mao oe E oo Bais J g zm i mes — Peco — rmronpene Figure 8. Comparison of chemiluminescence runs from saturated and unsaturated pure hydrocarbon polymers having tertiary carbons (Polypropylene, polyisoprene) in the main chain Scientific assessment of plastics degradation Measurement of polymer degradation by CL, TGA, DSC, Py-GCMS and SPME-GCMS The plastics used fr the degradation studies were either pure polymers used either a fils or powders, As a representative set of plastics in the study, commercially available the so-called Resinkit™ published and distributed by the ResinKit™ company (Woonsocket, Rl, USA. It includes 50 ofthe most frequent polymers containing different additives, Some plastics from ths set became the items in the SamCo collection described in the Chapter 1. Figure 8 shows how the elementary structure of the polymer affects the observed patterns of chemilur scence response in conygen at 120°C. AS expected, the degradabilty decreases inthe order poyisoprene < polypropylene in correspondance withthe amount of and tertiary carbon atoms in the polymer structure. The intensity of chemiluminescence signal from a heated polymer and the kinetics ofits change with time or temperature i determined i '* the quality of the polymer, its photo-oxidation and thermo- oxidation history expressed in concentration of hydroperoxides, carbonyl groups or of other oxidized structures and terminal sr0ups, The rate of an oxidative attack may then be related to the average molar mass and to its distribution, and to the ratio. of amorphous/crystalline structures. Polymers cannot be simply ordered according tothe intensity of light emission ata given temperature. The chemiluminescence-time pattems ae related with the rate of sample oxidation, but they may differ from one to the next polymer. ‘* the quality of the polymer surface; for example, for a particular polymer the intensity of the signal will be different for film and for powder. Measurement of polymer degradation 77 ry sumeoes fh we 5 i _—— 2 ome Fa Temperature os Figure 9, Non-isothermal chemiluminescence runs for polypropylene (PP, polyethylene (PE (part Aof the graph, polystyrene (PS) and poly(methyl methacrylate) (PMMA) (part B ofthe grap). Numbers 27, 25,1. and 9 denote the sample ofthe same polymers from the Resinkit™. The rate ofheating 5°C.min, oxygen 178 wee i 2 ee = 3 : rempeatese . —> — Pu ‘+ the temperature and concentration of oxygen in the atmosphere surrounding the oxidised sample. ‘+ the extent and quality of polymer stal + aditect oxidation of polymer additive, which may occassionaly give a much stronger signal than the oxidation of polymer itself. This is vey important as it may lead to an erroneous relation between the rate of polymer oxidation and chemiluminescence intensity. ‘An oxidation test involving temperature ramp experiments in ‘oxygen over a temperature range 40-220°C enables one to compare the degradabilty of different materials over a large temperature interval from the viewpoint of their chemiluminescence intensity. Inthe Figure 9 we may see the comparison of the chemiluminescence runs for pure polypropylene, polyethylene, polystyrene and poly(methyl methacrylate) and corresponding polymers from the Resinkit™ (27, 25, 1 and 5). The information that may be obtained from this comparison is as follows: as the line 27s shifted to significantly higher temperatures, PP 27 from the resin kit is heavily stabilised, At the same time polymer fragments which are formed in degradation in oxygen have substantially lower molar mass. PE 25 is stabilised only slightly but the degradation products have again lower molar mass than its in the case of pure polyethylene. The shift to higher temperatures and some extent of polymer stabilisation may be observed also in the case of PS 1 but the products of degradation have higher molar mass than ry sumexterm Temperature ITI Figure 10. DSC course of oxidation of polypropylene powder, ‘oxygen. Numbers denote the rate of heating in °C.min* non pen, ° —> cmon. 7 Cd 7 0H ory —> wo, + AX ow, ro of a Figure 2s. Primary step of formation of acids from elementary units of cellulose acetate, cellulose nitrate and poly(vinyl chloride) Scientific assessment of plastics degradation the reference sample while PMMA 9 from the resin kit denoted as acrylics degrades even faster than the reference sample. The important parameter which affects the position and shape of the corresponding non-isothermal records of degradation of polymers isthe rate of sample heating (Rychly 2011, b). As it may be seen in the Figure 10 the DSC exotherms of polypropylene in oxygen are shifted to higher temperatures and become more distinct due to the increasing rate of heating. However, this is valid for case of non-isothermal thermogravimetry in nitrogen and DSC and chemiluminescence from polyolefins in oxygen. In the former case itis even used for calculation of activation energy of polymer degradation. There are polymers such as poly(methyl methacrylate), polystyrene and others for which this shift is insignificant and/or cannot be observed at al. 3.4. Polymers that can damage other objects (“malignant” polymers) 3.4.1.1. Cellulose derivatives and poly(vinyl chloride) (Pvc) The first step of the unzipping process leading to the formation of cethanoic acid, nitric acid and hydrogen chloride from such kind of polymers is seen on the Figure 11. Nitric acid is a strong oxidising acid which causes tendering and decomposition of cellulose and proteins, and corrosion of metals. Cellulose acetate produces ethanoic acid (vinegar odor, hence the term “vinegar syndrome” to describe cellulose acetate degradation) and cellulose butyrate and cellulose acetate butyrate produce butyric acid which has a distinctive and characteristic vomit odour. These organic gases are not as harmful as nitric acid, but they also cause tendering, decomposition, and corrosion. Particularly cortosive effects are performed by hydrogen chloride. Cellulose nitrate (nitrocellulose, flash paper) is a highly flammable compound formed