Mechanical behavior of fire-resistant biocomposite

Composites: Part B 40 (2009) 206–211

Contents lists available at ScienceDirect

Composites: Part B journal homepage: www.elsevier.com/locate/compositesb

Mechanical behavior of fire-resistant biocomposite James Giancaspro a,*, Christos Papakonstantinou b, P. Balaguru c a b c

Department of Civil, Architectural, and Environmental Engineering, University of Miami, 1251 Memorial Drive, Coral Gables, FL 33146-0630, United States Department of Civil and Environmental Engineering, University of Massachusetts Dartmouth, 285 Old Westport Road, North Dartmouth, MA 02747, United States Department of Civil and Environmental Engineering, Rutgers, The State University of New Jersey, 623 Bowser Road, Piscataway, NJ 08854, United States

a r t i c l e

i n f o

Article history: Received 10 November 2008 Accepted 10 November 2008 Available online 21 November 2008 Keywords: A. Wood B. Mechanical properties D. Mechanical testing E. Casting

a b s t r a c t Biocomposites are typically formed by binding natural fibers derived from plants or cellulose using organic binders. The fibers that are used are normally industrial by-products and, hence, they are abundant and inexpensive. One such material is sawdust, and varieties of composite boards are being manufactured utilizing sawdust as filler material. Two major drawbacks of this system are their vulnerability to fire and very low bending strength. Both the matrix and the sawdust are flammable and this paper deals with using an inorganic matrix to improve the fire resistance. The inorganic matrix can resist temperatures up to 1000 °C and it provides protection to sawdust for short durations. The strength of these boards was increased by reinforcing with a very low percentage of high strength glass and carbon fibers. Since these fibers provide up to a fifteen-fold increase in strength, the cost increase is justifiable. Prisms were made using various proportions of sawdust ranging from about 11% to 38% by mass. The prisms were tested in compression and flexure to obtain the basic mechanical properties and determine the optimal sawdust content. Prisms with optimal sawdust content were also strengthened with glass or carbon fiber reinforcements to increase flexural capacity. The results indicate that it is possible to manufacture and engineer these types of composite beams to obtain a specified strength without using any specialized equipment, heat, or pressure, thus, producing an environmentally conscious biocomposite material. Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction For many decades, the residential construction field has used timber as its main source of building material for the frames of modern American homes. The American timber industry produced a record 49.5 billion board feet of lumber in 1999, and another 48.0 billion board feet in 2002. At the same time that lumber production is peaking, the home ownership rate reached a record high of 69.2%, with over 977,000 homes being sold in 2002 [1]. Because residential construction accounts for one-third of the total softwood lumber use in the United States, there is an increasing demand for alternate materials. Use of sawdust not only provides an alternative but also increases the use of the by-product efficiently. Wood plastic composites (WPC) is a relatively new category of materials that covers a broad range of composite materials utilizing an organic resin binder (matrix) and fillers composed of cellulose materials. The new and rapidly developing biocomposite materials are high technology products, which have one unique advantage – the wood filler can include sawdust and scrap wood products. Consequently, no additional wood resources are needed to manufacture biocomposites. Waste products that would tradi* Corresponding author. Tel.: +1 305 284 1006; fax: +1 305 284 3492. E-mail address: [email protected] (J. Giancaspro). 1359-8368/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.compositesb.2008.11.008

tionally cost money for proper disposal, now become a beneficial resource, allowing recycling to be both profitable and environmentally conscious. The use of biocomposites and WPC has increased rapidly all over the world, with the end users for these composites in the construction, motor vehicle, and furniture industries [1–4]. One of the primary problems related to the use of biocomposites is the flammability of the two main components (binder and filler). If a flame retardant were added, this would require the adhesion of the fiber and the matrix not to be disturbed by the retardant. The challenge is to develop a composite that will not burn and will maintain its level of mechanical performance [5,6]. In lieu of organic matrix compounds, inorganic matrices can be utilized to improve the fire resistance. Inorganic-based wood composites are those that consist of a mineral mix as the binder system. Such inorganic binder systems include gypsum and Portland cement, both of which are highly resistant to fire and insects [2]. The main disadvantage with these systems is the maximum amount of sawdust or fibers than can be incorporated is low [7,8]. One relatively new type of inorganic matrix is potassium aluminosilicate, an environmentally friendly compound made from naturally occurring materials. The Federal Aviation Administration has investigated the feasibility of using this matrix in commercial aircraft due to its ability to resist temperatures of up to 1000 °C without generating smoke, and its ability to enable carbon composites to withstand temperatures of 800 °C and maintain 63% of

