Kraft lignin and polyethylene terephthalate blends: effect on thermal and mechanical properties

1Laboratório de Química Analítica, Instituto de Química, Instituto Federal de Educação, Ciência e Tecnologia do Espírito Santo – IFES, Vila Velha, ES, Brasil 2Laboratório Petroleômico e Forense, Departamento de Química, Universidade Federal do Espírito Santo – UFES, Vitória, ES, Brasil 3Divisão de Metrologia de Materiais, Instituto Nacional de Metrologia, Qualidade e Tecnologia – Inmetro, Duque de Caxias, RJ, Brasil 4Universidade Federal do Rio de Janeiro – UFRJ, Duque de Caxias, RJ, Brasil


Introduction
Synthetic or petroleum-based polymers have many practical uses; however, their low biodegradability causes serious environmental problems. Consequently, several strategies to replace or reduce the use of synthetic polymers have been developed [1][2][3][4] . Sustainable development requirements open new perspectives for products obtained from polymer recycling processes, as they significantly contribute to a reduction in plastic waste [5] .
Polyethylene terephthalate (PET) is a semi-crystalline thermoplastic aromatic polyester known for its mechanical properties, lightness, strength, and high transparency, which ensure its widespread use in food and cosmetic packaging materials [2,[6][7][8][9][10] . Currently, recycled PET is often mixed with other polymers or fillers to produce polymer blends or composites with different mechanical and thermal properties compared to neat polymers, thus adding value to raw materials. For added-value processes to be consistently efficient, the compatibility between mixture components is highly essential to achieve satisfactory thermal and mechanical properties for a specific application [11][12][13] . Mechanical enhancements may be useful to the automobile and civil construction industries, while improvements of thermal stability would be useful for applications in the packaging and electronics industries.
Lignin is one of the three main constituents of a plant. It is the second most abundant polymer in the world after cellulose. Generally, its structure depends on the species of wood and the processing conditions. Kraft, sulfite, and soda are the main processes used in chemical wood pulping for extracting cellulose from wood by dissolving the lignin that binds the cellulose fibers together [14,15] .
The kraft pulping process involves digesting wood chips and moiled paper at elevated temperature and pressure in a water solution of sodium sulfide and sodium hydroxide, called "white liquor." The white liquor chemically dissolves the lignin that binds the cellulose fibers together. Spent "white liquor," containing suspended particulate solids and organic compounds, is concentrated to a mostly solid pulp, called "black liquor," which contains between 10 to 50 wt.% of dissolved lignin [14,16,17] . Today, most of the lignin produced by the pulp and paper industry as a constituent part of the black liquor by-product is burned to provide heat for electric power generation. However, as lignin is a complex polyfunctional macromolecule composed of a large number of polar functional groups, it has the potential for use in several technological applications and can be used to produce high-added-value products [18][19][20] .
Blending lignin with a polymeric matrix is a secure way to develop polymer-based products with desirable properties. However, the eventual incompatibility on a chemical level between the components may require the chemical modification of lignin before mixing to improve its dispersion in plastic or to increase interfacial adhesion [14,17,21,22] . This work aimed to develop a new engineering material with enhanced mechanical and thermal properties using waste PET bottles and lignin, which was obtained as a by-product of the kraft process. The PET/lignin blends were produced by melt extrusion and injection molding and contained chemically modified and unmodified lignin. The thermal and mechanical properties of the blends were compared with those of the PET R matrix.
Lignin was extracted from the black liquor via a precipitation method that consisted of the following steps: reduction of the liquor pH value (start solution was pH >13) with CO 2 injection, filtration of the precipitated lignin, suspension of the filtered lignin in a H 2 SO 4 solution (pH 2.5), and filtration and washing of the lignin with an acidic solution (pH 2.5 and 60 °C).
The bottle-grade PET (PET R ) was obtained by grinding colorless PET bottles in a Retsch mill (model SM300) with a 2 mm sieve and 1500 rpm rotation. Prior to milling, the bottle labels were removed, and the areas with glue residue were cleaned.

Lignin modification
The lignin modification was performed as described by Silva et al. [23] . Briefly, 10.0 g of lignin was suspended in 270 mL of 95% ethanol (v/v) under continuous stirring in a mechanical stirrer (Ethik Technology, model 105) to which 27 mL of a 30% NaOH aqueous solution (wt/v) was added at a rate of 1 mL min -1 using an electronic pipette (Transferpette S) at room temperature. After the addition of the NaOH solution, the final solution was stirred for a further 60 min. Subsequently, 12.0 g of monochloroacetic acid was gradually added over the course of 30 min, without further agitation of the solution. The mixture was then stirred for an additional 210 min at 55 °C. The residue was suspended in 670 mL of 95% ethanol solution (v/v), which was neutralized with glacial acetic acid, and the separated product was filtered. After filtration, the product was washed three times with approximately 50 mL of 95% ethanol (v/v) to remove the impurities and by-products and then dried at 60 °C in an oven until a constant mass was achieved. The modified lignin is denominated in this work as ML [23] .

