of PHPA can induce dipoles inthe moderately polar PVC, taking advantage of the pres-ence of the carbon–chlorine bond. This explains the decayin the dispersive component in the block copolymers, inrelation to that of PVC.

3.3.2. Determination of the acidity and basicity of surfacesIf a Lewis acid–Lewis base interaction occurs, as is the

case with polar probes, there will be a corresponding spe-cific component contribution, in addition to the dispersivecomponent, to the overall free energy of adsorption, DGa

[42]. The overall free energy of adsorption, DGa, is relatedto the net retention volume by Eq. 7 [43]. Therefore, thefree energy of adsorption that is caused by specific interac-tions, DGsp

a , can be estimated by calculating the differencebetween the values of R � T � lnðVnÞ, obtained for the polarprobes and the corresponding estimation for the apolarprobe. The graphical method is represented in Fig. 6.

DGa ¼ �R � T � lnðVnÞ þ C ð7Þ

Fig. 6. Graphical determination of the free energy of adsorption of thespecific interactions for L1075-A.

The enthalpy of adsorption, DHspa , due to specific inter-

actions between a probe (characterised by its Guttman’selectron acceptor, AN�, and electron donor, DN, numbers)and a solid surface (characterised by its acidity KA (Lewisacidity constant) and its electron donor ability KB (Lewisbasicity constant)) is given by Eq. 8 [43]. Since DGsp

a is pro-portional to DHsp

a ;KA and KB can be graphically determinedby plotting DGsp

a =AN� versus DN=AN� [38,43]. According toEq. 8, KA is obtained as the slope of the linear fit, whereasKB is the origin of such plot.

DHspa ¼ KA � DN þ KB � AN� ð8Þ

The results obtained for the different samples are pre-sented in Table 7. As expected from the chemical structureof PHPA, both the acidity constant and the basicity con-stant increase in the copolymer. Block copolymers with ahigher molecular weight means there will be longer seg-ments of each co-monomer. Thus, both the acidity and ba-sicity increase with the length of the PHPA segments inPVC-b-PHPA-b-PVC.

Since the basicity of the copolymer is mostly due to theester groups, the results show that the ester functionalityis highlighted in copolymers with longer PHPA segments.The increase in the KB is greater from L405-A to L539-Athan from L539-A to L1075-A. Indeed, the copolymer chainseems to need a critical length to affect the specific basicityof the surface. Similar behaviour can be found for the Lewisacidity constant KA (determined by the hydroxyl groupsfrom the PHPA segments). However, the great increase inKA occurs for even higher molecular weight copolymers.

As predicted from the thermal stability studies, themolecular weight of the copolymer seems to have a con-siderable influence on the manner in which these func-tional groups are distributed and on the way in whichthey interact at the surface of the copolymer particles. Thischange in the way the functional groups that are intro-duced by PHPA interact with PVC may provide an explana-

Table 7Acidity and basicity constants obtained for the different samples in study.

Sample KA KB Correlation coefficient

S950 0.068 0.180 1.000L405-A 0.075 0.226 1.000L539-A 0.074 0.341 0.996L1075-A 0.106 0.360 1.000

3396 N. Rocha et al. / European Polymer Journal 45 (2009) 3389–3398

tion for the increase in the thermal stability in the blockcopolymers.

Table 8Lewis acid–base interaction number for composites containing 40 wt.% ofPine wood flour and different polymer matrices.

Matrix ISP

S950 0.011L405-A 0.014L539-A 0.021L1075-A 0.031

3.4. Influence of PHPA/PVC interactions on the thermalstabilisation of PVC

HCl acceptance and the ability to react with labile chlo-roallylic groups [20] are among the more important fea-tures of any chemical substance that is designed to beused as thermal stabiliser of PVC. Metal carboxylates andepoxy compounds are effective heat stabilisers in PVC for-mulations. They prevent the formation of conjugated poly-ene sequences [19].

Studies of PVC and poly(methyl methacrylate) (PMMA)blends [47] have established that, in early stages, there isdecomposition of PMMA, due to reaction with HCl thatarises from PVC dehydrochlorination. Afterwards, the pro-duction of HCl is delayed. According to these studies, asimilar reaction between HCl and PHPA is expected to oc-cur. The resulting products have a molecular structure thatis similar to that of epoxy (anhydride structure) and car-boxylate (carboxylic acid structure) compounds that arecommonly used as PVC thermal stabilisers.

