- Email: [email protected]
wood flour composites
Polymer Degradation and Stability 93 (2008) 1252–1258
Contents lists available at ScienceDirect
Polymer Degradation and Stability journal homepage: www.elsevier.com/locate/polydegstab
Accelerated weathering of polypropylene/wood flour composites F.P. La Mantia*, M. Morreale ` di Palermo, Viale delle Scienze, 90128 Palermo, Italy Dipartimento di Ingegneria Chimica dei Processi e dei Materiali, Universita
a r t i c l e i n f o
a b s t r a c t
Article history: Received 20 February 2008 Received in revised form 7 April 2008 Accepted 15 April 2008 Available online 18 April 2008
Wood–plastic composites (WPCs) have received increasing attention during the last decades, because of many advantages related to their use. Some of their main applications are represented by outdoor furnishing and decking; therefore, it is important to assess their behaviour under UV exposure. In this work, polypropylene/wood flour composites were prepared and their resistance to photooxidation investigated. The composites were prepared by extrusion and compression moulding, and were subjected to mechanical tests, FTIR analysis and molecular weight measurements. The results showed that the composites retained a higher fraction of the original mechanical properties after accelerated weathering; the wood flour did not significantly degrade throughout the irradiation time slot of the investigation and the composites kept a higher percentage of the original molecular weight. Ó 2008 Elsevier Ltd. All rights reserved.
Keywords: Polypropylene Wood Composites Photooxidation Molecular weight FTIR spectroscopy
1. Introduction Wood–polymer composites (WPCs) have attracted a significant interest in the last decades, thanks to the specific advantages they can grant in comparison with the classic mineral filler/plastic composites. These include mainly the improved environmental performance, due to the use of biodegradable materials and the reduction in the use of non-renewable (oil based) resources throughout the whole life cycle of the composite [1]; the low cost of wood flour and of natural-organic fillers in general (since they often come from wastes); the lower specific weight of these fillers, in comparison to the traditional mineral-inorganic ones; the improvement in safety for the production employees (reduced hazard in the case of accidental inhalation); the special aesthetic properties of the composites, which can be conveniently processed and refined, obtaining wood-like looking products; the full recyclability of the composites. These materials can be used for many indoor and outdoor applications (panels for the automotive industry, decking, furnishing, packaging, etc.) [1–8]. Polyolefins, in particular polypropylene, one of the most widely used plastics, have been extensively studied in combination with wood derivatives (flour, flakes, fibres) [9–12]. Several researchers have focused their attention on the improvement of the mechanical properties (usually deteriorated after the addition of the wood flour, particularly the ductility), achieved with the use of small amounts of coupling agents, which improve the interfacial polymer–filler adhesion and the dispersion of the filler within the matrix [13–15]. * Corresponding author. E-mail address: [email protected] (F.P. La Mantia). 0141-3910/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.polymdegradstab.2008.04.006
Significant interest has also arisen about the outdoor performance of these composites, in particular their resistance to photooxidation. Polyolefins modified with mineral fillers and exposed to accelerated weathering have shown somewhat conflicting trends, as reported by Rabello and White [16], who studied the photooxidation of PP–talc blends and found an inversion of the trend of molecular weight against exposure time for the neat and the filled PP, upon increasing the exposure time. This was explained by considering the screening effect of the mineral particles, which is contrasted by the presence of chromophore impurities. Similar results have been found by Benavides et al. [17] for PP–CaCO3 composites. Valadez-Gonzalez et al. [18] found a stabilizing effect of CaCO3 in HDPE composites, but for the first 300 h of exposure the neat polymer showed higher molecular weights. The influence of the filler was complex, involving variations in the morphology, the mechanical and thermal stresses during processing and then the sites where photodegradation can start. Yang et al. [19] compared several mineral fillers for HDPE-based composites and found that calcium carbonate and wollastonite had a stabilizing effect, while kaolin, diatomite and mica accelerated photodegradation. The proposed explanation for the different behaviours took into account the relative absorbances of the fillers for UV light in the 290–400 nm range. Similarly, studies regarding photooxidation of wood filled polymer composites reported a variety of results. Alexy et al. [20] reported a light stabilizing effect of lignin (which is one of the major components of wood) to PP when added in low concentrations (up to 10 wt%), while it accelerated photodegradation of PE at amounts higher than 10 wt%. Environmental degradation on HDPE/wood composites as investigated by Li [21] resulted in limited effects on
F.P. La Mantia, M. Morreale / Polymer Degradation and Stability 93 (2008) 1252–1258
1253
Fig. 1. Dimensionless elastic modulus as a function of the irradiation time.
the composites (confined to the surface layer) in dry environment conditions, while water had an important degradative action. Matuana and Kamdem [22] studied the photooxidation of PVC/ wood composites subject to accelerated weathering, finding that the composites retained a higher percentage of the original mechanical properties with the exposure time, in comparison to the neat polymer, explained by admitting that the photooxidation involves only the surface of the exposed composite; similar results were also found by Lundin et al. [23]. However, the increase of the carbonyl index suggested an accelerating effect caused by the wood particles, attributed to the presence of chromophoric groups. Stark et al. [24–27] extensively studied the behaviour of HDPE/wood flour composites exposed to accelerated weathering. They focused the attention especially on the discoloration of the composites, and found also that the presence of wood causes an increase of the carbonyl index, attributed to the formation of chromophore groups following the degradation of lignin; the degradation can be contrasted or at least reduced by using UV absorbers, stabilizers or maleated coupling agents. Selden et al. [28] found similar results
regarding the carbonyl index of the composites, however, the photodegradation was confined to the surface, due to the screening effect of the wood particles; the maleated coupling agents had no significant influence on the degradation. Pucciariello et al. [29] blended several polymers (HDPE, LDPE, LLDPE, polystyrene) with lignin, finding that the latter acts as a stabilizer against UV radiation for PS, LDPE and LLDPE, mainly on the basis of molecular weight measurements. Muasher et al. [30] focused on the efficiency of photostabilizers on the colour change of WF-filled HDPE subject to photooxidation. It was found that the main responsible factor in wood photodegradation is lignin, which breaks down to water soluble products, that eventually lead to the formation of chromophoric functional groups. In this work, polypropylene/wood flour composites were prepared and subjected to accelerated UV weathering. The effects of the exposure to UV radiation were studied analyzing the mechanical properties, the chemical changes and the variations of the molecular weight. The photooxidation mechanisms of both the polymer and the filler were taken into account.
