High - S trength P olylactic A cid (PLA) B iocomposites R einforced by E poxy - M odified P ine F ibers Xianhui Zhao a , b, * , Kai Li a , Yu Wang a , Halil Tekinalp b , Greg Larsen b , Daniel Rasmussen c , Ryan Ginder d , Lu Wang e , f , Douglas Gardner e , Mehdi Tajvidi e, f , Erin Webb g , Soydan Ozcan b , d , * a Chemical Science s Division, Oak Ridge National Laboratory, 1 Bethel Valley Road, Oak Ridge, Tennessee 37831, United States b Manufacturing Demonstration Facility, Energy and Transportation Science Division , Oak Ridge National Laboratory, 2350 Cherahala Blvd, Knoxville, Te nnessee 37932, United States c Engineering and Media Technologies , Pellissippi State Community College , 10915 Hardin Valley Road , Knoxville , Tennessee 37932 , United States d Department of Mechanical, Aerospace and Biomedical Engineering, University of Tennessee, Knoxville, Tennessee 37996, United States e Advanced Structures and Composites Center , University of Maine , 35 Flagstaff Road, Orono , Maine 04469 , United States f School of Forest Resources , University of Maine , 117 Nutting Hall , Orono, Maine 04469, United States g Environmental Sciences Division, Oak Ridge National Laboratory, 1 Bethel Valley Road, Oak Ridge, Tennessee 37831, United States *Corresponding author. Tel.: +1 8 65 - 241 - 215 8; E - mail address: ozcans@ornl.gov ( S Ozcan ) ; E - mail address: zhaox@ornl.gov (X. Zhao) ; ABSTRACT : 2 T he stiffness and tensile strength of biopolymers (e.g., polylactic acid (PLA)) are less than desirable for load - bearing applications in their neat form T he use of natural fibers as reinforcements for composites (for large - scale 3D printing) has expanded rapidly attributable to their low weight, low cost, high stiffness, and renewable nature. Silane and acid/alkali are typically used to modify the surface of natural fibers to improve fiber/polymer interfacial adhesion. In this study, a simple method of impregnation was developed to modify pine fibers ( Loblolly , mesh size of 90 – 180 μ m , 30 wt% ) with a solvent - borne epoxy to reinforce PLA As a benefit of the epoxy modification (0.5 – 10 wt%) , the tensile strength s and Young’s modul i of the epoxy - modified pine/PLA composites increased by up to 20% and 82% respectively, as compared to neat PLA. The epoxy - modified pine/PLA composites , with an optimum epoxy modification (1.0 wt%) , had fewer voids on the fracture surface as compared with pine/PLA composites without the modification of pine fibers via epoxy. Results confirmed that e poxy partially penetrate d the por e/ hollow inner structures of pine fibers and improve d the fiber/matrix interfacial adhesion. Epoxy modification is found to be a simple and effective technique to improve the properties of biocomposites. KEYWORDS : Natural fiber, P ine , Epoxy , Polylactic acid, Biocomposi te , Impregnation 3 INTRODUCTION D emand for biomas s has increased recently to produce renewable high - performance bio - derived materials and help mitigate global warming 1 - 2 Polylactic acid (PLA) , made from renewable biomass, is an environmentally friendly bio polymer and has relatively high mechanical properties 1, 3 - 5 The easy processability and superior mechanical properties make PLA a promising biopolymer for industrial composite applications 1 Synthetic fibers such as carbon and glass fibers have been added to PLA to further improve the mechanical properties Synthetic fiber reinforced composites are being used for automotive and aircraft industry 6 However, the use of synthetic fiber reinforcement has significant drawbacks such as dependence on non - renewable resources, lack of material biodegradability at end of life , and high CO 2 output in the manufacturing process. N atural fibers , on the other hand, are low cost, low density (light weight) , biodegradable , and recyclable. Furthermore , some of these natural fibers are agricultural and industrial waste/ by - products , meaning their use could also reduce waste and create new value - added streams for existing industries 6 Natural fibers , from crops such as p ine, poplar, corn stover, and switchgrass , have been widely developed in the United States from research efforts on the production of biomaterials and biofuels 7 For example , p ine is a wind - pollinated crop , and p ine plantations have covered more than 32 million acres in the southeastern United States , where L oblolly pine ( Pinus taeda L.) is a commercially important and native pine species 8 - 9 P ine consists of approximately 42 wt % cellulose, 21 wt % hemicellulose , 26 wt % ligni n , and 3 wt% extractive 7 Natural fibers like these have been found to be able to improve the thermal and mechanical properties of polymers for composite applications 1, 10 However, t he low dispersibility of natural fibers and low interfac ial adhesion between natural fibers and PLA has created unique challenges for producing viable industrial composite s 1 T he mechanical properties of natural fibers are 4 typically inferior to those provided by conventional carbon or glass fibers , which restricts natural fiber composite applications to those where high mechanical performance is not required 6 A number of physicochemical methods , such as using silane or acid/alkali as modifiers , have been utiliz ed to modify the surface of fiber s to improve fiber dispersion and fiber - matrix interfacial adhesion . Increased interfacial interactions between fibers and matrix can result in higher tensile strength and Young’s modulus 6 PLA might be used to modify the fiber surface to improve the interfacial compatibility between fibers and PLA matrix. However, typical PLA solvent s like dichloromethane (CH 2 Cl 2 ) are costly and toxic thereby limiting interest in this approach P hysical treatment s (e.g., steam explosion and mechanical hot - pressing ) can change the structural and surface properties of natural fibers , thus influenc ing their mechanical bonding to polym ers 11 Chemical s urface treatments can bridge the gap in compatibility between hydrophilic natural fibers and hydrophobic polymer matrices to improve the performance of natural fiber reinforced polymer composites 11 E poxy resins are a natural choice for th is type of surface treatment 12 Epoxy resins exhibit high adhesiveness and superior chemical and heat resistance , 13 making them an attractive modifier for high - performance fiber/polymer composites E poxy resins are typically reacted with a curing agent (i .e., hardener , e.g., amine, dicyandiamide (DICY) , anhydride, etc. ) to form crosslinked polymer networks with desirable mechanical and thermal properties 13 These r eactions generally utilize the ring opening of the epoxide group , 14 but may also engage direct cross - linking of epoxy monomers to one another. S everal studies have been carried out on natural fibers treated by epoxy for biocomposite application s For example, Sujaritjun et al. 12 used a flexible (rubber - modified) epoxy ( epoxidized polybutadiene, EPOLEAD PB 3600 ) to treat bamboo fibers for reinforcing PLA, which improved tensile strength by approximately 10 % , compared with untreated bamboo 5 fiber/PLA compo site s Kyutoku et al. 1 utilized cellulose fibers coated with epoxy - based agents to reinforce PLA and found the epoxy treatment improved the interfacial adhesion between cellulose fibers and PLA It seems the me chanism for the improve ment was not given in th eir study. Their results showed an increase in storage modulus of up to 30%, when compared to untreated cellulose fiber/PLA composite s For the ir surface treatment, the cellulose fibers were completely immersed in epoxy - based agent solution . After stirring for 1 day, the coated cellulose fibers were removed 1 However , thermo - mechanical propert y data on this class of composites , such as thermal stability and rheological behavior , remain