1 Effects of aging and different mechanical recycling processes on the 1 structure and properties of poly(lactic acid)-clay nanocomposites. 2 F.R. Beltrán*a,b, E. Ortegaa, A.M. Solvolla, V. Lorenzob, M.U. de la Ordenb,c, J. Martínez 3 Urreagaa,b. 4 5 a Departamento de Ingeniería Química Industrial y del Medio Ambiente, E.T.S.I. 6 Industriales, Universidad Politécnica de Madrid, José Gutiérrez Abascal 2, 28006 7 Madrid, Spain 8 b Grupo de Investigación ‘‘Polímeros: Caracterización y Aplicaciones (POLCA)’’, E.T.S.I. 9 Industriales, Universidad Politécnica de Madrid, José Gutiérrez Abascal 2, 28006 10 Madrid, Spain 11 c Departamento de Química Orgánica I, Universidad Complutense de Madrid, Facultad 12 de Óptica y Optometría, Arcos de Jalón 118, 28037 Madrid, Spain 13 14 Corresponding author information: 15 Mailing address: Departamento de Ingeniería Química Industrial y del Medio Ambiente, 16 E.T.S.I. Industriales, Universidad Politécnica de Madrid, José Gutiérrez Abascal 2, 17 28006 Madrid, Spain 18 Telephone: +34 913363183 19 e-mail: f.beltran@upm.es 20 21 Suggested running head: Structure and properties of recycled PLA 22 nanocomposites 23 24 Abstract 25 The growing use of poly(lactic acid) (PLA) and PLA-based nanocomposites in packaging 26 has raised the interest of studying the mechanical recycling of the wastes and the 27 properties of the recycled materials. The main objective of this work was to study the 28 effect of two different mechanical recycling processes on the structure and properties of 29 a PLA-montmorillonite nanocomposite. The two recycling processes included 30 accelerated thermal and photochemical aging steps to simulate the degradation 31 experienced by post-consumer plastics during their service life. One of them also 32 included a demanding washing process prior to the reprocessing. A decrease in the 33 molecular weight of PLA was observed in the recycled materials, especially in those 34 subjected to the washing step, which explained the small decrease in microhardness 35 and the increased water uptake at long immersion times. Water absorption at short 36 immersion times was similar in virgin and recycled materials and was accurately 37 described using a Fickian model. The recycled materials showed increased thermal, 38 optical and gas barrier properties due to the improved clay dispersion that was observed 39 by XRD and TEM analysis. The results suggest that recycled PLA-clay nanocomposites 40 can be used in demanding applications. 41 mailto:f.beltran@upm.es 2 Keywords: 1 Mechanical recycling; Poly(lactic acid); Nanocomposites; Structure; Properties 2 3 1. Introduction 4 Poly(lactic acid) (PLA) represents an interesting alternative to the fossil-fuel 5 based polymers in packaging applications, especially for food products. This is due to 6 PLA being considered as biodegradable, safe in food contact and comparable to some 7 commodity plastics, such as polyethylene terephthalate (PET), in optical and mechanical 8 properties [1-3] . However, other PLA properties, such as impact strength and gas 9 permeability, show only moderate values and therefore the addition of different 10 nanometric reinforcing agents, such as organically modified clays, has been studied over 11 the past years. The nanocomposites obtained show, in general, improved impact 12 strength, fire resistance and gas barrier properties [4-6]. These characteristics have 13 placed PLA as one of the most important bioplastics in the market, with a global 14 production capacity of 205000 t in 2014, which is expected to exceed 500000 t in 2020 15 [7]. 16 The expected growth in production of PLA could generate some environmental 17 and social issues, especially if the origin of the raw materials necessary for its production 18 is taken into account. PLA is produced by the ring opening polymerization of lactide, the 19 cyclic dimer of lactic acid, which is obtained by the fermentation of the glucose present 20 in some food products such as corn, potato or sugar beet [8]. This situation could cause 21 an increasing demand of large amounts of land to produce such food products, also 22 increasing the environmental impact of the manufacture of PLA [9]. Furthermore, some 23 farmers could be tempted to exchange the cheap food production for the crops used in 24 the production of PLA, compromising the health and sustenance of developing countries 25 [10]. Lastly, there are some concerns regarding the end-of-life scenarios for this polymer, 26 since they play a very important role on the environmental impact of PLA. Some studies, 27 such as those conducted by Piemonte [11], Cosate de Andrade et al. [12] and Rossi et 28 al. [13] point out that composting is not necessarily the best alternative in the case of 29 PLA, suggesting that mechanical recycling might be a more interesting alternative. This 30 is especially important in the commercial grades used in packaging applications, since 31 they degrade at a slower rate than the accumulation of wastes [14-16]. 32 Considering the social and environmental issues derived from the massive 33 production of PLA and PLA-based materials, it can be concluded that mechanical 34 recycling is an interesting end-of-life scenario for PLA wastes coming from packaging 35 applications, since it allows the reduction in raw materials, energy and emissions,, thus 36 decreasing the environmental impact of the use of PLA. However, in order to evaluate 37 the feasibility of the mechanical recycling, it is necessary to compare the properties of 38 virgin and recycled materials, since if there is an important decrease in the properties of 39 the recycled materials, the mechanical recycling would be unfeasible. 