Thermal-oxidative induced degradation behaviour of

0 downloads 0 Views 276KB Size Report
... and thermooxidative ageing on the degradation behaviour of a commercial poly(oxymethylene) ... trioxane and a cyclic ether such as e.g. ethylene oxide,.
Polymer Degradation and Stability 91 (2006) 464e471 www.elsevier.com/locate/polydegstab

Thermal-oxidative induced degradation behaviour of polyoxymethylene (POM) copolymer detected by TGA/MS* S. Lu¨ftl*, V.-M. Archodoulaki, S. Seidler Institute of Materials Science and Technology, Vienna University of Technology, Favoritenstrasse 9-11, A-1040 Vienna, Austria Received 29 October 2004; received in revised form 12 January 2005; accepted 18 January 2005 Available online 18 March 2005

Abstract The influence of reprocessing and thermooxidative ageing on the degradation behaviour of a commercial poly(oxymethylene) (POM) copolymer was studied by means of thermogravimetric analysis (TGA) under nitrogen and air atmosphere. Five heating rates were used to evaluate activation energies at several degrees of conversion. TGA-measurements were accompanied by simultaneous monitoring of the evolved gases with a mass spectrometer (MS) coupled to the TGA-furnace outlet. The mass spectra showed that the main degradation product was formaldehyde and that in air further formation of water was detectable. In nitrogen atmosphere aged specimens emitted small amounts of carbon dioxide at the beginning of the mass loss. The activation energy for low degrees of conversion (!5%) increased in air and in nitrogen as a function of the conversion. For higher conversions a difference with progressing degradation emerged: in air, activation energies lowered continuously while they remained nearly constant under nitrogen. Ó 2005 Elsevier Ltd. All rights reserved. Keywords: POM; Polyacetals; Reprocessing; Ageing; Degradation; Thermogravimetric analysis (TGA)

1. Introduction Polyoxymethylene (POM) belongs to the group of engineering thermoplastics; it is also called polyacetal because of its basic molecular structure consisting of a repeating carboneoxygen linkage. Acetal homopolymer refers to resin containing solely the carboneoxygen backbone, while for the copolymer resin the oxymethylene structure is occasionally interrupted by a comonomer unit. Commercial acetal copolymers are typically produced by cationic polymerisation of 1,3,5trioxane and a cyclic ether such as e.g. ethylene oxide, * Based on a presentation at the 3rd International Conference on Modification, Degradation and Stabilisation of Polymers (MoDeSt), Lyon, 29 Auguste2 September, 2004. * Corresponding author. Tel.: C43 1 5880130850; fax: C43 1 5880130895. E-mail address: [email protected] (S. Lu¨ftl).

0141-3910/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.polymdegradstab.2005.01.029

1,3-dioxolane or 1,3-dioxepane. Acetal copolymer resin has greater stability but reduced crystallinity as a result of carbonecarbon bonded groups interspersed in its polymer chain. This polymer structure also imparts superior resistance to alkali, hot water, and other chemicals, as well as long life at elevated temperatures and more latitude in processing conditions. On the other hand, its tensile strength, rigidity, softening point and melting point are all lower than those found in the acetal homopolymer [1e3]. The use of POM is growing steadily in the automotive and electronic industries and therefore the problematic nature of the production waste as e.g. gates or faulty mouldings has to be resolved. The normal waste process for internal factory recycling is to grind the waste parts and use them as a re-pellet addition to the virgin pellet. The high demands of the automotive industry oblige suppliers to guarantee that no disadvantageous properties result from using reprocessed materials.

