7 ur may, ‘3' n. - \ a, 5"?!“ 'N, 'fl-MNKZH‘“ 3‘ ~ 4 ': . 4..-”-..5-5‘: mgr-‘1“? V. J. 4v (‘u 7 '1, 31-41.. ‘ ‘LJ f 3"“ 4 agm ‘4’ 3453:! 2;. w“ 1.. ,r 4: «‘4 .}:“% _...‘Lsx-“ a. - {5| ‘ tram I?“ “:7, $533; a: 72mm“ : r.“ - £plk5 W\‘ ‘4' $.43? ‘IK V. “L :74 4: w? ‘ . . $6143 {”4“ \— ‘fifi- r. .amrfi‘l 33$!" . I h .w .qga' I 32%;, 4: J: #41?“ Effig; ‘13:, .f 1! mall, ‘ “ "'Ej‘fiififlh .9: "i" “7; 4 z“. I‘m! '9“: \y . ”14$: ; yn’r"fl"\' 369,, f“ 4 hr" r P 14"”;4:; ‘1',” ’57:} 1,33"? 1;, ,- -.-_:4 {N .. 1‘". v! 1:941; .74.- r ""7. ”5544 (w... '31::4‘. -_.;-;1,v1‘3:,;' 1.15% , 14:73“ 44.: 2541;“ ‘ {Nix rlwemr“" pl.T;J-m" & '- w’, “a -;‘.'E;...!::; ”ital, ":‘4' y por -- t -1. :X‘ 231‘4" .fgr'vgg-g‘y 5/.- :I’c’ff' of ””1"“ 5’44 $§NW m 414' . .H. ""rx'l'9w. fCZ’ 3:3£";'('7”' "' r ' yak ”‘15:,” 4-24 4.42;; ‘IIIMMW V'I. way?- :34” a": 7.51: x. ""J'" .\‘r n- -5 Y—‘J‘ ~~ .. n.- .5(.-- 3.:5- - .- 112». .15 Idldai . -.,4.— u“..- " ”VJ-.46“- 4. .c 5-” f . .- Véé' ”3"th -v " ' x 7.. ; ,’;q.rr§;x'?c‘ 4 ' " .41" r. )4 Pi'i‘giw..‘;c§3:. rt _‘ 93‘444w ,4 v-y‘r::r:: ,dfigxsj... ,.:v~::~..w - .. "3-241‘1 1m wwmfimrfi .mm .9 n 'I: gun-‘1?” ”—3.5% 'WMSQ MICHIGAN STATE UNIV I II IIIIIIII IIIIIIIIIIII IIII I'IIIIIIII 3 12917298 II This is to certify that the thesis entitled COMPOSITE MATERIALS FROM RECYCLED MULTI-LAYER POLYPROPYLENE BOTTLES AND WOOD FIBERS presented by Rodney James Simpson has been accepted towards fulfillment of the requirements for Masters degree in Packaging 724M145. Major professor Date 5/] /9/ 0-7539 MS U is an Affirmative Action/Equal Opportunity Institution LIBRARY Ml‘chlgen State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE I ____I-I___I *7 I I II——7I I MSU Ie An Affirmative Action/Equal Opportunity Institution chG-p1 COMPOSITE MATERIALS PROM RECYCLED MULTI-LAYER POLYPROPYLENE BOTTLES AND WOOD FIBERS BY Rodney James Simpson A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE School of Packaging 1991 ABSTRACT COMPOSITE MATERIALS FROM RECYCLED MULTI-LAYER POLYPROPYLENE BOTTLES AND WOOD FIBERS BY Rodney James Simpson The feasibility of using recycled multi-layer polypropylene (PP) bottles in combination with untreated hardwood aspen fiber (thermomechanical pulp or TMP) was tested by evaluating the mechanical properties of the reclaimed polymer and virgin PP composites. Up to fifty weight-percent of wood fiber was incorporated into the matrix. Specimens were tested in both length and cross direction for Izod impact strength, tensile properties and flexural modulus. Specimens were subject to water absorption and also to creep testing in ambient and extreme environmental conditions. Orientation of the wood fiber resulted in significantly higher mechanical properties for both polymer-wood fiber composites, at each fiber concentration. The PP Reclaim-wood fiber composite was superior to the virgin PP composite, possibly due to multi- layer materials permitting an increase in adhesion at the interface. The multi-layer material also displayed better dimensional stability under extreme environmental conditions. to my wife, for her support and inspiration ACKNOWLEDGEMENTS I would like to thank my major professor, Susan Selke, PhD. (School of Packaging, Michigan State University), and my committee members, Gary Burgess PhD. (School of Packaging, Michigan State University), Jack Giacin, PhD. (School of Packaging, Michigan State University), and Otto Suchsland, PhD. (Department of Forestry, Michigan State University) for all their support and guidance. I would also like to thank Ruben Hernandez, PhD. (School of Packaging, Michigan State University) for his timely assistance, and Mike Rich from the Composite Research Center for instruction and use of the extruder and UTS equipment. A special thank you to Maria Keal and.Joanna Childress for their moral support and help with the extruder. I am also grateful to the State of Michigan Research Excellence Fund, the Composite Materials and Structures Center at Michigan State University, and.the‘USDA for their financial support. I would also like to thank my family for their blessing. II TABLE OP CONTENTS Page LISTOF TABLES OOOOOOOOOOOOOOOOOO0......OOOOOOOOOOOOOOOOOOOV LISTOF FIGURES 0.0.00.00.00.0000.0.0.000...OOOOOOOOOOOOOVii I. II. III. IV. INTRODUCTION OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOl LITERATURE REVIEW A. Composite Materials ..........................10 B. Prediction of Composite Properties ...........12 C. Review of Prior Research......................22 EXPERIMENTAL A. materials OOOOOOOOOOOOOOOOOOOOOO0.00.00.00.00029 B. Methads 0.0000000000000.0.0.0.000000000000000038 RESULTS A. Tensile Properties ............................43 B. Flexural