40 Despite the clear interest of the mechanical recycling, the properties of recycled 41 PLA and PLA-based materials have been scarcely studied up to the present. Different 42 authors have measured the thermal and mechanical properties of neat PLA after several 43 reprocessing steps. For instance, Scaffaro et al. [17], Badía et al. [18] and Żenkiewicz et 44 al [19] reported that PLA presented reduced mechanical properties after 1, 5 and 10 45 3 reprocessing cycles. However, Nascimento et al. [20] pointed out that a single 1 reprocessing step does not significantly affect the structure, thermal and mechanical 2 properties of PLA. 3 In general, these studies do not consider the degradation (thermal, 4 photochemical and hydrolytic) of the PLA during its useful life nor do stages of 5 accelerated aging, although the degradation of the polymer prior to reprocessing can 6 play a very important role in the decrease of the properties of the recycled plastic. In a 7 previous work we have studied the hydrolytic degradation of PLA subjected to different 8 recycling processes that include, in addition to the reprocessing step, accelerated aging 9 and washing steps to simulate the degradation of the material during the service life and 10 the cleaning of a plastic waste coming from food packaging, respectively [21]. The results 11 revealed that the recycled plastics show good resistance against the hydrolytic 12 degradation, although the behavior of the recycled materials depends on the conditions 13 of the recycling process. The impact of the different mechanical recycling processes on 14 the mechanical, optical, thermal and barrier properties of the recycled plastic was also 15 limited [1]. 16 As it has been mentioned, the mechanical recycling of unfilled PLA has already 17 been studied and reported in the literature. However, the behavior of PLA based 18 nanocomposites does not have to be the same, since the presence of clays could 19 promote the degradation of the polymer during its service life and also during the 20 recycling processes, thus affecting the final properties of the recycled materials. Despite 21 the growing use of clays in PLA based materials, in order to improve the properties of 22 the polymer, there is little data available regarding the behavior of mechanically recycled 23 PLA nanocomposites. In this regard, Kozlowski and Macyszyn studied the reprocessing 24 of a PLA-montmorillonite nanocomposite and found that PLA suffered increasing 25 degradation during the successive reprocessing cycles, which leaded to higher values 26 of the melt flow rate. However, oxygen permeability was reduced as a result of the 27 reprocessing [22]. In the same vein, Scaffaro et al. studied the effect of five reprocessing 28 cycles in nanocomposites of PLA and hydrotalcites. They reported that reprocessing 29 caused an improvement of the dispersion of the clay particles, which led to an increase 30 of the mechanical properties in the first three reprocessing steps. However, after four 31 and five reprocessing cycles, the degradation of PLA was severe, thus decreasing the 32 performance of the material [23]. In both cited works, the degradation of the material 33 during its service life and possible washing steps were not considered. 34 Consequently, the main aim of this work was to study the effect of different 35 mechanical recycling processes, including accelerated aging stages, on the properties 36 of PLA-clay nanocomposites. A commercial grade of PLA, especially designed for 37 packaging applications, and 2% wt. of an organically modified montmorillonite were melt 38 compounded and compression molded into films. A portion of these films were subjected 39 to two different mechanical recycling processes, one comprising an accelerated thermal 40 and photochemical aging before the second extrusion and compression molding step, 41 and other including a demanding washing step between the accelerated aging and the 42 reprocessing step. The effect of both processes on the structure and properties of the 43 material was followed by water absorption measurements, dilute solution viscosimetry 44 (IV), infrared spectroscopy (FTIR-ATR), X-ray diffraction (XRD), transmission electron 45 microscopy (TEM), differential scanning calorimetry (DSC), thermogravimetric analysis 46 (TGA), ultraviolet-visible spectroscopy (UV-Vis) and microhardness measurements. The 47 4 results indicate that, although the recycling processes modify the structure of the 1 nanocomposites, the recycled materials retain, to a great extent, the properties of the 2 virgin material. 3 4 2. Materials and methods 5 2.1. Materials and preparation of the samples 6 The PLA used was a commercial grade designed for packaging (Ingeo 2003D, 7 NatureworksTM), with a melt mass-flow rate of 6 g/10 min (2.16 kg at 210 °C). Prior to 8 processing, the material was dried according to the manufacturer's recommendations 9 (20 minutes at 100 ºC and 2 hours in a vacuum oven at 85 ºC). The clay used was 10 CloisiteTM 30B (C30), an organically modified montmorillonite supplied by Southern Clay 11 Products (USA), where the organic modifier is a methyl, dihydroxyethyl, dehydrogenated 12 tallow, quaternary ammonium chloride. The clay was dried in a convection oven at 100 13 ºC for 2 hours. 14 Neat PLA and the PLA nanocomposite with 2 wt.% C30 were melt compounded 15 in a Rondol Microlab twin-screw microcompounder, with L/D = 20, at 60 rpm. The barrel 16 temperature profile from hopper to die was 125,160,190,190,180 °C. The films with a 17 thickness of 230 ± 10 μm were obtained by compression in an IQAP-LAP hot-plate press 18 at 190 °C, beginning with a melting step, with no pressure, for 5 minutes, followed by a 19 degasification step for 2 minutes and cooling between cold plates at 14 MPa for 5 20 minutes. 21 22 2.2. Recycled materials 23 In order to simulate the degradation of the material during its service life, the films 24 of the virgin polymer were first subjected to an accelerated aging process, which included 25 468 h of thermal degradation in a convection oven at 50 ºC and 40 hours of 26 photochemical degradation in an Atlas UVCON chamber, equipped with eight F40UVB 27 lamps. A portion of these aged samples was washed at 85 ºC for 15 minutes, in a solution 28 of NaOH (1.0% by weight) and a surfactant (Triton X, 0.3% by weight), following the 29 method proposed by Chariyachotilert et al. [24]. Finally, all samples were reprocessed 30 by extrusion and compression in the above conditions. 31 According to the different recycling processes there were three different 32 nanocomposites: PLAV-C30, which was not reprocessed, PLAR-C30, which was 33 recycled without the washing step, and PLARW-C30 which was recycled with the 34 washing step at 85 °C. Before the characterization, all the samples were subjected to 35 physical aging at room temperature for 3 weeks. 