materials

Article Chemical Treatment of Waste Abaca for Natural Fiber-Reinforced Geopolymer Composite

Roy Alvin J. Malenab 1, Janne Pauline S. Ngo 1 and Michael Angelo B. Promentilla 1,2,*

1 Chemical Engineering Department, De La Salle University, Manila 0922, Philippines; [email protected] (R.A.J.M.); [email protected] (J.P.S.N.) 2 National Research Council of the Philippines, Taguig, Metro Manila 1631, Philippines * Correspondence: [email protected]; Tel.: +63-2-536-0223

Academic Editor: Nicole Zander Received: 29 March 2017; Accepted: 23 May 2017; Published: 25 May 2017

Abstract: The use of natural fibers in reinforced composites to produce eco-friendly materials is gaining more attention due to their attractive features such as low cost, low density and good mechanical properties, among others. This work thus investigates the potential of waste abaca (Manila hemp) fiber as reinforcing agent in an inorganic aluminosilicate material known as geopolymer. In this study, the waste fibers were subjected to different chemical treatments to modify the surface characteristics and to improve the adhesion with the fly ash-based geopolymer matrix. Definitive screening design of experiment was used to investigate the effect of successive chemical treatment of the fiber on its tensile strength considering the following factors: (1) NaOH pretreatment; (2) soaking time in aluminum salt solution; and (3) final pH of the slurry. The results show that the abaca fiber without alkali pretreatment, soaked for 12 h in Al2(SO4)3 solution and adjusted to pH 6 exhibited the highest tensile strength among the treated fibers. Test results confirmed that the chemical treatment removes the lignin, pectin and hemicellulose, as well as makes the surface rougher with the deposition of aluminum compounds. This improves the interfacial bonding between geopolymer matrix and the abaca fiber, while the geopolymer protects the treated fiber from thermal degradation.

Keywords: waste utilization; abaca (Manila hemp) fiber; chemical treatment; fiber-reinforced composite; geopolymer; tensile strength; definitive screening design of experiment

1. Introduction The growing concerns for the environment and sustainability have served as a strong drive for researches on developing eco-friendly composite materials [1]. Composite materials in general are formed by combining two or more constituent materials, a continuous medium called matrix and the dispersed phase/s, either fiber/s or particulate/s, in order to develop another material with desired combination of properties [2]. In fiber-reinforced composite (FRC) materials, such matrix is reinforced by fibers for their superior properties and load transfer characteristics. Recently, natural fibers such as cotton, flax, jute, sisal and hemp are also becoming more attractive than synthetic fibers because of their natural abundance, low cost, low density, good mechanical properties, nontoxicity, etc. [1–4]. For example, abaca fibers have been widely studied as a reinforcement in cement [5], polypropylene [6,7] and epoxy [8] matrices, among others. Abaca fiber or Manila hemp, known for its commendable mechanical strength, has been produced and widely exported by the Philippines. Commercial grade abaca has density 1.5 g/cm3 and tensile strength of about 980 N/mm2 [9]. In 2015, the annual abaca fiber production was 70,400 metric tons supplying around 87% of world requirements [10]. These fibers are extracted from a native banana species and are harvested mostly for their use in making ropes, in textile, and more recently,

Materials 2017, 10, 579; doi:10.3390/ma10060579 www.mdpi.com/journal/materials Materials 2017, 10, 579 2 of 19 in automotive [11]. The extraction processes from plant to usable fibers produce considerable amount of waste or scrap fibers, which are still underutilized [12]. This study thus explores the utilization of scrap abaca fiber as reinforcement for fly ash-based geopolymer matrix. Geopolymer is an inorganic polymer formed from the reaction of materials rich in reactive alumina and silica with an alkaline solution [13,14]. This material has also attracted widespread scientific and industrial interest for its potential to valorize waste or industrial by-product such as coal ash while producing cementitious material with performance comparable to that of Portland cement in many applications [15]. Furthermore, it has several other advantages such as superior heat and fire resistance, high chemical resistance, and lower carbon footprint and embodied energy. Accordingly, fiber-reinforced geopolymer composite could potentially provide better mechanical and thermal properties over a wide temperature range as compared to Portland cement or organic polymer-based matrix. The use of natural fibers such as bamboo, cornhusk, wool, cotton, or flax in geopolymer matrix has been mentioned [16–18] and shown with promising results in the recent critical review on fiber-reinforced geopolymers [19,20]. Another study also demonstrates the surface modification of such fiber which improves its alkali resistance and results to an improved flexural strength of a reinforced metakaolin-based geopolymer as compared to that of pristine geopolymer [21]. In contrast, one recent exploratory study [22] reported no significant differences in the flexural strength between that of pristine fly ash based-geopolymer and that of geopolymer reinforced with the untreated coconut coir, cotton, or sisal fibers. This suggests that surface modification of the natural fibers is essential to improve the performance of the reinforced geopolymer composites. To our knowledge, no studies have been reported yet on a chemical treatment of abaca fiber used for reinforcing fly ash-based geopolymer. Surface modification of abaca fibers could improve the interfacial adhesion with such geopolymer matrix. In general, the matrix and fibers retain their physical and chemical identities in FRCs, forming boundaries or interfaces [23]. Effectiveness of fiber reinforcement in composites depends on the degree or mechanism of bonding at the fiber–matrix interface and the ability to transfer stress from matrix to the fiber [24]. Stress transfer among the components should be effective in order to attain desired strength. The developed interface between the fiber and the matrix is thus a crucial factor of the composite’s static and dynamic stability [2]. This interface enables load transfer between the components. When the fiber is brought in contact with the matrix, the chemical groups on the fiber surface can bond with those present on the matrix. Electrostatic, hydrogen bonds and dipole-dipole interaction can develop depending on the nature and compatibility of the fiber and matrix. This is the basis for the fiber’s adhesion to the matrix. Surface treatment, which can modify or introduce chemical groups that could aid in bonding, is often applied to render the fiber compatible with the matrix. Meanwhile, the presence of hydroxyl and carboxylic groups in natural cellulosic fibers’ structure make them intrinsically hydrophilic and polar. When such fibers are used as reinforcement in hydrophobic matrices, like polymers, the resulting composite is a heterogeneous system that could have inferior properties due to lack of adhesion and chemical affinity. Some of the organic components of natural fibers can inhibit its compatibility with the targeted organic or inorganic matrix. Since the interface quality and bondage between matrix and fiber is insufficient, surface modification of these fibers through chemical treatment and other techniques have been pursued in research for years such as alkali, chemical coupling, oxidation, plasma, ultrasound, and enzyme treatments [23,25]. The most common treatment for natural fibers discussed in literature is alkali treatment, where the fiber is soaked in sodium hydroxide (NaOH) solution in order to strip off some of the unwanted waxes, hemicellulose and lignin. Due to the hydrophilic nature of some of the fiber components and thus, tendency for water absorption, natural fibers may exhibit weak interfacial bonding with hydrophobic matrices [26,27]. Furthermore, “swelling” may occur when a hydrophilic fiber within a hydrophobic matrix absorbs moisture within the matrix, making it susceptible to weakening and deterioration [28]. This may occur in the event that water absorption due to the fiber breaks hydrogen bonds between the fiber and matrix or among the fibers themselves. In addition, alkali treatment induces rougher surfaces for better adhesion between the fiber and matrix. Materials 2017, 10, 579 3 of 19

Materialsfiber and2017 matrix., 10, 579 Prior studies have also shown that alkali treatment of these fibers have improved3 of 19 the performance of reinforced composites in terms of thermal stability [29], water absorption and tensile strength [30,31]. Prior studies have also shown that alkali treatment of these fibers have improved the performance of Degradation of such natural fibers when used in cementitious or geopolymer matrix due to the reinforced composites in terms of thermal stability [29], water absorption and tensile strength [30,31]. harsh environment is another concern. The decay of plant fibers when used as reinforcement in Degradation of such natural fibers when used in cementitious or geopolymer matrix due to the alkaline matrices such as that of concrete has been observed [32]. Therefore, plant-based fibers must harsh environment is another concern. The decay of plant fibers when used as reinforcement in alkaline be treated not only to be compatible to its matrix but also to avoid or delay its degradation ensuring matrices such as that of concrete has been observed [32]. Therefore, plant-based fibers must be treated that its reinforcing function is sustainable. Aluminum sulfate treatment of natural fibers is a potential not only to be compatible to its matrix but also to avoid or delay its degradation ensuring that its technique to address the possible deterioration of plant-based fibers exposed to cementitious matrix. reinforcing function is sustainable. Aluminum sulfate treatment of natural fibers is a potential technique This technique has been employed for pulped wood fibers to reinforce cementitious matrices [33]. to address the possible deterioration of plant-based fibers exposed to cementitious matrix. This technique The treatment involves deposition of insoluble aluminum hydroxide on the fiber surface via has been employed for pulped wood fibers to reinforce cementitious matrices [33]. The treatment neutralization process. These ionic deposits are expected to have better affinity with the inorganic or involves deposition of insoluble aluminum hydroxide on the fiber surface via neutralization process. cementitious matrix when compared to the affinity of organic, hydrophilic nature of cellulosic fibers These ionic deposits are expected to have better affinity with the inorganic or cementitious matrix to such matrices. when compared to the affinity of organic, hydrophilic nature of cellulosic fibers to such matrices. In the present work, we investigate the effect of successive chemical treatment via NaOH In the present work, we investigate the effect of successive chemical treatment via NaOH pretreatment followed by an aluminum sulfate treatment of waste abaca fibers on their tensile pretreatment followed by an aluminum sulfate treatment of waste abaca fibers on their tensile strength strength and morphology. Fourier transform infrared spectroscopy (FTIR), scanning electron and morphology. Fourier transform infrared spectroscopy (FTIR), scanning electron microscopy with microscopy with energy-dispersive X-ray spectroscopy (SEM-EDX), X-ray diffraction (XRD), and energy-dispersive X-ray spectroscopy (SEM-EDX), X-ray diffraction (XRD), and thermo-gravimetric thermo-gravimetric analysis (TGA) were used to analyze the structural and chemical changes in the analysis (TGA) were used to analyze the structural and chemical changes in the fibers after chemical fibers after chemical treatment. Preliminary characterization of the fiber-reinforced geopolymer using treatment. Preliminary characterization of the fiber-reinforced geopolymer using SEM-EDX and TGA SEM-EDX and TGA combined with mechanical tests provides an impetus for future work on the combined with mechanical tests provides an impetus for future work on the development of chemically development of chemically treated abaca fiber-reinforced geopolymer composites. treated abaca fiber-reinforced geopolymer composites.