Specimen preparation
Milled PET R , KL, and ML samples were dried in a vacuum oven at 60 °C for 24 h prior to melting extrusion and injection molding.
The PET R /lignin blends were fabricated by extrusion using a Thermo Scientific Haake MiniLab II extruder (processing temperature 275 °C and screw rotation speed 100 rpm). The reference specimens (PET R ) and its blends were injected into a Thermo Scientific Haake MiniJet II injector (injection temperature 275 °C, injection pressure 450 bar, injection time 4 s and molding temperature 25 °C). Table 1 displays the formulations of the specimens.

FTIR
The attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectra were obtained on a Frontier spectrometer from Perkin Elmer. Each spectrum was recorded as the mean of 16 consecutive scans, with a resolution of 4 cm -1 in a working range of 4000 to 630 cm -1 .

Thermogravimetry (TGA)
Thermogravimetry analysis (TGA) was performed on an SDT Q600 from TA Instruments. Approximately 10 mg of the sample was heated in the alumina crucible (25 °C to 900 °C) at a heating rate of 10 °C min -1 under a nitrogen flow of 20 mL min -1 .

Differential Scanning Calorimetry (DSC)
Differential scanning calorimetry (DSC) was carried out using a TA Instruments Q200. Approximately 5 mg of the Table 1. Composition of the formulations used to fabricate the tensile specimens. sample (injected material) was heated in the alumina crucible (25 °C to 400 °C) at a heating rate of 10 °C min -1 under a nitrogen flow of 50 mL min -1 .

Tensile measurements
The tensile tests were performed according to the ISO 527-1:2012 (type 5A specimen) standard. The tests were conducted at the facilities of the Federal University of Espírito Santo (UFES)/LABPETRO using a Lloyd Instruments LR5K Plus universal testing machine. The measurement conditions were a load cell of 5 kN, a 50 mm strain gauge with a measurable deformation of 25 mm, and a crosshead speed of 1 mm min -1 . Figure S1a (see Supplementary Material) shows the ATR-FTIR spectra of the KL and ML samples, and the main bands and respective assignments are listed in Table 2 and  Table 3. Figure S1a shows significant differences between the ATR-FTIR spectra of ML and KL samples -one of them is the absence of the band at 1710 cm -1 in the ML spectrum, exhibited by unmodified lignin (KL), which is attributed to the vibration of the carbonyl group conjugated to the aromatic ring. Instead, ML has two high-intensity bands, at 1598 and 1416 cm -1 , assigned to the carboxylate anion. The appearance of these intense bands demonstrates the efficiency of the carboxylation process in the modification reaction. The KL and ML samples both exhibited bands related to the stretching and in-plane bending of the CH 2 group, at 2938 and 1453 cm -1 , respectively, (Tabs. 2 and 3), indicating that these groups did not participate in the chemical reaction. The main bands of the PET R polymer matrix as well as its PET R /KL and PET R /ML lignin blends are assigned in Table S1. Figure 1a depicts the TGA curves of KL and ML. Mass losses of approximately 6 wt.% for KL and 22 wt.% for ML were observed at around 100 °C. This weight drop can be attributed to moisture loss. The chemical decomposition of KL and ML occurs over a wide temperature range, with the most intense mass loss being observed between 200 °C and Table 2. Assignments of the vibrational bands of the lignin (KF) [23][24][25] .  Table 3. Assignments of the vibrational bands of the modified lignin (ML) [23][24][25] .    [23,26,27] . The wide range of thermal degradation of lignin can be explained by the different oxygenated functional groups in their structure, which provides different thermal stabilities [28] . The thermal degradation of lignin generally occurs in three main steps: 0 to 120 °C, related to the evaporation of absorbed water, 180 to 350 °C, associated with the degradation of carbohydrates, which are converted into volatile gases such as CO, CO 2 and CH 4 , and above 350 °C, related to the degradation of lignin-derived products together with the removal of produced gases [28] .