36 37 2.3. Experimental techniques 38 Infrared spectra were recorded at a resolution of 4 cm-1, with a total of 16 scans, 39 using a Nicolet iS10 spectrometer equipped with a diamond Attenuated Total 40 Reflectance (ATR) accessory. 41 5 The diffraction patterns of the materials were obtained using a X’PERT-MPD 1 diffractometer, equipped with a CuKα generator (λ1 = 0.154056 nm and λ2=0.154439 2 nm) at 45 kV and 40 mA, in a 2θ range from 1.7 to 50°. The basal spacing of the clay 3 was calculated using the mathematical expression of Bragg’s law, given in Eq. (1): 4 𝜆𝜆 = 2𝑑𝑑 ⋅ sin 𝜃𝜃 (1) 5 where λ is the wavelength of the rays, d is the spacing between the clay platelets and θ 6 is the Bragg angle. 7 The intrinsic viscosity (IV) was measured in chloroform at 25 ± 0.5 °C using an 8 Ubbelohde viscosimeter. Intrinsic viscosity was obtained using the Kraemer equation 9 [25], measuring four concentrations for each sample. The uncertainty was determined 10 by linear regression with a 95% confidence level. 11 Differential scanning calorimetry (DSC) analysis was performed using a TA 12 Instruments Q20 calorimeter on samples of 5-7 mg, in standard aluminum pans, at 5 13 ºC/min under nitrogen atmosphere. The samples were first heated from 20 to 200 ºC, 14 and then kept at 200 ºC for 3 minutes to erase thermal history. After that, the samples 15 were cooled to 0 ºC, and finally a second heating scan was performed until 200 ºC. 16 Thermogravimetric analysis was carried out using a TA Instruments TGA2050 17 thermobalance. Samples of 12-14 mg were heated at 10 °C/min from room temperature 18 to 800 °C in dry nitrogen (30 cm3/min). 19 The overall transmittance in the visible light region was measured according to 20 the ISO 13468 standard, using a Shimadzu 2401 PC UV-Vis spectrophotometer 21 equipped with a Shimadzu integrating sphere, using a scan speed of 200 nm/min. 22 The microhardness measurements were measured using a Type M Shimadzu 23 microhardness tester equipped with a Vickers pyramidal indenter and applying a load of 24 25 g for 10 seconds. Each measurement was repeated six times. 25 Water absorption was measured gravimetrically. Dried samples of 40 x 25 x 0.2 26 mm were dipped in flasks containing 100 ml of 0.05 M phosphate-buffered solution (PBS) 27 at pH 7.4 ± 0.2. The flasks were located in an oven set at 37 ºC. Specimens were 28 removed at selected immersion periods, gently wiped and weighed, at room temperature, 29 in a laboratory balance with a precision of 0.1 mg. The percentage of absorbed water at 30 any time t, Mt, was determined by this expression: 31 𝑀𝑀𝑡𝑡(%) = 𝑊𝑊𝑡𝑡−𝑊𝑊0 𝑊𝑊0 × 100 (2) 32 where W0 and Wt denote, respectively, the weight of the specimens before and after the 33 immersion in PBS. 34 The permeability measurements were performed in a homemade permeation cell 35 that has been described elsewhere [26]. The device consists of a permeation cell with 36 two chambers, separated by a PLA membrane of known thickness. Nitrogen, oxygen 37 and carbon dioxide permeability of the different samples has been determined at 30 ºC 38 by means of diffusion experiments trough an initially purged membrane. After purging, a 39 0.2 MPa step variation of the pressure was imposed on the high-pressure side of the 40 membrane, and the pressure on the low-pressure side was monitored. The permeability 41 6 coefficient, P, was estimated from the slope of pressure vs. time line after reaching 1 steady state [27]. 2 3 3. Results and discussion 4 3.1. Effects of recycling processes on the structure 5 The structure of PLA nanocomposites can be significantly altered by the 6 degradation that takes place during the service life and the recycling process, which 7 includes a melt reprocessing at high temperature and shear stress. On one hand, it is 8 known that PLA and PLA-based materials are susceptible to thermal, photochemical and 9 hydrolytic degradation processes that lead to a decrease of the average molecular mass, 10 which can negatively affect the thermal, optical and mechanical properties of the material 11 [28]. On the other hand, the dispersion of the nanoparticles in the polymer can be also 12 altered during the recycling process due to the shear stress, resulting in a different 13 morphology of the material, which also leads to changes in the properties. 14 The degree of dispersion was studied by means of XRD and TEM. Fig. 1 shows 15 the diffraction patterns of the PLA-C30 nanocomposite subjected to the different 16 mechanical recycling processes, along with the pattern of pure C30. There were no 17 differences in the region between 8 and 20º (not shown), thus indicating that not 18 important crystalline domains were formed in the polymer during the mechanical 19 recycling of the nanocomposite. However, the changes observed at angles lower than 20 8º reveal that the morphology of the composite was altered during the recycling 21 processes. 22 3 4 5 6 7 8 PLARW-C30 PLAR-C30 Re la tiv e In te ns ity 2θ C30 PLAV-C30 23 Fig. 1. XRD patterns of pure C30 and PLA-C30 nanocomposite subjected to different 24 mechanical recycling processes (V = virgin; R= recycled; RW = recycled with washing 25 step). 26 27 7 Firstly, Fig. 1 shows the structure of the virgin nanocomposite. It can be seen that 1 the characteristic (001) diffraction peak of the clay located at 2θ = 4.8º, which 2 corresponds to a basal spacing d001 = 1.8 nm, is located at lower angles in PLAV-C30 3 (2θ = 2.6º), indicating the intercalation of the polymer chains within the galleries of the 4 clay, and thus the formation of an intercalated structure. The nanocomposite also 5 presents a broad second diffraction peak located at 2θ = 5.4º, which may be related, at 6 least partially, to the second-order reflection of the main diffraction. However, the 7 formation of some clay aggregates with lower basal spacing during the melt processing 8 could also contribute to this peak. A similar result was observed by Pluta et al. in PLA 9 nanocomposites with 3 wt.% of C30 [29]. 10 Regarding the effect of the different mechanical recycling on the degree of 11 dispersion in the nanocomposite, Fig. 1 shows that the (001) peak disappears almost 12 completely in PLAR-C30 and PLARW-C30. The remaining weak (001) diffraction is 13 displaced towards lower angles, which can be attributed to an improved dispersion of 14 the clay, forming exfoliated structures in which the individual silicate layers are no longer 15 close enough to interact with each other. These results are very important, since an 16 improvement of the clay dispersion could lead to enhanced mechanical, optical and gas 17 barrier properties. These results agree with those obtained by Scaffaro et al. in PLA 18 nanocomposites with 5% hydrotalcites, where an improvement of the clay dispersion 19 was observed scanning electron microscopy (SEM) [23]. 