2. Materials and Method

2.1. Sources of Raw Materials Abaca fibers fibers are extracted by hand stripping and are subsequentlysubsequently segregated into standard grades basedbased onon color color and and texture. texture. Residual Residual grade grade abaca abaca fibers fibers shown shown in Figure in Figure1a were 1a were collected collected from Dumaguetefrom Dumaguete City, Philippines. City, Philippines. The average The diameteraverage diameter of the collected of the abaca collected strands abaca is 162 strands micrometers. is 162 Themicrometers. fibers were The cut fibers into approximately were cut into 25approximat cm long andely 25 segregated cm long toand 100 segreg strandsated per to sample 100 strands as shown per insample Figure as1 b.shown The flyin Figure ash used 1b. to The produce fly ash geopolymer used to produce matrix geopolymer was obtained matrix from was a coal obtained fired powerfrom a plantcoal fired situated power at theplant central situated region at the of Luzon,central Philippines.region of Luzon, X-ray Philippines. fluorescence X-ray (XRF, fluorescence Thermo Scientific, (XRF, Waltham,Thermo Scientific, MA, USA) Waltham, analysis ofMA, the USA) fly ash analysis used is of summarized the fly ash in used Table is1 .summarized Chemicals used in Table to treat 1. abacaChemicals fibers, used sodium to treat hydroxide abaca fibers, micropearls sodium hydrox (99 wtide %) micropearls and aluminum (99 wt sulfate %) and powders aluminum (99 sulfate wt %), 2 2 andpowders the water (99 wt glass %), solutionand the (moduluswater glass = 2.5;solution 34.1 wt (modulus % SiO2, 14.7= 2.5; wt 34.1 % Nawt2 %O) SiO for geopolymer, 14.7 wt % Na synthesisO) for weregeopolymer supplied synthesis by a local were company supplied (Just-in-One by a local Marketing)company (Just-in-One and were used Marketing) as received. and were used as received. Table 1. X-ray fluorescence (XRF) analysis of fly ash. Table 1. X-ray fluorescence (XRF) analysis of fly ash.

Oxide SiO2 Fe2O3 CaO Al2O3 MgO K2O Others Oxide SiO2 Fe2O3 CaO Al2O3 MgO K2O Others Mass % Mass33.9 % 33.926.5 26.513.6 13.6 13.513.5 7.9 7.9 1.2 1.2 3.4 3.4

Figure 1. WasteWaste abaca abaca from from Dumaguete, Dumaguete, Philippines: ( a) as as received; received; and ( (b) segregated segregated into 100 strands.

Materials 2017, 10, 579 4 of 19

2.2. Chemical Treatment of Waste Abaca Fiber Two types of chemical treatments of abaca fibers were used in this study: (1) alkali pretreatment; and (2) aluminum sulfate treatment, which involves deposition of aluminum hydroxide on the fiber surface via precipitation. Using SAS JMP 11.1.1 software, 10 experimental runs were prescribed using Definitive Screening Design (DSD). DSD works for factor screening containing a combination of continuous and two-level categorical factors. It can estimate unbiased main effects that are completely independent of two-factor interactions and avoid confounding of any pair of quadratic effects [34,35]. Table2 shows the factors, their levels and corresponding codes used in this design of experiment. There are three factors considered, one is a categorical factor (NaOH pretreatment) and the other two (soaking time and pH) are continuous. For the alkali pretreatment, fibers were soaked in 6% by weight NaOH solution for 48 h in a 500 mL beaker with constant agitation at room temperature. The said treatment can remove impurities of the fiber resulting to rough fiber surface [36,37]. Fibers were then removed from the NaOH solution and washed with distilled water several times. Alkali pretreated fibers were then air dried for 24 h. For the Al2(SO4)3 treatment, untreated and alkali pretreated fibers were soaked in 10% by weight of Al2(SO4)3 solution in different beakers. The initial pH of Al2(SO4)3 solution is 3.5. Final pH of the abaca fiber and aqueous Al2(SO4)3 mixture were varied based on the solubility and precipitation of Al(OH)x species as function of pH [38]. Soaking period was also varied to three levels, 30 min, 6.25 h and 12 h, to allow the penetration of ions into the fiber. Then, the pH was adjusted to a desired level (4.5, 6.0 or 7.5) by gradually adding 2 M NaOH solution. Subsequently, the fibers were left soaked in the mixture for another 24 h to allow further precipitation of Al(OH)3 on the fiber surface. Samples were then drawn from the mixture, washed with distilled water and air-dried for at least 24 h before being stored in sealed plastics. Figure2 illustrates the flow of the chemical treatment procedure and characterization performed for the treated fiber.

Table 2. Factors and levels used for the chemical treatment.

Factor/Level (Coded) −1 0 +1 NaOH pretreatment - No Yes Soaking time in Al2(SO4)3, hrs 0.5 6.25 12 Final pH 4.5 6.0 7.5

2.3. Preparation of Abaca Fiber-Reinforced Geopolymer Fly ash and abaca fibers (99:1 by weight) were dry-mixed to ensure uniform fiber distribution. For every 1 kg of fly ash–fiber mixture, 400 g of alkaline activator (well-mixed 80% water glass solution, 20% 12 M NaOH by weight) was added gradually while mixing using a laboratory motorized mixer. The mass ratio was chosen to achieve a geopolymer mix with considerable compressive strength and good workability [39]. Mixing was continued for about 15 min until the mixture appeared homogenous. About 30 mL of distilled water was added to attain the desired consistency. The mixture was then poured in 5 cm × 5 cm × 5 cm cube molds (for compressive strength test) and 5 cm × 5 cm × 18 cm molds (for flexural strength test), and placed on a shaker for 1 min to remove trapped bubbles. After 24 h, samples were demolded and cured in the oven at 75 ◦C for 24 h and then cured at ambient conditions for 28 days. A specimen of pristine geopolymer, i.e., with no abaca fiber reinforcement was also prepared to serve as the reference material. The specimens were then tested for compressive strength and flexural strength. Materials 2017, 10, 579 5 of 19 Materials 2017, 10, 579 5 of 19

FigureFigure 2. Chemical2. Chemical treatment treatment andand characterizationcharacterization of of waste waste abaca abaca fiber. fiber.

2.4. Tensile Strength Test of Abaca Fibers 2.4. Tensile Strength Test of Abaca Fibers Average tensile strengths of the abaca fibers (untreated and treated) were determined by taking Averagefive strands tensile from strengthseach bundle of of the 100 abaca strands fibers randomly. (untreated Each andstrand treated) was cut were to 5 determinedcm then both byends taking five strandswere mounted from each to cardboard bundle of(2 100cm × strands 2 cm) using randomly. quick-dry Each cyanoacrylate-based strand was cut toadhesive. 5 cm then Then, both the ends wereprepared mounted specimen to cardboard was (2held cm by× 2the cm) universal using quick-dry testing machine cyanoacrylate-based (Instron model adhesive. 3324, Instron, Then, the preparedNorwood, specimen MA, wasUSA) held with by gauge the universallength of 3 testingcm. The machine cardboard (Instron and the model hardened 3324, adhesive Instron, layer Norwood, in MA, USA)contact with with gauge the fiber length protect of 3 the cm. fiber The from cardboard possible and deformation the hardened due adhesiveto clamping layer pressure. in contact The with the fiberuniversal protect testing the fibermachine from (UTM) possible applies deformation tensile load due on the to clampingfiber strand pressure. equivalent The to universalan extension testing machinerate of (UTM) 5 mm appliesper min and tensile recording load onthe thestress fiber until strand it breaks. equivalent Prior to fracture, to an extension each strand rate was of placed 5 mm per under an optical microscope (Amscope microscope, AmScope, CA, USA) linked to a computer to min and recording the stress until it breaks. Prior to fracture, each strand was placed under an optical measure each of their diameters. Fiber strands were assumed to have approximately circular cross- microscopesectional (Amscope areas and microscope, were computed AmScope, accordingly. CA, USA) Tensile linked strength to a computerof the fiber to measureis defined each as the of their diameters.maximum Fiber stress strands it can were withstand, assumed and to is have computed approximately by dividing circular the load cross-sectional at failure by the areas original and were computedcross-sectional accordingly. area per Tensile fiber. strength The five ofreadings the fiber were is definedaveraged as to theobtain maximum the final stressvalue recorded. it can withstand, and is computed by dividing the load at failure by the original cross-sectional area per fiber. The five readings were averaged to obtain the final value recorded. Materials 2017, 10, 579 6 of 19