sparse In addition, the exact mechanism responsible for mechanical property improvements in epoxy surface treat ed fiber reinforc ed biocomposites has not been thoroughly explained. In the present study, a simple impre gnation method was developed to introduce solvent - borne epoxy for modifying pine fiber reinforce d PLA composites. It is hypothesized that using epoxy to modify the structure of the pine fibers can improve the strength of the fiber/PLA interfac e yielding increased composite performance To optimize the epoxy content of the composites , different amounts of epoxy ( 0.5 – 10 wt%) were investigated The mechanism behind the enhancement of epoxy treated composites was investigated and discussed herein Different investigative techniques , including tensile testing , dynamic mechanical analysis (DMA), F ourier - transform infrared spectroscopy (FT - IR) , rheolog ical testing , differential scanning calorimetry (DSC), t hermogravimetric analysis (TGA), and scanning electron microscopy (SEM) , were used for composite characterization As a benefit of the epoxy modification, the tensile strength and Youn g’s modulus of the epoxy - modified pine/PLA bio composites increase d by up to 20% and 82% respectively, compared with neat PLA. A simple, cost - effective, environmentally friendly, and industrially scalable approach for the preparation of strong biocomposites ( reinforced with 30 wt% of natural fiber s ) wa s demonstrated. 6 EXPERIMENTAL SECTION Materials Pine wood chips ( Loblolly , Pinus taeda , low ash and de - barked ) were provided from Screven County , GA They w ere milled with a hammer mill ( screen size of 1.59 mm, Model 5, Meadow s Mills Inc., North Wilkesboro, NC ) and then sieved into pine fibers ( or flours or particles, mesh size: 90 – 180 μ m ). PLA pelle ts ( CAS: 9051 - 89 - 2, Ingeo biopolymer 4043D) w ere purchased from NatureWorks LLC (Minnetonka, MN ). Poly( b isphenol A - co - epich lo rohydrin), glycidyl end - capped ( PBG, number average molecular weight: ~355, CAS: 25036 - 25 - 3 ) , and DICY ( molecular weight : 84.08 g/mol , CAS: 461 - 58 - 5 ) were both purchased from Sigma - Aldrich Co. ( St. Louis, MO) The structures of PBG and DICY are shown in Figure S1 Methanol (CAS: 67 - 56 - 1) was purchased fro m VWR International, LLC (Radnor, PA) Preparation of E poxy - modified P ine F ibers The epoxy - modified pine fibers were prepared using a n impregnation method The epoxy system ( hereafter simply called epoxy, unless otherwise noted ) was a two - component system combining PBG and DICY ( curing agent ) at a molar ratio of 1:1 The epoxy was dissolved in methanol to form a n epoxy solution with a concentration of 7.7 wt% (i .e., the sum of PBG and DICY counted 7.7 wt% , and methanol counted 92.3 wt% ) For the epoxy dissolution, PBG was first added into a glass bottle, followed with methanol , and then DICY The epoxy concentration of the impregnation solution was chosen to balance both the time needed to dissolve the epoxy in methanol and time needed for final evaporation of the methanol. During impregnation, the epoxy solution was doped gradually and evenly o n the pine fibers. The epoxy solution was absorbed on the surfaces of the pine fibers, as shown in Figure S 2 The mass ratio s of the epoxy solution to the pine fibers were varied 0.22 – 4.33 : 1 to test the impact of different epoxy loadings on final composite 7 performance The epoxy solution and pine fibers were mixed well and then dried at room temperature for approximately 2 h to allow for methanol evaporat ion . The mixture was then dried at 100 °C for 1 h , followed by 80 °C overnight. The final dried mixture was considered as epoxy - modified pine fibers. The drying temperatures 100 °C and 80 °C were chosen to evaporat e the solvent and remov e absorbed moisture without activating the epoxy cur ing T he optimization of the molar ratio of PBG to DICY , epoxy concentration in the epoxy