20 The effect of the recycling on the morphology can be also observed in the TEM 21 micrographs corresponding to the virgin and recycled plastics (Fig. 2). It can be seen 22 that, although some clay aggregates are still observed, the recycled material presents 23 smaller particles than the virgin material, thus indicating and enhancement of the clay 24 dispersion during the mechanical recycling, thus confirming the results observed in the 25 XRD data. 26 27 8 1 2 Fig. 2. TEM images of (a) PLAV-C30 and (b) PLAR-C30 (V = virgin; R= recycled). 3 4 The degradation of the polymer in the recycled materials was investigated by 5 using intrinsic viscosity (IV) measurements, FTIR spectroscopy and differential scanning 6 calorimetry. 7 (b ) (a) 9 The IV values are presented in Table 1. Firstly, it can be noticed that PLAR-C30, 1 the nanocomposite recycled without the washing step, presents a reduction of around 7 2 % on its intrinsic viscosity when compared with the virgin nanocomposite. This reduction 3 in the intrinsic viscosity, and hence in the average molecular weight, is a consequence 4 of chain scission processes in the polymer matrix due to the high temperatures and the 5 shear stresses to which the polymer was subjected during the reprocessing, and could 6 also be attributed to the degradation during the accelerated aging. Other authors have 7 reported similar decreases in the molecular weight of unreinforced PLA samples 8 subjected to a single reprocessing step [17,18,24]. This result is similar to that reported 9 in neat PLA subjected to the same recycling process, in which a decrease of 10 approximately 5% was obtained [1]. These results indicate that the clay does not have a 11 very detrimental effect on the polymer degradation during the recycling of the 12 nanocomposite. Moreover, the results reveal that the reduction of the intrinsic viscosity 13 is small in this recycling process, so only limited effects on the mechanical and thermal 14 properties of this recycled material must be expected. 15 Regarding the material subjected to the recycling process that included a washing 16 step, PLARW-C30, Table 1 shows that the decrease of IV is more important in this case 17 (near 20 % when compared with the virgin nanocomposite). This result reveals that a 18 demanding washing step such as the used in this work, with NaOH at high temperature, 19 might play an important role on the degradation of PLA during its mechanical recycling, 20 and thus could largely affect the properties of the recycled material. Again, unfilled PLA 21 showed a very similar descent in the intrinsic viscosity, close to 20 %, when it was 22 subjected to the recycling process including the washing step [1]. It was proposed that 23 the washing in severe conditions weakened the structure of the polymer, generating acid 24 groups that catalyzed the degradation during the following melt reprocessing. 25 The IV results indicate that the effects of the recycling strongly depend on the 26 conditions selected for the recycling process. PLA-C30 nanocomposites can withstand 27 a single recycling process without relevant decreases in the molecular weight of the 28 polymer. However, the introduction of a demanding washing step in the recycling process 29 promotes a further degradation of PLA during the reprocessing step, which might affect 30 both the structure and final properties of the polymer. 31 After studying the effect of the different recycling processes on the molecular 32 weight of PLA in the recycled nanocomposites with C30, the degradation suffered by the 33 polymer was also studied by means of FTIR-ATR spectroscopy and DSC. The IR spectra 34 (not shown) revealed that there are no major differences between the spectra of the 35 studied materials. Although the IV values show the existence of chain scission during 36 the recycling processes, which implies the formation of new carbonyl compounds, the IR 37 spectra only show a minor difference in the carbonyl stretching band (centered at 1756 38 cm-1). This result is in good agreement with those reported by Badía et al. [18], which 39 found that the generation of new carbonyl compounds during the reprocessing of PLA 40 causes only a slight displacement of the carbonyl stretching band toward higher 41 wavenumbers. Therefore, our results seem to indicate that, despite the degradation 42 observed by means of intrinsic viscosity measurements, only small amounts of new 43 carbonyl compounds are formed and the chemical structure of the polymer remains 44 almost unchanged during the mechanical recycling. 45 10 DSC measurements can provide information about the structural changes of the 1 polymer during the mechanical recycling of the nanocomposites. The first heating scans 2 presented in Fig. 3 reveal that all the samples present a glass transition around 60 °C, 3 followed by an endothermic peak related to the densification of the amorphous regions 4 of PLA, that is, the physical aging of the polymer [20]. The samples also show a cold 5 crystallization peak (Tcc) at approximately 110 ºC and finally, a double peak melting 6 endotherm (Tm1 and Tm2) near 150 ºC. The presence of this double melting peak has 7 been explained as the result of a melt recrystallization process, which includes the 8 melting of the less ordered crystals at a lower temperature (Tm1), their reorganization into 9 more perfect structures and their subsequent melting at a higher temperature (Tm2) 10 [30,31]. 11 12 35 70 105 140 175 H ea t f lo w (W /g ) Temperature (ºC) PLAR-C30 PLARW-C30 PLAV-C30 en do 13 Fig. 3. DSC first heating scans of nanocomposites subjected to different mechanical 14 recycling processes (V = virgin; R= recycled; RW = recycled with washing step). 