2.5. Scanning Electron Microscopy—Energy Dispersive Spectroscopy (SEM-EDS) Measurements 2.5. Scanning Electron Microscopy—Energy Dispersive Spectroscopy (SEM-EDS) Measurements Surface morphology and elemental microanalyses of samples were performed using Field-emission scanningSurface electron morphology microscope and (FESEM elemental Dual microanalysesBeam Helios Nanolab of samples 600i, FEI, were Hillsboro, performed OR, usingUSA) Field-emissionequipped with electron-dispersive scanning electron microscope X-ray spectroscopy (FESEM (EDS, Dual BeamFEI, Hillsboro, Helios Nanolab OR, USA). 600i, The FEI, accelerating Hillsboro, OR,voltage USA) and equipped beam current with for electron-dispersive SEM analysis are X-ray2.0 kV spectroscopyand 43 pA, respectively. (EDS, FEI, Hillsboro,For EDS analysis, OR, USA). the Theaccelerating accelerating voltage voltage and beam and beamcurrent current are 15.0 for kV SEM and analysis0.69 nA, respectively. are 2.0 kV and 43 pA, respectively. For EDS analysis, the accelerating voltage and beam current are 15.0 kV and 0.69 nA, respectively. 2.6. X-ray Diffraction (XRD) Measurements 2.6. X-ray Diffraction (XRD) Measurements XRD (Maxima XRD-7000 Shimadzu, Tokyo, Japan) analyses were carried out using X-ray beams generatedXRD (Maximafrom a Cu XRD-7000 Kα radiation Shimadzu, source Tokyo, (40 kV; Japan) 30 mA) analyses with were wavelength carried outof 1.54 using Å X-ray to scan beams the generatedsamples from from 3.00° a Cu to 70.00° Kα radiation 2-theta angles. source The (40 resolu kV; 30tion mA) for with this analysis wavelength was ofset1.54 at 0.020, Å to with scan scan the ◦ ◦ samplesspeed maintained from 3.00 atto 2° 70.00 per min.2-theta The angles. fiber’s Thecrystallinity resolution index for this(CI) analysiswas calculated was set by at 0.020,XRD peak with ◦ scandeconvolution speed maintained method [40]. at 2 Inper this min. method, The a fiber’s curve-fitting crystallinity program index (e.g., (CI PeakFit) was calculated v4.12) was by used XRD to peakidentify deconvolution and separate method the different [40]. In overlapping this method, di affraction curve-fitting spectrum program of both (e.g., the PeakFit crystalline v4.12) wasand usedamorphous to identify contributions, and separate and the then different the area overlapping was calculated diffraction under spectrumeach fitted of Gaussian both the peaks. crystalline It is andassumed amorphous that the contributions, observed peak and broadening then the area in th wase spectra calculated is mainly under eachattributed fitted to Gaussian the increased peaks. Itamorphous is assumed content that the in observed the sample. peak For broadening example, infour the crystalline spectra is peaks mainly (101, attributed 101, 002 to and the increased040) and amorphousbroad amorphous content peaks in the between sample. For18° example,to 22° were four identified crystalline [41] peaks and (101,assumed 101, 002as Gaussian and 040) andfunctions broad ◦ ◦ amorphousas shown in peaks Figure between 3. These 18 Gaussianto 22 were peaks identified were [extracted41] and assumed from the as GaussianXRD spectrum functions through as shown the incurve-fitting Figure3. These process Gaussian with at peaksleast 5000 were iterations, extracted which from thecorresponds XRD spectrum to a coefficient through theof determination curve-fitting 2 process(r2) of 0.98. with Based at least from 5000 the iterations, results of whichcurve fitting, corresponds CI is calculated to a coefficient as the of ratio determination of the sum (rof )areas of 0.98. of Basedall crystalline from the peaks results to of the curve sum fitting, of areasCI is calculatedof all peaks as the(crystalline ratio of the and sum amorphous) of areas of allas crystallineshown in peaksEquation to the (1) sum[40]. of areas of all peaks (crystalline and amorphous) as shown in Equation (1) [40].

A ∑ Acrystallinecrystalline CICI= (1)(1) ∑ Acrystalline + Aamorphous  AAcrystalline amorphous

Figure 3. X-ray diffraction (XRD) peak deconvolution method to measure crystallinity index.

2.7.2.7. Fourier Transform Infrared (FTIR)(FTIR) SpectroscopySpectroscopy MeasurementsMeasurements Attenuated total reflectance reflectance Fourier transform transform infrared infrared (ATR-FTIR) (ATR-FTIR) spectroscopy spectroscopy (Perkin (Perkin Elmer, Elmer, Waltham, MA, USA) was carriedcarried outout toto qualitativelyqualitatively identify the constituents of untreated and untreated abacaabaca fibers.fibers. Test Test results results were were obtained obtained using using Perkin Perkin Elmer Elmer FTIR FTIR Spectrometer Spectrometer Frontier Frontier with ATRwith accessoryATR accessory in the rangein the ofrange 4000–650 of 4000–650 cm−1. The cm ATR−1. The cell ATR is equipped cell is equipped with a trapezoidal with a trapezoidal diamond crystaldiamond as crystal the internal as the reflection internal element.reflection Baseline element. correction Baseline wascorrection applied was to theapplied spectrum to the to spectrum improve itsto improve quality without its quality distorting without the distorting band intensities the band in intensities the final spectrum.in the final spectrum.

Materials 2017, 10, 579 7 of 19 Materials 2017, 10, 579 7 of 19 2.8. Thermogravimetric Analysis (TGA) Measurements Thermal analysis was performed using a thermogravimetric analyzer (TA Instruments TGA 2.8. Thermogravimetric Analysis (TGA) Measurements Q50, New Castle, DE, USA). Samples were cut into small pieces (around 5 mm) of 20–25 mg and were placedThermal on a platinum analysis pan, was performedthen heated using from a 30 thermogravimetric to 900 °C at a rate analyzer of 10 °C (TA per Instrumentsmin under Argon TGA Q50, gas Newat 40 Castle,mL per DE,min USA).purge Samplesflow. Calculations were cut intoof percentage small pieces mass (around loss and 5 mm) derivative of 20–25 plots mg were and weredone ◦ ◦ placedusing the on TA a platinum Universal pan, Analysis then heated software. from 30 to 900 C at a rate of 10 C per min under Argon gas at 40 mL per min purge flow. Calculations of percentage mass loss and derivative plots were done using the2.9. TACompressive Universal Strength Analysis and software. Flexural Strength Test of Geopolymer Composites

2.9. CompressiveFor compressive Strength strength and Flexural testing, Strength cured cubic Test of (5 Geopolymer cm × 5 cm Composites × 5 cm) samples were subjected to a constant stress rate of 0.25 MPa/s until failure. The compressive strength (CS) of a sample was calculatedFor compressive using Equation strength (2): testing, cured cubic (5 cm × 5 cm × 5 cm) samples were subjected to a constant stress rate of 0.25 MPa/s until failure. The compressive strength (CS) of a sample was P calculated using Equation (2): CS  max (2) PA CS = max (2) A where Pmax is the total load on the sample at failure and A is the calculated area of the bearing surface whereof the specimen.Pmax is the total load on the sample at failure and A is the calculated area of the bearing surface of theFor specimen. flexural strengths or modulus of rupture (MOR) measurement, samples were subjected to a constantFor flexuralstress rate strengths of 0.25 MPa/s or modulus using a of three-point rupture (MOR flexure) measurement, loading until samplesfailure as were shown subjected in Figure to 4. a constantMOR was stress calculated rate of 0.25based MPa/s on the using load a at three-point failure, cross flexure sectional loading area until of failurethe sample as shown and distance in Figure of4. MORsupportswas using calculated Equation based (3): on the load at failure, cross sectional area of the sample and distance of supports using Equation (3): 3FL 3FL MORMOR= 2 (3)(3) 22bdbd2 wherewhere FFis is load load at at failure, failure,L is L the is distancethe distance between between support support or span, or bspan,is the b width is the and widthd is theand thickness d is the ofthickness the specimen of the beingspecimen tested. being tested.

Figure 4. Three-point flexural strength test. Figure 4. Three-point flexural strength test.

3. Results and Discussion 3. Results and Discussion

3.1. Tensile StrengthStrength ofof ChemicallyChemically TreatedTreated AbacaAbaca Fiber Fiber Table3 3summarizes summarizes the the measured measured tensile tensile strength strength of 10 of chemically 10 chemically treated treated waste waste abaca abaca fibers usingfibers theusing treatment the treatment combinations combinations according according to the designedto the designed experiment experiment from DSD from (see DSD Section (see Section2.2). Tensile 2.2). strengthTensile strength of treated of abaca treated fiber abaca in this fiber study in wasthis observedstudy was to observed range from to 210range to 450from MPa. 210 Comparedto 450 MPa. to theCompared average to tensile the average strength tensile of untreated strength abaca of untreate fiber (500d abaca MPa), fiber all treated (500 MPa), fibers all have treated tensile fibers strength have lowertensile than strength that oflower the untreated. than that Theof the sample untreated. with the The lowest sample tensile with strength the lowest (210 MPa)tensile was strength alkali (210 MPa) was alkali pre-treated, soaking in Al(OH)3 solution for 12 h and partially neutralized up to pre-treated, soaking in Al(OH)3 solution for 12 h and partially neutralized up to 4.5 pH while the sample4.5 pH while with thethe highestsample measuredwith the highest tensile measured strength (450 tensile MPa) strength did not (450 undergo MPa) alkalidid not pretreatment undergo alkali but

Materials 2017, 10, 579 8 of 19

pretreatment but was directly soaked in Al(OH)3 solution also for 12 h with final pH of 6. Thus, the Materials 2017, 10, 579 8 of 19 tensile strength of the fiber is observed to be significantly affected by the nature of chemical treatment. was directly soaked in Al(OH)3 solution also for 12 h with final pH of 6. Thus, the tensile strength of the fiber is observedTable to be 3. Measured significantly response affected (tensile by the strength) nature at of different chemical treatment treatment. conditions. NaOH Soaking Tensile Table 3. Measured response (tensile strength)Final at different pH treatment conditions. Pretreatment Time (h) Strength (MPa)

NaOH PretreatmentNo Soaking Time 12 (h) Final 6.0 pH Tensile 450 Strength ± 160 (MPa) No 0.5 7.5 430 ± 130 No 12 6.0 450 ± 160 NoNo 0.5 6.25 7.5 6.0 420 430 ± 140± 130 NoYes 6.25 12 6.0 7.5 370 420 ± ±82140 YesNo 12 0.5 7.5 4.5 340 370 ± 97± 82 NoYes 0.5 6.25 4.5 4.5 340 340 ± 140± 97 Yes 6.25 4.5 340 ± 140 Yes 6.25 6.0 300 ± 82 Yes 6.25 6.0 300 ± 82 YesYes 0.5 0.5 6.0 6.0 230 230 ± 82± 82 NoNo 6.25 6.25 7.5 7.5 220 220 ± 69± 69 YesYes 12 12 4.5 4.5 210 210 ± 83± 83

To quantifyTo quantify the effect,the effect, regression regressi analysison analysis was was conducted conducted to model to model the the response response variable variable (i.e., (i.e., the tensilethe tensile strength) strength) as a as function a function of these of these factors factor namelys namely the alkalithe alkali pretreatment, pretreatment, soaking soaking time time and and finalfinal pH. pH. Results Results from from the statisticalthe statistical analysis analysis are summarizedare summarized in Figure in Figure5. At 5. 5% At significance5% significance level, level, the significantthe significant main-effect main-effect factors factors to tensile to tensile strength streng wereth were the presencethe presence of NaOH of NaOH pretreatment pretreatment and and the finalthe final pH atpH which at which the Althe2(SO Al24(SO)3-treated4)3-treated samples samples were were neutralized neutralized (p-value (p-value = 0.0117 = 0.0117 and and 0.0169, 0.0169, respectively).respectively). NaOH NaOH pretreatment pretreatment and and higher higher final final pH werepH were generally generally observed observed to cause to cause a decline a decline in in tensiletensile strength. strength. The The soaking soaking time time of fiber of fiber samples samples in the in aluminum the aluminum salt solution salt solution was foundwas found to have to ahave relativelya relatively insignificant insignificant effect effect to the to tensile the tensile strength strength (p-value (p-value = 0.0806). = 0.0806). Nonetheless, Nonetheless, the interaction the interaction of the soakingof the soaking time with time the otherwith twothe other factors two was factors found towas affect found the tensileto affect strength the tensile most significantly strength most (p-valuesignificantly = 0.0109). (p-value Furthermore, = 0.0109). the Furthermore, model was ablethe model to explain was 98.7%able to of explain the observed 98.7% of variation the observed in tensilevariation strength in duetensile to thesestrength factors. due to these factors.