207

J. Giancaspro et al. / Composites: Part B 40 (2009) 206–211

its original flexural strength [9]. Potassium aluminosilicate matrices are compatible with many common building material including clay brick, masonry, concrete, steel, titanium, balsa, oak, pine, and particleboard [10–19]. The primary objective of the research reported in this paper was to fabricate a biocomposite particleboard by combining the aforementioned potassium aluminosilicate matrix with waste sawdust. The effect of varying the proportion of constituents on the biocomposite was measured using compressive strength specimens. Small beams of this biocomposite were reinforced with glass and carbon fibers to obtain higher flexural strengths since the boards are relatively weak in flexure. Fire tests recently conducted at the Federal Aviation Administration (FAA) Technical Center show that the beams satisfy the stringent FAA fire requirements [20]. 2. Experimental design The biocomposite material under investigation in this study consisted of the potassium aluminosilicate matrix mixed with waste sawdust obtained from commercial home improvement stores. While traditional particleboard typically requires both pressure and heat to bond the constituents [4], the authors opted for a much simpler and economical manufacturing system by casting the biocomposite by hand under standard laboratory conditions. For the first part of the experimental program, several formulations were mixed to examine the effect of varying the relative proportions of sawdust and resin. The proportion of sawdust for the eight formulations ranged from approximately 11–38% by mass and the primary response variables were workability, density, and compressive strength of the particleboard. The prismatic-shaped specimens were 100 mm long with a cross-section of 25 mm  25 mm.

The details of each of these specimens (C1 through C8) are presented in Table 1. In the second phase of experimental work, the particleboards were evaluated for flexural loading using beams 500 mm long, 50 mm wide, and 25 mm thick. Based on the test results of workability and compressive strength, these prisms were fabricated using 29% sawdust by mass. Some of the specimens were also strengthened with either glass or carbon fiber reinforcement on both the tension and compression faces of the beam using the inorganic matrix to impregnate the fibers. The primary variables for the reinforcement were the amount and type of reinforcement.  woven carbon and glass fabric with 3k carbon tows in the warp direction (‘‘3k Woven C&G”); reinforcements were 3k carbon tows (3000 filaments) of 234 GPa modulus with area of 0:748 mm2 =cm;  3k unidirectional carbon tape (‘‘3k Uni C tape”); reinforcements were 3k carbon tows (3000 filaments) and area of fibers was 0:985 mm2 =cm, while the modulus of elasticity was 230 GPa;  12k high-modulus carbon tows (‘‘12k HMC tow”) consisting of 12,000 filaments per tow with a total reinforcement area of 1:14 mm2 and a modulus of 640 GPa;  2k alkali-resistant glass roving (‘‘2k AR-glass roving”) consisting of 1566 filaments with a total reinforcement area of 0:444 mm2 and a modulus of 72 GPa;  4k standard glass roving (‘‘4k E-glass roving”) consisting of 4000 filaments with a total reinforcement area of 0:262 mm2 and a modulus of 72 GPa. For each of the nine configurations, two identical specimens (‘‘A” and ‘‘B”) were fabricated and tested, resulting in 18 beams, Table 2.

Table 1 Details of compressive strength specimens. Specimen ID

Mass fraction of sawdust (%)

Density (g=cm3 )

Compressive strength (MPa)

Compressive modulus (GPa)

Strain at failure (%)

C1 C2 C3 C4 C5 C6 C7 C8

10.9 12.5 14.0 16.9 29.0 33.8 35.5 38.0

1.741 1.642 1.636 1.633 1.254 1.092 1.067 0.929

39.6 23.3 18.0 20.2 6.8 2.8 4.3 2.1

2.52 1.27 1.01 1.43 0.64 0.39 0.40 0.18

2.4 2.8 5.0 8.5 8.0 6.0 7.8 6.8

Table 2 Details of flexural specimens. Specimen ID

1A 1B 2A 2B 3A 3B 4A 4B 5A 5B 6A 6B 7A 7B 8A 8B 9A 9B

Core beam density (g=cm3 )

