Waste Management 32 (2012) 847–851

Contents lists available at SciVerse ScienceDirect

Waste Management

journal homepage: www.elsevier .com/locate /wasman

Kinetics of thermal de-chlorination of PVC under pyrolytic conditions

Alexandra Castro a,b,⇑, Delfim Soares b, Cândida Vilarinho b, Fernando Castro b

a CVR – Centro para a Valorização de Resíduos (Center for Waste Valorization), Guimarães, Portugalb CT2M – Centre for Mechanical and Materials Technologies, Mechanical Engineering Department of Minho University, Guimarães, Portugal

a r t i c l e i n f o

Article history:Received 25 February 2011Accepted 6 January 2012Available online 7 February 2012

Keywords:De-chlorinationPVCPyrolysisKinetics

0956-053X/$ - see front matter � 2012 Elsevier Ltd. Adoi:10.1016/j.wasman.2012.01.004

⇑ Corresponding author at: CVR – Centro para a Valfor Waste Valorization), Guimarães, Portugal. Tel.:253510029.

E-mail address: acastro@cvresiduos.pt (A. Castro).

a b s t r a c t

Although PVC-containing wastes are an important potential source of energy they are frequently dis-posed in landfill. In thermal treatment processes such as pyrolysis and gasification, the presence of poly(-vinyl chloride) (PVC), a compound with 56.7% of chlorine, may cause problems concerned withenvironmental protection, as consequence of the formation of hydrochloric acid, chlorine gas and dioxins,as well as corrosion phenomena of the reactor/equipment materials. Thus, a possible solution mayinvolve a previous removal of the chlorine from PVC containing waste through a pyrolysis process atlow temperatures before the material being submitted to a subsequent thermal treatment, for energeticvalorization. In this work, a kinetic model for the thermal decomposition of PVC has been developed, inview of its de-chlorination. DTA/TGA testing at different temperatures indicated a first order reaction andan activation energy of 133,800 J/mol. An almost completed de-chlorination reaction was obtained at340 �C under an inert atmosphere. The resulted material is a CnHn type polymer with potential to be usedin an energy recovery process. Validation test performed at laboratory scale indicate that the temperatureof 340 �C enables the removal of �99.9% of the chlorine present in PVC. The chloride can be fixed in theform of an aqueous solution of HCl or calcium chloride, driving to an alternative full process with envi-ronmental benefits and reduction of the costs associated to the PCV – containing materials/wastesmanagement.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction the waste composition, the temperature, the pressure and the reac-

The scarcity of the petroleum resources constitutes the drivingforce for the development of innovative methodologies for theenvironmentally sustainable production of fuels and energy. It be-comes necessary to harness the potential energy of wastes,whether industrial or municipal, through the use of valorizationtechniques, for the production of both energy and alternative fuels.

Despite the incineration is the most applied technique when itconcerns to energetic valorization, two other techniques, pyrolysisand gasification, can be used for this purpose presenting someadvantages when compared to the previous one.

Those technologies differ from each other mainly by the pres-ence and amount of oxidizing agents and the operational parame-ters implemented influence the final products of reaction as well asits subsequent use or application.

Pyrolysis is a process of irreversible chemical modification ofcompounds under the action of heat and in the absence of oxygen,an energy valorization technique, which causes thermal degrada-tion (Lewis and Ablow, 1976). The nature of the products obtainedthrough the pyrolysis process is function of several factors, such as

ll rights reserved.

orização de Resíduos (Center+351 253510025; fax: +351

tion/residence time in the pyrolysis reactor. In the pyrolysis pro-cess, the reaction is endothermic and therefore requires anexternal source of heat to initiate the reaction, always on an indi-rect way, neither oxygen nor water vapor are being ever injectedon the reactor.

The presence of organic compounds on solid wastes, especiallypolymers, is considered an important potential source of energy.However, some of these materials contain polyvinyl chloride(PVC) in its constitution. The presence of PVC in wastes results inrecycling constraints when a thermal valorization process is con-sidered as an option for its treatment (Borgianni et al., 2002), pre-venting their use for energy recovery (Kamo et al., 1996;Zevenhoven et al., 2002). On wastes treatment through thermalprocesses, such as pyrolysis, the presence of PVC in wastes contain-ing high levels of chlorine, promotes serious pollution problems,mainly due to the formation of hydrochloric acid, chlorine gasand dioxins (Borgianni et al., 2002). Therefore, for this type ofwastes, the management approach usually considered is the land-fill disposal, which is an avoidable solution in terms of environ-mental policy and economic issues (Kim, 2001; Ma et al., 2002).The process of energy recovery from PVC-containing polymericwastes would imply, therefore, the prior removal of chlorine, fol-lowed by the valorization of the de-chlorinated fraction, throughan energetic process.