Fig. 2. Dimensionless tensile strength as a function of the irradiation time.
1254
F.P. La Mantia, M. Morreale / Polymer Degradation and Stability 93 (2008) 1252–1258
Fig. 3. Dimensionless elongation at break as a function of the irradiation time.
2. Experimental 2.1. Materials The polypropylene (PP) used in this work was an injection moulding grade commercially known as ‘‘Moplen X30G’’ (Basell Polyolefins, Italy), with a density of 0.9 g/cm3, melt index z 8 g/ 10 min (at 230 C and 21.6 N), melting point about T ¼ 170 C. This class of general purpose polypropylenes usually contains small amounts of phenolic primary antioxidants and phosphite secondary antioxidants. The wood flour, kindly provided by La.So.Le (Italy) and marketed as ‘‘150/200’’ type, has an average particle diameter of 150–200 mm and a humidity content (as delivered) around 5%; it is further indicated throughout the paper as ‘‘SDF’’.
2.2. Compounding and testing The wood flour was dried in an oven at 70 C overnight in order to drastically reduce the humidity content, and processed with the polypropylene in a co-rotating twin screw extruder (OMC, Italy, D ¼ 19 mm, L/D ¼ 35), with the nominal composition of 30 and 60% (by weight). The temperature profile was set at 120–130–140–150– 160–170–180 C and the screw speed at 200 rpm. All the materials were pelletized after extrusion by a rotating blade system. The specimens for the tensile and accelerated weathering tests were obtained by compression moulding, using a Campana (Italy) laboratory press (set at 180 C, pressure about 200 bar, compression time 2–3 min) and a 300 mm thick mould. This thickness value was chosen in order to not compromise the internal structure of wood particles. The sheets were placed in a Q-UV (USA) chamber, mounting eight UV-B lamps, which reproduces the damage caused by sunlight, rain and dew. The weathering cycle conditions were 8 h of light at T ¼ 55 C and 4 h of condensation at T ¼ 35 C. Tensile tests were performed by an universal Instron (USA) model 3365 apparatus on specimens (thickness z 0.3 mm, width ¼ 10 mm) cut out from the previously described sheets, following ASTM D882. At least five replicates were tested for each tensile property and the average values are reported with satisfactory reproducibility (7%). The changes in the sample chemistry upon UV exposure were analyzed by means of a Perkin–Elmer (UK) Spectrum One FT-IR spectrophotometer (in the range 450–4000 cm1, absorbance units). The carbonyl index (CI) was evaluated as the ratio of the area of the
peak between 1680 and 1800 cm1 and the area of the peak centred at 2722 cm1 (which is the characteristic vibration of stretching band of PP). At least four replicates were tested for each FTIR spectrum and the average values of the CI are reported with a good reproducibility (5%). The variations of the molecular weight were monitored by intrinsic viscosity measurements. The polymer was extracted from the samples (unweathered and weathered) by dissolving them in hot tetrahydronaphthalene (THN, 135 C) and separating the wood fibres using a vacuum filtering unit, made up of a flask connected to a water vacuum pump and a sintered glass filter with disposable paper filters (Gelman Sciences, USA). Ethanol was added to the filtered solution which was allowed to cool in order to cause precipitation of the PP, which was then recovered by a second filtration stage and used for the intrinsic viscosity measurements. These were performed using a Schott-Gerate viscometer in an automatic apparatus, SchottGerate model AVS 300. The solutions were prepared in hot THN at a concentration of 0.4%. The intrinsic viscosity was calculated using the Salomon–Ciuta equation [31]. pffiffiffiqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 hsp ln hr ½h ¼ (1) c where c is the concentration, and hsp and hr the specific and relative viscosities, respectively. At least four replicates of each sample were tested and average values (reproducibility 7%) are reported. 3. Results and discussion 3.1. Mechanical properties The results of the tensile tests are shown in Figs. 1–3, where the main tensile properties are plotted as a function of the irradiation time. It is worth pointing out that it was difficult to obtain the data for the 280 h irradiated neat samples, since they tended to break up easily when manipulated, so several specimens had to be excluded from the calculations. Table 1 Mechanical properties of the unweathered materials Sample
Elastic modulus (MPa) Tensile strength (MPa) Elongation at break (%)
PP 1225 PP30SDF 1990 PP60SDF 2040
32.8 23.3 13.4
9.4 2.9 1.2
F.P. La Mantia, M. Morreale / Polymer Degradation and Stability 93 (2008) 1252–1258
1255
Fig. 4. FTIR spectra of some neat and filled samples.
The values have been reported in dimensionless form – value at a given time divided by the value for the unweathered samples (Table 1) – to better highlight the kinetics of the variations of the properties. It can be observed that the elastic modulus changes are limited during the first 150 h of irradiation, while it starts to decrease significantly for longer exposure times. However, thanks to the stiffening effect of the filler, it stays higher in the filled samples. The tensile strength decreases significantly in the neat PP, because of the chain scission reaction induced by the irradiation, while the changes are less pronounced in the filled samples. The explanation can be easily found considering that filled samples have a lower amount of polymer subjected to the photooxidative degradation; in fact, the decrease of tensile strength upon increasing the irradiation time is, on a percent scale, smaller in the 60% filled samples. The effect of wood particles is even more pronounced for the elongation at break. The decrease of the deformability upon increasing the UV exposure time is dramatic in the neat polypropylene, while is much less significant in the filled samples,
where it stays almost constant. However, it is worth noting that the filled materials are very fragile even in the virgin conditions. These results show, on average, an advantage of the composites in comparison to the unfilled, neat polymer, when subjected to accelerated weathering. However, it has been pointed out that this is probably due to the lower amount of polymer used and to the fact that the photooxidation mainly affects the polymer itself. More precisely, mechanical tests alone could not be sufficient to assess the real influence of the wood flour on the photooxidation mechanisms. This could be understood by specifically analyzing the chemical modifications of the materials and in fact, as highlighted in Section 1, many literature studies have been reported which use spectroscopy to detect the formation or disappearance of specific chemical groups. However, in order to have a real understanding of the photooxidation the material has undergone, these investigations should be coupled with the measurement of the molecular weight. As hinted in Section 2, both investigations have been carried out on the materials studied in this work.