by nitrating cellulose through exposure to nitric acid or another powerful nitrating agent. When used as a propellant or low order explosive, itis also known as guncotton. Nitrocellulose plasticised by camphor was used by Kodak, and other ‘suppliers, from the late 1880's as a film base in photographs, X-ray films and motion picture films; it was known as “Nitrate film”. After numerous fires caused by unstable nitrate films, safety film started Measurement of polymer degradation 179 ry sume380 to be used from the 1930's in the case of X-ray stock and from 1948 for motion picture film, as well. Cellulose nitrate is found in many older artefacts, including sheets or films (e.g., photographic film base), vamishes and lacquers, and solid objects, especially those which imitate natural ‘materials like ivory (often called “French ivory"), tortoiseshell, and horn. Cellulose nitrate degrades to produce acidic and nitrogen oxide ‘gases which can seriously damage objects that are nearby o contact. This deterioration is accelerated by increased temperatures, by elevated relative humidity, and by acidic conditions. it may be even said that itis a merit of cellulose nitrate artefacts in museums which brought about an increased attention to museum's objects from plastics as such. Cellulose acetate is commonly encountered in two grades characterised by different degrees of substitution, namely, cellulose triacetate most commonly found in sheets like photographic film base and fibers, and cellulose diacetate in thicker sheets and three dimensional shapes and objects often simulating tortoiseshell, ivory, wood, and mother-of-pearl. The cellulose triacetate and diacetates are easily confused with cellulose nitrate when compared by visual appearance alone. Cellulose acetate objects usually contain plasticisers. Cellulose acetate degrades primarily by acid hydrolysis, which causes deacetylation (Figure 11). Deacetylation cleaves, pendant acetate groups from the cellulose polymer backbone and depolymerises the backbone. Formation of ethanoic acid from the plastic creates the acidic surfaces on the plastic and acidic atmospheres in enclosures. This process is analogous to that which happens with cellulose nitrate. Depolymerisation leads to decrease in mechanical strength and fracture plus deformations and warpage. Ethanoic acid diffuses through the display or storage space, and can cause corrosion of metals, or acidic catalysed degradation of other paper and textiles. Additives, especially plastcisers, migrate and may be lost, or are hydrolysed or oxidised to acidic compounds, This leads to warpage, embrittlement, and fracture, and to the development of acidic and sticky surfaces, sometimes with surface deposits of plasticiser or acidic degradation products. Pure poly(vinyl chloride) (PVC) degrades to produce hydrogen chloride at temperatures needed to form it into usable products by molding or extrusion. As a consequence, heat stabilisers are always added to overcome this problem. As pure poly(vinyl chloride) is a rigid plastic compounds called plasticisers are added to convert itinto flexible plastic. Plasticisers are typically oily potar organic liquids, which are very good solvents for many materials. ry sumePolymer ‘9, predominating action of volatiles, kK 40%, 5 karst kp 100% 5% ky 100% 5* Nitrate of ettutose 053 aes 66025 2605 64ex2 Polyiny chloride 054 28e7 75012 2006 ares Polyurethane, ester 084 a7en aaexg 3008 Soest Acetate of cellulose ost 3808 sge23 2ae7 79017 Figure 12, Rate constants of decomposition of some malignant The release of volatiles from the above polymers is seen in the polymers into volatiles as determined by nonisothermal Figure 13. Nitrate of cellulose is the least stable, then itis followed thermogravimetry in nitrogen (k) and by chemiluminescence in by poly(vinyl chloride), polyester urethane and acetate of cellulose, ‘oxygen (k,)for 40 and 100°C The rate constants k, determined from the theoretical fits by eq. 14 for j+3 and 40 and 100°C are in Figure 12 while activation energies for the release of predominating faction of volatiles are 337, 126, 133 and 248 mol’ forritrate of cellulose, poly(vinyl chloride), polyester urethane and acetate of cellulose, respectively. It may be of interest that in the sequence of stability at low temperatures according to the rate constants polyester urethane is more stable than acetate Nivogenthe rate ofheatng of cellulose. However, this is not the case when we consider the - mn’ stability atthe higher temperature region involving the fraction of the polymer which yields the prevailing amount of volatile. High . activation energies ofthe volatile release indicates the run typical for the explosion like course of cellulose nitrate degradation. . The non-isothermal chemiluminescence curves in oxygen for the ‘same polymers are seen in the Figure 14. The sequence of stability is: . cellulose nitrate < poly(vinyl chloride) < polyester urethane < acetate of cellulose and corresponds to that in the nitrogen atmosphere » from thermogravimetry experiments regardless of the fact that the average rates constants determined by means of eq. 1 and 12 . are higher because of the oxygen.The chemiluminescence runs of degradation of nitrocellulose involve typicaly the decomposition peak A ofthe cellulose nitrate itself which is followed by the peak B monitoring the oxidation ofthe char residue being formed from = : original cellulose nitrate, SB rrr In the Figure a5 we see the comparison of stabilities of diferent — Paving tore) Ree acetates of cellulose from the Resinkit™ collection as determined by chemiluminescence and thermogravimetry under non-isothermal Figure 33, Nonisothermal thermogravimetry runs in nitrogen of conditions; the sequence is in both kinds of experiments the same ‘malignant polymers. Points denate te theoretical fit according tothe and corresponds to the order: Eq. 14, Cellulose acetate was powdered sample, polyester urethane and cellulose nitrate ae refered to below and poly(vinyl chloride) is cellulose acetate