TGA
Due to the complexity of lignin's structure, the chemical decomposition of this material involves several competing reactions. Lignin contains several chemical subunits within its macromolecule. During the thermal degradation process, chemical bonds of different bond energies are broken [23,26,27] .
The dehydration of lignin prevails in its thermal degradation pathways. Dehydration of the lignin structure results in pyrolysis products with unsaturated side chains [28] . Carbon monoxide, carbon dioxide, and methane are also formed during these processes [29,30] . The decomposition of aromatic rings occurs above 400 °C [31] . Prolonged heating above 400 °C leads to saturation of aromatic rings, the disruption of C-C bonds, and the release of smaller molecules, such as water, CO 2 , and CO, accompanied by the rearrangement and condensation of the aromatic rings within the lignin structural units [29] . Figure 1a shows l a plot of the first derivative of the mass loss versus temperature for the KL and ML samples. One can conclude that the KL sample is thermally more stable than the ML sample due to the highest temperature of maximum weight loss related to the lignin carbohydrate degradation appeared at 320 °C for the ML sample but appeared in the interval between 320 °C to 380 °C for the KL sample.
According to Kindsigo and Kallas [32] , at ambient temperature in the presence of oxygen and water, it is possible to observe the degradation of lignin via wet oxidation, which is increased with elevated temperatures. In the analyzed temperature range of 110-190 °C, the lignin degradation was 20% at 110 °C and 53% at 190 °C; thus, it is expected that at higher temperatures the degradation rate would continue to increase. In this way, the chemical modification of lignin through carboxylate anion incorporation favors the incorporation of water molecules into its structure due to the higher hydrophilicity of the carboxylated lignin sample (ML sample), making it more unstable and resulting in faster degradation. The total mass losses of the KL and ML samples when heated to 900 °C were 96 wt.% and 88 wt.%, respectively. One can also see from Figure 2a that the lignin continued to decompose at temperatures higher than 900 °C. Figures 1b and 1c compare the TGA curve of PET R with those of the PET R /KL (Figure 1b) and PET R /ML (Figure 1c) blends. The results of the TGA measurements demonstrate that the admixing of KL and ML into the PET R to form a blend caused a shift in the onset temperature of thermal degradation (450 °C), whereas the PET R sample exhibited two mass loss events: the first one in the temperature range of 50 °C to 100 °C (mass loss of 12.5 wt.%) and the second one in the temperature range of 400 °C to 450 °C (mass loss of 75 wt.%). As a consequence, these blends presented a higher residual mass than PET R (20-25 wt.% versus 12.5 wt.%) at 450 °C.
The increase in the onset temperature of thermal degradation of the PET R /lignin blends in relation to neat PET R may be attributed to the chemical compatibility between the PET R and KL and ML fillers due to the aromatic structure present in both sample types. However, the moisture present in the lignin blends increases the mass loss content in this temperature range, as our results show (see Figure 1a) with degradation temperatures above 450 °C. Figure 2a shows the DSC curves of KL and ML. Typically, the T g values of unmodified lignins vary from 90 to 180 °C [33,34] . The significant enthalpy relaxation process that usually occurs in polymers during DSC scanning makes it challenging to determine the T g value of lignin from the DSC measurements. Complex hydrogen bonding interactions and the highly amorphous structure of kraft lignin favor enthalpy relaxation. This is detected in the interval where a slope change occurs in the heating curve [35][36][37] . The T g values of KL and ML are 100 °C and 127 °C, respectively (Figure 2a). Table 4 summarizes the thermal properties of the studied samples obtained from the DSC measurements, such as the glass transition temperature (T g ), crystallization temperature (T c ), melting temperature (T m ), melting enthalpy (ΔH m ), and degree of crystallinity (X c ).