S. Lu¨ftl et al. / Polymer Degradation and Stability 91 (2006) 464e471

For technical and safety reasons re-granulate is not used until now in parts made by POM because of the negligible amount of information about this topic. Although degradation mechanisms of the both POMtypes have been studied by several researchers they were basically obtained from laboratory grade polyacetals (with and without additional stabilisers) or from just one-time processed commercial material. Unfortunately, not much information about the degradation of thermooxidative aged POM and the application of hyphenated thermal analysis techniques to determine its effects on the degradation progress is available. Thus the aim of the present study is to present changes occurring in thermal degradation behaviour of a commercial available heat stabilised POM-copolymer after multiple processing and additional exposure to artificial oven ageing at 140  C.

2. Experimental The investigated heat stabilised POM copolymer UltraformÒ W2320 003 (BASF) is a commercial distributed resin for injection moulded parts. Therefore, standard tensile test specimens (length of 145 mm, width of 10 mm and thickness of 4 mm (ISO 527-2)) were produced by injection moulding in accordance with processing conditions recommended by the producer (temperature of the nozzle 215  C, mould temperature 95  C, moulding cycle approximate 1 min). Some of the specimens were put aside (labelled 1st processing step conditions) while the rest was ground in an industrial cutting mill to pellets of about 5 mm size. The grinding stock was processed again under the same conditions in order to obtain specimens consisting of 100% recycled material. The procedure was repeated up to six times to get specimens in a range from the 1st processing step to the 7th processing step. In addition, specimens of the 1st and the 7th processing step were stored for up to 8 weeks in an air circulating oven at 140  C for accelerated thermooxidative ageing in accordance with long-term heat stability test conditions reported in Ref. [4] for POM. A number of samples (n) of about 20 mg was scraped from the specimens in the area of the gate by the use of a scalpel, put in a ceramic crucible and then analysed by means of thermogravimetry (TGA 2050, TA Instruments) in the dynamic mode. The samples were scraped within a depth of about 1 mm from the specimen surface. Different heating rates (2.5, 5, 10, 20 and 40  C/ min) were selected to evaluate activation energies in the different ageing conditions and at several conversion degrees. The thermogravimetric analysis (TGA) was performed under (pressurised) air and nitrogen atmosphere (sample compartment gas flow 90 ml/min, balance purge gas flow (only nitrogen) 10 ml/min).

465

Gases evolving during the decomposition of the sample were registered every 10  C temperature increase by a coupled quadrupole mass-spectrometer (MS) equipped with electron ionisation and channeltron detector (Thermostar, Balzers). The coupling consisted of a heated quartz capillary tube (120  C) connecting the TGA furnace outlet with the MS gas inlet through a pin hole diaphragm.

3. Results and discussion After the artificial ageing the visual aspect of the specimens shows changes in comparison to the initial condition. As reported in Ref. [4] the specimens are discoloured as a result of the oven ageing. The yellowing, the loss of surface gloss and the appearance of cracks increase with the duration of the storage at 140  C. The thermal and thermooxidative degradation of POM homo- and copolymers have been studied in detail by several researchers and reviews of the different proposed degradation mechanisms can be found in [1,3,5]. According to Kern [5] degradation reactions to be found in POM-homopolymers are also occurring in copolymers but in a reduced scale as the copolymer units hinder the unzipping of the complete chain. The depolymerisation is stopped at the adjacent CeCbond of the copolymer unit and a new CeO-bond scission step is needed to reinitiate depolymerisation in the remaining polymer chain. Nevertheless to prevent autooxidation, stabilisers against an attack of molecular oxygen on the sensitive methylene groups have to be added to the polymer [4,5]. In the following figures the results of the TGA investigations are plotted as percentage of the remaining sample mass to the initial one (designated as TG) as a function of the temperature for a heating rate of 10  C/min if not otherwise mentioned. TGA/MS measurements at the different heating rates give best results for a heating rate of 10  C/min. Slower heating rates improve the resolution of the TG curves but lower the ion current signal in the mass spectra (MS). While faster heating rates give stronger ion current but different degradation steps are no longer separated, especially when the decomposition reaction itself is relatively fast as it is the case in the presence of oxygen. The tangent method recommended in TGA standards (e.g. EN ISO 11358 [6]) for the onset temperature determination gives unsatisfactory results: (1) comparisons between the various conditions are not possible due to dissimilar degrees of conversion corresponding to the onset temperature determined by means of the tangent method. (2) The degree of conversion corresponding to the commonly used onset evaluation is