”Total reinf. ratio” (%)

Reinforcement on each face # Layers

Type

Area (mm2 )

1.017 0.997 1.028 1.009 1.027 1.012 0.995 0.998 1.017 1.012 1.014 1.021 1.024 1.036 1.008 1.002 1.033 1.017

0.00 0.00 0.58 0.58 1.15 1.15 0.18 0.18 0.55 0.55 0.07 0.07 0.14 0.14 0.77 0.77 0.14 0.14

0 0 1 1 2 2 1 1 3 3 1 1 2 2 1 1 1 1

None (control) None (control) 3 k Woven C&G 3 k Woven C&G 3 k Woven C&G 3 k Woven C&G 12 k HMC tow 12 k HMC tow 12 k HMC tow 12 k HMC tow 2 k AR-glass roving 2 k AR-glass roving 2 k AR-glass roving 2 k AR-glass roving 3 k Uni C tape 3 k Uni C tape 4 k E-glass roving 4 k E-glass roving

N/A N/A 3.600 3.600 7.200 7.200 1.140 1.140 3.420 3.420 0.445 0.445 0.890 0.890 4.800 4.800 0.844 0.844

Average maximum moment (N m)

Deflection at peak load (mm)

Flexural stiffness, EI (N m2)

9.32

3.30

59

121.16

16.16

179

140.80

12.26

389

34.02

2.56

259

85.48

3.90

461

21.25

2.58

156

27.62

2.65

209

119.75

9.26

379

22.52

3.08

161

208

J. Giancaspro et al. / Composites: Part B 40 (2009) 206–211

3. Specimen preparation Both sets of specimens (compressive strength prisms and beams) were prepared in a similar manner. First, the waste sawdust used for the fabrication was screened for large wood fragments (greater than 10 mm) and non-cellulose material. Once this debris was removed, the sawdust was then mixed with the appropriate amount of inorganic matrix in a high-shear mixer for one minute. The resulting mixture was then poured into wooden â molds lined with one layer of non-porous Teflon fabric to facilitate easy removal of the specimen, Fig. 1a. The matrix cures in about 24 h at 20 °C but they were kept in the mold for 4 days to avoid variability in strength gain. After curing was complete, the ends of the compressive strength specimens were cut with a diamond-tipped circular saw to ensure the faces were parallel. This ensured that the load would be applied uniformly to the specimen during compression testing. The beams were lightly scoured with a wire brush to improve adhesion between the substrate (particleboard) and the reinforcement. Next, a thin layer of inorganic matrix was applied to one face of the beam. The dry reinforcements were then hand-impregnated with matrix and applied to the beam. After reinforcing one face of the

beam, the matrix was allowed to cure in room temperature/ambient conditions for three days before the process was repeated on the opposite face of the beam. After both beam faces were covered, the matrix was allowed to cure for three weeks at 21 °C in the laboratory. Fig. 1b shows each of the reinforced beams after curing. 4. Test procedure The prisms used to determine compressive strength were tested in a materials testing workstation in accordance with the guidelines of ASTM D1037 [21]. The specimen was placed between two steel platens, which lightly contacted the specimen prior to test commencement. The compression test was conducted using a constant displacement rate of 0.5 mm/min. Both the load and the deformation of the specimen were recorded continuously using an electronic data acquisition system. The reinforced beams were tested using a three-point flexure test setup, as shown in the inset of Fig. 5. The span length for each test was 457 mm, with the applied load located directly at midspan. Each test was conducted using a constant mid-span deflection rate of 2 mm/min. The load versus deflection behavior was recorded until failure was reached.

Fig. 1. (a) Cores being cast in wooden molds, and (b) flexural specimens after curing and prior to testing.

40 C1

35

C2 25 mm

Compressive Stress (MPa)

30

25 mm

C3 C4

25

100 mm

20 15 10 5 0 0.0

1.0

2.0

3.0

4.0 5.0 6.0 Compressive Strain (%)

7.0

Fig. 2. Compressive strength versus strain for specimens C1–C4.