FigureFigure 5. Actual 5. Actual vs.Predicted vs. Predicted Plot Plot (---- (- 0.05- - - 0.05 significance significance curve; curve;---- - mean)- - -mean) for thefor regressionthe regression model model andand the modelthe model coefficients. coefficients.

FigureFigure6 shows 6 shows a matrix a matrix of interaction of interaction plots plots between between the factorsthe factors considered considered in the in chemicalthe chemical treatment.treatment. Each Each cell cell of of the the matrix matrix shows shows the the interaction interaction of of the the rowrow effecteffect withwith the column column effect. effect. For Forexample, thethe left-most-lowestleft-most-lowest cellcell represents represents the the interaction interaction of of pH pH (row) (row) and and NaOH NaOH pretreatment pretreatment (column)(column) effects. effects. A line A line segment segment is plotted is plotted for eachfor each level level of row of row effect effect using using response response values values (tensile (tensile strength)strength) predicted predicted by the by model. the model. As observed As observed in the interactionin the interaction profile, profile, abaca fiber abaca samples fiber thatsamples were that notwere pretreated not pretreated with NaOH with showed NaOH a muchshowed less a significant much less decline significant in tensile decline strength in tensile with increasingstrength with finalincreasing pH, whereas final alkali-treated pH, whereas fibers alkali-treated steeply declined fibers insteeply strength declined with increasing in strength final with pH. increasing This could final be explainedpH. This bycould the extendedbe explained contact by the with extended NaOH solution contact usedwith forNaOH neutralization, solution used resulting for neutralization, to further stripping of cellulosic components. It was found that when the Al2(SO4)3-treated abaca fibers were Materials 2017, 10, 579 9 of 19 resulting to further stripping of cellulosic components. It was found that when the Al2(SO4)3-treated abacaadjusted fibers to lowerwere finaladjusted pH (4.5),to lower the tensilefinal pH strength (4.5), improvedthe tensile withstrength increased improved soaking with time. increased When soakingadjusted time. to a pHWhen of 7.5, adjusted the tensile to a strengthpH of 7.5, decreased the tensile with strength increasing decreased soaking with time increasing in the aluminum soaking salt solution. Note that the main purpose of Al (SO ) treatment is to form deposits that could protect time in the aluminum salt solution. Note that2 the4 main3 purpose of Al2(SO4)3 treatment is to form depositsthe fiber fromthat could the harsh protect environment the fiber from of geopolymer the harsh matrix.environment It is possible of geopolymer that soaking matrix. in the It is aluminum possible thatsalt solutionsoaking in inherently the aluminum improves salt solution tensile strengthinherently but improves was counteracted tensile strength by the but addition was counteracted of NaOH bysolution the addition in adjusting of NaOH the solution final pH, in resulting adjusting instead the final to pH, a decline resulting in strengthinstead to when a decline excessive in strength alkali whensolution excessive is added alkali to attain solution a high is added final pH. to attain a high final pH.

Figure 6. InteractionInteraction plot plot matrix matrix of of the the factors factors fo forr chemical treatment of the abaca fibers.fibers.

3.2. Fourier Transform Infrared Spectroscopy (FTIR)(FTIR) ofof RawRaw andand Alkali-TreatedAlkali-TreatedAbaca AbacaFiber Fiber FTIR allows the qualitative analysis of variations in surface chemistry of natural fibers fibers after treatment. FTIR spectra of untreated, alkali-treated and NaOH-then-Al2(SO4)3 treated abaca fibers are treatment. FTIR spectra of untreated, alkali-treated and NaOH-then-Al2(SO4)3 treated abaca fibers shownare shown in Figure in Figure 7a–c.7a–c. In Inaddition, addition, Figure Figure 7d7d desc describesribes the the spectra of drieddried precipitateprecipitate oror residues residues from the spent solution after the Al2(SO4)3 treatment of abaca fibers. Note that the spectra in these from the spent solution after the Al2(SO4)3 treatment of abaca fibers. Note that the spectra in these figuresfigures were offset vertically to observe the differ differenceence more clearly. Table 4 4 shows shows thethe observedobserved bandsbands in FTIR spectra and their assignments to functional groups. In In general, general, untreated untreated abaca fiber fiber showed loss of functional groups based on the comparison of its FTIR spectrum to that of the NaOH-treated −1 −1 −1 fibers.fibers. Peaks thatthat recededreceded inin thethe treated treated fibers fibers over over frequencies frequencies 1510 1510 cm cm−1,, 1430 1430 cm cm−1 andand 12501250cm cm−1 could entail loss of phenolic groups, signifying removal of large amount of lignin and pectin whose −1 structure is aa phenolicphenolic polymer.polymer. AnotherAnother visiblevisible differencedifference waswas observedobserved atat 17401740 cmcm−1 due to C=OC=O stretching vibrationvibration which which is is a characteristica characteristic band band of hemicelluloseof hemicellulose [42]. [42]. The The disappearance disappearance of this of band this bandindicates indicates the dissolution the dissolution of hemicellulose of hemicellulose after the af alkaliter the pretreatment. alkali pretreatment. However However there is no there evidence is no evidenceindicating indicating complete complete removal ofremoval these impurities.of these impu Therities. decrease The decrease in tensile in strength tensile strength of the samples of the samples could be attributed to these extractions. In addition, characteristic peak of Al(OH)3 at 990 cm−−1 could be attributed to these extractions. In addition, characteristic peak of Al(OH)3 at 990 cm corresponding to Al–O bonds [43] was observed in the spectra of NaOH-then-Al2(SO4)3 treated abaca corresponding to Al–O bonds [43] was observed in the spectra of NaOH-then-Al2(SO4)3 treated abaca fiber,fiber, and the precipitateprecipitate or residuesresidues fromfrom thethe spentspent solution.solution. Thus, the FTIR results confirmconfirm the possible formation of Al(OH)3 during neutralization and its deposition on the abaca surface. possible formation of Al(OH)3 during neutralization and its deposition on the abaca surface.

Materials 2017, 10, 579 10 of 19 Materials 2017, 10, 579 10 of 19

Figure 7. Fourier Transform Infrared Spectroscopy (FTIR) spectra of: (a) untreated abaca; (b) NaOH- Figure 7. Fourier Transform Infrared Spectroscopy (FTIR) spectra of: (a) untreated abaca; treated abaca; (c) NaOH + Al2(SO4)3 treated abaca; and (d) precipitate or residue from the spent (b) NaOH-treated abaca; (c) NaOH + Al2(SO4)3 treated abaca; and (d) precipitate or residue from the solution of Al2(SO4)3 treatment. spent solution of Al2(SO4)3 treatment. Table 4. Functional Groups of Observed Bands in FTIR spectra of abaca fiber [25,42,43]. Table 4. Functional Groups of Observed Bands in FTIR spectra of abaca fiber [25,42,43]. Wavenumber (cm−1) Vibration Source Wavenumber3690 (cm−1) VibrationFree OH moisture Source 3350 O–H linked stretching Polysaccharide 3690 Free OH moisture 2920 C–H stretching Cellulose, Hemicellulose 3350 O–H linked stretching Polysaccharide 1740 C=O stretching Hemicellulose 2920 C–H stretching Cellulose, Hemicellulose 17401630 OH C=O in H stretching2O, bending Hemicellulose Moisture 1510 C=C aromatic symmetrical stretching Lignin 1630 OH in H2O, bending Moisture 15101430 C=C aromaticCH2 symmetric symmetrical bending stretching Pectin, Lignin lignin 14301250 CH2 C–Osymmetric aryl group bending Pectin, Lignin lignin 1250980 Al–O C–O arylstretching group Al(OH) Lignin3 980900 Glycosidic bonds Al–O symmetric stretching ring stretching Polysaccharide Al(OH)3 900 Glycosidic bonds symmetric ring stretching Polysaccharide 3.3. XRD Analysis of Untreated and Alkali-Treated Abaca Fibers 3.3. XRD Analysis of Untreated and Alkali-Treated Abaca Fibers Figure 8 shows the X-ray spectra or diffractograms of untreated and NaOH-treated abaca fibers. NoteFigure that8 shows the first the two X-ray crystalline spectra peaks or diffractograms cannot be observed of untreated separately and because NaOH-treated they overlapped. abaca fibers. NoteMoreover, that the firstdue to two the crystallinevery broad XRD peaks spectrum cannot of be the observed amorphous separately cellulose, becauseit overlapped they with overlapped. the first three crystalline peaks (101, 101, and 002). In this figure, one can notice the broader peak around Moreover, due to the very broad XRD spectrum of the amorphous cellulose, it overlapped with the the peak 002 of untreated fiber as compared with that of treated fiber. Accordingly, fitted Gaussian first three crystalline peaks (101, 101, and 002). In this figure, one can notice the broader peak around crystalline peaks were obtained using the XRD peak deconvolution method described in Section 2.6 the peakto quantify 002 of the untreated individual fiber contributions as compared of these with components that of treated to the overall fiber. Accordingly,XRD spectrum fittedof the fiber Gaussian crystallinesample. peaks This was were then obtained used to usingcompute the the XRD relative peak crystallinity deconvolution values method to interpret described the changes in Section in the 2.6 to quantifycellulose-based the individual structure contributions of the fiber after of these treatmen componentst. Results tofrom the the overall Gaussian XRD peak spectrum fitting showed of the fiber sample.the Thispresence was of then the usedfour major to compute crystalline the relativepeaks ofcrystallinity cellulose (101, values 101, 002 to interpretand 040) thewhich changes can be in the cellulose-basedobserved at the structure following of 2-theta the fiber angl afteres, 15.1°, treatment. 16.8°, 22.9° Results and 34.7°, from respectively. the Gaussian The peak broad fitting peak showedfor the presencethe amorphous of the cellulose four major was crystallineobserved at peaks2-theta ofof 20.2°. cellulose Thus, (101, the calculated 101, 002 andcrystallinity 040) which indices can be observed(CI) from at the the following XRD peak 2-theta deconvolution angles, 15.1 method◦, 16.8 using◦, 22.9 Equation◦ and 34.7 (1)◦ ,for respectively. the untreated The and broad NaOH- peak for treated abaca fibers were found to be 61.0% and 78.8%, respectively. It is anticipated that after the amorphous cellulose was observed at 2-theta of 20.2◦. Thus, the calculated crystallinity indices (CI) sufficient alkali pretreatment, the fiber crystallinity index is bound to increase due to the extraction from the XRD peak deconvolution method using Equation (1) for the untreated and NaOH-treated or removal of its amorphous components including hemicelluloses, lignin and other non-cellulosic abacacomponents fibers were which found allow to be the 61.0% cellulosic and fibers 78.8%, to respectively. adopt a more Itcrystalline is anticipated structure that [31]. after The sufficient observedalkali pretreatment,increase in thecrystallinity fiber crystallinity due to the removal index is of bound lignin and toincrease hemicellulose due tois also the in extraction agreement or with removal the of its amorphousresults of FTIR components analysis. including hemicelluloses, lignin and other non-cellulosic components which allow the cellulosic fibers to adopt a more crystalline structure [31]. The observed increase in crystallinity due to the removal of lignin and hemicellulose is also in agreement with the results of FTIR analysis. Materials 2017, 10, 579 11 of 19 Materials 2017, 10, 579 11 of 19