15 16 Concerning the recycling of the nanocomposite, it can be seen on Fig. 3 that 17 PLAR-C30 shows a higher Tcc than the virgin nanocomposite, while PLARW-C30 18 presents a lower value of Tcc. This behavior is different from that previously observed in 19 the unreinforced polymer, where the mechanical recycling processes always led to lower 20 Tcc values in the recycled materials, which was explained as a consequence of the 21 greater mobility of the shorter chains produced by the degradation during the recycling 22 process [1,21]. The different behavior observed in the recycled nanocomposites 23 suggests that the mechanical recycling presents in this case two opposing effects on the 24 crystallization behavior of the polymer. On the one hand, the generation of shorter 25 polymer chains facilitates the crystallization of PLA, while, on the other hand, the 26 improvement of the dispersion of the clay platelets (observed by means of XRD) hinders 27 11 the packaging of the polymer chains, hence obstructing the formation of crystalline 1 domains. In the recycling without the washing step (PLAR-C30), the degradation was 2 small, as it was observed in the intrinsic viscosity measurements, thus prevailing the 3 hindering effect of the better dispersion of the clay. However, the degradation observed 4 when the demanding washing step was included (PLARW-C30) was significantly bigger, 5 counteracting the effect of the better dispersion of the clay and thus decreasing the Tcc 6 of the material. 7 The presence of two opposite effects of the mechanical recycling in the 8 nanocomposites is also noticeable in the shapes of the melting endotherms of the 9 different materials. Fig. 3 shows that PLAR-C30 presents a larger low temperature 10 melting peak, which can be explained by the formation of less perfect crystalline structure 11 and the hindering of the melt recrystallization mechanism, caused by the better 12 dispersion of the clays in the polymeric matrix. On the other hand, PLARW-C30 shows 13 a larger high temperature melting peak, due to the presence of shorter polymer chains, 14 which can rearrange during the melting, forming more stable crystalline structures. 15 Despite the changes observed in the Tcc and melting endotherms, Fig. 3 shows that the 16 crystallization and melting enthalpies are similar for each material (the values, measured 17 as peak areas, are shown in Table 1), thus indicating that most of the crystals that melt 18 during the heating are previously formed in the cold crystallization. This behavior 19 suggests that all the materials are essentially amorphous, and that the different recycling 20 processes do not cause significant changes on the crystallinity of PLA, thus supporting 21 the results obtained from the XRD data. This conclusion is important, since the 22 development of crystalline structures could affect the mechanical, optical and gas barrier 23 properties of PLA. The fact that only marginal changes in crystallinity were observed 24 during the different mechanical recycling processes might imply that various important 25 properties of PLA remain unchanged despite the degradation of the polymer during the 26 mechanical recycling. 27 28 Table 1. Properties of the virgin and recycled nanocomposites (V = virgin; R= recycled; 29 RW = recycled with washing step). The given uncertainties were determined with a 95 30 % confidence level. 31 Material Intrinsic viscosity (mL/g) ΔHC (J/g) ΔHM (J/g) T10 (ºC) Tmax (ºC) Visible light transmission (%) Vickers hardness (MPa) PLAV-C30 126 ± 3 27.0 29.2 340.0 367.3 83.4 189 ± 5 PLAR-C30 117 ± 4 25.9 28.3 338.7 377.7 85.6 184 ± 4 PLARW- C30 103 ± 4 27.7 28.2 339.4 374.1 85.9 180 ± 4 32 3.2. Properties of the recycled materials 33 12 After studying the effects of the different recycling processes on the structure and 1 morphology of the nanocomposite with C30, the effects of the structural changes on 2 some of the properties of the recycled materials were analyzed. 3 The thermal stability was studied by means of TGA. Table 1 gives the values of. 4 T10 and Tmax, two characteristic temperatures commonly used as indicatives of the 5 thermal stability of the materials, which are defined as the temperature at which 10% of 6 the total mass is volatilized and the temperature of maximum rate of decomposition, 7 respectively. Regarding the effect of the mechanical recycling on the thermal stability of 8 the nanocomposite, it can be seen that both recycled materials present higher values of 9 Tmax, despite the degradation observed in the intrinsic viscosity measurements. This 10 behavior can be related at least in part to the better dispersion of the clay platelets, 11 observed in the XRD measurements, which effectively act as a barrier for the liberation 12 of the decomposition products of PLA, thus increasing the characteristic temperatures of 13 the recycled nanocomposites. Furthermore, the chain scission of PLA during the 14 recycling produces a decrease of the average molecular weight of the polymer, which 15 reduces the thermal stability, but also the generation of new carboxyl groups, which leads 16 to stronger interactions between the polymer chains and, hence, increased thermal 17 stability of the recycled materials. 18 The changes in the morphology reported in the previous section can result in 19 changes in the optical properties of the materials, which play a key role in the field of 20 food packaging. The light transmission of the different materials was studied by means 21 of UV-Vis spectroscopy, according to the ISO 13468 standard. The results are shown in 22 Fig. 4 and Table 1. 23 265 365 465 565 665 765 0 20 40 60 80 100 Tr an sm itt an ce (% ) Wavelength (nm) Neat PLA PLAV-C30 PLAR-C30 PLARW-C30 24 Fig. 4. UV-Vis spectra of the neat PLA and PLA-C30 nanocomposites subjected to 25 different mechanical recycling processes (V = virgin; R= recycled; RW = recycled with 26 washing step). 27 28 13 Fig. 4 shows the UV-Vis spectra of neat PLA, and the virgin and recycled 1 nanocomposites. When compared with neat polymer, nanocomposites present higher 2 absorption in the UV region of the spectrum, which confirms that the incorporation of C30 3 provides some protection against the photochemical degradation caused by the UV 4 radiation, which may be interesting from the point of view of food packaging applications. 