848 A. Castro et al. / Waste Management 32 (2012) 847–851

In this context, pyrolysis is considered by several authors(Zevenhoven et al., 2002; Kim, 2001; Ma et al., 2002; Jakslandet al., 2000; Qiao et al., 2006; Qing-lei et al., 2007; Saeed et al.,2004), as the most promising technique for the treatment of wastescontaining PVC, through the thermal degradation of the chlorinemolecule. PVC pyrolysis involves significant cross-linked reactionswith the formation of polyaromatic structures (possibly chlori-nated) and a carbonaceous residue (char) (Blazevska-Gilev andSpaseska, 2010). Thus, it is possible to break down this molecule,allowing the chlorine recovery as hydrochloric acid or chloride (Ta-naka et al., 2007), with potential economic gains.

The thermal degradation process of PVC occurs through a seriesof chemical chain reactions, catalyzed by HCl, which is formed dur-ing the process (Rodolfo and Mei, 2007).

Several studies refer that in the pyrolitic degradation of PVC twostage reactions are detected (Kim, 2001; Wu et al., 1994a,b; Miran-da et al., 1999; Marcilla and Beltrán, 1995). The first stage (up to360 �C) corresponds to the de-chlorination and to the formationof hydrocarbons namely unsubstituted aromatics such as benzene,naphthalene and anthracene. In the second stage the degradationof the polyene chains takes place (Miranda et al., 1999).

According to Wu et al. (1994a,b) in the first stage two parallelreactions of de-chlorination from two components with differentconjugations may occur, with the production of the same carbona-ceous intermediate.

Considering thermogravimetric analysis, it is assumed that thedegradation of PVC occurs between 200 and 400 �C (Zevenhovenet al., 2002; Karayildirim et al., 2006). At 250 �C, the decompositionof PVC has already been initiated, reaching a maximum at approx-imately 300 �C. At 350 �C the amount of chlorine present in PVCwaste is less than 0.1%, which means that at this temperature,99.5% of the whole chlorine has already been released (Zevenhovenet al., 2002). At the end of the process of chlorine removal, a resid-ual amount of chlorine remains on the waste (Tanaka et al., 2007).

The C–Cl bonds in the structure of PVC have a relatively lowerbinding energy than the C–C and C–H bonds, which justifies thatthe bonds of chlorine are the first to be broken, thus starting the

Fig. 1. Percentage of weight loss curves obtained for PVC samples heated up to 250

thermal degradation of PVC. The de-chlorination of PVC is a freeradicals chain reaction therefore requiring low activation energyto start, occurring at low temperatures (Ma et al., 2002; Qing-leiet al., 2007).

De-chlorination of PVC-containing wastes is an advisable previ-ous step for any treatment process, avoiding the deleterious effectsassociated with the release of chlorine and additionally allowingenergy recovery from these wastes. In fact, from the decompositionof PVC according to Eq. (1) a polymeric fraction with high energeticvalue is produced:

C2H3Cl! HClþ 2=nCnHn ð1Þ

Considering the above discussion, for any process seeking theenergetic valorization of PVC-containing wastes, the key factorsare the de-chlorination step, as well as the extent of this reaction.For a subsequent processing of the de-chlorinated fraction mate-rial, it is normally desirable that the chlorine content is as low aspossible.

In the present work, the thermal de-chlorination of PVC wasstudied, by simultaneous DTA/TGA testing, under inert atmo-sphere, allowing the development of a kinetic model of the thermalde-chlorination reaction of PVC under pyrolitic conditions and theestablishment of the optimized operating conditions for this reac-tion. An experimental validation at a laboratory scale has been per-formed by analysis of the pyrolysis products.

2. Experimental procedure

One of the most widely used techniques in the study of decom-position reactions of polymers is the thermogravimetric analysis(TGA) (Marcilla and Beltrán, 1996). This was the technique usedin the present work, in order to determine the relationship be-tween the rate of PVC de-chlorination and the temperature of thethermal treatment, under an inert atmosphere.

For that purpose samples of pure PVC, accurately weighted,were subjected to heat cycles, thereby recording their weight lossover a programed thermal cycle (Slapak et al., 2000). The experi-

, 275, 300, 325 and 400 �C (weight curves were shifted to avoid overlapping).

Table 1Initial temperature for the beginning of weight loss process (de-chlorinationreaction), determined by DSC/TGA analyses.