Fig. 5. Carbonyl index of the materials as a function of the irradiation time.
1256
F.P. La Mantia, M. Morreale / Polymer Degradation and Stability 93 (2008) 1252–1258
Fig. 6. FTIR spectra of neat SDF, SDF extracted from 120 h irradiated 30% filled samples, and SDF extracted from 120 h irradiated 60% filled samples.
3.2. Fourier-transform infrared spectroscopy The IR spectra of the neat and filled samples are shown in Fig. 4. In particular, only the spectra of pure PP (weathered for 72 and 120 h) and of 30% filled PP (weathered for 72 and 120 h) are reported to make the observation of the graphs easier. The spectra showed, as predictable, an increase in the carbonyl peak upon increasing the irradiation time. This is a well-known effect directly linked with the photooxidation of the polypropylene. At the same time, an increase in the carbonyl peak was observed upon increasing the wood flour content. A quick, rough explanation could be that the wood flour accelerates the photodegradation rate of the material, resulting in a higher amount of carbonyl groups. However, a more rigorous way to assess the presence of carbonyl groups consists in taking into account the actual amount of polypropylene, and then calculating a carbonyl index according to the method described in Section 2. Therefore, a diagram has been prepared, plotting the carbonyl index versus the irradiation time for all the samples, and is shown in Fig. 5 (where all the values are normalized by subtracting the values of the unweathered samples). It can be clearly seen from the plot of the photooxidation rate that the latter is slower upon increasing the wood flour content. These results upset the rough considerations which could be made just upon calculating the carbonyl peak area (even if normalized with the thickness of the samples). This is in agreement with previous studies, as highlighted in Section 1, but a deeper explanation must be provided and therefore the role of the wood flour has been analyzed thoroughly. The FTIR spectra of the wood flour (before being compounded with the polymer and processed and after THN extraction from the samples) are shown in Fig. 6 (only some representative spectra are reported).
Besides the carbonyl peak, other significant peaks are in the 1425–1510 cm1 region linked with lignin. It is known from the literature [30,32–34] that the main factor responsible for wood photodegradation is lignin, which absorbs UV/vis light because of its chromophoric groups. The degradation mechanism is complex, with different paths leading to water soluble products and, finally, to chromophoric groups like carboxylic acids, quinones or hydroperoxides. In order to accurately assess the photodegradation of wood flour, therefore, it is first necessary to find one (or more) reference peak. According to previous studies [33], the phenomena can be monitored by comparing the carbonyl peak area (here having the maximum at approximately 1738 cm1), and the area of the 1505 cm1 peak (which is the aromatic band of lignin), to the peaks at 1375 and 1158 cm1, which are related to carbohydrates and are not significantly affected from degradation. Therefore, four indices (two for lignin, two for the carbonyl region) have been calculated and are reported in Table 2. The results show clearly that there are quite small differences in the indices of the neat, ‘‘as-delivered’’ wood flour and the same extracted from processed and irradiated samples. The same trend was observed also in the other samples (which therefore are not reported here). Therefore, it can be concluded that the wood flour does not undergo significant degradation within the time scale of observation adopted in this investigation. This means, furthermore, that no significant amounts of chromophoric groups are produced and therefore the wood flour does not accelerate the photooxidative processes of polypropylene, confirming the previous comments regarding the composite spectra. The lignin does not degrade significantly in this experimental study probably because the wood flour used comes from beech, a hardwood, and it is known that hardwoods degrade at a slower rate than softwoods [33]. Furthermore, the lignin, staying almost unaltered, can exert a protective action as suggested by Pucciariello et al. [29] and Alexy et al. [20].
Table 2 Variations of relative intensities of lignin (C]C) peak at 1505 cm1 and carbonyl (C]O) peak at 1738 cm1 for different SDF samples
Table 3 Intrinsic viscosities of neat and composite samples at different irradiation times
Sample
A (1505/1161)
A (1505/1375)
A (1738/1161)
A (1738/1375)
Sample
0h
72 h
120 h
280 h
SDF SDF-30%-120 h SDF-60%-120 h
1.21 1.11 1.18
1.35 1.17 1.24
4.51 4.82 4.65
5.21 5.37 5.05
PP PP30SDF PP60SDF
1.42 0.77 0.47
0.75 0.65 0.41
0.49 0.29 0.37
0.36 0.24 0.30
F.P. La Mantia, M. Morreale / Polymer Degradation and Stability 93 (2008) 1252–1258 Table 4 Weight average molecular weights of neat and composite samples at t ¼ 0
Mw at t ¼ 0 h
PP
PP30SDF
PP60SDF
237 000
90 000
41 000
3.3. Molecular weight measurements The intrinsic viscosity of the polypropylene (neat and extracted from the composites) was calculated according to the previously described method, providing the results shown in Table 3. The intrinsic viscosities can be quickly converted into calculated values of the weight average molecular weights by the following equation [35]: Mw ¼
½h 5:77 104
1=0:631
(2)
Table 4 reports the calculated values of molecular weight (Mw) of the materials at t ¼ 0, while Fig. 7 reports the dimensionless values at different irradiation times. The normalization was carried out dividing the molecular weight at time t by the molecular weight of the same material at t ¼ 0. A first noteworthy observation concerns the values of the unirradiated samples. The composites have lower molecular weights, as lower as the filler content increases. This is due to the higher mechanical stresses which arise during processing, related to the higher viscosity caused by the presence of the filler, as already discussed in our previous studies on similar systems [15,36,37]. In fact, the thermo-mechanical degradation during processing also affects the neat polypropylene, since the weight average molecular weight of the virgin PP is about 300 000 and it decreases, after processing, to the value reported in Table 4. The higher mechanical stress generated during the processing of the composites leads, therefore, to a significantly greater thermomechanical degradation. The different samples (neat, 30% filled, 60% filled) actually have polypropylene matrices with different molecular weights exposed to accelerated weathering. This fact may significantly affect the degradation paths in the case of thermo-mechanical degradation, since different molecular weights give rise to different viscosities and thus to different extents of chain scission; but in the case of UV exposure, the different chain length does not seem to affect the photooxidation reactions. This is