8.0

9.0

209

J. Giancaspro et al. / Composites: Part B 40 (2009) 206–211

5. Test results The compression test results including compressive strength, modulus, and strain at failure are presented in Table 1. The strain was calculated by dividing the deformation of the specimen by the specimen’s original length (nominally, 100 mm), and the modulus calculated as the slope of the initially linear portion of the stress–strain curve. This portion was identified as the points between 5% and 40% of the ultimate load. These points were then fitted with a regression line, the slope of which was taken as the modulus. Since the deformation was measured between platens, the deformation might include some extension deformations. Therefore, the modulus values have to be considered lower bound values. The mechanical response of specimens C1 through C4 are shown in Fig. 2, while specimens C5 through C8 are illustrated in Fig. 3. The relationships between density, compressive strength, and modulus are presented in Fig. 4. For the flexure specimens, the maximum moment was calculated as PL=4, where P is the maximum-recorded load and L is the span length of the beam, namely, 457 mm. These values along with flexural stiffness and deflection at failure are presented Table 2. The flexural stiffness was calculated using the equation

content results in a much lighter composite, as shown in Table 1. However, the higher sawdust content reduced the cohesion between the constituents of the composite, resulting in a mixture difficult to manipulate and press into the molds. Ease of mixing and molding were used as criteria to choose optimum sawdust content of 29% and 71% inorganic binder by mass. These proportions of sawdust and inorganic matrix yielded a biocomposite material (specimen 3 C5) with a density of approximately 1254 kg=m , which classifies it as a high-density particleboard based upon the terminology provided in the ASTM D1554 standard [22]. 6.2. Compressive strength A careful examination of the results presented in Table 1 and Figs. 2 through 4 lead to the following observations. As expected, increases in sawdust content resulted in both lower densities and lower compressive strengths. The increase in compressive strength and modulus of elasticity follow an exponential path, Fig. 4. The compressive strength varied from 2.1 to 39.6 MPa and modulus of elasticity varied from 0.18 to 2.52 GPa. The behavior becomes more ductile with increase in sawdust content. 6.3. Flexural response

ðDPÞL3 EI ¼ ðDdÞ48

ð1Þ

where ðDPÞ=ðDdÞ is the initial slope of the load–deflection curve. The load–deflection behavior of the control beams and beams reinforced with glass, carbon fabric, and carbon tows are presented in Fig. 6. In these figures, each curve represents the average of two specimens. 6. Discussion 6.1. Manufacturing approach The fabrication was done using mold-pour and hand impregnation techniques. Use of external pressure, heat, and vacuum bagging could improve the mechanical properties. Since the matrix is nontoxic and odorless, no special personal protective equipment other than latex gloves was used during the manufacture. Higher sawdust

E-glass and AR-glass, as well as carbon fibers were effectively used to increase the strength of the biocomposite core material when subjected to flexural loading. As expected, carbon fibers provide a much higher increase in strength. The average strength of specimens reinforced with two layers of 3k Woven C&G fabric (specimens ‘‘3A” and ‘‘3B”) was approximately 140.8 N m, more than fifteen times higher than the average strength of control samples, only 9.32 N m. The 12k high-modulus carbon fibers provided the largest increase in flexural stiffness compared to other carbon fabrics tested, as shown in Table 2. Hence, high-modulus carbon may be effectively used to satisfy requirements for high stiffness and low deflection. High-modulus carbon tows were able to provide significant increases in both strength and stiffness without adding significant mass to the sandwich beam core. Beams reinforced with only one carbon tow were, on average, 265% stronger and 338% stiffer than control beams while providing a reinforcement ratio of only 0.18%.

8 7

25 mm 25 mm

Compressive Stress (MPa)

6 100 mm

5 4 3

C5

2

C6 C7

1

C8

0 0.0

1.0

2.0

3.0

4.0 5.0 6.0 Compressive Strain (%)

7.0

Fig. 3. Compressive strength versus strain for specimens C5–C8.

8.0

9.0

10.0

210

J. Giancaspro et al. / Composites: Part B 40 (2009) 206–211

40

3.0 Compressive Strength Compressive Modulus

2.5

30 Strength = 2.605 (Density)4.365 R2 = 96.3%

25

2.0

20

1.5

15

1.0

10 0.5

(Density)3.403

Modulus = 0.27

5

Compressive Modulus (GPa)

Compressive Strength (MPa)

35

R2 = 93.4% 0

0.0 0.8

1.0

1.2 1.4 Density (g/cm3)

1.6

1.8

Fig. 4. Relationship between specimen density, strength, and modulus.