5 Respecting the visible part of the spectrum, between 400 and 800 nm, the addition of 6 clays causes a decrease on the light transmission of the polymer, due to the presence 7 of clay aggregates in the polymer matrix, which scatter and reflect the light, as it was 8 pointed out by Cele et al. [32]. However, it is worth to note that both recycled 9 nanocomposites present a higher visible light transmission than the virgin one. This 10 result can be explained by the better dispersion of the clay nanoparticles in the recycled 11 nanocomposites, observed by means of XRD and TEM, which reduces the amount of 12 clay aggregates. Table 1 shows that the two recycled nanocomposites present good 13 visible light transmission, higher than 85 %. In these materials, mechanical recycling not 14 only does not worsen properties but even improves optical clarity. 15 The effect of the different recycling processes under consideration on the 16 structure could also affect the mechanical properties and, hence, the potential use of 17 recycled materials in various applications. In order to study the effects of the mechanical 18 recycling, microhardness indentation tests were performed because it has been 19 suggested that hardness measurement is more sensitive to molar mass changes than 20 other mechanical characterization techniques [33]. The results that are shown on Table 21 1 point out that the different mechanical recycling processes causes a small decrease, 22 lower than 10%, in the Vickers hardness of the nanocomposite. The decrease of the 23 hardness can be related to the lower molecular weight of PLA in the recycled materials 24 [34]. However, the decrease is very small because the improved dispersion of the clay 25 platelets in the recycled nanocomposites causes an increase on the hardness of the 26 samples, thus partially counteracting the negative effect of the degradation of PLA during 27 the different mechanical recycling processes. A similar behavior was observed by 28 Scaffaro et al. in PLA/hydrotalcites nanocomposites, which showed an increase of the 29 Young modulus even after 3 reprocessing steps [23]. The moderate impact of the 30 recycling processes on the hardness seems to indicate that the mechanical properties 31 of the recycled materials should not be a limiting factor when considering the use of 32 recycled PLA based materials in packaging applications. 33 The permeability is one of the most important properties of the materials used in 34 food packaging applications, because this property is one the main factors for 35 determining the shelf life of the packed products. In this work, the effect of the aging and 36 recycling processes on the permeability of the material to O2, N2 and CO2 was evaluated. 37 The results, shown in Fig. 5, indicate that the permeability of the nanocomposite clearly 38 decreases as a result of the mechanical recycling. Again, this behavior must be 39 explained by considering two opposing effects of the recycling processes. On one hand, 40 the polymer degradation during the service life and the recycling process must favor the 41 permeation, because fractional free volume, vf, grows as a consequence of the 42 molecular weight reduction and, according to Doolitle's equation, gas diffusion through a 43 glassy polymer becomes easier if the vf increases [35]. On the other hand, XRD and 44 TEM experiments indicate that the melt reprocessing improves the dispersion of the 45 impermeable clay platelets into the polymer, thus increasing the tortuosity of the paths 46 followed by the gas molecules when diffusing through the polymer, which leads to 47 14 decreased permeability. In our case, the low permeability values corresponding to the 1 recycled materials reveal that the improvement of the dispersion of the clay is the 2 predominant effect of the recycling processes. 3 4 PLAV-C30 PLAR-C30 PLARW-C30 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Pe rm ea bi lity (B ar re r) Sample O2 N2 CO2 5 Fig. 5. Permeability of the virgin and recycled nanocomposites (V = virgin; R= recycled; 6 RW = recycled with washing step). 7 8 In summary, the results indicate that the mechanical recycling of PLA-C30 9 nanocomposites cause an improvement on the gas barrier properties as a consequence 10 of the better dispersion of the clay due to the reprocessing. 11 12 0 10 20 30 40 50 60 70 80 90 0 2 4 6 W at er u pt ak e (% ) PLAV-C30 PLAR-C30 PLARW-C30 Time (days) 13 15 Fig. 6. Water absorption curves for PLAV-C30, PLAR-C30 and PLARW-C30 (V = 1 virgin; R= recycled without washing step; RW = recycled with washing step). 2 3 Water absorption and hydrothermal degradation tests are widely used to 4 measure the quality and degradability of materials [36]. In addition, these tests are 5 especially important in materials such as those studied in this work, because it is likely 6 for these materials to be in contact with liquids or wet products during their service life. 7 Fig. 6 presents the water absorption curves of the virgin and recycled 8 nanocomposites. It can be seen that for all materials, the amount of water absorbed does 9 not reach equilibrium but grows continuously with the immersion time. This behavior was 10 observed in a previous study with unfilled PLA [21], and was explained as the result of a 11 two-stage sorption process. In the first stage, the fast water absorption is driven by the 12 concentration gradient of the diffusant, while in the second stage the absorption is driven 13 by the swelling and relaxation of the polymer [21]. Similar results have been reported by 14 Davis et al. [21,37]. Furthermore, the hydrolytic degradation of PLA after long immersion 15 time generates holes and hydrophilic products, which may facilitate the absorption of 16 water [38]. Balart et al. also reported the appearance of cracks and holes in 17 nanocomposites of PLA and hazelnut shell flour immersed in distilled water [36]. 18 The complex nature of the sorption kinetics prevents the whole absorption 19 process from being accurately described using a Fickian model, since this model 20 assumes that the water transport is controlled by a concentration gradient and does not 21 consider other phenomena such as the polymer degradation or the molecular relaxation. 22 However, the first stage of the absorption process has great importance in practice for 23 PLA-based materials, because an important fraction of these materials is used in the 24 packaging of fresh food, so it would be interesting to elucidate if that first stage is 25 controlled by a concentration gradient and can be accurately described using a Fickian 26 model. An accurate model would allow the estimation of the water absorption 27 parameters, and thus the study of the effect of the mechanical recycling on the kinetics 28 of the first stage of the water absorption of the different samples. 