propionate < cellulose acetate butyrate < cellulose acetate Samco 37 sample ‘othe mass Temperturesc eluose acetate Cellulose nitrate examined in the next experiments was a white membrane used in biology; itis produced by GE Healthcare (US). Scientific assessment of plastics degradation Measurement of polymer degradation 181 ro 181 ry sumeyen the rate of hating smi nes const me piridi Temperatura. — Coos acetate — Celoseniate — Polyester urethane — Potvin tinge) Rexble Figure 14 (Left). Nonisothermal chemiluminescence runs of polymers in oxygen from the Figure 33 Figure a5 (right). Nonisothermal chemiluminescence in oxygen and thermogravimetry in nitrogen for different acetates of cellulose 382 3 ae _— toe * 3 See of 2 3 ic “ ‘Temperature s¢ — Resinkie™ scelulose acetate —— Resinki™ 1-cllulose acetate butyrate — Resinkiem 1-cllulse acetate propionate The quantitative elemental analysis indicated a weight percent of carbon, hydrogen and nitrogen 27.20, 2.93 and 11.76%, respectively. Cellulose nitrate contained two nitro groups per one glucopyranosyl unit of the cellulose. The samples were aged for periods varying from 6 hours until 26 days at 130° in open conditions of ambient air and uncontrolled relative humidity. The aged samples were collected at intervals 6 hours, 1 day, 2 days, 4 days, 8 days and 16 days, respectively. Visual observation confirmed the increase of the yellowing with time of ageing already after 2 days of ageing. The samples became slightly brown after 4 days of ageing and dark brown when aged longer. Due to ageing nitrocellulose becomes brittle. The formation of acid volatiles was evaluated by pH ‘measurements of aqueous extract (1.5 mg of sample was immersed inas0 plof pure water with a stirring bar, during 3 hours). This extract became significantly acidic with the duration of ageing. The pH changed from 6.5 (for the reference sample) to 2.5 (for 16 days- aged sample). The pH values obtained after 1 day, 2 days, 4 days and B days were 4.8, 4.7,3:5 and 3.0, respectively. From the Figure 16 it may be seen how nitration destabilises the original cellulose. The effect of oxygen on the degradation of nitrocellulose is not almost visible on thermogravimetry experimental runs while that for pure cellulose itis significant. This may be explained by the intervention of oxygen into the initiation reaction of oxidation via the carbon atom 6 of glucopyranosy unit of the cellulose while in the case of nitrocellulose the initiation occurs ry sumete Mroctsose cellulose x nitrogen g ‘ongen é nitogen » conyeen| ‘Temperature, °¢ Figure 36 (let). Comparison of nonisothermal thermogravimetry runs of cellulose and nitrocellulose in nitrogen and oxygen. The rate of heating 10°C.min* Figure 47 (right). Comparison of non-isothermal thermogravimetry runs of aged nitrocellulose samples, nitrogen, the rate of heating, 20°C.min® Scientific assessment of plastics degradation Swofthe mass referees « Smee shoe Af, 16577 hays Temperature independently via the scission of bonds -O-NO, . Oxygen affects the oxidation in the subsequent stages of the reaction. The loss of volatiles is also much sharper for nitrocellulose than for cellulose. At high temperatures there remains a certain amount of char residue which under nitrogen looses the weight slowly. In oxygen an acceleration of the char residue oxidation may be observed above 420°C. Non-isothermal thermogravimetry runs ofthe set of cellulose nitrate in nitrogen at the rate of heating 10°C.min* are seen in the Figure 17. Itis worth of noticing that with the progress of the sample ageing at 130°C in ar the temperature ofthe inflexion point shifts to higher values being 197.7°C for reference sample and 207.7°C for the most aged sample. This corresponds with the gradual loss of nitro ‘groups from cellulose nitrate due to ageing. At the same time the percentage of the char residue which remains after thermogravimetry experiment at 550°C slowly increases. ‘The set of parameters obtained from non-isothermal thermogravimetry experiments (eq. 14) such as the activation energy and fractional values a of respective components released a volatiles or remaining as a char residue is given in the Figure 18. Activation energy estimated according to eq.14 for release of volatiles decreases with the progress of nitrocellulose ageing being 260 K).mol for the reference and 130 kj.mol* for the sample aged 16 days in air at 130°C. At the same time, the rate constants k, calculated from Arthenius parameters for 130°C increase with the extent of nitrocellulose ageing. ‘The Figure 19 brings an overview of the rate constants of nitrocellulose decomposition determined from TG, DSC and Measurement of polymer degradation 183 ry sume@ ° ons zo. 78 1 0618 uaa a 4 0.596 1246 062 6 489 664 omy Figure 18, Parameters of degradation of aged nitrocellulose fom non- Isothermal thermogravimetry (eg 14)~a, isthe predominating faction of ‘volatiles formed, ris the corresponding hal. (¢,=0.693/k), k, and E, are corresponding rate constants andthe activation energy, respectively ageas 2607 0.0229 reference sample o, 1G in nitrogen 2009 4506 0.0032 reference sample o, Lin nitrogen : a2e6 : Surface below DSC curves of aged samples, air 5.0¢9 ses 2.0015 ‘sample 16 days aged nat, onygen, slower reaction 4.2010 below 8°) ———_2.9¢6 (above 100°C) 7863 (ottrauer and Fontin 1983) 7009 5764 0.489 Method by Ozawa (Rong et al. 1998) Figure 19. Rate constants ofthe fist order ofthe nitrocellulose cchemiluminescence runs for 50, 130 and 200°C, respectively, Whi Aegradation exposed to ageing obtained by the diferent approach __relatively good agreement may be seen for 200°C with the exception ‘of data from the papers (Pourmortazavi 2009; Rong 1999) which provide much higher values, the temperature close to ambient (50°C) gives the span over several orders of magnitude which are typical particularly for non-isothermal TG and DSC measurements. Even data corresponding to temperature 130°C (the temperature of ageing) are not very far each from the other. It may be of interest that the rate