(a) (b)

Figure 8. XRD spectra of untreated (a) and NaOH-treated (b) abaca fibers. Figure 8. XRD spectra of untreated (a) and NaOH-treated (b) abaca fibers. 3.4. Morphology and Elemental Microanalysis of Untreated and Treated Abaca Fiber Surface through SEM-EDS 3.4. Morphology and Elemental Microanalysis of Untreated and Treated Abaca Fiber Surface through SEM-EDSEffect of alkali pretreatment to the morphology and chemistry of fiber surface is important in the development fiber–matrix interaction. SEM-EDS can also provide evidences of the surface modificationEffect of alkali in terms pretreatment of morphological to the morphology and elemen andtal microanalyses. chemistry of Based fiber surfaceon the results is important of FTIR in the developmentanalysis, there fiber–matrix are fiber components interaction. from SEM-EDS the untrea canted also fiber provide that were evidences removed of theafter surface its contact modification with in termsNaOH of solution, morphological thus it is and expect elementaled to observe microanalyses. difference with Based thei onr morphology. the results ofFigure FTIR 9 shows analysis, the there are fiberSEM components images of raw from and thetreated untreated abaca fibers. fiber thatFigure were 9a removedshows impurities after its on contact the textured with NaOH surface solution, of thusuntreated it is expected abaca to observefiber. After difference treatment with with their Na morphology.OH solution, Figureimpurities9 shows were the almost SEM imagesremoved of raw resulting to a more uniform but rougher or corrugated surface as shown in Figure 9b. This may be a and treated abaca fibers. Figure9a shows impurities on the textured surface of untreated abaca fiber. result of stripping off the lignin and hemicellulose from fibers in alkali solution as indicated in the AfterFTIR treatment analysis. with It is NaOHevident solution,in the images impurities that corrugation were almost on the removed fiber surface resulting became to sharper a more after uniform but roughertreatment. or This corrugated was observed surface among as shown all the in five Figure treated9b. samples This may that be underwent a result of SEM stripping analysis. off the ligninCorrugation and hemicellulose on the fiber from surface fibers indicates in alkali a solution better fiber–matrix as indicated interface in the FTIR by virtue analysis. of mechanical It is evident in the imagesinterlocking. that corrugationSuch corrugation on the in fiber the surface indicates became sharperhigher frictional after treatment. bond between This was the observedtwo amongcomponents, all the five which treated has previously samples that shown underwent to be effective SEM in analysis. resisting shear Corrugation and bending on thestress fiber [44]. surface indicatesIn a betterFigure fiber–matrix9c, SEM image interface of a treated by virtueabaca fiber of mechanical (no alkali pretreatment, interlocking. 12 Suchh soaking; corrugation final pH in the surface= 6) indicates indicates higherthe Al(OH) frictional3 deposits bond which between almost the covered two components, the fiber surface. which At higher has previously magnification shown to be effective(5000×; Figure in resisting 9c1), the shear Al(OH) and3 bendingparticles were stress observed [44]. to have porous structure. Figure 9d1 shows the SEM image of another treated fiber (NaOH pretreated, 12 h soaking; final pH = 6). Similar to the In Figure9c, SEM image of a treated abaca fiber (no alkali pretreatment, 12 h soaking; final pH = 6) previous image, it can be observed that the fiber is almost completely covered with Al(OH)3 particles. × indicatesAt higher the Al(OH) magnification3 deposits (5000×; which Figure almost 9d2), covered porous thestructure fiber surface.is still present At higher with magnificationadditional crystal (5000 ; Figurelike9c1), features. the Al(OH) However,3 particles in Figure were 9e,f, observed coarse and to havefine Al(OH) porous3 structure.particles having Figure porous9d1 shows and non- the SEM imageporous of another structures treated were fiberobserved (NaOH on the pretreated, treated abaca 12 fiber h soaking; covering final partially pH = the 6). surface. Similar This to the implies previous image,that it canvarious be observed forms of thatAl(OH) the3 fiberwere is deposited almost completely on the surface covered of abaca with Al(OH)fibers that3 particles. were treated At higher magnificationdifferently. (5000 ×; Figure9d2), porous structure is still present with additional crystal like features. EDS analysis was also performed on the surface of alkali treated abaca surface as shown in However, in Figure9e,f, coarse and fine Al(OH) 3 particles having porous and non-porous structures wereFigure observed 10. A on region the treated of interest abaca was fiber selected covering to meas partiallyure the elemental the surface. composition This implies in the that said various surface. forms However, it should be noted that the quantification error could be large as sample volume for EDS is of Al(OH)3 were deposited on the surface of abaca fibers that were treated differently. small in an order of few micrometers in diameter. Nevertheless, the results of EDS is indicative of the EDS analysis was also performed on the surface of alkali treated abaca surface as shown in presence of compound with known chemical structure. For example, in Figure 10a, the measured Figurecarbon 10. A to region oxygen of mass interest ratio was (C/O) selected is around to measure 1.3 whereas the elemental theoretically composition (based on in the the chemical said surface. However,formula) it shouldthe C/Obe of notedcellulose that is 0.9 the and quantification C/O of lignin error is 2.9 could[45,46]. be This large is indicative as sample that volume the lignin for EDS is smalland inhemicellulose an order of fewwere micrometersalmost removed in diameter. from abaca Nevertheless, fiber after alkali the resultstreatment. of EDS Moreover, is indicative the of the presencecomputed of O/Al compound values in with regions known such chemical as that of structure.Figure 10b Forrange example, from 2.8 into Figure3.2. This 10 isa, indicative the measured carbonof the to oxygenpresence mass of aluminum ratio (C/O) compounds is around in 1.3the whereasform of aluminum theoretically hydroxide (based which on the was chemical confirmed formula) the C/Ofrom ofthe celluloseFTIR results. is 0.9Note and that C/O the said of ligninO/Al calc isulation 2.9 [45 is,46 based]. This on the is indicativeassumption that that theall oxygen lignin and hemicelluloseatoms detected were are almost from removedthe organic from components abaca fiber of abaca after fiber alkali and treatment. aluminum Moreover, hydroxide the deposits computed O/Al values in regions such as that of Figure 10b range from 2.8 to 3.2. This is indicative of the presence of aluminum compounds in the form of aluminum hydroxide which was confirmed from the FTIR results. Note that the said O/Al calculation is based on the assumption that all oxygen atoms detected Materials 2017, 10, 579 12 of 19

Materials 2017, 10, 579 12 of 19 are from the organic components of abaca fiber and aluminum hydroxide deposits only, and that the only, and that the carbon to oxygen mass ratio (C/O) of fiber is 1.3 to determine the oxygen mass from carbon to oxygen mass ratio (C/O) of fiber is 1.3 to determine the oxygen mass from the organic. the organic.

500x 100μm 500x 100μm 500x 100μm

impurities

a1 b1 c1 5000x 10μm 5000x 10μm 5000x 10μm Porous Al(OH)3

a2 b2 c2 500x 100μm 500x 100μm 500x 50μm

coarse particles

fine particles

d1 e1 f1 5000x 10μm fine particles

10,000x 1 μm 1 μm 10,000x d2 e2 f2

Figure 9. Scanning electron microscopy (SEM) images at low and high magnification of raw Figure 9. Scanning electron microscopy (SEM) images at low and high magnification of raw (untreated), (untreated), NaOH-treated and Al2(SO4)3-treated abaca fibers: (a) raw (untreated); (b) NaOH treated; a b c NaOH-treated(c) without NaOH and Al pretreatment;2(SO4)3-treated 12 h abaca soaking; fibers: final ( pH) raw = 6; (untreated); (d) with NaOH ( ) NaOH pretreatment; treated; 12 ( )h; without NaOHpH = pretreatment; 6; (e) without12 NaOH h soaking; pretreatment; final pH 12 = h; 6; pH (d) with= 4.5; NaOHand (f)pretreatment; with NaOH pretreatment; 12 h; pH = 6; 0.5 (e )h, without NaOHpH = pretreatment;6. Low and high 12 magnification h; pH = 4.5; are and labeled (f) with as “1” NaOH and “2”, pretreatment; respectively. 0.5 h, pH = 6. Low and high magnification are labeled as “1” and “2”, respectively.