Based on the mechanisms presented in the literature, forPMMA decomposition in the presence of HCl and on themechanism of PVC stabilisation by reaction of labile chloro-allylic groups with epoxy [48] and carboxylate [49] com-pounds, the improved thermal stability of PVC-b-PHPA-b-PVC, in relation to that of PVC, can be explained by the reac-tion of the PHPA degradation products with the labile chloro-allylic groups. The initiation of dehydrochlorination in PVC-b-PHPA-b-PVC, the rate constants of which are in the samerange of that of PVC alone (see Table 3), would lead to thedegradation of PHPA. PHPA-degraded structures are effectivePVC thermal stabilisers through the substitution of labilegroups. This explains the decrease in the propagation rateconstant of the dehydrochlorination of PVC-b-PHPA-b-PVC.However, it cannot explain the decrease that occurs withchange in the block copolymer molecular weight.

Braun et al. [21] have studied PVC/PMMAs blends. Theyhave shown that the PVC thermal stability increases withthe length of the ester group sequence. For longer esterside chains, the PMMAs particles would be more able tocapture the HCl that arises from PVC dehydrochlorination.A similar explanation, based on the availability of stabilis-ing functional groups, was given by Monteiro and Thau-maturgo [22]. Since the thermal stability of PVC blends isincreased by the length and availability of the ester groups,the increase in the thermal stability of PVC-b-PHPA-b-PVC,with its molecular weight, could be predicted. It is ex-pected that when the copolymers have longer PHPA cen-tral blocks, more PHPA units would be available for PVCstabilisation, either by the formation of anhydride struc-tures (extremely favoured by the proximity of acrylategroups) or by the formation of carboxylic acid structures.The increased availability of such ester groups, in thecopolymers that are of higher molecular weight, is high-lighted by the increase in the Lewis basicity constant,determined in IGC experiments. The increase in the pre-dominance of the ester functionality leads to a significant

decrease in the dehydrochlorination propagation rateconstant.

Copolymers containing methacrylate and hydroxylgroups have successfully been used as co-stabilisers [50].The stabilising effect was related to interactions with thethermal stabiliser. The acidic behaviour of hydroxyl groupsthat are present in PHPA can also contribute to the stabili-sation of chloroallylic labile groups of PVC, though to a les-ser extent. The mechanism of operation is similar to that ofthe carboxylate groups. The increase in the Lewis acidityconstant shows that there is an increased tendency to-wards losing the proton at the hydroxyl functionality,improving the thermal stabilisation of the PVC, either asthe main stabiliser or as a co-stabiliser.

3.5. Lewis acid–base interaction parameters

Considering that PVC-b-PHPA-b-PVC has been success-fully used in PVC-wood flour composites, to improve thecompatibility between PVC and wood flour [12], then theincreases in the Lewis acidity constant and in the Lewis ba-sicity constant may be a good indication of stronger inter-actions with wood flour.

Using the same IGC method as that described for thePVC and for the copolymer samples, KA and KB were deter-mined for the Pine wood flour sample. The values obtainedwere 0.050 and 0.058, for KA and KB, respectively, with acorrelation coefficient of 0.996.

According to Matuana et al. [29], the magnitude of thespecific interaction, given by the Lewis acid–base interac-tion number ISP , between the wood flour and the polymermatrix in wood and polymer composites, can be estimatedby use of Eq. (9). Here, the superscripts m and f representthe matrix and the filler, respectively. P is the proportionof the filler (wood flour) or the matrix (polymer) in thecomposite. Stronger interactions are expected for higherinteraction numbers.

ISP ¼ 2 � ðPmÞ2 � ðKmA � K

mB Þ

þ 12� ðPmÞ � ðPf Þ � ðKm

A � KfB þ Km

B � KfAÞ ð9Þ

þ 2 � ðPf Þ2 � ðKfA � K

fBÞ

Table 8 presents the estimated Lewis acid–Lewis baseinteraction number for composites with 40 wt.% of Pinewood flour and S950, or different copolymers, as the ma-trix of the composite. The results show that, as expected,the chemical adhesion between the wood flour and matrixincreases when using the copolymer as the matrix insteadof PVC. The copolymer with the higher molecular weight,L1075-A, should give a greater degree of adhesion to wood

N. Rocha et al. / European Polymer Journal 45 (2009) 3389–3398 3397

flour among the considered matrices quoted in Table 8.L1075-A has a value that nearly triples the value of ISP ofthe composite that contains the PVC homopolymer matrix.