constants from thermogravimetry increase with the extent of sample ageing. This fact should be considered when treating different cellulose nitrate samples ofthe different storage history.Kis aoe Figure 20 (left). The correlation ofthe rate constants approximated from the set of experiments (Figure 17 and Figure 28) with a char residue xa, Figure 21 (right). Example of non-isothermal DSC runs in nitrogen of aged samples of cellulose nitrate, the rate of heating 5°C.min* ‘The numbers denote the days of sample ageing at 130°C Scientific assessment of plastics degradation scm Temperatures —« The rate constants ofthe first order determined from the parameters of the fit using eq, 14 for 130 and 200°C correlate with the percentage ofthe char residue determined from non-isothermal TG runs; those at 130°C with an increasing percentage of the char residue increase (Figure 20). Although the extrapolation of rate constant to 130°C from non-isothermal TG is affected by the faster auto accelerating decomposition process occurring above 190°C, the increasing tendency of the rate constants indicate that there may occur the effect ofthe forming char on the autocatalysis of decomposition of nitrocellulose. Until now the autocatalysis in decomposition of cellulose nitrate was ascribed to the effect of nitric acid. From the flammability experiments on some energetic materials we know that the glowing char may act as a spark ofthe ignition of the explosive course. With the extent of nitrocellulose ageing there also occurs a shift of the maximum of DSC exotherm to higher temperatures while the peak height is reduced (Figure 21). The essential data from DSC measurements in nitrogen forthe rates of heating 5 and 40°C.min® are summarized in the Figure 22 indicating quite a good repeatability of respective experiments. As expected, the maximum of the exotherm is shifted to higher temperatures for the higher rate of heating Provided that the surface below DSC exotherm is proportional to the concentration of unreacted nitro groups and the kinetics of nitrocellulose decay in membrane during its ageing at 130°C is approximated by the first order scheme (Figure 23) we see that the rate constant of nitrocellulose degradation is somewhat lower than Measurement of polymer degradation 185 ry sume@ B70 1723 665 35 20007 ©2078 & 381 655 5a 19946 207.10 g a 562109 200572065 1746895 5.02 9.26 po103 2075 174148 330 15 pou 2077 E 1292 239 aa 203.2 n08 739.40 653 a 22 2066255 Figure 22, DSC parameters forthe aged nitrocellulose — that found by extrapolation from chemiluminescence measurement. comparison of parallel experiments in nitrogen at therate of ‘However, the difference is not as large (Figure 19). heating 5 and 10°C.min* Regardless of the observations provided by thermogravimetry namely that the effect of oxygen on nitrocellulose degradation is not as distinct as its in the case of cellulose it still exists. In Figure 24 ‘we may see the comparison of DSC exotherms for the rate of heating, '5°C.min’; the surface below DSC exotherm in nitrogen is 66% of that for oxygen. When comparing DSC and chemiluminescence runs in oxygen, we see another peak (See also peak B in the Figure 14) appearing inthe latter case (Figure 25) which is likely to correspond with the ‘oxidation of char residue. In nitrogen experiments this peak was not observed. We may also see that in the presence of nitrogen the chemiluminescence signal, that is lower than in oxygen, is also important. Figure 26 shows how chemiluminescence intensity changes with the stepwise increase of temperature. The experiment was performed so that temperature was kept constant for a certain time= , oxygen fe 2° a z: = ie i: i- 2 i fe : ‘nitrogen a rs = = = x Figure 23, The decay ofthe surface below DSC exotherms of the Figure 24. Comparison af DSC runs for nitrocellulose reference sample o) cellulose nitrate aged ina at 130°C in oxygen and in nitrogen. The rate of heating 5°C.min FE swe % J i 3 oe 3 f= i : ¢ : £ = ee i 1) Ew é Temperature,°C Tea Figure 25, Non-sothermal chemiluminescence runsin oxygen and _Figure 26. Chemiluminescence-stepwise increase of temperature for nitrogen (the rate of heating :C.min’ and thermogravimety run nitrocellulose sample In oxygen (the rate of heating 10°C.min’) for reference sample (Sample 0) of nitrocellulose and the chemiluminescence run in oxygen occurred isothermally. The fast increase of chemiluminescence intensity took place already at 180°C. At lower temperatures the isothermal parts of the decomposition curve gave the stationary levels of the intensity of light emission eventually being accompanied by the slow decay. This, experiment also serves as an elegant combination of isothermal and norvisothermal experiments for a fast assessment of plastic stability. ‘The temperature interval well below 190°C which corresponds to the activation energy 89 kJ.mol” gives evidently good rate constants Scientific assessment of plastics degradation Measurement of polymer degradation 187 ro 07 ry sumewom Figure 27. Initial stages of the decomposition of cellulose nitrate 188 when extrapolating to lower temperatures (Figure 19). This allows to predict the remaining service life of the cellulose nitrate more bly. There are no doubts that the initial step in degradation of nitrocellulose is the splitting of -0-NO, bonds of the secondary nitrate group joined to carbon atoms 2 or 3 ofthe glucopyranosyl ring (Figure 27). These bonds have the dissociation energy 167, Kj.mol" while those at the primary position of carbon atom 6 have dissociation energy about 330 kl.mole* (Shashoua 2008). The subsequent steps are more the matter of speculation, especially those leading ultimately to a char. One possibility is that in the sequence of B-scissions glucopyranosyl ring may open, the free radicals R’ split out another molecule of NO, and finally there appear aldehydes, which are prone to oxidation due to direct reaction with oxygen or in its absence with NO.. The transfer reaction to CH, bonds on carbon atoms 6 may induce the splitting off the formaldehyde (Figure 27). Nitric acid which is formed from NO, due to the presence of air humidity will contribute to the cation induced cleavage of glycosidic bonds C-0-C linking glucopyranosyl units and the molar mass of nitrocellulose is reduced. At the same time, after decarbonylation of aldehydic groups the following sequential moieties in the nitrocellulose macromolecules may be formed which form the possible skeleton of the char formed. 