DTA/TGA Max. temp. (�C) Initial temp. (�C) for the weight loss

250 250.0275 251.7300 254.1325 253.8400 254.3

A. Castro et al. / Waste Management 32 (2012) 847–851 849

ments have been performed in DTA/TGA 2960 TA Instrumentsequipment, under controlled inert atmosphere, obtained by a con-stant flow of helium over the samples of pure PVC. The resultingDTA and TGA curves are simultaneously plotted on a dual Y-axisgraph so the DTA’s fingerprint and the TGA’s weight loss/gain char-acteristics are directly compared as the test samples are heatedand cooled. The DTA identifies the temperature regions and themagnitude of critical events during a drying or firing process. TheTGA measures the weight loss of a material from drying or fromchemical reactions.

Experiments have been conducted from room temperatures upto five different maximum temperatures: 250, 275, 300, 325, and400 �C, with a heating rate of 10 �C/min, to verify the degradationof PVC as mentioned in the literature. After reaching the test tem-peratures, a stage has been done during 360 min, in order to verifyif the reaction proceeds at the selected temperature. Heat flux(weight corrected heat flow in W/g) and sample weight, have beencontinuously recorded. The amount of sample used in the DTA/TGAtests ranged from 29.8 to 44.9 mg.

To characterize the PVC material used in the samples for DTA/TGA tests and the carbonaceous residue produced as result of itsde-chlorination under pyrolytic conditions, a TruSpec ElementalDeterminator, model TruSpec CHN, of Leco and a Philips Analyticalsequential X-ray fluorescence (XRF) Spectrometer model X’UniqueII were used.

The material sample used in this test, was a commercial purePVC powder with the chemical formula C2H3Cl, in which 56.7%

Fig. 2. DTA/TGA at the temperature of 400 �C with weight loss

(by wt.%) is chlorine. The reference of the material is VICIR S 950and it is a vinyl chloride homopolymer produced by a suspensionpolymerization process. It has high bulk density and a very narrowparticle size distribution.

At room temperature, it is presented as a white powder, odor-less and chemically stable, without causing decomposition or reac-tion with other products under normal conditions and is thereforeclassified as non-corrosive, non-explosive and non-flammableproduct. It is insoluble in water, and is considered a nontoxic andharmless product, when in contact with skin.

3. Results

The DTA/TGA results obtained for the five different tested tem-peratures are presented in Fig. 1. They were the basis used for thedevelopment of the kinetic model, considering only the data ob-tained when the temperature was stabilized.

Through the analysis of all DTA/TGA results, presented in Fig. 1(weight curves were shifted to avoid overlapping), it is possible toconclude that the release of chlorine starts at 252 ± 2 �C, more de-tailed in Table 1. There was no significant mass release until thestarting temperature of the PVC thermal decomposition. In fact amaximum of 0.2 mg of mass variation was verified which may bedue to the release of some water or organic contamination.

Fig. 2 shows the heat flow and the weight change from PVC sam-ple heated up to 400 �C. Up to 58% of the mass is removed in the formof hydrochloric acid between 252 and 374 �C. In this temperaturerange it seems that two types of reactions take place as shown bythe change on the heat flow curve at 270 �C. These results may be ex-plained by the two parallel reactions of de-chlorination proposed byWu et al. (1994a,b) were the reactions sequences will be: a conver-sion of PVC into intermediate species, HCl and the decomposition ofintermediates into polyene and some volatiles.

Considering the data from the weighing of PVC sample, as afunction of time, the rate of the weight loss has been determined.As temperature was continuously recorded, a set of values (tem-perature; rate of weight change) were obtained.

es (full line) and weight corrected heat flow (dashed line).

Fig. 3. Detail of the DTA, in which after the temperature has remained a stage of 360 min to check continuity of the reaction.

850 A. Castro et al. / Waste Management 32 (2012) 847–851

Assuming the Arrhenius equation:

K ¼ Ae�Ea=RT ð2Þ

or

ln ðkÞ ¼ ln ðAÞ � Ea=RT ð3Þ

and

k ¼ rjHCljn ð4Þ

where n is the order of the reaction, thus we obtain the expression:

ln ðrÞ ¼ ln ðAÞ � Ea=RT � n ln ðjHCljÞ ð5Þ

where r is the% change of weight per min:

r ¼ ðwi �wi�1Þ=ðti � ti�1Þ ð6Þ

being w the weight of the polymeric sample in mg and t is the timein min. A – kinetic constant, Ea – activation energy, in J/mol, R – gasconstant, J/mol.K, T – absolute temperature in K, n – order of thereaction, |HCl| – concentration of HCl in PVC sample, during theexperiment, by weight%. |HCl| has been calculated at each timeassuming that all the decomposition product is HCl, according toreaction (1).