From the standpoint of the optimisation of the perfor-mance properties of solid composite mixtures, Lewis acid–Le-wis base interaction numbers can give important leads withrespect to optmising the adhesion in a range of compositions.

4. Conclusions

PVC homopolymers and PVC-b-PHPA-b-PVC blockcopolymers were synthesised via SET–DTLRP, on a pilotscale, using standard industrial conditions. Low contentsof PHPA were used in the block copolymer in order to cre-ate products with most of the properties being close tothose of PVC homopolymers, but with enough change tomodifify relevant properties, such as the thermal stabilityand the surface interactions.

Small differences were found between the thermal sta-bility of the SET–DTLRP synthesised PVC homopolymersand the commercially available PVC. However, the syn-thesised block copolymers have a thermal stability that ismuch greater than that of conventional PVC. The stabilitytime increases linearly with the molecular weight, calcu-lated on the basis of its K-value. For molecular weights thatare close to that of the conventional PVC that is used in ri-gid applications, the block copolymer reaches between 2and 3 times the stability time of conventional PVC. Calcu-lation of the rate constants shows that differences in thedehydrochlorination are more evident at the propagationstage than at the initiation stage.

IGC studies have shown that dispersive component ofthe surface energy, due to London interactions, decreasesin the copolymers. This decrease is attributed to the inter-action of polar PHPA with moderately polar PVC. Regardingthe specific component of the surface energy, the copoly-mers have greater Lewis acidity and Lewis basicity con-stants than those of PVC. This effect is attributed to thehydroxyl and the ester functionalities, respectively. Thedominance of these functionalities increases with thePVC-b-PHPA-b-PVC molecular weight.

According to proposed mechanisms for acrylate func-tion degradation, during PVC dehydrochlorination, thePHPA degraded structures become effective PVC thermalstabilisers by reacting with PVC labile structures. There-fore, the predominance of PHPA functionalities in thosecopolymers with higher molecular weights increases itseffectivness in the thermal stabilisation of PVC.

The increase in the Lewis acidity and Lewis basicity con-stants of the block copolymers, in comparison with thoseof the PVC homopolymer, suggests that stronger interac-tions with the surface of other materials, such as woodflour will arise. This indicates that such block copolymerswould be suitable for use as coupling agents in PVC andwood flour composites.

Acknowledgements

The authors gratefully acknowledge the support ofCires, S.A., Portugal and funding from Fundação para aCiência e Tecnologia, Portugal (SFRH/BD/25701/2005).

References

[1] Titow W. PVC technology. Elsevier Applied Science; 1984.[2] Wypych G. PVC formulary. ChemTec Publishing; 2009.[3] Hadjichristidis N, Pispas S, Floudas G. Block copolymers: synthetic

strategies, physical properties, and applications. John Wiley & Sons,Inc.; 2003.

[4] Wannemacher T, Braun D, Pfaender R. Novel copolymers vianitroxide mediated controlled free radical polymerization of vinylchloride. Macromol Symp 2003;202:11–23.

[5] Percec V, Popov AV, Ramirez-Castillo E, Monteiro M, Barboui B,Weichold O, et al. Aqueous room temperature metal-catalysed livingradical polymerization of vinyl chloride. J Am Chem Soc2002;124:4940–1.

[6] Percec V, Popov AV, Ramirez-Castillo E, Weichold O. Living radicalpolymerization of vinyl chloride initiated with iodoform and catalysedby nascent cu0/tris(2-aminoethyl)amine or polyethyleneimine in waterat 25 �C proceeds by a new competing pathways mechanism. J Polym SciA 2003;41:3283–99.

[7] Percec V, Popov AV, Ramirez-Castillo E, Coelho JFJ, Hinojosa-FalconLA. Non-transition metal-catalyzed living radical polymerization ofvinyl chloride initiated with iodoform in water at 25 �C. J Polym Sci A2004;42:6267–82.

[8] Percec V, Ramirez-Castillo E, Popov AV, Hinojosa-Falcon LA,Guliashvili T. Ultrafast single electron transfer (SET) – degenerativechain transfer (DT) mediated living radical polymerization (SET–DTLRP) of acrylates initiated with iodoform in H2O at roomtemperature and catalysed by Na2S2O4. J Polym Sci A 2005;43:2178–84.