1257
also confirmed by the FTIR analysis, which did not indicate different photooxidation mechanisms in neat and filled materials. An important feature that arises while observing the variations the molecular weight undergoes upon increasing the exposure times is that the composites retain a higher percentage of the molecular weight in comparison to the pure PP, and this trend is even more pronounced in the 60% filled materials. This suggests that the wood flour exerts a stabilizing effect by slowing down the photooxidation and this explanation is in agreement with the conclusions drawn from the FTIR analysis. A further possible explanation of this stabilizing effect may be found considering also that wood contains a series of ‘‘extractives’’ (resins, tannins, polyphenols, waxes, esters of the C16–C18 fatty acids, terpenes, etc.) which it seems could act as antioxidants and radical quenchers [34]. Furthermore, the fact that the commercial PP used in this work has a general purpose phenolic stabilizing system may also have contributed to the substantial stability of lignin discussed above. Finally, the stabilizing effect of wood flour could be explained also by taking into account the possibility for the filler to act as a ‘‘screen’’, confining the photodegradation to the outer layers of the samples [16,22,23,28]. 4. Conclusions Polypropylene/wood flour composites were prepared and their resistance to oxidative photodegradation was investigated. The mechanical tests showed that the composites retain a higher fraction of the original mechanical properties after accelerated weathering, especially with regard to elongation at break, whereas the neat polymer undergoes a dramatic drop of the mechanical properties. FTIR analysis showed that the increase in the carbonyl peaks of composites is mainly due to the presence of wood flour, but this does not provide any acceleration of the photooxidation rate in comparison to the neat polymer upon increasing the irradiation time, since the wood flour (and, in particular, lignin) does not degrade significantly throughout the irradiation times of this study; indeed, wood flour effectively acts as a stabilizer, slowing the photooxidation. This was confirmed also by molecular weight determination. This effect is probably due to the action of lignin, and maybe of the extractives. At higher irradiation times, it is likely that the trend would be reversed, since the degradation of wood leads to the formation of chromophoric groups which accelerate photooxidation and yellowing; in this case, the inversion time would depend on the relative
Fig. 7. Dimensionless weight average molecular weights of neat and composite samples as a function of the irradiation time.
1258
F.P. La Mantia, M. Morreale / Polymer Degradation and Stability 93 (2008) 1252–1258
weight of the two opposing phenomena and therefore would strongly depend both on the processing and weathering conditions. References [1] Joshi SV, Drzal LT, Mohanty AK, Arora S. Compos Part A 2004;35:371–6. [2] Rozman HD, Lai CY, Ismail H, Mohd Ishak ZA. Polym Int 2000;49:1273–8. [3] Gachter R, Muller H. Plastics additives. 3rd ed. Munich: Hanser Publishers; 1990. [4] Canche`-Escamilla G, Rodriguez-Laviada J, Cauich-Cupul JI, Mendizabal E, Puig JE, Herrera-Franco PJ. Compos Part A 2002;33:539–49. [5] Coutinho FMB, Costa THS, Carvalho DL. J Appl Polym Sci 1997;65:1227–35. [6] Balasuriya PW, Ye L, Mai YW, Wu J. J Appl Polym Sci 2002;83:2505–21. [7] Raj RB, Kokta VB, Maldas D, Daneault C. J Appl Polym Sci 1989;37:1089–103. [8] Netravali AN, Chabba S. Mater Today 2003;6:22–6. [9] Carroll DR, Stone RB, Sirignano AM, Saindon RM, Gose SC, Friedman MA. Resour Conserv Recycl 2001;31:241–51. [10] Selke SE, Wichman I. Compos Part A 2004;35:321–6. [11] Avella M, Bazzi C, dell’Erba R, Focher B, Mazzetti A, Martuscelli E. Angew Makromol Chem 1995;233:149–66. [12] Marcovich NE, Reboredo MM, Aranguren MI. Polymer 1999;40:7313–20. [13] Harper D, Wolcott M. Compos Part A 2004;35:385–94. [14] Keener TJ, Stuart RK, Brown TK. Compos Part A 2004;35:357–62.
[15] La Mantia FP, Morreale M. Compos Interface 2007;14:685–98. [16] Rabello MS, White JR. Polym Compos 1996;17:691–704. [17] Benavides R, Gonzalez-Hernandez R, Gonzalez-Cantu` MC, Reyes-Vielma B, Billingham NC. J Vinyl Addit Technol 2003;9:41–9. [18] Valadez-Gonzalez A, Cervantes-Uc JM, Veleva L. Polym Degrad Stab 1999;63: 253–60. [19] Yang R, Yu J, Liu Y, Wang K. Polym Degrad Stab 2005;88:333–40. [20] Alexy P, Kosikova B, Podstranska G. Polymer 2000;41:4901–8. [21] Li R. Polym Degrad Stab 2000;70:135–45. [22] Matuana LM, Kamdem DP. Polym Eng Sci 2002;42:1657–66. [23] Lundin T, Cramer SM, Falk RH, Felton C. J Mater Sci Eng 2004;16:547–55. [24] Stark NM, Matuana LM. J Appl Polym Sci 2003;90:2609–17. [25] Stark NM, Matuana LM. Polym Degrad Stab 2004;86:1–9. [26] Stark NM, Matuana LM. J Appl Polym Sci 2004;94:2263–73. [27] Stark NM, Matuana LM. Polym Degrad Stab 2007;92:1883–90. [28] Selden R, Nystrom B, Langstrom R. Polym Compos 2004;25:543–53. [29] Pucciariello R, Villani V, Bonini C, D’Auria M, Vetere T. Polymer 2004;45:4159–69. [30] Muasher M, Sain M. Polym Degrad Stab 2006;91:1156–65. [31] Solomon OF, Ciuta LZ. J Appl Polym Sci 1962;6:683–6. [32] Colom X, Carrillo F, Nogues F, Garriga P. Polym Degrad Stab 2003;80:543–9. [33] Pandey KK. Polym Degrad Stab 2005;90:9–20. [34] George B, Suttie E, Merlin A, Deglise X. Polym Degrad Stab 2005;88:268–74. [35] Romanini D, Pezzin G. Rheol Acta 1982;21:699–704. [36] La Mantia FP, Morreale M, Mohd Ishak ZA. J Appl Polym Sci 2005;96:1906–13. [37] La Mantia FP, Tzankova Dintcheva N. Macromol Symp 2003;194:277–86.