With all the experimental data obtained, a multivariate regres-sion of ln (r) has been performed in function of 1/T and ln (|HCl|). Avery good fit has been obtained for all experimental points afterthe desired temperature has been reached, except for the experi-ment up to 400 �C. In this experiment, the decomposition rate ismuch higher than expected, a worse fit being obtained by the mod-el. However, points obtained during heating cycle, until the tem-perature of 340 �C give a good fit with the proposed model. Forthis reason, kinetic model has been calculated only for pointswhere temperature was lower than 340 �C. It is assumed that, forhigher temperatures, decomposition of PVC does not follow reac-tion (1) or, as shown in the weight loss curves presented inFig. 3, other reactions may take place, like the volatilization of low-er molecular weight CnHn compounds. From Fig. 3 it is clear that a

different type of reaction starts to occur after the chloride removalreactions with a transition temperature between 325 and 400 �C.

The obtained model is:

ln r ¼ 31:3� 16100=T þ 1:020 ln C ðHClÞ

with r2 = 0.9912.This allows to consider that reaction (1) is a first order reaction

with an activation energy of 133,800 J/mol ± 800 J/mol, value thatis in accordance to those obtained by other researchers (Maet al., 2002). Although the first stage of PVC decomposition pre-sents the two reactions mechanisms, the weight loss analysis didnot evidence the mechanism in the temperature range of the pro-posed kinetic model (252–340 �C). This behavior was alreadyfound by Miranda et al. (1999) who proposed a similar kineticmodel for both reactions.

4. Validation test

Subsequent experimental test of PVC thermal decompositionhave been performed on laboratory scale for validation/confirma-tion of the optimal operating conditions determined at the DTA/TGA tests.

In the performed test a sample of pure PVC was introduced in aglass tube, which was heated by electrical resistance. The sealedglass tubes were connected to a polyurethane tube submerged inwater. During the test the pH was measured continuously, allow-ing the control of the formation of hydrochloric acid, through thechloride rich gases bubbled in aqueous solution.

For the validation test 15.0 g of pure PVC in each glass tube (sixglasses were used). The experimental test occurs for 2 h taking25 min to reach 340 ± 1 �C and kept at this temperature until theend of the test. The formation of hydrochloric acid was visuallyverified with the release of chlorine in aqueous solution.

A final residue of 6.9 g was obtained in each glass tube, corre-sponding to 46.0% of the initial sample. These results are in accor-dance with total weight loss of 47.8%, up to 340 �C, obtained in theDTA/TGA experiments and confirm the proposed kinetic model.

Table 2Chemical composition (by wt.%) by elemental analysis and X-ray fluorescencespectrometry PVC sample used and the carbonaceous residue obtained from pyrolysisat 340 �C.

Pure PVC (initial sample) Carbonaceous residue

Carbon 38.4 89Hydrogen 4.9 7Chlorine 56.7 0.07

A. Castro et al. / Waste Management 32 (2012) 847–851 851

The chemical composition of the carbonaceous residue (char)obtained after the pyrolysis at 340 �C was evaluated by X-rayfluorescence spectrometry and elemental analysis. The obtainedresults are presented on Table 2, as well as the characterizationof the initial sample of pure PVC, highlighting the significantdecrease of chlorine and increase of carbon in the final residue.The chemical composition of the carbonaceous residue is consis-tent with a CnHn polymer, showing potential for energeticvalorization.

Through the analysis of the carbonaceous residue, it was possi-ble to conclude that the chlorine content is less than 0.1% (of origi-nal sample) at the temperature of 340 �C.

5. Conclusions

The kinetics of the reaction of thermal decomposition of PVCwas studied, leading to the development of a kinetic model, withthe expression ln r = 31.3–16,100/T + 1.020 ln C (HCl).

This model was obtained for the decomposition temperatureslower than 340 �C, in which almost all chlorine is removed fromthe pure PVC through the chemical reaction described, with anactivation energy of 133,800 J/mol.

The kinetic model was verified at a laboratory scale, at a tem-perature up to 340 �C, and a reduction of 99.88% of the chlorinecontained in the PVC sample has been observed. The final carbona-ceous residue formed after the pyrolysis at 340 �C seems appropri-ate to be submitted to a subsequent energetic valorization processin order to produce a synthesis gas with high energetic potential.

The results obtained in the present work, were the basis for anongoing project called ‘‘PVC4GAS-Material and energy valorizationof PVC containing wastes’’, that aims the validation at a pilot plantand the practical implementation of a developed technology basedin the two stages above mentioned for the treatment of PVC-con-taining wastes from textile industry and polymeric sectors.