3.4.2. Polyurethanes (PUR) Polyurethane foams are widely present in museum collections either as a constituent of the museum artefacts or as a material for their conservation (stuffing, protection, packing and storage) (Waentig 2008, 301; Quye and Williamson 1999). Among museum collections they feature predominantly as sculptures, design objects, cushioning materials, textiles, and toys. Today many of polyurethane foam artefacts are in poor condition and often exhibit specific conservation issues. Instability of polyurethane foams has been recognized although it mainly depends on the raw materials used during manufacturing, There are two types of polyurethane foams: ‘one based on polyether polyols and the other based on polyester polyols. The polyether polyurethanes are particularly susceptible to oxidation. Oxidation is initiated and accelerated by exposure to light, especially UV radiation (Szycher 1999). Foam degradation is particularly devastating usually leading to complete crumbling of the foam object starting at its surface. Polyester urethanes are much less susceptible to oxidative degradation, but are subject to hydrolytic degradation at high relative humidity. Coated and painted foams ry sumeos, ae Riipom ew {O)-mingeoccic bon, ° } Polyther urethanes (oreo emo on onmgeonienit ona Pobetereetanes swine A obi ocmid po { rwtrenesoiclitons Polytherurethane 3 Figure 28.The predominant structures in polyurethanes examined Scientific assessment of plastics degradation were found to be more resilient because they have a protective barrier against oxygen. Three structures of polyurethanes were examined as seen below (Figure 28), namely polyether urethane having toluene (TD) and diphenyl methane diisocyanate (MDI) moieties (Sample 1 and 3) and polyester urethane (Sample 2) with toluene diisocyanate (TD)) units (Rychly 20x). These polyurethanes foams were supplied by RAJA Company, France. The foams were subjected to light ageing under daylight 1000 Wm2, 25°C//50% relative humidity or thermal ageing under dry (90°C, <10% RH) or humid (90°C, 50% RH) conditions. Thermal oxidation of polyester based polyurethanes starts Inpolyisocyanate segments, probably on methylene units adjacent to NH groups. This was confirmed by experiments of chemiluminescence in oxygen of either polyisocyanate or polyol in comparison to polyurethane prepared from the two components. The light emission uptake from polyisocyanate precedes that from polyester polyol. A similar pattern, however, shifted to higher temperatures, was revealed by nonisothermal thermogravimetry in nitrogen (Ravey 1997). Polyester urethane storage during 18 months leads to a considerable change of the chemiluminescence versus temperature patter. The sample became less thermally stable in oxygen and the Intensity ofthe light emission was also lower. At the same time, the ‘molar mass distribution of the polyol did not change. The contact with tap water of a polyurethane film cast on glass caused a reduction of the themooxidative stability of polyurethane. This was attributed to the combined effect of hydrolysis and ions present in the water (Gajewski 1990). The simple and fast differentiation between polyether and polyester urethanes realized through the thermooxidation ‘experiment performed on chemiluminescence device may be seen in the Figure 29. Polyether urethanes due to the ether structures ~CH.,-O- and much easier oxidation perform considerably higher light intensity at final stages than polyesther urethane. However, the degradation ofthe latter starts earlier. In one of our papers (Malikové 2010) we have shown that the development of the chemiluminescence follows after the release of the nitrogen containing moieties and is linked with the oxidation of the crosslinked structures in the polymer. Differentiation of degradation pattem of polyether and polyester, urethanes as seen by thermogravimetry in oxygen and nitrogen is shown inthe Figure 30. Polyester urethane in oxygen undergoes the extensive crosslinking which is accompanied by carbonisation Measurement of polymer degradation 189 ry sumeLiners, counts Temperature — polesterurethane — —— poyether thane Figure 29. Chemiluminescence from polyurethanes, non-aged ‘samples of polyether urethane and polyester urethane, oxygen, the rate of heating 5°C.min* Figure 30. The comparison of nonisothermal thermogravimetry for polyether urethane (line 3) and polyester urethane (ine 2) in nitrogen and oxygen, the ate of heating 5°C.min* 190 2 5 Zo Temperature, °C —torygen | —anivogen — 2onygen — anivogen of the superficial layers and sudden release of volatiles at elevated temperature. The mutual links of the outputs of respective methods (thermogravimetry, DSC, chemiluminescence) are seen in Figures 31 and 32. DSC endotherms in nitrogen (Figure 33) represent the records of the active decomposition of both polyurethanes into volatiles and again confirm the essential differences in polyether and polyester urethanes degradation. Figure 34 shows non-isothermal thermogravimetry records of original polyurethane samples 1 and 2 and those aged by light. The differences are significant, indeed. While original polyurethane foams give in nitrogen two waves ofthe formation of volatiles which were ascribed to the decomposition of polyisocyanate (the first) and polyol or/polyester (the second) moieties, aged polyether samples give only one wave of the decay of the mass (1a). The aged polyether samples (sa) are already disintegrated. Polyester urethanes (2) start to loose volatiles from polyisocyanate moieties at lower temperatures but the mass loss