466

S. Lu¨ftl et al. / Polymer Degradation and Stability 91 (2006) 464e471

generally above 10%. So the degradation is advanced and one can no longer speak about a degradation beginning that is normally characterised by the term of onset. Therefore, for comparison purposes the onset temperature is defined as the temperature at which a mass loss of 3% of the initial sample mass is reached (Fig. 1). TG profiles in nitrogen generally show a one-step degradation behaviour with a residual mass at 500  C lower than 1% of the initial mass. The residue consists mainly of thermally degraded processing aids and stabilising compounds, but no further investigation was carried out to determine the nature of the residue. The average onset temperature in nitrogen (10  C/ min heating rate) for the material taken from the specimen that has been solely one-time processed was 314  C (n Z 6). It does not differ statistically (one-way ANOVA, significance level 0.05) from the average onset temperature of 311  C (n Z 11) of the material taken from the specimen that has been made from material having been reprocessed six times. It follows that multiple processing does not affect the thermal resistance of the material as neither the onset nor the complete TG curve shape were different for the two investigated processing conditions. The corresponding MS reveals that the main degradation product is formaldehyde (m/z 30), but there are also some fragments belonging probably to trioxane (m/z 31 and 61) that can be detected in a reduced scale. Also in a reduced scale fragments that can be assigned to the comonomer (m/z 43, 45, 60) and the stabilising system (m/z 57 and

77 corresponding to hindered phenols) can be found (Fig. 2). It should be mentioned that in the MS additional ion fragments are detected but only the most interesting ones are represented in the figures. Investigations on POM (DelrinÒ) performed by pyrolysis/gas chromatography/mass spectrometry could identify up to 54 compounds formed during thermal degradation. But the major product which accounted for over 93% of the detected volatile products was in this case also formaldehyde. Two additional compounds (dimethylformaldehyde and formic acid) had an abundance over 1% while the other compounds were detected in a lesser amount. They consisted of a wide variety of low molecular weight methoxy oligomers and 1,3,5-trioxanes with methoxy substituents [7].

Fig. 1. TG curve (solid line) in nitrogen of a sample taken from a specimen after the 6th reprocessing step and an oven ageing duration of 8 weeks at 140  C with the corresponding MS for selected ions: m/z 18-H2O, 29-fragment of H2C]O and other components, 30-H2C]O, 31-fragment of CH3eOH and trioxane, 43-CH3eCOH, 44-CO2, 45fragment of HCOOH and CH3eCOOH, 46-HCOOH, 57-fragment additives, 60-acetic acid, 61-fragment trioxane, 77-fragment additives (phenyl-group). Heating rate 10  C/min.

Fig. 2. TG curve (solid line) in nitrogen of a sample taken from a specimen after the 1st processing step (a) and the 6th reprocessing step (b) with the corresponding MS for selected ions: m/z 18-H2O, 29fragment of H2C]O and other components, 30-H2C]O, 31-fragment of CH3eOH and trioxane, 43-CH3eCOH, 44-CO2, 45-fragment of HCOOH and CH3eCOOH, 46-HCOOH, 57-fragment additives, 60acetic acid, 61-fragment trioxane, 77-fragment additives (phenylgroup). Heating rate 10  C/min.