The implementation of the described process for the treatment/valorization of the PVC-containing wastes has many advantagessince it promotes the generation of three products with addedvalue, namely concentrated hydrochloric acid, carbon-rich residue(carbonaceous residue – char) used as feedstock for production of

heat and a synthesis gas that can be used in turbines or gas enginesinstead of natural gas.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.wasman.2012.01.004.

References

Blazevska-Gilev, J., Spaseska, D., 2010. Formal kinetic analysis of PVC thermaldegradation. Journal of the University of Chemical Technology and Metallurgy45 (3), 251–254.

Borgianni, C., De Filippis, P., Pochetti, F., Paolucci, M., 2002. Gasification process ofwastes containing PVC. Fuel 81, 1827–1833.

Jaksland, C., Rasmussen, E., Rohde, T., 2000. A new technology for treatment of PVCwaste. Waste Management 20, 463–467.

Kamo, T., Yamamoto, Y., Miki, K., Sato, Y., 1996. Conversion of waste polyvinylchloride (PVC) to useful chemicals. Resources and Environment. 305, Japan.

Karayildirim, T., Yanik, J., Yuksel, M., Saglam, M., Vasile, C., Bockhorn, H., 2006. Theeffect of some fillers on PVC degradation. Journal of Analytical and AppliedPyrolysis 75, 112–119.

Kim, S., 2001. Pyrolysis of waste PVC pipe. Waste Management 21, 609–616.Lewis, F., Ablow, C., 1976. Pyrogas from biomass. Presented to a conference on

capturing the sun through bioconversion, Washington, DC, ShorehamAmericana Hotel, Stanford Research Institute.

Ma, S., Lu, J., Gao, J., 2002. Study of the low temperature pyrolysis of PVC. Energy &Fuels 16, 338–342.

Marcilla, A., Beltrán, M., 1996. Kinetic models for the thermal decomposition ofcommercial PVC resins and plasticizers studied by thermogravimetric analysis.Polymer Degradation and Stability 53, 251–260.

Marcilla, A., Beltrán, M., 1995. Thermogravimetric kinetic study of poly(-vinylchloride) pyrolysis. Polymer Degradation and Stability 48, 219–229.

Miranda, R., Yang, J., Roy, C., Vasile, C., 1999. Vacuum pyrolysis of PVC I. Kineticstudy. Polymer Degradation and Stability 64, 127–144.

Qiao, W., Song, Y., Yoon, S., Korai, Y., Mochida, I., Yoshiga, S., Fukuda, H., Yamazaki,A., 2006. Carbonization of waste PVC to develop porous carbon material withoutfurther activation. Waste Management 26, 592–598.

Qing-lei, S., Xin-gang, S., Yun-liang, L., He, Z., Xiao, W., Chuan-ge, C., Jian-hua, L.,2007. Thermogravimetric-Mass Spectrometric Study of the Pyrolysis Behaviorof PVC. Journal of China University of Mining & Technology 17 (2), 242–245.

Rodolfo, A., Mei, L., 2007. Mecanismos de degradação e estabilização térmica doPVC. Faculdade de Engenharia Química, UNICAMP (in Portuguese).

Saeed, L., Tohka, A., Haapala, M., Zevenhoven, R., 2004. Pyrolysis and combustion ofPVC, PVC-wood and PVC-coal mixtures in a two-stage fluidized bed process.Fuel Processing Technology 85, 1565–1583.

Slapak, M., Kasteren, J., Drinkenburg, A., 2000. Determination of the pyrolyticdegradation kinetics of virgin-PVC and PVC-waste by analytical andcomputational methods. Computational and Theoretical Polymer Science 10,481–489.

Tanaka, Y., Tsuji, T., Shibata, T., Uemaki, O., Itoh, H., 2007. Dehydrochlorination Ratein Thermal Degradation of PVC. School of Engineering, Hokkaido University,Japan 060-8628.

Wu, C., Chang, C., Hor, J., 1994a. Pyrolysis kinetics of waste PVC at high conversion.The Chemical Engineering Journal and the Biochemical Engineering Journal 55(1–2), 87–92.

Wu, C., Chang, C., Hor, J., 1994b. Two-stage pyrolysis model of PVC. The CanadianJournal of Chemical Engineering 78 (4), 644–650.

Zevenhoven, R., Axelsen, E., Hupa, M., 2002. Pyrolysis of waste-derived fuelmixtures containing PVC. Fuel 81, 507–510.