29 For evaluating the accuracy of the Fickian model at short absorption times and 30 calculating the parameters of the diffusion process, i.e., the diffusion coefficient and the 31 moisture content at equilibrium, the experimental water absorption data for 120 minutes 32 were fitted using a numerical solution of Fick’s second law, given by Eq. (3) [39]: 33 34 𝑀𝑀𝑡𝑡 𝑀𝑀∞ = 1 − 8 𝜋𝜋2 ∑ 1 (2𝑛𝑛+1)2 ∞ 𝑛𝑛=0 exp �−(2𝑛𝑛+1)2𝜋𝜋2𝐷𝐷𝑡𝑡 ℎ2 � (3) 35 36 where Mt is the moisture content at time t, M∞ is the moisture content at equilibrium or 37 saturation mass, D is the apparent diffusion coefficient and h is the thickness of the 38 sample. The serial expansion of Eq. (3) was truncated and only the first 10 terms were 39 considered. Finally, the values of D were corrected by using Eq. (4), in order to consider 40 the real dimensions of the samples used in the experiment, as indicated by Gupta and 41 Pawar [40]: 42 43 16 𝐷𝐷𝐷𝐷 = 𝐷𝐷 × �1 + ℎ 𝑥𝑥 + ℎ 𝑦𝑦 � (4) 1 2 where Dc is the corrected diffusion coefficient and h, x and y are the thickness, length 3 and width of the samples, respectively. 4 5 0 40 80 120 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 PLAV-C30 PLAR-C30 PLARW-C30 W at er u pt ak e (% ) Time (minutes) 6 Fig. 7. Water absorption data, at short times, for the different samples. The lines 7 correspond to the fitting of the data using Eq. 3 (with n = 10) (V = virgin; R= recycled; 8 RW = recycled with washing step). 9 10 The results obtained by fitting the experimental data using Eq. 3 are summarized 11 in Fig. 7 and Table 2. It can be seen that, for short immersion times, the water absoprtion 12 can be adequately fitted using Eq. 3, thus revealing that the diffusion of water is driven 13 by the concentration gradient in the first stages of the absorption. 14 15 Table 2. Water absorption parameters calculated at short times for the different 16 materials using Eq. 3 (with n = 10) 17 Sample Dc ×108 (cm2/s) M∞ (%) χ2 PLAV-C30 4.2 ± 1.2 0.61 ± 0.01 0.0001 PLAR-C30 3.8 ± 1.5 0.62 ± 0.01 0.00003 PLARW-C30 5.7 ± 1.6 0.65 ± 0.01 0.0002 18 17 Table 2 and Fig. 7 reveal that virgin and recycled nanocomposites behave 1 similarly during the first stages of the absorption process. Only the material obtained 2 when the recycling process includes the washing step shows slightly higher values of 3 the diffusion coefficient and the saturation water. In order to explain these results, the 4 structural changes taking place during the recycling processes must be again 5 considered. The degradation of the recycled materials, responsible for the decrease in 6 the intrinsic viscosity that has been shown above, generates hydrophilic groups and an 7 increase of free volume that favors the diffusion and absorption of water. However, the 8 improved dispersion of clay nanoparticles slows the diffusion of water, so that recycling 9 causes a negligible net effect on the absorption of water at short times. The slightly higher 10 values of D and M∞ in PLARW-C30 are explained as a consequence of the higher 11 degradation observed in this material. 12 13 0 10 20 30 40 50 60 70 80 90 40 60 80 100 120 140 In tri ns ic vis co sit y (m L/ g) Time (days) PLAV-C30 PLAR-C30 PLARW-C30 14 Fig. 8. Evolution of the intrinsic viscosity of the materials during the immersion (V = 15 virgin; R= recycled; RW = recycled with washing step). 16 17 The difference in water absorption between the nanocomposites is higher in 18 prolonged immersion times, as can be seen in Fig. 6. The increase in absorption at long 19 immersion times in PLA-based materials has been related not only to the degradation of 20 the polymer but also to a leaching process with the formation of pores and cracks [38]. 21 Fig. 6 shows that there is a different behavior between PLAR-C30 and PLARW-C30. The 22 former shows a behavior very similar to that of the virgin nanocomposite, while the latter 23 shows a slightly higher absorption. These results indicate that the polymer degradation 24 during the accelerated aging, washing step and reprocessing favors the formation of 25 pores and cracks at long immersion times. To determine if the initial degradation of the 26 18 polymer also favors the decrease in the average molecular weight during the dipping, 1 the intrinsic viscosity values were measured after different immersion times. The results 2 shown in Fig. 8 indicate that the intrinsic viscosity decreases during immersion in water 3 and that the decrease rate is similar in virgin and recycled nanocomposites. 4 The above results show that recycled nanocomposites have acceptable stability 5 against hydrolytic degradation and confirm that recycled nanocomposites have good 6 properties, comparable to those of the virgin material, so they could be used in 7 demanding applications. 8 9 4. Conclusions 10 The effects of accelerated aging and mechanical recycling processes on the 11 structure and properties of PLA-clay nanocomposites have been studied. The results 12 show two main differences between the structure of the recycled nanocomposites and 13 that of the virgin. Firstly, mechanical recycling causes an improvement in the dispersion 14 of the clay nanoparticles into the polymer, which has been observed by XRD and TEM. 15 Secondly, intrinsic viscosity measurements reveal a decrease in the average molecular 16 weight of the polymer in the recycled materials, especially when a demanding washing 17 step is included during the recycling process. These two differences, which have also 18 been observed in DSC curves, have opposite effects on the properties of recycled 19 materials. 20 The virgin and recycled materials behave similarly in contact with water. The 21 water absorption curves are complex, showing at least two distinct stages. In the first 22 stage, at short immersion times, the absorption can be accurately described using a 23 Fickian model. At long immersion times the absorption grows again due to the leaching 24 and formation of voids and cracks in the polymer. The water absorption is slightly higher 25 in the recycled material when the washing step is included in the recycling process, due 26 to the higher molecular weight decrease in this case, which cannot be fully compensated 27 by the better dispersion of the clay in the polymer. However, the decrease in intrinsic 28 viscosity due to hydrolytic degradation is similar for virgin and recycled nanocomposites. 