Contents lists available at ScienceDirect
Polymer Degradation and Stability journal homepage: www.elsevier.com/locate/polydegstab
Accelerated weathering of polypropylene/wood flour composites F.P. La Mantia*, M. Morreale ` di Palermo, Viale delle Scienze, 90128 Palermo, Italy Dipartimento di Ingegneria Chimica dei Processi e dei Materiali, Universita
a r t i c l e i n f o
a b s t r a c t
Article history: Received 20 February 2008 Received in revised form 7 April 2008 Accepted 15 April 2008 Available online 18 April 2008
Wood–plastic composites (WPCs) have received increasing attention during the last decades, because of many advantages related to their use. Some of their main applications are represented by outdoor furnishing and decking; therefore, it is important to assess their behaviour under UV exposure. In this work, polypropylene/wood flour composites were prepared and their resistance to photooxidation investigated. The composites were prepared by extrusion and compression moulding, and were subjected to mechanical tests, FTIR analysis and molecular weight measurements. The results showed that the composites retained a higher fraction of the original mechanical properties after accelerated weathering; the wood flour did not significantly degrade throughout the irradiation time slot of the investigation and the composites kept a higher percentage of the original molecular weight. Ó 2008 Elsevier Ltd. All rights reserved.
Keywords: Polypropylene Wood Composites Photooxidation Molecular weight FTIR spectroscopy
1. Introduction Wood–polymer composites (WPCs) have attracted a significant interest in the last decades, thanks to the specific advantages they can grant in comparison with the classic mineral filler/plastic composites. These include mainly the improved environmental performance, due to the use of biodegradable materials and the reduction in the use of non-renewable (oil based) resources throughout the whole life cycle of the composite [1]; the low cost of wood flour and of natural-organic fillers in general (since they often come from wastes); the lower specific weight of these fillers, in comparison to the traditional mineral-inorganic ones; the improvement in safety for the production employees (reduced hazard in the case of accidental inhalation); the special aesthetic properties of the composites, which can be conveniently processed and refined, obtaining wood-like looking products; the full recyclability of the composites. These materials can be used for many indoor and outdoor applications (panels for the automotive industry, decking, furnishing, packaging, etc.) [1–8]. Polyolefins, in particular polypropylene, one of the most widely used plastics, have been extensively studied in combination with wood derivatives (flour, flakes, fibres) [9–12]. Several researchers have focused their attention on the improvement of the mechanical properties (usually deteriorated after the addition of the wood flour, particularly the ductility), achieved with the use of small amounts of coupling agents, which improve the interfacial polymer–filler adhesion and the dispersion of the filler within the matrix [13–15]. * Corresponding author. E-mail address: [email protected] (F.P. La Mantia). 0141-3910/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.polymdegradstab.2008.04.006
Significant interest has also arisen about the outdoor performance of these composites, in particular their resistance to photooxidation. Polyolefins modified with mineral fillers and exposed to accelerated weathering have shown somewhat conflicting trends, as reported by Rabello and White [16], who studied the photooxidation of PP–talc blends and found an inversion of the trend of molecular weight against exposure time for the neat and the filled PP, upon increasing the exposure time. This was explained by considering the screening effect of the mineral particles, which is contrasted by the presence of chromophore impurities. Similar results have been found by Benavides et al. [17] for PP–CaCO3 composites. Valadez-Gonzalez et al. [18] found a stabilizing effect of CaCO3 in HDPE composites, but for the first 300 h of exposure the neat polymer showed higher molecular weights. The influence of the filler was complex, involving variations in the morphology, the mechanical and thermal stresses during processing and then the sites where photodegradation can start. Yang et al. [19] compared several mineral fillers for HDPE-based composites and found that calcium carbonate and wollastonite had a stabilizing effect, while kaolin, diatomite and mica accelerated photodegradation. The proposed explanation for the different behaviours took into account the relative absorbances of the fillers for UV light in the 290–400 nm range. Similarly, studies regarding photooxidation of wood filled polymer composites reported a variety of results. Alexy et al. [20] reported a light stabilizing effect of lignin (which is one of the major components of wood) to PP when added in low concentrations (up to 10 wt%), while it accelerated photodegradation of PE at amounts higher than 10 wt%. Environmental degradation on HDPE/wood composites as investigated by Li [21] resulted in limited effects on
F.P. La Mantia, M. Morreale / Polymer Degradation and Stability 93 (2008) 1252–1258
1253
Fig. 1. Dimensionless elastic modulus as a function of the irradiation time.
the composites (confined to the surface layer) in dry environment conditions, while water had an important degradative action. Matuana and Kamdem [22] studied the photooxidation of PVC/ wood composites subject to accelerated weathering, finding that the composites retained a higher percentage of the original mechanical properties with the exposure time, in comparison to the neat polymer, explained by admitting that the photooxidation involves only the surface of the exposed composite; similar results were also found by Lundin et al. [23]. However, the increase of the carbonyl index suggested an accelerating effect caused by the wood particles, attributed to the presence of chromophoric groups. Stark et al. [24–27] extensively studied the behaviour of HDPE/wood flour composites exposed to accelerated weathering. They focused the attention especially on the discoloration of the composites, and found also that the presence of wood causes an increase of the carbonyl index, attributed to the formation of chromophore groups following the degradation of lignin; the degradation can be contrasted or at least reduced by using UV absorbers, stabilizers or maleated coupling agents. Selden et al. [28] found similar results
regarding the carbonyl index of the composites, however, the photodegradation was confined to the surface, due to the screening effect of the wood particles; the maleated coupling agents had no significant influence on the degradation. Pucciariello et al. [29] blended several polymers (HDPE, LDPE, LLDPE, polystyrene) with lignin, finding that the latter acts as a stabilizer against UV radiation for PS, LDPE and LLDPE, mainly on the basis of molecular weight measurements. Muasher et al. [30] focused on the efficiency of photostabilizers on the colour change of WF-filled HDPE subject to photooxidation. It was found that the main responsible factor in wood photodegradation is lignin, which breaks down to water soluble products, that eventually lead to the formation of chromophoric functional groups. In this work, polypropylene/wood flour composites were prepared and subjected to accelerated UV weathering. The effects of the exposure to UV radiation were studied analyzing the mechanical properties, the chemical changes and the variations of the molecular weight. The photooxidation mechanisms of both the polymer and the filler were taken into account.