Differential Scanning Calorimetry (DSC)
The variation in the values of T g , T c , and T m of the blends in relation to those of PET R show a clear dependence on both the type and the amount of lignin added. The PET R sample exhibited T g , T c , and T m values of 63.3, 119.5, and 253.8 °C, respectively, which are typical for this polymer. The DSC curves (Figure 2b, c) for the blends also showed well-defined endothermic peaks of glass transition, exothermic cold crystallization, and endothermic melting peaks, which is characteristic of PET.
In general, PET R /KL blends exhibited higher T g and T m values (approximately 70 °C and 257 °C, respectively) compared with those of PET R . This shift of the glass and melting transitions to higher temperatures was somewhat larger in the case of PET R /ML for T m (approximately 258 °C) and was notably larger for T g (about 75 °C), except for the PET R /ML 5.0 wt.% sample. On the whole, the PET R /ML blends showed more stable melt characteristics (melting temperatures and melting enthalpies) compared to those of the PET R /KL blends.
The value of T c presented a maximum decrease in three units (119.5 → 116.4 °C) of temperature for the PET R /KL 1.0 wt.% blend, while in the PET R /ML blends, the most significant variation was T c = 114 → 110 ° C ( Table 5). The cold crystallization temperatures of the blends showed an opposite trend from the melting temperatures, where the cold crystallization temperatures of the PET R /KL and PET R /ML samples were decreased in comparison to that of PET R . This temperature reduction was notably larger in the PET R /ML blends.
The lignin macromolecule contains polar groups capable of producing chemical interactions to become closer to the PET R chains. These secondary forces may have contributed to the increase of T g and T m for the studied blends. The PET R /ML blends presented higher increases in these two parameters in relation to the PET R /KL blends, probably due to the incorporation of carboxylate groups in the lignin, favoring its chemical interaction with the PET R matrix, and consequently, increasing the T g and T m values [38,39] .
Miscibility is a crucial parameter to be achieved in polymer blends to improve the properties of homopolymers. In our study, a single T g was observed in the DSC curves of the blends, which is indicative of miscibility. The polarity of the lignin molecules results in strong interactions between them that hinder their miscibility with other polymers. To maintain miscibility, interaction forces between the polymer matrix (PET in our case) and lignin are required. Generally, PET has the ability to interact with lignin through π electronic interactions favoring the miscibility between them. Hydrogen bonds that eventually form between PET and lignin polar groups also facilitate miscibility. For this reason, lignin surface chemical modification is typically performed to reduce the interaction forces between the lignin molecules, attaching them to hydrophilic polymer matrices, such as PET R .
The degree of crystallinity is another important property of semi-crystalline thermoplastics that is directly related, among others, to the mechanical properties of plastics. The degree of crystallinity by DSC of the studied blends was assessed using the following Equation 1:  In this equation, the difference between the measured heats of melting, ΔH m , and the cold crystallization, ΔH c , define the fusion enthalpy of a sample. The term ΔH m 0 is a reference value and represents the heat of melting if the polymer were 100% crystalline. The value ΔH m 0 = 140 J g -1 was used [38] .
All samples exhibited low values of X c (see Table 4). Two different factors could have affected the overall crystallinity of the samples. The first one is related to the production process of the specimens, which was by injection molding. Significant differences between the injection and molding temperatures resulted in a predominantly amorphous sample structure because the polymer chains did not have enough time for periodic ordering. On the other hand, the presence of KL and ML lignin favored the crystallization of PET R , indicating that lignin acted as a nucleating agent to the PET domains [39] .

Tensile tests
The results of the mechanical tensile tests, including the modulus of elasticity ( ), maximum tensile strength (σ), and strain at break (ε ), are shown in Table 5. The σ parameter (maximum tensile strength) of a polymer blend indicates the tension transfer capacity from the matrix to the filler, where the higher the interaction between phases, the higher the maximum tensile strength. The chemical modification of lignin did not produce efficient interfacial molecular interactions between the PET and ML components or, consequently, an enhancement in the mechanical properties of the PET R /ML blends, as is seen from the experimental data (Table 5). Moreover, higher contents of KL in polymer blends (<3.0 wt%) also seemed to be unfavorable for the strength and deformability of the PET R /KL blends. Formulations containing 3 wt.% and 5 wt.% of ML could not be tested at all, as the mixture of ML with PET R in these proportions resulted in specimens forming cracks even before the mold was withdrawn.
In contrast, Young's modulus of all PET R /KL blends followed the fundamental law of mixtures and was nearly the additive function of the composition. On the other hand, lignin, which is more rigid than PET R , has the capacity to support the applied tension transferred from the polymeric matrix to itself, resulting in a higher Young's module of the respective blends in comparison with PET R [40] . Distinct from Young's modulus, the strength, and deformability of blends show more complex behavior. It is generally thought that these characteristics primarily depend on the strength of interfacial adhesion of the components, which is determined, in turn, by the mutual contact surface and the respective strength of interaction [41] . Water molecules can act as plasticizers to increase the mobility of lignin's polymer chains during its dispersion in the PET R matrix. There was a much finer dispersion of the KL phase than the ML phase in the respective blends, resulting in a higher contact surface between the KL and PET R molecules. However, the positive effects of KL on the tensile properties of the PET R /KL blends seemed to diminish after the content of stiff KL chains exceeded a critical limit. We believe that the analogous critical limit was already attained for the lowest content of ML in the PET R /ML samples with respect to Young's modulus of the material. As a result, a decreasing trend in the stiffness of PET R /ML with the content of ML was observed.

Conclusions
In this study, kraft lignin (KL) and chemically modified kraft lignin (ML) were used to produce PET R /lignin blends. The results of ATR-FTIR, TGA, and DSC analyses verified the presence of chemical modifications in the ML samples. The TGA measurements indicated that KL was thermally more stable than ML, which is intrinsically linked to the ability of ML to absorb water, thus, increasing this material's susceptibility to degradation. Additionally, the water molecule absorption impacts the mechanical properties of the PET R /lignin blends, making the PET R /ML blends more fragile than the PET R /KL blends. The ATR-FTIR spectra of the blends showed no significant differences between them, while the DSC curves exhibited higher glass transition temperatures for the PET R /lignin blends compared with the PET R material.
PET R /KL blends with small amounts of KL (up to approximately 1%) showed improved mechanical properties. Both the modulus of elasticity and the maximum tensile strength of PET R benefited from the addition of lignin in this situation; this is the expected behavior of compatible polymer blends.