Modulus...............................46 C. Izod Impact Strength...........................47 D. Creep Test ....................................48 E. Water Absorption ..............................51 F. Results Summary ...............................51 G. LinearRegreSSion 0.0.0....OOOOOOOOOOOOOOOOOOOOSZ III VI. VII. DISCUSSION Page A. Tensile Strength ..............................53 B. Percent Elongation at Break .... ....... ........55 C. Tensile Modulus ...............................58 D. Flexural Modulus...............................61 E. Izod Impact Strength... ......... ...............64 F. Creep Extension ...............................66 G. Water Absorption ..............................70 H. Discussion Summary ..... ...... .................70 SUMMARY A. conC1uSions OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO0.072 B. Recommendations for Further Research ..........73 APPENDICES A. Composite Contents ............................75 BO Test Data OOOOOOOOOOOOOOOOOOOOOOOOOOOOO ..... .0076 C. Statistical Analysis of Data ..................84 BIBLIOGRAPHY 00.00.0000...OOOOOOOOOOOOOOOOOOOOOOO0.99 IV 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. LIST OF TABLES Page Properties of Fortilene PP ................. ...... ..30 Ketchup bottle multi-layer structure (weight %) ....32 Composition of Trembling aspen (percent of extractive-free wood) .................... ........ ..35 Cell type proportions and fiber dimensions . ........ 37 Actual composite contents by weight-percent ........38 Tensile strength at break (PSI) ....... . ............ 44 Percent elongation at break (%) ....................44 Young's modulus of elasticity (PSI) . ..... ..........45 Flexural modulus (PSI) .............................46 Izod impact strength (ft.lb/in) .......... .......... 47 Effect of fiber content on creep extension (500 h) .0...O0.0.0....OOOOOOOOOOOOOOOOOOOO000......48 Water absorption test (%) ..........................51 Linear Regression .......................... ........ 52 Physical and mechanical properties of aspen fiber ................................. ....... 58 Theoretical tensile moduli for the tested composite structures ...............................61 Effect of fiber content on creep extension (500 h): Actual vs. predicted .....................68 Actual composite composition by weight ....... ..... .75 Equations for linear regression ....................75 Tensile strength data from tensile test (PSI) ......76 V 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. Page Percent elongation data from tensile test (%) ......77 Young's modulus of elasticity data (PSI) ...........78 Flexural modulus of elasticity data (PSI) .. ...... ..79 Data from Izod impact test (ft.lb/in) ..............80 Creep elongation data (in) in ambient conditions ...81 Creep elongation data (in) in extreme conditions ...82 Water absorption data (%) ................. ......... 83 Two-way analysis of variance for tensi14 strength ..... ..... . ..... ................ ........... 84 Two-way analysis of variance for percent elongation at break ...OOOOOOOOOOOOOOO0.0.0.0000000085 Two-way analysis of variance for tensile mOdulus O0.0.00.0...OOOOOOOOOOOOOOOOOOOOO. .......... 86 Two-way analysis of variance for flexural mOduluS 0.0.I...OOOOOOOOOOOOOOOOOOOO0.0.00.00.00.00087 Two-way analysis of variance for Izod impact strength 0.00.00...OOOOOOOOOOOOOOOOOOOOOOOO ......... 88 One-way analysis of variance for tensile strength O......OOOOOOOOOOOOOOOOOO0...... ........... 89 One-way analysis of variance for percent elongation at break OOOOOOOOOOOOOOOOOOOOO0.000.0.0.090 One-way analysis of variance for tensile mOdulus O0.0.0.0....OOOOOOOOOOOOOOOOOO... ........... 91 One-way analysis of variance for flexural mOdulus I0......O...OCOOOOOOOOOOOOOOOOOOOO ......... .92 One-way analysis of variance for Izod impact strength ..................................... ...... 93 Coefficients for the partitioning of the sum of squares among fourteen treatments into fifteen independent (orthagonal) comparisons (1-15) .. ...... 94 VI LIST OF FIGURES Figure Page 1. Molecular structure of EVOH ........................31 2. Chain formulas of cellulose and xylan . ....... ......34 3. Creep extension in ambient conditions (23°C, 37% RH) OOOOOOOOOOOIOOOOOOOOOOOO 0000000000000 49 4. Creep extension in extreme conditions (37°C, 94% RH) ...OOOOOOOOOOOOOOOOOOOOO00...... ..... 50 5. Tensile strength as a function of wood fiber content and orientation ............................54 6. Percent elongation at break as a function of wood fiber content and orientation ..... ......... 57 7. Young's modulus as a function of wood fiber content and orientation ..................... ....... 