solution , and drying conditions can be investigated in a future study Preparation of B iocomposites The bio composite preparation process mainly consisted of im pregnation , melt compounding, hot pressing , and compression molding PLA with untreated and epoxy - modified fibers , as shown in Figure 1 (A) During the impregnation process, epoxy was utilized to reduce the defects and modify the surface of the pine fiber s via penetrating the pores and hollow inner channels of pine fibers. It was observed that some of the outer pores and hollow channels of pine fibers were partially filled by the epoxy molecules , shown in Figure 1 (B - F ) , Figure S3 and Figure S4 . The outer pores and hollow channels of pine fibers have diameters (major axis for elliptical channel) of approximately 6 μ m and 7 – 23 μ m, respectively, so that they can be penetrated by the epoxy/methanol solution during the preparation of epoxy - modified pine fibers. During melt compounding, PLA pellets were gradually loaded in to a shear m ixer (C.W. Brabende r Instruments, Inc. , So. Hacken sack, NJ ) at 7 0 revolutions per minute ( rpm ) and 180 °C D ried , untreated or epoxy - modified pine fibers were loaded gradually in to the shear m ixer a fter melting and mixing the pure PLA pellets for 3 min . The compounding of fibers and PLA matrix was run for 5 min 8 Figure 1. The (A) pine fiber reinforced PLA composite production process, including epoxy impregnation , melt compounding, hot pressing, and final compression molding. (B) SEM image of outer surface of pine fibers . (C) SEM image of cross - section surface of pine fibers. (D) and (E) SEM images of outer surface of epoxy - modified pine fibers. (F) SEM image of cross - section surface of epoxy - modified pine fibers. ( G ) T ensile strength and Young’s modulus of (a) neat PLA, (b) pine/PLA, (c) 0.5 epoxy/pine/PLA, (d) 1.0 epoxy/pine/PLA, (e) 2.0 epoxy/pine/PLA, and (f) 4.0 epoxy/pine/PLA. The resulting biocomposite material s w ere collected and loaded in to a square mold (length × width × thickness : 100 × 100 × 1.65 mm) , heated at 1 8 0 °C for 5 min , and then pressed at approximately 4,536 kg (i.e., 10,000 pounds) for another 5 min in a Carver Laboratory Press ( model number: 2878, Fred S. Carver Inc. ) T he pressed plaques w ere cooled at room temperatu re under a heavy metal plate (~ 20 kg) for 2 min After that , the test plaques w ere cut 9 into multiple slit - shaped bars using a premium guillotine trimmer ( model number: 561 , Dahle ). The obtained slit - shaped bars were further compression - mold ed in the Carver Laboratory Press at 1 8 0 °C in to uniform bars , based on American Society for Testing and Materials ( ASTM ) standard D470 3 15 The total pine fiber content in all composites was fixed a t 30 wt% with epoxy content (0.5 – 10 wt%) varied For example, t he 1.0 epoxy / pine /PLA composite was a combination of 1.0 wt% of epoxy , 3 0 wt% of pine fibers, and the remainder PLA. Material Characterization Mechanical properties: Tensile tests of dog - bone specimens were performed on a servo - hydraulic testing machine to determine the tensile strength and Young’s modulus, according to ASTM standard D638. 3 A Universal MX7. Vi software was used with a strain channel of Stroke 3inch. The strain rate of 1.524 mm/ min , gage length of 12.7 mm, and load cell of approximately 907 kg were used. An extensometer was used, and t he test frame wa s a custom 4 post servo hydraulic frame using an MTS actua tor. The testing of each sample was repeated around three times to ensure the data reliability , and the average value s were reported TGA test : The thermal stability of the samples was inves tigated using a TGA ( Q500 , TA Instruments) in N 2 at a purge flow rate of 20 mL/min 3 The dry samples were heated from 35 to 70 ° C at a heating rate of 10 ° C/ min and kept at 70 ° C for 20 min to remove moisture absorbed in the air T hen , samples were heated