S. Lu¨ftl et al. / Polymer Degradation and Stability 91 (2006) 464e471

467

After an oven storage duration of 8 weeks at 140  C a slight decrease in the average onset temperature of about 10  C can be observed for both material conditions (1st processing step 299  C (n Z 5) and 7th processing step 301  C (n Z 5)). But statistical data processing reveals that the difference in the average between the initial condition and the artificially aged one is not significant. Also after ageing no statistical significant difference in the onset of the material of the 1st and the 7th processing step can be found. Apparently the thermal stability of the initial macromolecular structure seems not to be affected as no real change in the TG-curve shape could be determined between the initial condition and the oven aged one. The formation of lower molecular weight compounds through accelerated degradation should result in a start of the mass loss in a lower temperature range, while cross-linking should increase the degradation begin. Nevertheless, in the corresponding MS of the oven aged materials a change in the m/z 44 curve occurs in comparison to the initial condition of MS. A small peak starting at about 210  C can be observed before the main degradation starts (see Fig. 1). This peak could correspond to carbonyl, carboxyl or hydroperoxide groups formed in the main chain during thermooxidative degradation according to the mechanism below [1,5,8].

a function of the temperature a temperature decrease can be observed due to the delayed response of the temperature controlling system to fast temperature changes (Fig. 3b). Therefore, TGA temperature data obtained in air atmosphere have to be considered as to be doubtful for conversion degrees above about 40e 50%. This is particularly the case for the samples taken from specimens of the 1st and 7th processing step at the initial condition. In every case the onset- and derivative TGmaxtemperatures are higher in nitrogen than in air (Fig. 4). In nitrogen atmosphere the mass loss starts at temperatures above 270  C, which is an indicator that mainly thermal degradation as result of bond cleavage in the carboneoxygen backbone takes place [1]. In air the average onset temperature is 259  C (n Z 11) for material of the 1st processing step condition. As mentioned above the reaction with oxygen induces chain scissions and formation of compounds which accelerate the degradation. So as one can see on the TGA-results in Fig. 4 the presence of oxygen in combination with high temperature leads to an earlier breakdown of the POM-chain than only high temperatures in an inert atmosphere. Also the corresponding degradation rate expressed by the DTG curves is remarkably higher in air than in nitrogen.

The hydroperoxides on the methylene groups can induce chain scission followed by the formation of formaldehyde, but they can also oxidise the later to formic acid. Both the reactions are responsible for the acidolytic degradation of the polyacetal chain. Additionally, another reaction with acid, the transacetalisation, can occur with formation of cyclic compounds or new linear POM chains [5].

For the initial condition no statistical significant difference in the onset temperature can be found for samples extracted from a solely one-time processed resin in comparison to such from multiple processed pellets (average onset 255  C (n Z 10)). The corresponding MS shows that as in nitrogen one of the major degradation product is formaldehyde (m/z 30) but in the presence of oxygen additional formation of water (m/z 18) occurs

The investigations in air atmosphere show that the degradation is accompanied by a strong exothermic reaction (Fig. 3a) which seems to be the result of autooxidation [1] after the consumption of the stabilising additives. In TGA plots, where TG is plotted as

(Fig. 5a). Carbon dioxide (m/z 44) emission is clearly enhanced in air atmosphere as the ion current signal exceeds the signal for formaldehyde. It seems that the formaldehyde released by the unzipping reaction of the POM-chain is oxidised to formic acid. The later

468

S. Lu¨ftl et al. / Polymer Degradation and Stability 91 (2006) 464e471

Fig. 4. Comparison of the average TG (black line) and derivative TG (DTG) (dark grey line) of a sample taken after the 1st processing step in nitrogen (n Z 6) and in air (n Z 11) atmosphere.

Fig. 3. TG curves in air of a sample taken after the 1st processing step showing a deviation in the temperature curves due to a strong exothermic reaction during decomposition. (a) Plot of TG and temperature versus time for different heating rates. (b) Plot of the same TG curves as in Fig. 3a versus temperature showing a temperature decrease due to delayed response of the temperature control system.