Materials 2017, 10, 579 13 of 19 Materials 2017, 10, 579 13 of 19

Figure 10. Sample EDS spectra of: (a) NaOH-treated; and (b) Al2(SO4)3-treated abaca Figure 10. Sample EDS spectra of: (a) NaOH-treated; and (b) Al2(SO4)3-treated abaca. 3.4. Thermogravimetric Analysis of Raw and Treated Abaca Fiber 3.5. Thermogravimetric Analysis of Raw and Treated Abaca Fiber Thermal stability of chemically treated abaca fibers was determined using thermogravimetric Thermalanalysis (TGA). stability In ofTGA, chemically thermal treatedstability abacawas studied fibers wasin terms determined of weight using losses thermogravimetric under argon analysisatmosphere (TGA). with In TGA,respect thermal to temperature stability ramping was studied(10°/min) infrom terms 30 °C of to weight900 °C. Note losses that under natural argon atmospherefibers are with mainly respect composed to temperature of hemicellulose, ramping cellulose (10◦/min) and fromlignin. 30 Thermal◦C to 900 degradation◦C. Note of that these natural fiberscomponents are mainly was composed reported to of take hemicellulose, place in different cellulose temperature andlignin. ranges [25,47 Thermal]. The degradationobserved thermal of these decompositions are summarized in Table 5. The value of Tmax represents the temperature at which components was reported to take place in different temperature ranges [25,47]. The observed thermal the maximum decomposition rate was observed based from the differential thermogravimetry (DTG) decompositionsdata. are summarized in Table5. The value of T max represents the temperature at which the maximum decomposition rate was observed based from the differential thermogravimetry (DTG) data. Table 5. Decomposition Temperature of Fiber Components [25,47] Table 5. Decomposition Temperature of Fiber Components [25,47]. Components Temperature of Decomposition, °C Tmax Based on DTG, °C Moisture 30–100 ◦ 80 ◦ Components Temperature of Decomposition, CTmax Based on DTG, C Hemicellulose (Xylan) 160–350 245 (side chain), 298 (backbone) MoistureCellulose 240–365 30–100 335 80 Hemicellulose (Xylan) 160–350 245 (side chain), 298 (backbone) Lignin 300–500 337 Cellulose 240–365 335 Lignin 300–500 337 Thermograms of abaca samples treated differently in Figure 11 show the Tmax, percentage mass loss and rate of decomposition at Tmax for each stage of thermal decomposition. For the four abaca Thermogramssamples, the initial of abacamass loss samples ranging treated from 7.8 differently to 10.3 percentage in Figure mass 11 show loss from the 50 Tmax °C ,to percentage around 120 mass loss and°C is rate mainly of decomposition due to the evaporation at Tmax offor free each an staged absorbed of thermal moisture decomposition. on the fiber. Considering For the four the abaca samples,untreated the initial abaca fiber, mass the loss main ranging decomposition from 7.8 event to 10.3 was percentage observed from mass 200 loss °C to from 370 °C 50 equivalent◦C to around 120 ◦Cto ispercentage mainlydue mass to loss the of evaporation 78.2%. At this of temperat free andure absorbed range, a shoulder moisture before on thethe fiber.peak located Considering at 300 the °C and a peak was observed at 345 °C (DTG). Fastest rate of decomposition with respect to untreated abaca fiber, the main decomposition event was observed from 200 ◦C to 370 ◦C equivalent to temperature was recorded at 1.1 percentage mass loss per °C. From the onset of the main thermal percentage mass loss of 78.2%. At this temperature range, a shoulder before the peak located at 300 ◦C decomposition to the location◦ of shoulder, the mass loss is mainly due to the decomposition of and ahemicellulose. peak was observed As the attemperature 345 C (DTG). increases, Fastest mo ratere cellulose of decomposition decomposed with and respect lignin tostarted temperature its ◦ was recordeddegradation. at 1.1 Beyond percentage 370 °C, mass slower loss decomposition per C. From of the the onset remaining of the components main thermal proceeded decomposition up to to the location900 °C resulting of shoulder, to a total the mass mass loss loss of is87.1% mainly or re duesidual to mass the decompositionof 12.9%. On one ofhand, hemicellulose. thermogram of As the ◦ temperatureabaca fiber increases, treated with more Al cellulose2(SO4)3 only decomposed is shown in andFigure lignin 11b. The started features its degradation. of this plot are Beyond similar 370to C, slowerthat decomposition of untreated abaca of the except remaining for the less components obvious shoulder proceeded in DTG up plot to 900 and◦ Clower resulting percentage to a totalmass mass loss in the main decomposition event, and thus leaving more residues. These differences can be loss of 87.1% or residual mass of 12.9%. On one hand, thermogram of abaca fiber treated with Al2(SO4)3 only isassociated shown into Figurethe partial 11b. removal The features of hemicellulose of this plot and are lignin similar during to that the ofneutralization untreated abaca process except by for adding NaOH solution. Figure 11c,d show thermograms of abaca fiber treated with NaOH, and the less obvious shoulder in DTG plot and lower percentage mass loss in the main decomposition NaOH then with Al2(SO4)3, respectively. It can be observed that the shoulder feature is no longer event, and thus leaving more residues. These differences can be associated to the partial removal of present in the DTG plots of these samples. Since these samples were exposed to a higher hemicelluloseconcentration and NaOH lignin solution during for the a neutralizationlonger period, more process hemicellulose by adding particles NaOH were solution. removed Figure from 11 c,d showthe thermograms fiber during the of abacachemical fiber treatment. treated with NaOH, and NaOH then with Al2(SO4)3, respectively. It can be observed that the shoulder feature is no longer present in the DTG plots of these samples. Since these samples were exposed to a higher concentration NaOH solution for a longer period, more hemicellulose particles were removed from the fiber during the chemical treatment. Materials 2017, 10, 579 14 of 19 Materials 2017, 10, 579 14 of 19

◦ ThermalThermal decomposition decomposition of of Al(OH) Al(OH)3 3as as reportedreported from a previous study study [48] [48 ]starts starts at at about about 240 240 °C C upup to to about about 400 400 ◦°CC overlapping overlapping with with the the thermal thermal decomposition decomposition of the of organic the organic components components of abaca of abacafibers. fibers. Furthermore, Furthermore, due to due the tosmall the amount small amount of Al(OH) of3Al(OH) deposited3 deposited on abaca onsurface, abaca the surface, effect of the Al(OH)3 thermal decomposition to the overall thermograms of Al2(SO4)3 treated fibers was not effect of Al(OH)3 thermal decomposition to the overall thermograms of Al2(SO4)3 treated fibers was notobvious. obvious.

Figure 11. Thermograms of: (a) untreated; (b) treated solely with Al (SO ) ;(c) alkali-pretreated; and Figure 11. Thermograms of: (a) untreated; (b) treated solely with Al2(SO44)3;3 (c) alkali-pretreated; and (d) alkali and Al (SO ) solution-treated abaca fiber samples. (d) alkali and Al2 2(SO4 43)3 solution-treated abaca fiber samples

3.6.3.4. Preliminary Preliminary Characterization Characterization of of Fiber-Reinforced Fiber-Reinforced GeopolymersGeopolymers IncorporatingIncorporating abaca abaca fiber fiber (about (about 1%1% byby weightweight of flyfly ash) in in geopolymer geopolymer matrix matrix improved improved the the compressivecompressive strength strength of of untreated untreated fiber-reinforced fiber-reinforced (25.9(25.9 MPa) and and treated treated fiber-reinforced fiber-reinforced composite composite (22.2(22.2 MPa) MPa) by by 20% 20% and and 3%, 3%, respectively, respectively, comparedcompared to that that of of a a pristine pristine geopolymer geopolymer (21.6 (21.6 MPa). MPa). Likewise,Likewise, the the flexural flexural strength strength of of the the untreateduntreated fiberfiber reinforced (5.5 (5.5 MPa) MPa) and and the the fiber-reinforced fiber-reinforced compositecomposite (7.3 (7.3 MPa) MPa) was was also also improved improved by by 95%95% andand 161%, respectively, respectively, compared compared to to that that of of a pristine a pristine geopolymergeopolymer (2.8 (2.8 MPa). MPa). Such Such results results indicate indicate thatthat thethe fiber’sfiber’s main main contribution contribution is is to to improve improve primarily primarily thethe composite’s composite’s flexural flexural strength strength and and notnot muchmuch ofof itsits compressive strength strength since since the the fibers fibers control control the cracking that gives rise to a “graceful” fracture by bridging across the cracks. Figure 12 shows the the cracking that gives rise to a “graceful” fracture by bridging across the cracks. Figure 12 shows fracture surface of the pristine geopolymer and those of the abaca fiber-reinforced. Moreover, a the fracture surface of the pristine geopolymer and those of the abaca fiber-reinforced. Moreover, significant improvement in flexural strength of the composite reinforced with chemically treated a significant improvement in flexural strength of the composite reinforced with chemically treated abaca fiber suggests a better interfacial adhesion between the fiber and the matrix. However, the abaca fiber suggests a better interfacial adhesion between the fiber and the matrix. However, the difference of improvement between the mechanical properties of geopolymer reinforced with differenceuntreated of fiber improvement and treated between fiber still the needs mechanical further properties verification of in geopolymer future works. reinforced with untreated fiber and treated fiber still needs further verification in future works.

Materials 2017, 10, 579 15 of 19 Materials 2017, 10, 579 15 of 19

FigureFigure 12. 12. FractureFracture surface surface of of pristine pristine geopolymer, geopolymer, reinforced withwithuntreated untreated abaca abaca and and reinforced reinforced withwith treated treated abaca. abaca.