[9] Percec V, Guliashvilli T, Popov AV, Ramirez-Castillo E. Synthesis ofpoly(methyl methacrylate)-b-poly(vinyl chloride)-b-poly(methylmethacrylate) block copolymers by CuCl/2,2-bipyridine-catalyzedliving radical block copolymerization initiated from di(iodo)poly(vinyl chloride) prepared by single-electron-transfer/degenerative chain-transfer mediated living radical polymerization.J Polym Sci A 2005;43:1478–86.

[10] Percec V, Popov AV, Ramirez-Castillo E, Hinojosa-Falcon LA.Synthesis of poly(vinyl chloride)-b-poly(2-ethylhexyl acrylate)-b-poly(vinyl chloride) by competitive single-electron-transfer/degenerative-chain-transfer mediated living radical polymerizationof vinyl chloride initiated from di(iodo)poly(2-ethylhexyl acrylate)and catalyzed with sodium dithionite in water. J Polym Sci A2005;43:2276–80.

[11] Coelho JFJ, Silva AMFP, Popov AV, Percec V, Abreu MV, GonçalvesPMOF, et al. Synthesis of poly(vinyl chloride)-b-poly(n-butylacrylate)-b-poly(vinyl chloride) by the competitive single-electron-transfer/degenerative-chain-transfer-mediated living radicalpolymerization in water. J Polym Sci A 2006;44:3001–8.

[12] Rocha N, Coelho JFJ, Fonseca AC, Kazlauciunas A, Gil MH, GonçalvesPM, et al. Poly(vinyl chloride) and wood flour composites: newbonding strategies. J Appl Polym Sci 2009;113:2727–38.

[13] Taylor LD, Cerankowski LD. Preparation of films exhibiting abalanced temperature dependence to permeation by aqueoussolution - a study of lower consolute behavior. J Polym Sci PolChem 1975;13:2551–70.

[14] Eggenhuisen TM, Becer CR, Fijten MWM, Eckardt R, Hoogenboom R,Schubert US. Libraries of statistical hydroxypropyl acrylatecontaining copolymers with LCST properties prepared by NMP.Macromolecules 2008;41:5132–40.

[15] Perera DI, Shanks RA. Networks characteristics of homopolymer andsome copolymers of poly(2-hydroxyethyl methacrylate). Polym Int1995;37:133–9.

[16] Mühlebach A, Gaynor SG, Matyjaszewski K. Synthesis of amphiphilicblock copolymers by atom transfer radical polymerisation (ATRP).Macromolecules 1998;31:6046–52.

[17] Bian K, Cunningham MF. Nitroxide-mediated living radical polymerizationof 2-hydroxyethyl acrylate and the synthesis of amphiphilic blockcopolymers. Macromolecules 2005;38:695–701.

[18] Rocha N, Kazlauciunas A, Gil MH, Gonçalves PM, Guthrie JT.Poly(vinyl chloride)-wood flour press mould composites: theinfluence of raw materials on performance properties. Compos PartA-Appl S 2009;40:653–61.

[19] Wypych G. PVC degradation & stabilization. ChemTec Publishing;2008.

[20] Starnes WH. Structural defects in poly(vinyl chloride). J Polym Sci A2005;43:2451–67.

[21] Braun D, Böhringer B, Eidam N, Fischer M, Kömmerling S.Degradation and stabilization of PVC in blends. Macromol MaterEng 1994;216:1–19.

3398 N. Rocha et al. / European Polymer Journal 45 (2009) 3389–3398

[22] Monteiro EEC, Thaumaturgo C. Surface phenomena and polymermiscibility of PVC/EVA blends. Compos Sci Technol 1997;57:1159–65.

[23] Fowkes FM. Quantitative characterization of the acid–baseproperties of solvents, polymers, and inorganic surfaces. J AdhesSci Technol 1990;4:669–91.

[24] Su CS, Patterson D, Schreiber HP. Thermodynamic interactions andthe properties of PVC-plasticizer systems. J Appl Polym Sci1976;20:1025–34.

[25] Schreiber HP, Degrendele C, Bataille P. Component interactions andproperties of crosslinked vinyl systems. Polym Eng Sci1990;30:981–8.

[26] Fowkes FM, Tischler DO, Wolfe JA, Lanning LA, Ademu-John CM,Halliwell MJ. Acid–base complexes of polymers. J Polym Sci PolChem 1984;22:547–66.

[27] Chehimi MM, Abel ML, Perruchot C, Delamar M, Lascelles SF, ArmesSP. The determination of the surface energy of conducting polymersby inverse gas chromatography at infinite dilution. Synthetic Met1999;104:51–9.