from polyester partis shifted to higher temperature when compared with polyether urethanes 1. Typically much more carbon residue remains on the pan for aged ‘samples (Figure 35) which indicate the additional crosslinking in polyol or polyester part of polyurethane provoked by ageing. ‘Also chemiluminescence ~ temperature records in oxygen are ificantly changed by polyurethane ageing (Figure 36). Wi sample 1a shows a maximum, the samples 2a and 3a give an Increased light intensity at lower temperatures when compared ry sumei 7f ee a, z 2 ° x 7 reac 7 cemetrs 7 Figure 33. Comparison of nonisothermal thermogravimetry and DSC records for polyether (the lines 2) and polyester (the lines 2) urethane foams, oxygen, the rate of heating 5°C.min* : arse z x a9 368°C Temperature, *¢ Figure 33. Comparison of non-isothermal DSC records of polyether urethane (1) and polyester urethane in nitrogen. The rate of heating 5°C.min* Scientific assessment of plastics degradation Figure 32. Chemiluminescence and DSC measurements in oxygen for non-aged polyether urethane (3) and polyester urethane (2). The rate of heating 5°C.min* ofthe mass Temperature, °C Figure 34. The non-isothermal thermogravimetry runs for aged polyether and polyester urethanes samples 12-2a, nitrogen, 1-2 are original non-aged samples, the rate of heating 5°C.min*, The points denote the fit of the experimental run by eg. 14 for=3. The corresponding rate constants are in Figure 37 Measurement of polymer degradation @ Chamiuminescence Intent, couns/s/amg 391ee. J i: i. i. Tie of aging. days Figure 35. The amount of carbonaceous residue remaining atthe pan of TG instrument after sample heating in nitrogen from 4o to '550°C by the rate 5°C.min'. s-aged sample of polyurethane 1, 2- aged sample of polyurethane 2 Figure 36. The chemiluminescence intensity runs for aged polyether and polyester urethanes samples 12-3a (red), oxygen, lines 1-3 are original non-aged samples, the rate of heating '5°C.min*. Points denote the fit of experimental runs by eq. 21 ‘and 12, The corresponding rate constants are in Figure 37 192 ty rng I with reference. One may see that the reference polyether sample 3 sives the faster uptake of the light emission when compared with polyether sample 2 This may be brought about by the oxidation of ‘methylene group in MDI moiety of polyurethane which is converted to carbonyl group with a higher yield ofthe light emission. From the set of rate constants on degradation in oxygen (Cl) and in nitrogen (16) which were determined for 100, 200 and 250°C we may see that pre-ageing of polyurethanes leads to the increase of the respective rate constants which is very pronounced in the case of polyether urethane 1 and less significant in the case of polyester urethane 2 and polyether urethane 3. In the latter case the oxidative attack is probably more focused on methylene groups of MDI The rate constants in nitrogen (T6) are usually lower than those in oxygen (CL) (Figure 37). While the investigation of the changes of the solid matrix of polyurethanes is the subject of numerous papers, only a few studies are devoted to the formation of volatile organic compounds (VOCs) emitted by polyurethanes (Hakkarainen 2008, Lattuatl-Derieux et al. 2011; Thigbaut et a. 2007: Watanabe et al. 2007) during their degradation. Actually, overtime, low molecular weight compounds, either degradation products or manufacturing residual products are emitted from the polyurethane matrix and a relationship between the environmental factors (humidity, temperature and daylight), the volatile fraction formed and the bulk polymer composition should be established. From this viewpoint the sample 1 and sample 2 original and artificially aged were examined by using solid phase micro extraction coupled with gas chromatography mass spectrometry ry sumeSample 3 Figure 37. Average rate constants k,, determined from nonisothermal chemiluminescence (Cl) and thermogravimetry (16 ~ eq. 14) ‘measurements of original and light aged (a) samples of polyurethanes 1,2 nd 3. Temperatures 100, 200 and 250°C Kos a 39e7 aes aae6 7666 1206 3766 kaos 189 3806 4807 ares Scientific assessment of plastics degradation kaos a 98e5 2303 5404 1103 384 6104 kos ses 2364 2306 1484 kas a a7e3 asez 403 7783 ares me 16 5.64 aaea 5.604 ase (SPME-GCMS) and pyrolysis gas chromatography mass spectrometry (Py-GeMs). In addition to sample 1 and 2, four naturally aged polyurethane foams collected from various daily life objects were examined as well, Sample Sa was taken from a suitcase and visually seemed to be in poor state of conservation. Sample S2 was part of a conditioning box and appears to be in a good state of conservation. Sample $3 was taken from a much degraded chair stuffing. Sample S4 was collected from the back of a chair not exposed to light and also ‘seemed to be in a very good state of conservation. Samples Si and ‘53 were yellow and showed a total loss of integrity and a pronounced degradation resulting in powdering. Morphological changes were followed based on visual examinations (binocular); naturally aged foams were preliminary characterised using attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR). Infrared data were also collected from newly made foams after artificial ageing. ‘Similarly as from chemiluminescence, DSC and thermogravimetry data from all artificially and naturally aged samples, visual ith the literature (Kerr and Batcheller 1993; Szycher 1999) it was confirmed that PUR ‘observations and Py-GCMS data and in agreement esters are more sensitive to thermal ageing in humid conditions and hydrolysis and less sensitive to light ageing, while PUR ether degrade primarily by photo-oxidation and have higher resistance to hydrolysis. The main morphological aspects observed visually and the ‘main degradation compounds analytically characterised from the artificially and naturally aged PUR esters and PUR ethers are ‘summarized in Figure 38. Itis interesting to note that artificial and natural ageing were providing similar analytical results and thus, it