In this paper, we focus on the microstructure characterization of the abaca-fiber reinforced In this paper, we focus on the microstructure characterization of the abaca-fiber reinforced geopolymer composites through SEM and thermogravimetric analysis. Representative fractured geopolymer composites through SEM and thermogravimetric analysis. Representative fractured surfacessurfaces of ofthe the composites composites were were analyzed analyzed using SEM-EDSSEM-EDS asas shownshown in in Figure Figure 13 13.. Heterogeneous Heterogeneous geopolymergeopolymer of of different different morphology such such as as textured textur spheres,ed spheres, porous porous and crystal-likeand crystal-like structures structures were wereobserved observed for bothfor compositesboth composites samples. samples. Results ofResults EDS indicatedof EDS thatindicated Si and that Al in Si the and geopolymer Al in the geopolymerhave an overall have atomic an over ratioall ofatomic 2:1. Fibers ratio wereof 2:1. clearly Fibers embedded were clearly on the embedded geopolymer on matrix the geopolymer as shown matrixfrom as the shown micrographs. from the Formicrographs. composites For reinforced composit withes reinforced untreated with abaca untreated (Figure 13 abacaa,c), visible(Figure gaps 13a,c), visiblewere observedgaps were between observed the geopolymerbetween the matrix geopolymer and fiber matrix indicating andpoor fiber interfacial indicating adhesion. poor interfacial This is adhesion.likely due This to theis likely incompatibility due to the ofincompatibility the fiber and matrix. of the Circular-likefiber and matrix. particles Circular-like were also observedparticles were on alsothe observed surface, whichon the are surface, likely which to be geopolymer are likely to precursors be geopolymer or nuclei precursors that adhered or nuclei on the that abaca adhered surface. on theOn abaca the othersurface. hand, Onfor the the other composites hand, for reinforced the composites with treated reinforced abaca with fibers treated (Figure abaca 13b,d), fibers narrower (Figure 13b,d),gaps werenarrower observed gaps between were observed the matrix between and treated the ma fiber.trix The and zeolite-like treated fiber. particles The onzeolite-like the fiber surfaceparticles onindicate the fiber that surface reaction indicate took place that on reaction the surface took for place these on structures the surface to form. for Clusteringthese structures of abaca to fibers form. Clusteringwas also observedof abaca infibers other was region also of theobserved fractured in surface. other region Although of geopolymerthe fractured and surface. zeolite depositsAlthough geopolymerwere observed and which zeolite indicates deposits the interactionwere observed between which the geopolymer indicates matrixthe interaction and fiber surface,between pull the geopolymerout sites were matrix observed and fiber suggesting surface, the pull formed out sites interfacial were observed adhesion wassuggesting not enough the formed in some interfacial regions. adhesionNevertheless, was not the enough pulling outin some of these regions. fibers duringNeverthe theless, loading the processpulling absorbsout of these energy fibers that resultsduring to the loadingthe improvement process absorbs of the energy flexural that strength results of to composite. the improvement of the flexural strength of composite. After compressive strength testing, treated abaca fibers from the fractured composites were After compressive strength testing, treated abaca fibers from the fractured composites were removed from the composite and subjected to TG analysis using the same method of thermal stability removed from the composite and subjected to TG analysis using the same method of thermal stability analysis of free abaca fibers. Thermogram of abaca fibers collected from the composite is shown in analysis of free abaca fibers. Thermogram of abaca fibers collected from◦ the composite is shown in Figure 14b. Tmax of the main thermal decompositions was observed at 287 C and the total percentage Figure 14b. Tmax of the main thermal decompositions was observed at 287 °C and the total percentage mass loss was 43.3%. In comparison to free treated abaca fibers, the percentage mass loss for the main massdecomposition loss was 43.3%. event In iscomparison lower for fibers to free from treated the compositeabaca fibers, than the the percentage free fiber mass samples. loss for Moreover, the main decomposition event is lower for fibers from the composite than◦ the free fiber samples. Moreover, the measured rate of decomposition at Tmax is lower at 0.2% per C versus the rate at Tmax of free abaca thefiber measured at average rate of of 1.2% decomposition per ◦C. This at is T likelymax is due lower to sitesat 0.2% of geopolymer per °C versus growth the rate formed at T onmax theof free surface abaca fiberof theat average treated fibers.of 1.2% The per abaca °C. This fibers is werelikely covered due to withsites geopolymerof geopolymer particles growth and formed thus protecting on the surface the of fibersthe treated from thermal fibers. The degradation. abaca fibers were covered with geopolymer particles and thus protecting the fibers from thermal degradation.

Materials 2017, 10, 579 16 of 19

Materials 2017, 10, 579 16 of 19

Figure 13. SEM images of composite fractured surfaces: (a) reinforced with untreated abaca; Figure 13. SEM images of composite fractured surfaces: (a) reinforced with untreated abaca; (b) (b)reinforced reinforced with with treated treated abaca; abaca; (c) (interfacec) interface between between untreated untreated abaca abaca fiber fiberand geopolymer and geopolymer matrix matrix and and(d ()d interface) interface between between treated treated abac abacaa fiber fiber and and geopolymer geopolymer matrix. matrix.

FigureFigure 14. 14.SEM SEM image image (a ()a and) and thermogram thermogram ofof untreateduntreated abaca fibers fibers and and treated treated abaca abaca fibers fibers (b ()b from) from thethe geopolymer geopolymer composite. composite.

4. Conclusions 4. Conclusions Chemical treatment of waste abaca fibers modifies their structure and chemical composition, withChemical the modification treatment depending of waste abacaon the fiberstreatment modifies conditions. their structureThe high tensile and chemical strength composition,among the withtreated the modification fibers was achieved depending without on thealkali treatment pretreatme conditions.nt (6% by Thewt NaOH high tensilesolution strength for 48 h), among soaked the treatedfor 12 fibers h in aluminum was achieved sulfate without solution alkali and pretreatment neutralized to (6% pH by6. wtFTIR NaOH results solution indicate for that 48 the h), alkali soaked forpretreatment 12 h in aluminum successfully sulfate dissolved solution some and neutralizedof the amorphous to pH and 6. FTIR hydrophobic results indicate components that thesuch alkali as pretreatmentlignin, pectin successfully and hemicellulose, dissolved and some the XRD of the results amorphous indicate and an hydrophobicincrease in fiber components crystallinity such that as lignin,may pectinbe attributed and hemicellulose, to such dissolution. and the Results XRD resultsfrom thermogravimetric indicate an increase analysis in fiber (TGA) crystallinity confirm the that mayremoval be attributed of lignin, to pectin such dissolution.and hemicellulose Results from from the thermogravimetric fibers, and also suggest analysis that (TGA) the geopolymer confirm the 1 removalitself coat of lignin, the treated pectin andabaca hemicellulose fibers and fromprotect the them fibers, from and alsothermal suggest degradation. that the geopolymer The Al2(SO itself4)3

Materials 2017, 10, 579 17 of 19

coat the treated abaca fibers and protect them from thermal degradation. The Al2(SO4)3 treatment is effective to form Al(OH)3 deposits that roughen the surface for better fiber- matrix interface and could also protect the fiber from the harsh environment of the geopolymer matrix. Preliminary results on abaca fiber-reinforced geopolymer indicate better adhesion for the treated fiber/matrix which results in an improved flexural strength as compared to the pristine geopolymer. Future works will further investigate the effect of chemical treatment on other properties of fiber such as swelling, as well as the mechanical properties of such fiber-reinforced geopolymer composites.

Acknowledgments: The authors acknowledge the anonymous reviewers for their comments and suggestions that help improve the content of the paper. The authors would like also to acknowledge the financial support for the research project (P-015) from the National Research Council of the Philippines (NRCP), Philippine Council for Industry, Energy and Emerging Technology Research and Development (PCIEERD) and Advanced Device and Materials Testing Laboratory (ADMATEL). We would also to thank the Polymer Chemistry Laboratory and Synthetic Organic Chemistry Research Laboratory, Institute of Chemistry, University of the Philippines Diliman. Author Contributions: Michael Angelo Promentilla and Roy Alvin Malenab conceived the research and designed the experiment including the write-up of the main narrative of the paper. Janne Pauline Ngo helped in writing the manuscript, performing the experiments and analyzing the data. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Khalil, H.P.; Bhat, A.H.; Yusra, A.F. Green composites from sustainable cellulose nanofibrils: A review. Carbohydr. Polym. 2012, 87, 963–979. [CrossRef] 2. Balasubramanian, M. Composite Materials and Processing; CRC Press Taylor & Francis Group: Boca Raton, FL, USA, 2014; pp. 2–4. 3. Chandrasekar, M.; Ishak, M.R.; Sapuan, S.M.; Leman, Z.; Jawaid, M. A review on the characterisation of natural fibres and their composites after alkali treatment and water absorption. Plast. Rubber Compos. Process. Appl. 2017, 46, 119–136. [CrossRef] 4. Achour, A.; Ghomari, F.; Belayachi, N. Properties of cementitious mortars reinforced with natural fibers. J. Adhes. Sci. Technol. 2017, 31, 1938–1962. [CrossRef] 5. Coutts, R.; Warden, P. Air-cured abaca reinforced cement composites. Int. J. Cem. Compos. 1987, 9, 69–73. [CrossRef] 6. Punyamurthy, R.; Sampathkumar, D.; Ranganagowda, R.P.; Bennehalli, B.; Srinivasa, C.V. Mechanical properties of abaca fiber reinforced polypropylene composites: Effect of chemical treatment by benzenediazonium chloride. J. King Saud Univ. Eng. Sci. 2015.[CrossRef] 7. Bledzki, A.K.; Mamun, A.A.; Jaszkiewicz, A.; Erdmann, K. Polypropylene composites with enzyme modified abaca fibre. Compos. Sci. Technol. 2010, 70, 854–860. [CrossRef] 8. Liu, K.; Zhang, X.; Takagi, H.; Yang, Z.; Wang, D. Effect of chemical treatments on transverse thermal conductivity of unidirectional abaca fiber/epoxy composite. Compos. Part A Appl. Sci. Manuf. 2014, 66, 227–236. [CrossRef] 9. Bledzki, A.K.; Faruk, O.; Sperber, V.E. Cars from Bio-Fibres. Macromol. Mater. Eng. 2006, 291, 449–457. [CrossRef] 10. Philippine Statistics Authority, Republic of the Philippines. Selected Statistics on Agriculture; PSA: Quezon City, Philippines, 2016; pp. 13–18; ISSN 2012-0362. Available online: https://psa.gov.ph/sites/default/files/ Selected%20Statistics%20on%20Agriculture%202016.pdf (accessed on 15 March 2017). 11. Philippine Fiber Industry Development Authority, Department of Agriculture, Republic of the Phillipines. Available online: http://fida.da.gov.ph/index.php/archived-articles/19-philippine-abaca-helps-in-global- environment-conservation (accessed on 15 March 2017). 12. Dungani, R.; Karina, M.S.; Sulaeman, A.; Hermawan, D.; Hadiyane, A. Agricultural waste fibers towards sustainability and advanced utilization: A review. Asian J. Plant Sci. 2016, 15, 42–55. [CrossRef] 13. Davidovits, J. Geopolymers and geopolymeric materials. J. Therm. Anal. 1989, 35, 429–441. [CrossRef] 14. Duxson, P.; Fernández-Jiménez, A.; Provis, J.L.; Lukey, G.C.; Palomo, A.; Deventer, J.S. Geopolymer technology: The current state of the art. J. Mater. Sci. 2006, 42, 2917–2933. [CrossRef] Materials 2017, 10, 579 18 of 19