250 Control AR-Glass, 1 Tow AR-Glass, 2 Tows E-Glass, 1 Tow

200

Load (N)

150

100 High Strength Reinforcement

Applied Load, P

Particleboard Core

50

228.5 mm

228.5 mm 457 mm

0 0

1

2

3 Deflection (mm)

4

5

6

Fig. 5. Load versus deflection for specimens reinforced with glass tows.

1200

1000 3k Woven C&G, 2 Layers 3k Woven C&G, 1 Layer 3k Uni C Tape, 1 Layer 12k HMC, 3 Tows 12k HMC, 1 Tow Control

Load (N)

800

600

400

Particleboard Core

High Strength Reinforcement

Applied Load, P

200 228.5 mm

228.5 mm 457 mm

0 0

2

4

6

8 10 Deflection (mm)

12

14

16

Fig. 6. Load versus deflection for specimens reinforced with carbon fabric or tows.

18

J. Giancaspro et al. / Composites: Part B 40 (2009) 206–211

The load–deflection curves of samples reinforced with carbon have two distinct regions. There is an initial region of high stiffness, which decreases once the core fails at the tension side and the reinforcement is forced to bear the load. The beams reinforced with carbon tows failed by fracture of the tows. Other reinforced samples failed due to compression failure of the sandwich beam core. None of the failures occurred due to delamination, indicative of a strong bond between reinforcement and the substrate. The quantity of reinforcement played a significant role in increasing the flexural strength for specimens reinforced with high-modulus carbon tows. For example, the average moment capacity of beams reinforced with three high-modulus carbon tows per face was 151% higher than that of beams with only one carbon tow per face. Conversely, the addition of a second layer of 3k woven carbon and glass fabric to each face of the sandwich beam produced a mere 16.2% increase in strength compared to the average strength of beams reinforced with only one layer of fabric per face. The addition of a second AR-glass tow raised the average strength by only 30%, compared to using only one tow. Similar trends in modulus were also evident when examining the data. For example, when a second layer of 3k woven carbon and glass fabric was applied to the beam’s face, the average stiffness of the beam increased by 118% when compared with the average rigidity of the beams with one layer of fabric applied. The average stiffness of beams reinforced with high-modulus carbon tows increased by only 78% when three tows were applied to each face as opposed to one tow per face. When a second tow of AR-glass roving was applied to each beam face, the average stiffness increased by 34%.

7. Conclusions Based upon the observations and test results obtained during preparation and testing of the biocomposite material, a number of conclusions may be drawn regarding the feasibility and strength of this material. First, the material is viable in terms of production without using heat, pressure, or any specialized casting equipment. This yields a relatively environmentally benign, composite material that consumes little energy during manufacture and produces virtually no waste products. During the manual casting process, the wet mixture is easy to cast, and the shape and dimensions of the molds may be customized to fit any specified needs. The compressive strength, workability, and density are highly dependent upon the proportions of sawdust and resin in the mix. The composite density could be reduced by simply increasing the sawdust content. However, this resulted in substantially lower compressive strengths and moduli of elasticity. From a processing standpoint, the sawdust content could be increased up to 29% (with 71% inorganic matrix binder) without compromising workability. The 3 resulting biocomposite material had a density of 1254 kg=m , and a compressive strength and modulus of 6.8 MPa and 0.64 GPa, respectively. When biocomposite cores were reinforced with high strength fibers and tested in flexure, the following conclusions were reached:  Both glass and carbon are useful in obtaining increased strength and stiffness values in manufactured sandwich beams. The type of reinforcement applied to the core particleboard beam strongly influences the moment capacity of the beam. In general, carbon will provide better reinforcement than either AR-glass or E-glass.  Adding additional reinforcement to the sandwich beam will increase both its strength and its stiffness.  Carbon fibers significantly increased the strength by 151% (using high-modulus carbon) and stiffness by 118% (using carbon fabric) without significantly increasing the sandwich beam