to 7 00 ° C at a heating rate of 10 ° C/ min. DSC test : The thermal properties of the samples w ere measured using a DSC ( Q2000 , TA Instruments) 3 During the first heating cycle, the sample s (3 – 7 mg) w ere heated from 20 to 200 ° C at a heating rate of 5 ° C/min, and stabilized at 200 ° C for 1 min. Then the sample s were cooled to 20 ° C at a rate of 10 ° C/min and heated again to 200 ° C at a rate of 5 ° C/min for the second heating cycle Nitrogen with 50 mL/min was used as purge gas. The degree of crystallization ( χ c ) was measured from the first heating curve using Eq. (1), where w is the 10 weight percentage of PLA in the composite sample , Δ H m is the melting enthalpy of a material , Δ H 100 is the melting enthalpy of 100% crystalline PLA ( 93 J/g ) , 16 and Δ H c is the crystallization enthalpy of a material. The glass transition temperature ( T g ) , crysta llization temperature ( T c ), and melting temperature ( T m ) were measured from the second heating curve. 𝑥 ! = ∆ # ! $ ∆ # " % × ∆ # #$$ (1) DMA test : The damping (reported as tan 𝛿 ) and storage modulus (E ́) of the samples were determined using a DMA ( Q800 , TA Instruments) 3 The glass transition temperatures T g (E ́) and T g ( t an 𝛿 ) were observed in the onset of the storage modulus drop and the peak of the t an 𝛿 , respectively. A multi - frequency - strain mode was used, and t he strain was 0.05%. A dual cantilever clamp w as used on rectangular specimen s with dimensions approximately 6 4 × 9.6 × 2.9 mm (length × width × thickness ). The specimen s were tested from 25 to 120 ° C at a heat rate of 3 ° C/min. FT - IR test : The chemical structures of the samples were investigated using a FT - IR spectrometer (PerkinElmer ) , scann ing from 4000 cm - 1 to 600 cm - 1 . The accumulation and resolution were 32 scans and 2 cm - 1 , respectively. SEM observation : The outer, cross - section, and fracture surface morpholog ies of the samples w ere observed using a field emission scanning electron mic roscopy (FE - SEM, Merlin TM , Carl Zeiss NTS GmbH) at 1.0 0 kV The pine fibers’ cross - section surface was derived from being fractured in liquid nitrogen. The samples’ fracture surface was derived from being pulled to break during the tensile tests . The surface w as gold coated with a sputtering device (Cressington Sputter Coater : 208HR ) prior to SEM analysi s Rheological properties: The complex viscosities and storage moduli of the sample s were studi ed using a Discovery Hybrid Rheometer (DHR - 3 , TA Instruments ) at 1 8 0 ° C within the linear viscoelastic region 3 The ga p was set a t 600 μ m. Parallel plates (ETC aluminum 11 disposable) with 8 mm diameter were applied for the frequency sweep test ( from 0.1 to 100 rad/s ) C ontact angle test: The water contact angles of PLA, epoxy, and pine wood chip w ere determined using a Force Tensiometer ( KRÜSS GmbH, Germany ) at room temperature The sample was mounted on a holder and test ed with a detection speed of 6 mm/min. A laborator y desktop software LabDesk 3.2.2 was used. RESULTS AND DISCUSSION A nalysis of PLA - based C omposites Tensile P roperty R esults Figure s 1 ( G ) show s the measured tensile strength s and Young’s modul i for neat PLA and the PLA - based bio composites. With the addition of pine fibers into PLA, the Young’s modulus increased by 65% without sacrific ing tensile strength. This likely indicates that the pine fibers restrain PLA polymer chain movement in addition to imparting stiffness enhancement as composite strength would otherwise have been expected to decrease Using epoxy - modified fibers within the pine/PLA composite caused both tensile strength and Young’s modulus to increase. When 1.0 wt% epoxy was used, the tensile strength and Young’s modulus attain ed their highest values, 71 MPa and 5,378 MPa (i.e., 5.4 GPa) , respectively. While too low epoxy content (e.g., 0.5 wt%) was insufficient to fill the pores of pine