15. Singh, B.; Ishwarya, G.; Gupta, M.; Bhattacharyya, S. Geopolymer concrete: A review of some recent developments. Constr. Build. Mater. 2015, 85, 78–90. [CrossRef] 16. Ribeiro, R.A.; Ribeiro, M.G.; Sankar, K.; Kriven, W.M. Geopolymer-bamboo composite—A novel sustainable construction material. Constr. Build. Mater. 2016, 123, 501–507. [CrossRef] 17. Musil, S.S.; Keane, P.F.; Kriven, W.M. Green Composite: Sodium-based Geopolymer Reinforced with Chemically Extracted Corn Husk Fibers. In Development in Strategic Materials and Computational Design IV; The American Ceramic Society; John Wiley & Sons, Inc.: Hoboken, NJ, USA, 2014; pp. 123–133. [CrossRef] 18. Alzeer, M.; Mackenzie, K.J.D. Synthesis and mechanical properties of novel composites of inorganic polymers (geopolymers) with unidirectional natural flax fibres (phormiumtenax). Appl. Clay Sci. 2013, 75–76, 148–152. [CrossRef] 19. Yan, L.; Kasal, B.; Huang, L. A review of recent research on the use of cellulosic fibres, their fibre fabric reinforced cementitious, geo-polymer and polymer composites in civil engineering. Compos. Part B Eng. 2016, 92, 94–132. [CrossRef] 20. Davidovits, J. Reinforced Geopolymer Composites: A Critical Review. Available online: http://www.materialstoday.com/polymers-soft-materials/features/reinforced-geopolymer-composites- a-critical-review/ (accessed on 28 March 2017). 21. Alzeer, M.; Mackenzie, K.J. Synthesis and mechanical properties of new fibre-reinforced composites of inorganic polymers with natural woolfibres. J. Mater. Sci. 2012, 47, 6958–6965. [CrossRef] 22. Korniejenko, K.; Fr ˛aczek,E.; Pytlak, E.; Adamski, M. Mechanical properties of geopolymer composites reinforced with natural fibers. Procedia Eng. 2016, 151, 388–393. [CrossRef] 23. Mallick, P.K. Fiber-Reinforced Composites—Materials, Manufacturing and Design, 3rd ed.; CRC Press Taylor & Francis Group: Boca Raton, FL, USA, 2008; pp. 19–126. 24. Pickering, K.L.; AruanEfendy, M.G.; Le, T.M. A review of recent developments in natural fibre composites and their mechanical performance. Compos. Part A Appl. Sci. Manuf. 2016, 83, 98–112. [CrossRef] 25. Jabbar, A.; Militký, J.; Wiener, J.; Javaid, M.U.; Rwawiire, S. Tensile, surface and thermal characterization of jute fibres after novel treatments. Indian J. Fibre Text. Res. 2016, 41, 249–254. 26. Ramadevi, P.; Sampathkumar, D.; Srinivasa, C.; Bennehalli, B. Effect alkali treatment on water absorption of single cellulosic. BioResources 2012, 7, 3515–3524. 27. Onuaguluchi, O.; Banthia, N. Plant-based natural fibre reinforced cement composites: A review. Cem. Concr. Compos. 2016, 68, 96–108. [CrossRef] 28. Kabir, M.M.; Aravinthan, H.; Wang, H.; Cardona, T.; Lau, K.T. Effects of natural fiber surface on composite properties: A review. In Proceedings of the 1st International Postgraduate Conference on Engineering, Designing and Developing the Built Environment for Sustainable Wellbeing, Darling Heights, Australia, 27–29 April 2011; pp. 94–99. 29. Izani, M.N.; Paridah, M.; Anwar, U.; Nor, M.M.; H’Ng, P. Effects of fiber treatment on morphology, tensile and thermogravimetric analysis of oil palm empty fruit bunches fibers. Compos. Part B Eng. 2013, 45, 1251–1257. [CrossRef] 30. Ramadevi, P.; Sampathkumar, D.; Ranganagowda, R.; Bennehalli, B.; Badyankal, P.; Venkateshappa, S.C. Surface modification of abaca fiber by benzene diazonium chloride treatment and its influence on tensile properties of abaca fiber reinforced polypropylene composites. Ciênc. Tecnol. Mater. 2014, 26, 142–149. [CrossRef] 31. Cai, M.; Takagi, H.; Nakagaito, A.N.; Li, Y.; Waterhouse, G.I. Effect of alkali treatment on interfacial bonding in abaca fiber-reinforced composites. Compos. Part A Appl. Sci. Manuf. 2016, 90, 589–597. [CrossRef] 32. Lima, H.C.; Willrich, F.L.; Barbosa, N.P.; Rosa, M.A.; Cunha, B.S. Durability analysis of bamboo as concrete reinforcement. Mater. Struct. 2007, 41, 981–989. [CrossRef] 33. Cook, J.T.; Schoggen, H.L.; Bell, R.I.; Smith, D.J.; Morton, G.H. Cementitious Material Reinforced with Chemically Treated Cellulose Fiber. U.S. Patent No. US7357833, 2008. 34. Jones, B.; Nachtsheim, C.J. A class of screening designs robust to active second-order effects. J. Qual. Technol. 2010, 105–112. [CrossRef] 35. Jones, B.; Nachtsheim, C.J. Definitive screening designs with added two-level categorical factors. J. Qual. Technol. 2013, 45, 121–129; ISSN: 00224065. 36. Mwaikambo, L.Y.; Ansell, M.P. Hemp fibre reinforced cashew nut shell liquid composites. Compos. Sci. Technol. 2003, 63, 1297–1305. [CrossRef] Materials 2017, 10, 579 19 of 19

37. Sedan, D.; Pagnoux, A.; Smith, A.; Chotard, T. Mechanical properties of hemp fibre reinforced cement: Influence of the fibre/matrix interaction. J. Eur. Ceram. Soc. 2008, 28, 183–192. [CrossRef] 38. Peden, M.E.; Amankwah, S.A.; Keller, B.J.; Krug, E.C.; Peden, J.M. Evaluation of Aluminum Speciation Using Synthetics and Natural Samples: Final Report; SWS Contract Report 463; Illinois State Water Survey Division: Champaign, IL, USA, 1989; pp. 13–22. 39. Kalaw, E.M.; Culaba, A.; Hinode, H.; Kurniawan, W.; Gallardo, S.; Promentilla, M.A. Optimizing and Characterizing Geopolymers from Ternary Blend of Philippine Coal Fly Ash, Coal Bottom Ash and Rice Hull Ash. Materials 2016, 9, 580. [CrossRef] 40. Park, S.; Baker, J.O.; Himmel, M.E.; Parilla, P.A.; Johnson, D.K. Cellulose crystallinity index: Measurement techniques and their impact on interpreting cellulase performance. Biotechnol. Biofuels 2010, 3, 10. [CrossRef] [PubMed] 41. Ouajai, S.; Shanks, R. Composition, structure and thermal degradation of hemp cellulose after chemical treatments. Polym. Degrad. Stab. 2005, 89, 327–335. [CrossRef] 42. Le Troëdec, M.L.; Peyratout, C.S.; Smith, A.; Chotard, T. Influence of various chemical treatments on the interactions between hemp fibres and a lime matrix. J. Eur. Ceram. Soc. 2009, 29, 1861–1868. [CrossRef]

43. Goudarzi, M.; Ghanbari, D.; Salavati-Niasari, M.; Ahmadi, A. Syntheis and characterization of Al(OH)3, Al2O3 Nanoparticles and Polymeric Nanocomposites. J. Clust. Sci. 2016, 27, 25–38. [CrossRef] 44. Ranjbar, N.; Mehrali, M.; Mehrali, M.; Alengaram, U.J.; Jumaat, M.Z. High tensile strength fly ash based geopolymer composite using copper coated micro steel fiber. Constr. Build. Maters. 2016, 112, 629–638. [CrossRef] 45. Sgriccia, N.; Hawley, M.C.; Misra, M. Characterization of natural fiber surfaces and natural fiber composites. Compos. Part A Appl. Sci. Manuf. 2008, 39, 1632–1637. [CrossRef] 46. Brigida, A.I.S.; Calado, V.M.A.; Goncalves, L.R.B.; Coelho, M.A.Z. Effect of chemical treatments on properties of green coconut fiber. Carbohydr. Polym. 2010, 79, 832–838. [CrossRef] 47. Zhang, J.; Feng, L.; Wang, D.; Zhang, R.; Liu, G.; Cheng, G. Thermogravimetric analysis of lignocellulosic biomass with ionic liquid pretreatment. Bioresour. Technol. 2014, 153, 379–382. [CrossRef][PubMed] 48. Hapuarachchi, T.D.; Peijs, T. Aluminum triohydroxide in combination with ammonium polyphosphate as flame retardant for unsaturated polyester resin. Express Polym. Lett. 2009, 3, 743–751. [CrossRef]

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).