Fig. 2. Dimensionless tensile strength as a function of the irradiation time.
1254
F.P. La Mantia, M. Morreale / Polymer Degradation and Stability 93 (2008) 1252–1258
Fig. 3. Dimensionless elongation at break as a function of the irradiation time.
2. Experimental 2.1. Materials The polypropylene (PP) used in this work was an injection moulding grade commercially known as ‘‘Moplen X30G’’ (Basell Polyolefins, Italy), with a density of 0.9 g/cm3, melt index z 8 g/ 10 min (at 230 C and 21.6 N), melting point about T ¼ 170 C. This class of general purpose polypropylenes usually contains small amounts of phenolic primary antioxidants and phosphite secondary antioxidants. The wood flour, kindly provided by La.So.Le (Italy) and marketed as ‘‘150/200’’ type, has an average particle diameter of 150–200 mm and a humidity content (as delivered) around 5%; it is further indicated throughout the paper as ‘‘SDF’’.
2.2. Compounding and testing The wood flour was dried in an oven at 70 C overnight in order to drastically reduce the humidity content, and processed with the polypropylene in a co-rotating twin screw extruder (OMC, Italy, D ¼ 19 mm, L/D ¼ 35), with the nominal composition of 30 and 60% (by weight). The temperature profile was set at 120–130–140–150– 160–170–180 C and the screw speed at 200 rpm. All the materials were pelletized after extrusion by a rotating blade system. The specimens for the tensile and accelerated weathering tests were obtained by compression moulding, using a Campana (Italy) laboratory press (set at 180 C, pressure about 200 bar, compression time 2–3 min) and a 300 mm thick mould. This thickness value was chosen in order to not compromise the internal structure of wood particles. The sheets were placed in a Q-UV (USA) chamber, mounting eight UV-B lamps, which reproduces the damage caused by sunlight, rain and dew. The weathering cycle conditions were 8 h of light at T ¼ 55 C and 4 h of condensation at T ¼ 35 C. Tensile tests were performed by an universal Instron (USA) model 3365 apparatus on specimens (thickness z 0.3 mm, width ¼ 10 mm) cut out from the previously described sheets, following ASTM D882. At least five replicates were tested for each tensile property and the average values are reported with satisfactory reproducibility (7%). The changes in the sample chemistry upon UV exposure were analyzed by means of a Perkin–Elmer (UK) Spectrum One FT-IR spectrophotometer (in the range 450–4000 cm1, absorbance units). The carbonyl index (CI) was evaluated as the ratio of the area of the
peak between 1680 and 1800 cm1 and the area of the peak centred at 2722 cm1 (which is the characteristic vibration of stretching band of PP). At least four replicates were tested for each FTIR spectrum and the average values of the CI are reported with a good reproducibility (5%). The variations of the molecular weight were monitored by intrinsic viscosity measurements. The polymer was extracted from the samples (unweathered and weathered) by dissolving them in hot tetrahydronaphthalene (THN, 135 C) and separating the wood fibres using a vacuum filtering unit, made up of a flask connected to a water vacuum pump and a sintered glass filter with disposable paper filters (Gelman Sciences, USA). Ethanol was added to the filtered solution which was allowed to cool in order to cause precipitation of the PP, which was then recovered by a second filtration stage and used for the intrinsic viscosity measurements. These were performed using a Schott-Gerate viscometer in an automatic apparatus, SchottGerate model AVS 300. The solutions were prepared in hot THN at a concentration of 0.4%. The intrinsic viscosity was calculated using the Salomon–Ciuta equation [31]. pffiffiffiqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 hsp ln hr ½h ¼ (1) c where c is the concentration, and hsp and hr the specific and relative viscosities, respectively. At least four replicates of each sample were tested and average values (reproducibility 7%) are reported. 3. Results and discussion 3.1. Mechanical properties The results of the tensile tests are shown in Figs. 1–3, where the main tensile properties are plotted as a function of the irradiation time. It is worth pointing out that it was difficult to obtain the data for the 280 h irradiated neat samples, since they tended to break up easily when manipulated, so several specimens had to be excluded from the calculations. Table 1 Mechanical properties of the unweathered materials Sample
Elastic modulus (MPa) Tensile strength (MPa) Elongation at break (%)
PP 1225 PP30SDF 1990 PP60SDF 2040
32.8 23.3 13.4
9.4 2.9 1.2
F.P. La Mantia, M. Morreale / Polymer Degradation and Stability 93 (2008) 1252–1258
1255
Fig. 4. FTIR spectra of some neat and filled samples.
The values have been reported in dimensionless form – value at a given time divided by the value for the unweathered samples (Table 1) – to better highlight the kinetics of the variations of the properties. It can be observed that the elastic modulus changes are limited during the first 150 h of irradiation, while it starts to decrease significantly for longer exposure times. However, thanks to the stiffening effect of the filler, it stays higher in the filled samples. The tensile strength decreases significantly in the neat PP, because of the chain scission reaction induced by the irradiation, while the changes are less pronounced in the filled samples. The explanation can be easily found considering that filled samples have a lower amount of polymer subjected to the photooxidative degradation; in fact, the decrease of tensile strength upon increasing the irradiation time is, on a percent scale, smaller in the 60% filled samples. The effect of wood particles is even more pronounced for the elongation at break. The decrease of the deformability upon increasing the UV exposure time is dramatic in the neat polypropylene, while is much less significant in the filled samples,
where it stays almost constant. However, it is worth noting that the filled materials are very fragile even in the virgin conditions. These results show, on average, an advantage of the composites in comparison to the unfilled, neat polymer, when subjected to accelerated weathering. However, it has been pointed out that this is probably due to the lower amount of polymer used and to the fact that the photooxidation mainly affects the polymer itself. More precisely, mechanical tests alone could not be sufficient to assess the real influence of the wood flour on the photooxidation mechanisms. This could be understood by specifically analyzing the chemical modifications of the materials and in fact, as highlighted in Section 1, many literature studies have been reported which use spectroscopy to detect the formation or disappearance of specific chemical groups. However, in order to have a real understanding of the photooxidation the material has undergone, these investigations should be coupled with the measurement of the molecular weight. As hinted in Section 2, both investigations have been carried out on the materials studied in this work.