fibers and modify their surface, t oo high epoxy content (e.g., 4.0 wt%) caused a decrease in both tensile strength and Young’s modulus by 7% and 8%, respectively. The excess epoxy molecules might be acting as a plasticizer , 17 causing property degradation. The measured tensile property numerical values for the 10 epoxy/pine/PLA composite and other PLA - based composites are listed in Table S1 C omparison of example tensile strength s and Young’s modul i of PLA - based composites compounded with various natura l fibers (such as bamboo, poplar, pine, banana, and jute) from 12 other sources is presented in Table 1 The tensile strength s of these composites typically var y from ~23 to 73 MPa. The Young’s modul i of these composites are typically in the range of ~ 1 .0 – 9.0 G Pa. The natural fiber species, fiber content , fiber size , treatment method, and treating material content have an effect on the mechanical properties of the natural fiber - PLA composites. For example, t he 30 wt% bleached jute strand fiber/PLA composite s exhibited exceptionally high tensile performance In this instance, the composite performance likely stems from j ute being a long bast fiber in the form of a strand and that, by bleaching the jute strand fibers with NaClO solution to remove lignin , the f ibers were primarily comprised of cellulose A cellulose rich surface will have more hydroxyl groups that may contribute to enhanced interfacial bonding through hydrogen bonding Another possible reason is that bleaching has increased the fiber surface roughness to improve the fiber - matrix interface Otherwise, t he 1.0 epoxy/pine/PLA composites obtained in this study exhibited both higher tensile strength and Young’s modulus than most of the natural fiber - PLA composites. The developed impregnation method is simple and highly improve s the mechanical properties of the natural fiber - PLA composites. 13 Table 1. The C omparison on T ensile S trength and Young’s M odulus of PLA - based C omposites C ompounded with N atural F ibers F iber Treatment of natural fibers Tensile strength (MPa) Young’s modulus ( G Pa) Ref. 30 wt% bamboo fiber + 15 wt% t annic acid - crosslinked epoxidized soybean oil ( TA - ESO ) Spraying TA - ESO solution on bamboo fibers ~23 ~1 .0 18 30 wt% pine fiber + 3 wt% maleic anhydride grafted polyethylene ( MAPE ) Adding MAPE as a coupling agent 26 1 .5 19 30 wt% banana pseudo - stem water - soluble extract fiber Isolating water - soluble extract from banana pseudo - stem 30 2 .6 20 19.6 w t% poplar fiber + 0.4 wt% coupling agent Adding 3 - a minopropyltriethoxy silane (KH550) as a coupling agent ~46 ~1 .5 21 20 wt% poplar fiber No treatment 54 4 .3 3 30 wt% bamboo fiber + ~ 0.3 wt% epoxy Epoxy treatment (no details on treatment process ) ~57 ~4 .9 12 30 wt% bamboo fiber No treatment ~61 ~2 .1 18 30 wt% alkali - treated bamboo fiber Soaking bamboo fibers with NaOH solution 62 N/A 22 16 wt% poplar fiber + 4 wt% polymethyl methacrylate ( PMMA ) Adding PMMA as a compatibility agent , and varying poplar fiber size s ~(65 – 70) ~(1 .7 – 1 .8 ) 23 30 wt% bleached jute strand fiber Bleaching jute strand fibers with NaClO solution for 0.5 h 69 9.0 24 30 wt% bleached kraft pine fiber Bleaching kraft pine fiber s, and a dding diglyme (evaporated after compounding) as a dispersing agent 69 N/A 25 30 wt % pine fiber + 1 wt% epoxy Impregnation of epoxy solution on pine fibers 71 5 .4 This study 30 wt% bleached jute strand fiber Bleaching jute strand fibers with NaClO solution for 1.5 h 73 9.0 24 SEM R esults Figure 2 shows the fracture surface s of PLA - based composites after tensile testing While n eat PLA exhibited a relatively smooth fracture surface in line with its expected brittle behavior , the addition of pine fibers resulted in presence of large voids. These large voids indicate fiber pullout from the fiber - matrix interface failure during tens ile testing. There was also an obvious