211

mass. This is a very useful trait in fields such as aerospace technology. Future work on this biocomposite material will evaluate moisture resistance properties and mechanical behaviors including impact strength and flexural testing of large slabs to simulate residential deck construction. Waste by-products from specific tree species will be isolated to better understand the effect of wood type on the performance of the biocomposite material. Acknowledgements The authors wish to thank Richard Lyon and the Federal Aviation Administration for providing technical guidance and funding to support this research endeavor. In addition, the authors would like to express their gratitude towards Philip Feltz, who carried out the experimental portion of this project as part of the James J. Slade Scholars program for undergraduate researchers at Rutgers, the State University of New Jersey. References [1] Howard JL.US timber production, trade, consumption, and price statistics 1965–2002. Forest products laboratory research paper FPL-RP-615. United States Department of Agriculture; December 2003. [2] English B. Wastes into wood: composites are a promising new resource. Environ Health Perspect 1994;102(2):168–70. [3] Clemons C. Wood-plastic composites in the United States: the interfacing of two industries. Forest Prod J 2002;52(6):10–8. [4] Forest Products Laboratory. Wood particleboard and flakeboard: types, grades, and uses. General technical report FPL-GTR-53, United States Forest Service; November 1986. [5] Malvar LJ, Pendleton DE, Tichy R. Fire issues in engineered wood composites for naval waterfront facilities. SAMPE J 2001;37(4):70–5. [6] Stark NM. Wood fiber derived from scrap pallets used in polypropylene composites. Forest Prod J 1999;49(6):39–46. [7] Elinwa AU, Mahmood YA. Ash from timber waste as cement replacement material. Cem Concr Compos 2002;24:219–22. [8] Turgut P. Cement composites with limestone dust and different grades of wood sawdust. Building Environ 2007;42:3801–7. [9] Lyon RE, Balaguru PN, Foden AJ, Sorathia U, Davidovits J. Fire-resistant aluminosilicate composites. Fire Mater 1997;21(2):67–73. [10] Papakonstantinou CG, Balaguru PN, Lyon RE. Comparative study of hightemperature composites. Compos Part B: Eng 2001;32(8):637–49. [11] Giancaspro J, Balaguru PN, Lyon RE. Fire protection of flammable materials utilizing geopolymer. SAMPE J 2004;40(5):42–9. [12] Foden AJ. Mechanical properties and material characterization of polysialate structural composites. PhD thesis, Rutgers, The State University of New Jersey; 1999. [13] Giancaspro J, Balaguru PN, Lyon RE. Use of inorganic polymer to improve the fire response of balsa sandwich structures. J Mater Civil Eng, ASCE 2006;18(3):390–7. [14] Hammell JA, Balaguru PN, Lyon RE. Strength retention of fire-resistant aluminosilicate-carbon composites under wet-dry conditions. Compos Part B: Eng 2000;31(2):107–11. [15] Papakonstantinou CG. High temperature structural sandwich panels, PhD thesis. Rutgers, The State University of New Jersey; 2003. p. 251. [16] Papakonstantinou CG, Giancaspro J, Balaguru PN. Fire response and mechanical behavior of polysialate syntactic foams. Compos Part A: Appl Sci Manuf 2008;39(1):75–84. [17] Papakonstantinou CG, Balaguru PN. Fatigue behavior of polysialate structural composites. ASCE J Mater Civil Eng 2007;19(4):321–8. [18] Papakonstantinou CG, Balaguru PN. Use of geopolymer matrix for high temperature resistant hybrid laminates and sandwich panels. In: Davidovits J, editor. Geopolymers, green chemistry and sustainable development solutions, geopolymer institute, 2006. p. 201–7. [19] Giancaspro J. Influence of reinforcement type on the mechanical behavior and fire response of hybrid composites and sandwich structures. PhD thesis. Rutgers, The State University of New Jersey; 2004. 512p. [20] Giancaspro J, Papakonstantinou CG, Balaguru PN. Fire response of inorganic sawdust biocomposite. Compos Sci Technol 2008;68(7-8):1895–902. [21] American society for testing and materials. Test method D1037 – standard test methods for evaluating properties of wood-base fiber and particle materials. In: Annual book of ASTM standards, West Conshohocken, Pennsylvania 2002;4(10):142–71. [22] American society for testing and materials. Specification D1554 – standard terminology relating to wood-base fiber and particle panel materials. In: Annual book of ASTM standards, West Conshohocken, Pennsylvania 2002;4(10):222–4.