Fig. 5. Carbonyl index of the materials as a function of the irradiation time.
1256
F.P. La Mantia, M. Morreale / Polymer Degradation and Stability 93 (2008) 1252–1258
Fig. 6. FTIR spectra of neat SDF, SDF extracted from 120 h irradiated 30% filled samples, and SDF extracted from 120 h irradiated 60% filled samples.
3.2. Fourier-transform infrared spectroscopy The IR spectra of the neat and filled samples are shown in Fig. 4. In particular, only the spectra of pure PP (weathered for 72 and 120 h) and of 30% filled PP (weathered for 72 and 120 h) are reported to make the observation of the graphs easier. The spectra showed, as predictable, an increase in the carbonyl peak upon increasing the irradiation time. This is a well-known effect directly linked with the photooxidation of the polypropylene. At the same time, an increase in the carbonyl peak was observed upon increasing the wood flour content. A quick, rough explanation could be that the wood flour accelerates the photodegradation rate of the material, resulting in a higher amount of carbonyl groups. However, a more rigorous way to assess the presence of carbonyl groups consists in taking into account the actual amount of polypropylene, and then calculating a carbonyl index according to the method described in Section 2. Therefore, a diagram has been prepared, plotting the carbonyl index versus the irradiation time for all the samples, and is shown in Fig. 5 (where all the values are normalized by subtracting the values of the unweathered samples). It can be clearly seen from the plot of the photooxidation rate that the latter is slower upon increasing the wood flour content. These results upset the rough considerations which could be made just upon calculating the carbonyl peak area (even if normalized with the thickness of the samples). This is in agreement with previous studies, as highlighted in Section 1, but a deeper explanation must be provided and therefore the role of the wood flour has been analyzed thoroughly. The FTIR spectra of the wood flour (before being compounded with the polymer and processed and after THN extraction from the samples) are shown in Fig. 6 (only some representative spectra are reported).
Besides the carbonyl peak, other significant peaks are in the 1425–1510 cm1 region linked with lignin. It is known from the literature [30,32–34] that the main factor responsible for wood photodegradation is lignin, which absorbs UV/vis light because of its chromophoric groups. The degradation mechanism is complex, with different paths leading to water soluble products and, finally, to chromophoric groups like carboxylic acids, quinones or hydroperoxides. In order to accurately assess the photodegradation of wood flour, therefore, it is first necessary to find one (or more) reference peak. According to previous studies [33], the phenomena can be monitored by comparing the carbonyl peak area (here having the maximum at approximately 1738 cm1), and the area of the 1505 cm1 peak (which is the aromatic band of lignin), to the peaks at 1375 and 1158 cm1, which are related to carbohydrates and are not significantly affected from degradation. Therefore, four indices (two for lignin, two for the carbonyl region) have been calculated and are reported in Table 2. The results show clearly that there are quite small differences in the indices of the neat, ‘‘as-delivered’’ wood flour and the same extracted from processed and irradiated samples. The same trend was observed also in the other samples (which therefore are not reported here). Therefore, it can be concluded that the wood flour does not undergo significant degradation within the time scale of observation adopted in this investigation. This means, furthermore, that no significant amounts of chromophoric groups are produced and therefore the wood flour does not accelerate the photooxidative processes of polypropylene, confirming the previous comments regarding the composite spectra. The lignin does not degrade significantly in this experimental study probably because the wood flour used comes from beech, a hardwood, and it is known that hardwoods degrade at a slower rate than softwoods [33]. Furthermore, the lignin, staying almost unaltered, can exert a protective action as suggested by Pucciariello et al. [29] and Alexy et al. [20].
Table 2 Variations of relative intensities of lignin (C]C) peak at 1505 cm1 and carbonyl (C]O) peak at 1738 cm1 for different SDF samples
Table 3 Intrinsic viscosities of neat and composite samples at different irradiation times
Sample
A (1505/1161)
A (1505/1375)
A (1738/1161)
A (1738/1375)
Sample
0h
72 h
120 h
280 h
SDF SDF-30%-120 h SDF-60%-120 h
1.21 1.11 1.18
1.35 1.17 1.24
4.51 4.82 4.65
5.21 5.37 5.05
PP PP30SDF PP60SDF
1.42 0.77 0.47
0.75 0.65 0.41
0.49 0.29 0.37
0.36 0.24 0.30
F.P. La Mantia, M. Morreale / Polymer Degradation and Stability 93 (2008) 1252–1258 Table 4 Weight average molecular weights of neat and composite samples at t ¼ 0
Mw at t ¼ 0 h
PP
PP30SDF
PP60SDF
237 000
90 000
41 000
3.3. Molecular weight measurements The intrinsic viscosity of the polypropylene (neat and extracted from the composites) was calculated according to the previously described method, providing the results shown in Table 3. The intrinsic viscosities can be quickly converted into calculated values of the weight average molecular weights by the following equation [35]: Mw ¼
½h 5:77 104
1=0:631
(2)
Table 4 reports the calculated values of molecular weight (Mw) of the materials at t ¼ 0, while Fig. 7 reports the dimensionless values at different irradiation times. The normalization was carried out dividing the molecular weight at time t by the molecular weight of the same material at t ¼ 0. A first noteworthy observation concerns the values of the unirradiated samples. The composites have lower molecular weights, as lower as the filler content increases. This is due to the higher mechanical stresses which arise during processing, related to the higher viscosity caused by the presence of the filler, as already discussed in our previous studies on similar systems [15,36,37]. In fact, the thermo-mechanical degradation during processing also affects the neat polypropylene, since the weight average molecular weight of the virgin PP is about 300 000 and it decreases, after processing, to the value reported in Table 4. The higher mechanical stress generated during the processing of the composites leads, therefore, to a significantly greater thermomechanical degradation. The different samples (neat, 30% filled, 60% filled) actually have polypropylene matrices with different molecular weights exposed to accelerated weathering. This fact may significantly affect the degradation paths in the case of thermo-mechanical degradation, since different molecular weights give rise to different viscosities and thus to different extents of chain scission; but in the case of UV exposure, the different chain length does not seem to affect the photooxidation reactions. This is