interface gap ( typically ~ 0.9 – 3.3 μm ) between the PLA matrix and pine fiber without epoxy modification ( Figure 3A ). P ine fibers being pulled out of the PLA matrix 14 indicat es a poor fiber - matrix interfacial adhesion (causing composites susceptible to debonding) limiting mechanical p erformance Some of the smaller voids could also have been generated from slight polymer or fiber decomposition or off gassing which should be minimized for composite performance. With the initial addition of epoxy, fewer voids were observed in the epoxy - modified pine/PLA composites. The pine fibers were somehow connected with the PLA matrix, probably because the epoxy modified the surface of pine fibers to enhance the interface between hydrophilic pine fibers and hydrophobic PLA matrix ( Figure 3B ) The fiber - matrix interface gap range dec reas ed to approximately 0.1 – 1.0 μm at the fracture surface of the 1.0 epoxy/pine/PLA composites. These improvements (fewer voids and better fiber - matrix interface in the composites) with the addition of epoxy are consistent with the above higher tensile strength and Young’s modulus, proving that the epoxy improves the compatibility between pine fibers and PLA matrix. However, more voids were observed when the epoxy content was increased to 4 wt% in the epoxy - modified pine/PLA composites, probably because the excessive epoxy dispersed into the PLA matrix can be pulled out. Similar behavior has been obse rved by Zhang et al. 26 Their study reported that the mobility of PLA molecules increa sed sufficiently to fill gaps between wood fibers after the addition of PLA grafted with maleic anhydride (MAH - g - PLA), resulting in decreased voids/pores on the cross - section of the composites. 15 Figure 2 SEM images of fracture surfaces of neat PLA and PLA - based composites at two different scale s of 50 μ m and 5 μ m: (a , g) neat PLA, (b , h) pine/PLA, (c , i) 0.5 epoxy/pine/PLA, (d , j) 1.0 epoxy/pine/PLA, (e , k) 2.0 epoxy/pine/PLA, and (f , l) 4.0 epoxy/pine/PL A. 16 Figure 3 The SEM images of fracture surfaces of (A) pine/PLA and (B) 1.0 epoxy/pine/PLA composites showing fiber - matrix interface cracking The DMA analysis of (C) temperature dependence of storage modulus, (D) storage modulus obtained at 25 °C and tan 𝛿 peak intensity , (E) temperature dependence of tan 𝛿 , (F) T g (E ́) and T g (tan 𝛿 ): (a) neat PLA, (b) pine/PLA, (c) 0.5 epoxy/pine/PLA, (d) 1.0 epoxy/pine/PLA, (e) 2.0 epoxy/pine/PLA, and (f) 4.0 epoxy/pine/PLA. 17 Figure 1 and Figure 2 show that pores and hollow channels run through the pine fiber s ’ structure Following impregnation , some of the se pores and hollow channels became filled with epoxy. Epoxy fill ing of these structures was verified via SEM prior to melt compounding because polymer (e.g., PLA ) molecules are also somewhat able to penetrate into the pores and hollow channels of natural fibers (e.g., poplar fibers), as shown in Figure S 5 and related literature 27 - 28 The partial penetration of epoxy molecules into the hollow channels of pine fibers mainly took place during the preparation of epoxy - modifie d pine fibers. The evaporation of solvent during the epoxy - pine mixture drying process and low content of epoxy in the epoxy - modified pine/PLA composites might cause the partial penetration Bakri et al. 29 has also reported epoxy being absorbed inside the hollow lumen structure of banana fibers, yielding improvements to the mechanical properties of banana fiber/epoxy composites Th e above results suggest that the physical attachment and/or coating of epoxy on pine fibers can be optimized in the future research to improv e the mechanical properties of epoxy - modified pine/PLA composites further DMA R esults Figure 3(C - F) shows the measured temperature dependence of the storage modulus and tan 𝛿 of neat PLA and PLA - based composites. The storage modulus (obtained at 25 ° C) of pine/PLA composites was 