accelerates the degradation through acidolysis. Additionally, the formic acid is oxidised to water and carbon dioxide. These observations are in accordance with the different degradation mechanisms proposed by [1,5,9]. The TG-profiles of the oven aged specimens indicate a shift in the onset- and DTG maximum (DTGmax) temperature to lower temperatures as well as a change in the shape of the TG-curve only after a storage time of 5 weeks at 140  C [10]. After 8 weeks oven ageing the average onset temperature of the samples extracted from specimens of the 1st processing step and those extracted from specimens of the 7th processing step are equal to 245  C (n Z 5). The lowering in the onset temperature of the 8 weeks oven aged samples is statically significant in comparison to the initial conditions. So one can conclude that during thermooxidative ageing the stabilisers which act as radical scavenger [3,4] are almost consumed as in the MS the m/z 77 signal (phenyl group

from hindered phenols) is relatively weak after the artificial thermooxidative ageing. Therefore, the protection against oxygen attack is no longer effective and an earlier mass loss due to the induced thermooxidative bond cleavage with subsequent unzipping of the polymer chain results [10]. The degradation of the aged samples progresses no longer in one step, as after the exothermic process a slow degradation of apparently more thermally resistant compounds can be found (Fig. 5b). This could be an evidence that during ageing formation of macrocyclic POM happens, as described by Jaacks [11], Wegner et al. [12] and Hasegawa et al. [13]. The Arrhenius-plots of the logarithm of the heating rate versus the reciprocal absolute temperature for 0.1, 1, 2, 3, 4, 5, 6, 8, 10, 20, 30, 40 and 50% conversion are well fitted by a linear regression (correlation coefficient ÿ0.98 ! r % ÿ1.00). Therefore, the degradation can be assumed to be a first order process for these degrees of conversion. Since the temperature data in air atmosphere for conversion degrees above 50% are doubtful no activation energy is reported beyond this point even though the data are also well fitted by linear regression. Activation energies as a function of the conversion degree were evaluated according to the FlynneWall method [14]. Published activation energy values for polyacetals cover a relatively broad range and depend on the material (e.g. polyoxymethylene with hydroxyl, methoxy or acetate end groups, copolymer, molecular weight distribution, etc.), the dominating depolymerisation mechanism (thermal, anionic, cationic, etc.) and the methods (isothermal or dynamic mode, temperature range, kinetic approach, .) used to determine it. So the activation energies for the emission of organic compounds from POM in the temperature range of

S. Lu¨ftl et al. / Polymer Degradation and Stability 91 (2006) 464e471

469

energy for low degrees of conversion (!5%) increases principally in air (Fig. 6) and in nitrogen (Fig. 7) with advancing conversion. For higher conversion degrees a difference with progressing degradation emerges: in air activation energies lowered continuously while under nitrogen atmosphere they remained nearly constant. The higher activation energy values found in this study in comparison to the values cited above might result from the added stabiliser package and from the presence of copolymer units. The change of the activation energy with progressing degradation under the different conditions seems to reflect the degradation mechanisms involved. Until 5% conversion the radical scavenging ability of the stabilisers is systematically consumed and chain scissions are initiated. In nitrogen mainly thermal induced depolymerisation after the chain scission occurs.

Fig. 5. TG curve (solid line) in air atmosphere of a sample taken from a specimen after (a) the 6th reprocessing step and (b) an additional oven ageing duration of 8 weeks at 140  C with the corresponding MS for selected ions: m/z 18-H2O, 29-fragment of H2C]O and other components, 30-H2C]O, 31-fragment of CH3eOH and trioxane, 43CH3eCOH, 44-CO2, 45-fragment of HCOOH and CH3eCOOH, 46HCOOH, 57-fragment additives, 60-acetic acid, 61-fragment trioxane, 77-fragment additives (phenyl-group). Heating rate 10  C/min.