29 Despite the decrease of the molecular weight, the Vickers hardness remains 30 almost unchanged and the thermal stability of the recycled nanocomposites is even 31 slightly higher than that of the virgin, due to the better dispersion of the clay and the 32 formation of carboxyl groups. The better dispersion of the clay also explains the higher 33 optical clarity of the recycled nanocomposites, as well as the improved barrier properties 34 for oxygen, nitrogen and carbon dioxide. 35 In general, the effect of accelerated aging and mechanical recycling processes 36 on the properties of the nanocomposite is limited. The recycled nanocomposites have 37 similar or even better properties in some cases than those of the virgin material, so the 38 use of recycled nanocomposites in demanding applications, and even in the same 39 applications as the virgin material, should not be ruled out. 40 41 5. Acknowledgements 42 The authors would like to thank the Centro Nacional de Microscopía Electrónica 43 and the CAI Difracción de Rayos X of the Universidad Complutense de Madrid (Spain), 44 19 for the collaboration in the TEM and XRD measurements, respectively. The authors also 1 would like to acknowledge the funding from MINECO-Spain (project MAT2013-47972-2 C2-2-P), Universidad Politécnica de Madrid (project UPM RP 160543006) and 3 Ecoembes (project DEHIPLA-R). 4 6. Data availability 5 The datasets analyzed during the current study are available from the 6 corresponding author on reasonable request. 7 8 7. References 9 [1] F. R. Beltrán, V. Lorenzo, J. Acosta, M. U. de la Orden and J. Martínez Urreaga, J 10 Environ Manag (In press). 11 [2] J. D. Badia, L. Santonja-Blasco, A. Martínez-Felipe and A. Ribes-Greus, Polym. 12 Degrad. Stab. 97 (2012) 1881. 13 [3] R. Auras, L. Lim, S. E. M. Selke and H. Tsuji, Poly(Lactic Acid): Synthesis, 14 Structures, Properties, Processing, and Applications. (John Wiley & Sons, New Jersey, 15 2010). 16 [4] J. Raquez, Y. Habibi, M. Murariu and P. Dubois, Prog. Polym. Sci. 38 (2013) 1504. 17 [5] J. Rhim, H. Park and C. Ha, Prog. Polym. Sci. 38 (2013) 1629. 18 [6] P. M. S. Souza, A. R. Morales, M. Marin-Morales and L. H. I. Mei, J. Polym. 19 Environ. 21 (2013) 738. 20 [7] F. Aeschelmann and M. Carus, Ind. Biotechnol. 11 (2015) 154. 21 [8] M. M. Reddy, S. Vivekanandhan, M. Misra, S. K. Bhatia and A. K. Mohanty, Prog. 22 Polym. Sci. 38 (2013) 1653. 23 [9] T. Leejarkpai, T. Mungcharoen and U. Suwanmanee, J. Clean. Prod. 125 (2016) 95. 24 [10] R. Mülhaupt, Macromol. Chem. Phys. 214 (2013) 159. 25 [11] V. Piemonte, J. Polym. Environ. 19 (2011) 988. 26 [12] M. F. Cosate de Andrade, P. M. S. Souza, O. Cavalett and A. R. Morales, J. 27 Polym. Environ. 24 (2016) 372. 28 [13] V. Rossi, N. Cleeve-Edwards, L. Lundquist, U. Schenker, C. Dubois, S. Humbert 29 and O. Jolliet, J. Clean. Prod. 86 (2015) 132. 30 [14] M. Niaounakis, Biopolymers Reuse, Recycling, and Disposal (William Andrew 31 Publishing, Oxford, 2013). 32 [15] J. D. Badia and A. Ribes-Greus, European Polymer Journal 84 (2016) 22. 33 20 [16] J. D. Badia, O. Gil-Castell and A. Ribes-Greus, Polym. Degrad. Stab. 137 (2017) 1 35. 2 [17] R. Scaffaro, M. Morreale, F. Mirabella and F. P. La Mantia, Macromol. Mater. Eng. 3 296 (2011) 141. 4 [18] J. D. Badia, E. Strömberg, S. Karlsson and A. Ribes-Greus, Polym. Degrad. Stab. 5 97 (2012) 670. 6 [19] M. Żenkiewicz, J. Richert, P. Rytlewski, K. Moraczewski, M. Stepczyńska and T. 7 Karasiewicz, Polym. Test. 28 (2009) 412. 8 [20] L. Nascimento, J. Gamez-Perez, O. O. Santana, J. I. Velasco, M. L. Maspoch and 9 E. Franco-Urquiza, J. Polym. Environ. 18 (2010) 654. 10 [21] F. R. Beltrán, V. Lorenzo, M. U. de la Orden and J. Martínez-Urreaga, Polym. 11 Degrad. Stab. 133 (2016) 339. 12 [22] M. A. Kozlowski and J. Macyszyn, Recycling of Nanocomposites in: Polymer 13 Nanomaterials for Food Packaging, eds. C. Silvestre and S. Cimmino (CRC Press, 14 Florida, 2013), pp. 313-336. 15 [23] R. Scaffaro, F. Sutera, M. C. Mistretta, L. Botta and F. P. La Mantia, Express 16 Polym. Lett. 11 (2017) 555. 17 [24] C. Chariyachotilert, S. Joshi, S. E. M. Selke and R. Auras, J. Plast. Film Sheeting 18 28 (2012) 314. 19 [25] E. O. Kraemer, Ind. Eng. Chem. 30 (1938) 1200. 20 [26] J. Arranz-Andrés, V. Lorenzo, M. U. de la Orden, E. Pérez and M. L. Cerrada, J. 21 Membr. Sci. 377 (2011) 141. 22 [27] S. W. Rutherford and D. D. Do, Adsorption 3 (1997) 283. 23 [28] P. E. Le Marec, L. Ferry, J. Quantin, J. Bénézet, F. Bonfils, S. Guilbert and A. 24 Bergeret, Polym. Degrad. Stab. 110 (2014) 353. 25 [29] M. Pluta, J. K. Jeszka and G. Boiteux, Eur. Polym. J. 43 (2007) 2819. 26 [30] M. L. Di Lorenzo, J. Appl. Polym. Sci. 100 (2006) 3145. 27 [31] M. Yasuniwa, S. Tsubakihara, Y. Sugimoto and C. Nakafuku, J. Polym. Sci. B 28 Polym. Phys. 42 (2004) 25. 29 [32] H. M. Cele, V. Ojijo, H. Chen, S. Kumar, K. Land, T. Joubert, M. F. R. de Villiers 30 and S. S. Ray, Polym. Test. 36 (2014) 24. 31 [33] I. Pillin, N. Montrelay, A. Bourmaud and Y. Grohens, Polym. Degrad. Stab. 93 32 (2008) 321. 33 [34] V. Lorenzo and J. M. Pereña, Curr. Trends Polym. Sci. 4 (1999) 65-76. 34 21 [35] M. H. Cohen and D. Turnbull, J. Chem. Phys. 31 (1959) 1164. 1 [36] J. F. Balart, N. Montanes, V. Fombuena, T. Boronat and L. Sánchez-Nacher, J. 2 Polym. Environ. (2017) 1. 3 [37] E. M. Davis, G. Theryo, M. A. Hillmyer, R. A. Cairncross and Y. A. Elabd, ACS 4 Appl. Mater. Interfaces 3 (2011) 3997. 5 [38] M. Deroiné, A. Le Duigou, Y. Corre, P. Le Gac, P. Davies, G. César and S. 6 Bruzaud, Polym. Degrad. Stab. 108 (2014) 319. 7 [39] J. Crank, The Mathematics of Diffusion (Clarendon Press, Oxford England, 1975). 8 [40] K. M. Gupta and S. J. Pawar, Materials Science and Engineering: A 412 (2005) 78. 9 10 1. Introduction 2. Materials and methods 2.1. Materials and preparation of the samples 2.2. Recycled materials 2.3. Experimental techniques 3. Results and discussion 3.1. Effects of recycling processes on the structure 3.2. Properties of the recycled materials 4. Conclusions 5. Acknowledgements 6. Data availability 7. References