51% higher than that of neat PLA. This indicates an increase in the stiffness, which is consistent with the reinforcing effect of pine fibers and measured tensile moduli The addition of pine fibers might be caus ing a restriction in the segment mobility of the PLA chains Similarly, Saba et al. 30 found that cellulose nanofiber filler can increase the storage modulus of the polymer matrix attributable to the constrained movement of polymeric chains With fiber e poxy - modification , the composite storage modulus increased further up to 6% The pores and hollow channels of pine fibers were partial ly filled by the epoxy molecules during the process of preparing epoxy - modified pine fibers The epoxy 18 helped bonding individual fibers together, making them more resistant to debonding internally when under load in the transverse direction. The reinforcement effect of natural fibers in polymers is strongly affected by the structure of the natural fibers ; 6 the reduction of pores in pine fibers can help increase the composite integrity Figure 3(D) shows that storage modulus appeared to peak with t he 1.0 epoxy/pine/PLA composite and then falle n off at higher epoxy loadings, agreeing with the trend in stiffness response observed in tensile testing. The intensity of the t an 𝛿 peak of PLA - based composites was 48 – 57% lower than that of neat PLA. The decrease in intensity of the PLA - based compo sites t an 𝛿 peak indicates that the segmental chain motion of PLA was stunt ed by the pine fibers and/or epoxy during the transition. The t an 𝛿 second peak temperatures of neat PLA and PLA - based composites varied between 104 and 108 °C , shown in Table S 2 A sharp drop (approximately at 5 7 – 5 9 °C), attributed to relaxation of the PLA amorphous region, was observed on the storage modulus curves of n eat PLA and PLA - based composites. The low T g for neat PLA is likely because of the presence of flexible PLA polymer chains 30 When pine fibers and 1.0 wt% of epoxy w ere added, the T g (E ́) and T g ( t an 𝛿 ) were respectively 4% and 3% higher than those of neat PLA. The incorporation of pine fibers and epoxy to the PLA matrix might limit the PLA polymer chain mobility, which resulted in an increase of T g for the epoxy - modified pine/PLA composites. It is also possible that the T g changes were within the margin of error. DSC R esults Figure 4 ( A ) shows DSC second heat flow curves for neat PLA and PLA - based composites. The T g , T c , T m , χ c , D H C , and D H m of neat PLA and PLA - based composites are shown in Table 2 . With the addition of pine fibers into PLA, the T c and T m both decreased slightly The T g values of neat PLA and PLA - based composites were slightly different from those determined by DMA because of the different mechanisms in determining the T g in these two methods. When 1.0 wt% of epoxy was added, the 1.0 epoxy/pine/PLA composite had 19 higher T g , T c , and T m values than those of pine/PLA composite. The increase in T g could be attributed to the enhanced interfacial adhesion between the pine fibers and PLA because of the epoxy modification restricting the segmental chain movement of PLA molecules The other possible reason is that the changes in T g were within the margin of error. The T g of the composite s c an be affected by the compatibility of their components The crystallinity, χ c , of the pine/PLA composite was higher than that of ne at PLA. The increase in crystallinity is p ossi bly attributable to a nucleating effect from the pine fibers in the PLA polymer 31 20 Figure 4 The ( A ) DSC second heat flow curves , ( B ) TGA analysis , (C) FT - IR curves ( 4,000 – 600 cm - 1 ) , ( D ) FT - IR curves (1,400 – 600 cm - 1 ) , ( E ) FT - IR curves (4,000 – 2,600 cm - 1 ) , and ( F ) FT - IR curves (1,800 – 1,400 cm - 1 ) of : (a) neat PLA, (b) pine/PLA, (c) 0.5 epoxy/pine/PLA, (d) 1.0 epoxy/pine/PLA, (e) 2.0 epoxy/pine/PLA, and (f) 4.0 epoxy/pine/PLA.