40e80  C were indicated as 16e60 kJ/mol by Choczyn´ski et al. [15]. Ra¨tzsch and Eckhard [16] reported an activation energy for thermal degradation from about 40 to 110 kJ/mol, for cationic depolymerisation of about 80 kJ/mol and for anionic depolymerisation of about 60 kJ/mol. Grassie and Roche [17] found that energy of activation for the thermal degradation of hydroxyl group terminated polymer increases from 20 to 30 kJ/mol in the early stages to 210e220 kJ/mol from about 40% volatilisation onwards. In the present study, generally a lower activation energy is found for the same degree of conversion for measurements carried out in air atmosphere (80e210 kJ/ mol) than in nitrogen (90e290 kJ/mol). The activation

Fig. 6. Evolution of the activation energy in function of the conversion degree and ageing condition in air atmosphere. (a) 1st processing step, (b) 7th processing step.

470

S. Lu¨ftl et al. / Polymer Degradation and Stability 91 (2006) 464e471

(e.g. hydroperoxide and/or macro-cation formation) in the polymer chain preceding the definitive chain rupture. One would have expected a lowering of the activation energies with progressing ageing as a result of thermooxidatively induced damage in the polymer chain and therefore advancing susceptibility for degradation processes, but a clear correlation between the ageing condition and the activation energy evolution could not be found.

4. Conclusions

Fig. 7. Evolution of the activation energy in function of the conversion degree and ageing condition in nitrogen atmosphere. (a) 1st processing step, (b) 7th processing step.

The depolymerisation stopped at the adjacent CeCbonds needs a new chain scission for further progression. So from about 5% conversion, termination and reinitiating of the depolymerisation will keep the activation energy relatively constant. In air the presence of oxygen enhances the thermal effect as oxygen insertion in the macromolecular chain is favoured at higher temperatures. The oxygen containing groups formed induce chain scission and formaldehyde release as described above. Additional secondary products of the autooxidation, which can act as an accelerator for the degradation, are formed. So in air because of the development of multiple mechanisms favouring degradation (oxidation, thermal, acidolysis, etc.) the activation energy will drop as a result of pre-damage

The present TGA/MS investigations in nitrogen and in air atmosphere show a different degradation behaviour depending on the selected experimental conditions. In N2 the onset- and DTGmax-temperature are mainly higher than in air and the temperature range of the degradation process let us suppose that it is mainly due to thermally induced scission of the polymer chain followed by depolymerisation. In the corresponding MS the major degradation product is formaldehyde, in a reduced scale carbon dioxide and fragments which belong to formic acid, trioxane and methoxy oligomers are detected. No real change in the TG-curve shape can be determined even after a storage duration of 8 weeks at 140  C, so apparently at the first look the thermal stability of the initial macromolecular structure seems not to be affected by accelerated thermooxidative ageing at 140  C. Nevertheless, in the MS a release of carbon dioxide at about 210  C precedes the formation of formaldehyde for the artificially oven aged samples indicating that changes in the polymer chain happened. No difference in the thermal degradation behaviour emerges between material which was processed only once and material which was ground and reprocessed up to six times. Also by means of accelerated ageing no change in the degradation behaviour results among the samples from the different processing conditions. The degradation in air is accompanied by an exothermic process especially for non-aged material and the degradation rate (DTG) is remarkably higher than in N2. After only 5 weeks oven storage the onset temperature of oven aged specimens shifts to about 10  C lower temperatures and the shape of the TGcurves becomes flatter. The lowering in the onset in air is an indicator that the stabilising system is consumed during thermooxidative ageing and that pre-damage (e.g. hydroperoxide substituents) in the molecular backbone happened [10]. Samples extracted from specimens after 5 and 8 weeks oven storage no longer exhibit an one step degradation. At the end of the degradation a small amount of thermally more resistant compounds appears. Among the different processing conditions even in air no differences in the degradation behaviour emerge.

S. Lu¨ftl et al. / Polymer Degradation and Stability 91 (2006) 464e471

The MS shows that one of the major degradation product is formaldehyde; in the presence of oxygen additional formation of water occurs and the carbon dioxide ion current signal exceeds the signal for formaldehyde. So in air, as reported in the literature [5] there are several degradation mechanisms which encroach in the decomposition of POM. Consequently, the activation energies for several degrees of conversion are lower than those evaluated for the degradation in nitrogen. Neither in air nor in nitrogen atmosphere a clear correlation between the ageing condition and the activation energy progress can be found.