1257
also confirmed by the FTIR analysis, which did not indicate different photooxidation mechanisms in neat and filled materials. An important feature that arises while observing the variations the molecular weight undergoes upon increasing the exposure times is that the composites retain a higher percentage of the molecular weight in comparison to the pure PP, and this trend is even more pronounced in the 60% filled materials. This suggests that the wood flour exerts a stabilizing effect by slowing down the photooxidation and this explanation is in agreement with the conclusions drawn from the FTIR analysis. A further possible explanation of this stabilizing effect may be found considering also that wood contains a series of ‘‘extractives’’ (resins, tannins, polyphenols, waxes, esters of the C16–C18 fatty acids, terpenes, etc.) which it seems could act as antioxidants and radical quenchers [34]. Furthermore, the fact that the commercial PP used in this work has a general purpose phenolic stabilizing system may also have contributed to the substantial stability of lignin discussed above. Finally, the stabilizing effect of wood flour could be explained also by taking into account the possibility for the filler to act as a ‘‘screen’’, confining the photodegradation to the outer layers of the samples [16,22,23,28]. 4. Conclusions Polypropylene/wood flour composites were prepared and their resistance to oxidative photodegradation was investigated. The mechanical tests showed that the composites retain a higher fraction of the original mechanical properties after accelerated weathering, especially with regard to elongation at break, whereas the neat polymer undergoes a dramatic drop of the mechanical properties. FTIR analysis showed that the increase in the carbonyl peaks of composites is mainly due to the presence of wood flour, but this does not provide any acceleration of the photooxidation rate in comparison to the neat polymer upon increasing the irradiation time, since the wood flour (and, in particular, lignin) does not degrade significantly throughout the irradiation times of this study; indeed, wood flour effectively acts as a stabilizer, slowing the photooxidation. This was confirmed also by molecular weight determination. This effect is probably due to the action of lignin, and maybe of the extractives. At higher irradiation times, it is likely that the trend would be reversed, since the degradation of wood leads to the formation of chromophoric groups which accelerate photooxidation and yellowing; in this case, the inversion time would depend on the relative
Fig. 7. Dimensionless weight average molecular weights of neat and composite samples as a function of the irradiation time.
1258
F.P. La Mantia, M. Morreale / Polymer Degradation and Stability 93 (2008) 1252–1258
weight of the two opposing phenomena and therefore would strongly depend both on the processing and weathering conditions. References [1] Joshi SV, Drzal LT, Mohanty AK, Arora S. Compos Part A 2004;35:371–6. [2] Rozman HD, Lai CY, Ismail H, Mohd Ishak ZA. Polym Int 2000;49:1273–8. [3] Gachter R, Muller H. Plastics additives. 3rd ed. Munich: Hanser Publishers; 1990. [4] Canche`-Escamilla G, Rodriguez-Laviada J, Cauich-Cupul JI, Mendizabal E, Puig JE, Herrera-Franco PJ. Compos Part A 2002;33:539–49. [5] Coutinho FMB, Costa THS, Carvalho DL. J Appl Polym Sci 1997;65:1227–35. [6] Balasuriya PW, Ye L, Mai YW, Wu J. J Appl Polym Sci 2002;83:2505–21. [7] Raj RB, Kokta VB, Maldas D, Daneault C. J Appl Polym Sci 1989;37:1089–103. [8] Netravali AN, Chabba S. Mater Today 2003;6:22–6. [9] Carroll DR, Stone RB, Sirignano AM, Saindon RM, Gose SC, Friedman MA. Resour Conserv Recycl 2001;31:241–51. [10] Selke SE, Wichman I. Compos Part A 2004;35:321–6. [11] Avella M, Bazzi C, dell’Erba R, Focher B, Mazzetti A, Martuscelli E. Angew Makromol Chem 1995;233:149–66. [12] Marcovich NE, Reboredo MM, Aranguren MI. Polymer 1999;40:7313–20. [13] Harper D, Wolcott M. Compos Part A 2004;35:385–94. [14] Keener TJ, Stuart RK, Brown TK. Compos Part A 2004;35:357–62.
[15] La Mantia FP, Morreale M. Compos Interface 2007;14:685–98. [16] Rabello MS, White JR. Polym Compos 1996;17:691–704. [17] Benavides R, Gonzalez-Hernandez R, Gonzalez-Cantu` MC, Reyes-Vielma B, Billingham NC. J Vinyl Addit Technol 2003;9:41–9. [18] Valadez-Gonzalez A, Cervantes-Uc JM, Veleva L. Polym Degrad Stab 1999;63: 253–60. [19] Yang R, Yu J, Liu Y, Wang K. Polym Degrad Stab 2005;88:333–40. [20] Alexy P, Kosikova B, Podstranska G. Polymer 2000;41:4901–8. [21] Li R. Polym Degrad Stab 2000;70:135–45. [22] Matuana LM, Kamdem DP. Polym Eng Sci 2002;42:1657–66. [23] Lundin T, Cramer SM, Falk RH, Felton C. J Mater Sci Eng 2004;16:547–55. [24] Stark NM, Matuana LM. J Appl Polym Sci 2003;90:2609–17. [25] Stark NM, Matuana LM. Polym Degrad Stab 2004;86:1–9. [26] Stark NM, Matuana LM. J Appl Polym Sci 2004;94:2263–73. [27] Stark NM, Matuana LM. Polym Degrad Stab 2007;92:1883–90. [28] Selden R, Nystrom B, Langstrom R. Polym Compos 2004;25:543–53. [29] Pucciariello R, Villani V, Bonini C, D’Auria M, Vetere T. Polymer 2004;45:4159–69. [30] Muasher M, Sain M. Polym Degrad Stab 2006;91:1156–65. [31] Solomon OF, Ciuta LZ. J Appl Polym Sci 1962;6:683–6. [32] Colom X, Carrillo F, Nogues F, Garriga P. Polym Degrad Stab 2003;80:543–9. [33] Pandey KK. Polym Degrad Stab 2005;90:9–20. [34] George B, Suttie E, Merlin A, Deglise X. Polym Degrad Stab 2005;88:268–74. [35] Romanini D, Pezzin G. Rheol Acta 1982;21:699–704. [36] La Mantia FP, Morreale M, Mohd Ishak ZA. J Appl Polym Sci 2005;96:1906–13. [37] La Mantia FP, Tzankova Dintcheva N. Macromol Symp 2003;194:277–86.