seems well relevant to use such accelerated ageing to simulate natural degradation processes. The results are illustrated by the commercial polyurethane ester (Sample 2) and polyurethane ester samples S1 and S2. After eight weeks of ageing in humid condition, the commercial polyurethane Measurement of polymer degradation @ 193,Samples and ageing ‘Main morphological aspects Main degradation compounds Humid ageing of commercial PUR ester ample 2) seweeks ageing time ‘Slight weakening and darkening EG, TOA, TA ‘.weeks ageing time Important weakening and darkening. Presence of adi acid rystals DEG, TDA,TAL AA Light ageing of commercial PUR ether ‘zorhous ageing time Sight ylowing and racking on the surface 6 24o-hout ageing time Important yellowing cracking and crumbling resulting in powdeing 6 ‘Si: PUR ester from suitease Poor state of conservation: very yellowed total los of coherence DEG, TOA, AA resutng in powering nd presence of adipic ci rsa S2:PURester fom condoning box Good state of conservation DEG, Ta 53: PUR ether fom cai stung Poot state of conservation ver yellowed and total loss of coherence 6 resutng in powering S4:PURetherffom the back of achalr Good state of conservation 6 Figure 38. Main morphological aspects visually observed and main degradation compounds analytically characterised by Py-GCMS and SPME-GCMS from the selected artificially and naturally aged foam samples 194 ester, besides the important loss of mechanical properties and darkening, gives white crystals of adipic acid (AA) which appear inside the pores and on the surface of the foam. Adipic acid crystals were also observed in naturally aged foam Si. Such crystals have been also detected and identified from one polyurethane museum artefact (van Oosten 2002). The Py-GCMS analysis permits to identify potential chemical markers ofthe polyurethane ester degradation. Figure 39 shows the pyrograms of commercial Sample 2 unaged and artificially aged in humid conditions during 4 and 8 weeks. In the unaged Sample 2, we identified the compounds such as carbon dioxide, adipic ketone (K), diethylene glycol (DEG), 2,4- and 2,6- toluene diisocyanate isomers (TDI) and some adipate derivatives (A) of which chemical structures have not been identified. The adipic ketone (C,H,0, MW 84) can originate from the adipic acid raw ‘material (Font 2001, Ohtani 1987). The DEG (C,H,,0,, MW 106), a low ‘molecular weight polyol is used in the polyurethane synthesis either a chain extenders or component of the polyester polyol synthesis. During the artificial ageing, the chromatographic fingerprints become simple and the isocyanate moiety was modified. After four weeks of ageing, in addition to the relative increase of diethylene slycol, the occurrence of toluene amino isocyanate isomers (TAl) and 2,4- and 2,6-toluene diamine isomers (TDA) may be noted. TAI may undergo hydrolysis to TD! (Berlin 2005) and 2,4- and 2,6- toluene diamines are being formed. An increasing release of 2,4- and 2,6- TDAwith duration of 2,4- and 2,6-TDI hydrolysis have been reported ry sume‘un ae ay 586) Figure 39 (left). Pyrograms of commercial unaged and artificially ‘aged Gn humid conditions) polyester-based polyurethane foams {Sample 2). Compounds marked () correspond to: (K adipic Ketone, {(O€6) diethylene giyco, (TDD toluene diisocyanate isomers, (TDA) toluene (AA) adipic acid, (A) adinate derivatives, the chemical structures of hich have not been identified iamine isomers, (TA) toluene amino isocyanate isomers, Figure 4o (right). Pyrograms of naturally aged polyester-based polyurethane foams collected (A) from a sultcase $1 and (8) froma conditioning box $2, Compounds marked () correspond to: (0) adipic ketone, (DEG) diethylene glycol, (TD! toluene diisocyanate isomers, (TDA) toluene diamine isomers, (TAD toluene amino isocyanate ‘isomers, (AA) adipic acid, (A) adipate derivatives, the chemical structures of which have not been identified, (P) phthalate esters, Figure 42. Chromatograms of the SPME extracts from naturally aged polyester-based polyurethane foams collected (A) from a suitcase ‘St and (6) from a conditioning box S2. Compounds marked () correspond to: (DES) diethylene glyco, (P) phthalate esters Scientific assessment of plastics degradation ounce ae ° ; ‘i es —- © . te es Tine (nnd Measurement of polymer degradation 195 ry sume196 previously (Lind 1996). After eight weeks of ageing, adipic acid was identified. re 40 presents two Py-GCMS fingerprints of Sx and S2 foams Which were in different states of degradation. Pyrograms previously obtained from commercial samples helped us in their identification; they are from poly(diethylene glycol adipate) polyurethanes. Phthalate derivatives (.e. phthalic acid diisobutyl ester) could inate from plasticisers introduced during the manufacturing process. Sample S1 gave a relatively high amount of diethylene slycol (DEG), 2.4- and 2,6-toluene diamines (TDA) and of adipic acid (AA). Pyrogram of the foam S2 which was ina relatively good state showed a lower amount of diethylene glycol (DEG), presence of toluene amino isocyanate isomers (TAl) and no adipic acid. These results were coherent with those obtained from artificial ageing and therefore the presence of these compounds can be used as degradation markers. Chromatograms of the SPME extracts from the commercial ester foam artificially aged in humid conditions did not indicate the obvious changes within the VOCs composition except for the presence of diethylene giycol (DEG). Chromatograms of the SPME extracts from naturally aged PUR ester foams were very similar but, the presence of diethylene glycol in different relative amount was noted. Figure 41 presents two SPME fingerprints from foams S1 and 2. The relative increase of diethylene glycol could be correlated with the physical state ofthe foams. 