Acknowledgement The authors are grateful to the project partner the former HB-Plastic GmbH, Austria, (now SFS intec GmbH) for the injection moulding and reprocessing of the specimens and to the Austrian Industrial Research Promotion Fund (FFF) for the financial contribution to this project.

References [1] Kern W, Cherdron H, Jaacks V, Baader H, Deibig H, Giefer A, et al. Polyoxymethylene, Angew Chem 1961;73(6):177e86. [2] Weißermel K, Fischer E, Gutweiler K, Hermann HD. Zur Copolymerisation des Trioxans. Kunststoffe 1964;54(7):410e5. [3] Sabel HD, Schlaft H, Unger P, Ziegler U. In: Bottenbruch L, editor. Technische Thermoplaste: Polycarbonate, Polyacetale, Polyester, Celluloseester. Kunststoff-Handbuch 3. Mu¨nchen Wien: Carl Hanser Verlag; 1992. p. 300e95.

471

[4] Stohler FR, Berger K. Stabilization of polyacetals. Angew Makromol Chemie 1990;176/177(3074):323e32. [5] Kern W. Chemische Elementarvorga¨nge bei der Alterung von Kunststoffen. Chemiker-Ztg/Chem Apparatur 1967;91(8): 255e62. [6] EN ISO 11358. Plastics e thermogravimetry (TG) of polymers e general principles; 1997. [7] Day M, Cooney JD, Touchette-Barrette C, Sheehan SE. Pyrolysis of mixed plastics used in the electronics industry. J Anal Appl Pyrol 1999;52(2):199e224. [8] Sedla´rˇ J, Kucera M. Der thermooxydative Abbau von Polyformaldehyd. Makromol Chem 1967;102(2243):245e58. [9] Dudina LA, Agayants LA, Kamilova LV, Enikolopyan NS. Thermal and thermo-oxidative degradation of polyformaldehyde. V. Role of formic acid in thermal-oxidative degradation. Polym Sci USSR 1964;5(2):316e22. [10] Archodoulaki VM, Lu¨ftl S, Seidler S. Thermal degradation behaviour of poly(oxymethylene): 1. Degradation and stabilizer consumption. Polym Degrad Stab 2004;86(1):75e83. [11] Jaacks V. U¨ber die Mo¨glichkeit der Bildung makrocyclischer Polyoxymethylene. Makromol Chem 1966;99(2449):300e2. [12] Wegner G, Rodriguez-Baeza M, Lu¨cke A, Lieser G. Kinetik und Mechanismus der kationischen Polymerisation von Trioxan: Ein katalysierter Kristallwachstumsprozeß. Makromol Chem 1980; 181:1763e90. [13] Hasegawa M, Yamamoto K, Shiwaku T, Hashimoto T. Formation of macrocyclic Poly(oxymethylene) with narrow molecular weight distribution on the lamellar crystals. Macromolecules 1990;23(10):2629e36. [14] Flynn JH, Wall LA. A quick, direct method for the determination of activation energy from thermogravimetric data. J Polym Sci B Polym Lett 1966;4:323e8. [15] Choczyn´ski M, Necki JM, Juszkiewicz A. Determination of rate constants and activation energies for emission of organic compounds from polyoxymethylene. Acta Chromatogr 2004;14(1):37e48. [16] Ra¨tzsch M, Eckhardt G. Abbaureaktionen von Polyoxymethylenen und A¨nderung der Molekulargewichtsverteilung durch diese Reaktionen II. Plaste und Kautschuk 1973;20(6):424e32. [17] Grassie N, Roche RS. The thermal degradation of polyoxymethylene. Makromol Chem 1968;112(2581):16e33.