[28] Slimane AB, Boukerma K, Chabut M, Chehimi MM. An inverse gaschromatographic characterization of polypyrrole-coated poly(vinylchloride) powder particles. Colloid Surface A 2004;240:45–53.

[29] Matuana LM, Woodhams RT, Balatinecz JJ, Park CB. Influence ofinterfacial interactions on the properties of PVC/cellulosic fibercomposites. Polym Composite 1998;19:446–55.

[30] Coelho JFJ, Silva AMFP, Popov AV, Percec V, Abreu MV, GonçalvesPMOF, et al. Single electron transfer-degenerative chain transferliving radical polymerization of n-butyl acrylate catalyzed byNa2S2O4 in water media. J Polym Sci A 2006;44:2809–25.

[31] Conder J, Young C. Physicochemical measurement by gaschromatography. Wiley; 1979.

[32] Ballistreri A, Foti S, Montaudo G, Scamporrino E. Evolution ofaromatic compounds in the thermal decomposition of vinylpolymers. J Polym Sci Pol Chem 1980;18:1147–53.

[33] Fisch MH, Bacaloglu F. Kinetics and mechanism of the thermaldegradation of poly(vinyl chloride). J Vinyl Addit Techn1995;1:233–40.

[34] Alberty R. Physical chemistry. John Wiley & Sons; 1987.[35] Atakul S, Balköse D, Ülkü S. Synergistic effect of metal soaps and

natural zeolite on poly(vinyl chloride) thermal stability. J Vinyl AdditTechn 2005;11:46–56.

[36] Dean L, Dafei Z, Deren Z. Mechanism and kinetics of thermo-dehydrochlorination of poly(vinyl chloride). Polym Degrad Stabil1988;22:31–41.

[37] Chehimi MM, Abel ML, Sahraoui Z. An inverse gas chromatographicstudy of the PMMA/conducting polypyrrole interface. J Adhes SciTechnol 1996;10:287–303.

[38] Balard H, Brendle E, Papirer E. Acid–base interactions: relevance toadhesion science and technology, vol. 2. VSP International SciencePublishers; 2000. pp. 299–316.

[39] Riddle FL, Fowkers FM. Spectral shifts in acid–base chemistry. 1. Vander Waals contributions to acceptor numbers. J Am Chem Soc1990;112:3259–64.

[40] Kamdem DP, Bose SK, Luner P. Inverse gas chromatographycharacterization of birch wood meal. Langmuir 1993;2:3039–44.

[41] Chtourou H, Riedl B, Kokta BV. Surface characterization of modifiedpolyethylene pulp and wood pulps fibers using xps and inverse gaschromatography. J Adhes Sci Technol 1995;9:551–74.

[42] Santos J, Guthrie J. Analysis of interactions in multicomponentpolymeric systems: the key-role of inverse gas chromatography. MatSci Eng R 2005;50:79–107.

[43] Mukhopadhyay P, Schreiber H. Aspects of acid–base interactions anduse of inverse gas chromatography. Colloid Surface A1995;100:47–51.

[44] Michalski MC, Hardy J, Saramago BJV. On the surface free energy ofPVC/EVA polymer blends: comparison of different calculationmethods. J Colloid Interf Sci 1998;208:319–28.

[45] D’yakova A, Trifonov SA, Sosnov EA, Malygin AA. Effect of chemicalmodification on structural and energy characteristics of the surfaceof polyethylene and polyvinyl chloride films. Russ J Appl Chem2009;82:622–9.

[46] Wilkes CE, Summers JW, Daniels CA. PVC handbook. Hanser Verlag;2005.

[47] McNeill IC, Neil D. Degradation of polymer mixtures. III. Poly(vinylchloride)/poly(methyl methacrylate) mixtures, studied by thermalvolatilization analysis and other techniques. The nature of thereaction products and the mechanism of interaction of the polymers.Eur Polym J 1970;6:569–83.

[48] Anderson DF, McKenzie DA. Mechanism of the thermal stabilizationof poly(vinyl chloride) with metal carboxylates and epoxyplasticizers. J Polym Sci A-1 1970;8:2905–22.

[49] Frye AH, Horst RW. The mechanism of poly(vinyl chloride)stabilization by barium, cadmium, and zinc carboxylates. I.Infrared studies. J Polym Sci 1959;40:419–31.

[50] Braun D, Belik P, Richter E. Polymeric costabilisers for poly(vinylchloride). Macromol Mater Eng 1999;268:81–6.