3.2, The progress of degradation in normalized coordinates {As the maximum intensity of chemiluminescence changes from the polymer to the polymer an idea was proposed to plot the respective ‘uns in the normalised coordinate of the intensity. Normalised intensity is the actual intensity divided by the maximum intensity. In ssuch a way intensity at the maximum is equal toa and the differences in the lower temperature regions due to sample ageing may well be differentiated, From the viewpoint of long term stability of polymers, the region of lower temperatures is of higher interest. Uniform approach may be introduced, namely to compare the temperature of the section of normalised graphs up to 10% of the maximum chemiluminescence intensity measured at the beginning and at the end of the polymer exposition. The example may be seen on artifical doll monitoring before and that after 17 months of its exposition in laboratory ry sumeMpa Tepeture, _2r-polyrpytene homopolymer) ‘ap aletytne (medium dns) _24-paletylne (ow density) 25 alatylne (igh density) 2s polypropylene copolyme) 2plonomer 2apotyuryiene Figure 42. Chemiluminescence from some polyolefins in the Resinkit™ in normalized coordinates Scientific assessment of plastics degradation conditions. The most significant changes were observed for polyester urethane foam (shoe of the dolD -32°C, polystyrene head -31°C, polyether urethane foam (shoe of the doll) -23°C and PET body 22°C. The changes on polyurethane foams were already visible (yellowing). The items least touched by ageing were polyamide skirt, acrylic halts and low density polyethylene arms, The more detailed observations related to the long term ageing ofthe doll at different locations are reported in the Chapter 2. 3.43. CLof polymers in the ResinKit™ Atlas of non-isothermal thermogravimetry, DSC and cchemiluminescence was prepared for those wanting to have the fast, orientation in the degradation of the respective polymer. Here we present the data on chemiluminescence from the polymers of the ResinKit™ extracted from this Atlas. Al chemiluminescence experiments were carried out in oxygen atmosphere and at the rate of heating 5°C.min*. The measurements were performed with powders obtained by grinding of corresponding Resinkit™ sample. The example of the chemiluminescence measurements on Resinkit™ hydrocarbon polymers in normalised coordinates is seen in Figure 42 and parameters determined for each iter of the Resinkit™ are in the Figure 43. 3.4.4, TGA of ResinKit™ polymers after two years of storage Polymers in Resinkit™ are usually considered to be rather stable over years when kept under laboratory conditions. This assumption, however, should be taken with some reservation. As tis shown in Figure 44 where non-isothermal DSC and thermogravimetry records are presented for polyolefins from the Resinkit™, thermogravimetry in nitrogen shows the observable shift of corresponding line to lower temperature region after two years of Resinkit™ storage under laboratory conditions. This shift appears the most significant for polypropylene homopolymer which belongs to polymers which are rather sensitive to the oxygen effect. Thermogravimetry line observed in oxygen may be found at temperatures more than 100°C below that in nitrogen, Jozef Rychly, Lyda Rychla and Agnés Lattuati-Derieux Measurement of polymer degradation 197 ry sumeResinkit™ No, Polymer 198 ie Ble Bs 3 3 35 37 39 ra a 2 4 a 45 6 a ~° 50 Polystyrene (general purpose) Polystyrene (medi impact) Polystyrene (High impact) Styrene Acryonitile (SAN) ‘Acryloitle-butadiene-styene (ABS)—transparent ‘Acrylnitlebutadiene-styene (ABS)-medium impact ‘Acylritlebutadiene-styrene (ABS)-high impact Styrene butadiene block copolymer Acrylic Modified ary Cellulose acetate Cellulose acetate butyrate Cellulose acetate propionate Nylon (transparent Nyion type 66 Nylon ype 6 Thermoplastic polyester PBT) ‘Thermoplastic polyester (PETG) Polyphenylene oxide (PPO) Polycarbonate Palysyufone Polybutylene lonomer Polyethylene (ow density) Polyethylene (High density) Polypropylene (Copolymer) Polypropylene (Homopolyme) Polypropylene (Baryum sulfate reinforced) Polyvinyl chiovie) (PVC exible) Polyviny chore) (PVC rigid) ‘Acetal resin (homopolymer) ‘Acetal resin (copolyme) Polyphenylene sulfide (PPS) Eylene vinyl acetate copolymer (EVA) Synthetic elastomer (Styrene block copolymer) Polypropylene (ass filed) Urethane elastomer, thermoplastic (TPU) Polypropylene (Fame retardant) Polyester elastomer -Acrylritle butadiene styrene (ABS) flame retardant Polyaliomer Styrene terpolymer Palymethy! pentene Polypropylene (Talc reinforced) Polypropylene (Calium carbonate reinforced) Polypropylene (Mica reinforced) Nylon type 66 ~ 33% glass ‘Thermoplastic rubber (TPY) Polyethylene (medium density) ‘Acylnitle butadiene styrene (ABS) nylon alloy Temperature ofthe so% uptake of CLintensity Maximum CLntensity 1. compared tothe maximum, aay 1836 1800 1942 817 1925 967 234 75 1950 2099 2083 1890 1949 50 165.0 2258 1942 2017 1583 1608 208 25 257 283 237 238 198.0 1823 1883 2092 2032 466 2287 1792 25 205.0 2464 1858 m7 1967 2375 208. 1928 sw7a 1558 242 ma wo counts/s/sm 24861 aasee 50597 5383 2240%6 707 708 143 846 7328 5905 ss7ai2 a3922 499309 49070 703760 206 2.0806 79407 122604 134552 20067 15389 109036 2091 12406 006 0906 18306 1406 29628 18825 903 323 8507 84 062 399982 93228 0406 461 651158 885 600 25133 1.2906 269 442 1.0306 as72 551219 m7 703 760 238 687 1.66 907 328s,m sc, Resinki™ 25 ~ Polythene ow density) trogen, the ate of eating Cin “Temperature, °¢ Resnkt™ 49 ~ Polyethylene (esium density) trogen, the ate of heating, 58min Temperature, °C Figure 43 (preceding page). Parameters from non-isothermal chemiluminescence of polymers from Resinkit™ Scientific assessment of plastics degradation wottne mass ‘ofthe maze Resin ™ 25 ~ Polyethylene igh density) i if ¥ z : a a eg Tonpetrc tein -Polpapere ones aren] z “i # zg a Nvogen the rat of eating. si Temperature, °C Figure 44. Nonisothermal thermogravimetry in nitrogen (magenta tine) ‘of some Resinkit™ polymers after two years (blue lines) ofthe storage In airat laboratory conditions. Greenline represents the corresponding, DSC recotd. The orange line shows the comparison with degradation in oxygen Measurement of polymer degradation 199 @