59 8. Flexural modulus as a function of wood fiber content and orientation .................... ........ 62 9. Izod impact strength as a function of wood fiber content and orientation ............. ......... 65 10. Effect of fiber concentration on creep extenSion (500 h) OOOOOOOOOOOOOOOOOOOOOOO 000000 00.0.67 11. Predicted vs. actual creep extension as a function of fiber concentration ............ ........ 69 12. Water sorption (%) as a function of wood fiber content ...OOOOOOOOOOIOOOOOO0.000.000.0000... ....... 71 VII INTRODUCTION Our throwaway society that has emerged during the late twentieth century produces so much pollution and waste that it is slowly strangling itself. The historically cheap and simple solution of burying trash in a sanitary landfill is no longer workable. Opposition to landfills, often referred to as the not-in-my-backyard or Nimby syndrome, is slowing the opening of new sites (Nulty, 1990). In addition, a third of the 6,500 municipal landfills in the U.S. are expected to reach capacity by the mid-19905 (Fahey, 1990). Along with the opposition to new landfill sites, stricter government-mandated environmental standards and controls have accelerated the cost of landfill operation. Tipping fees, the price of dumping municipal solid waste in landfills, rose 30% in 1988 to a national average of $27 per ton (Nulty, 1990). One solution, as practiced by a Florida landfill, is to mine the landfill for materials to recycle which creates additional space for garbage and reduces the cost of operation (Fahey, 1990). With the annual production of refuse in the U.S. to increase to 193 million tons by the end of the century, Cook (1990) suggests only four ways to deal with solid waste: "Bury it, burn it, recycle it.- or don't make as much in the first place." Incineration, a "resource recovery" method popular in the 2 1970s, creates as many problems as it solves. Incinerators produce an ash with such a high ratio of toxic metals to harmless substances that it often legally qualifies as a hazardous waste, and they toxify the atmosphere with traces of heavy metals and dioxins (Luoma, 1990). Yet, with all these problems, incinerator operators say that pollution controls such as high-temperature furnaces, scrubbers and bag houses virtually eliminate harmful emissions. The problem becomes economic: construction costs run as high as $500 million, and even though public utilities are required by law to purchase the power, the energy produced is not cost-competitive (Beck et al., 1989). Incineration has blinded the public to the opportunities that waste reduction and recycling offer. Waste reduction is the most desirable environmental alternative for waste disposal in the hierarchy of waste management options (Resource Integration Systems Ltd., 1987). A reduction in the quantity of material used per unit of product has been used effectively to reduce waste in the past (e.g., light weight glass beverage bottles), although for economic rather than solid waste considerations. The method with the greatest potential for waste reduction, according to Selke (1990), is to persuade people to moderate their needs and desires, thus creating a direct reduction of material goods. This, however, is unlikely because our society in the United States places a greater emphasis on convenience. 3 Recycling has long been viewed as part of the solution to the problem of solid waste management (Spang, 1990). An industry spokesperson from Du Pont concurs that a greater emphasis should be placed on recycling, as the real solution to the solid waste problem (Chemical Marketing Reporter, May 1990). The U.S. Environmental Protection Agency (EPA) estimated.that 13% or 23.5 million tons of the country's municipal solid waste (MSW) in 1988 was recovered (not necessarily recycled) (Recycling Times, July 1990). Recovered materials are those that are separated from the waste stream and are categorized as recycled, only if markets exist. The recovery of materials is expected to reach between 20% and 28% of MSW generated in 1995 (Recycling Times, July 1990). In addition, Brown et al. (1990) predict that in the year 2030, waste reduction and recycling industries will have largely replaced the garbage collection and disposal companies of today. In contrast, a new national average recycling goal of 40% by 1996 is being considered by the 0.8. Environmental Protection Agency (EPA), heavily stressing market development for secondary materials and ensured supply of recovered materials to reach this goal (Recycling’ Times, August 1990a). Roughly eight :million American homes now sort refuse into recyclable and nonrecyclable piles for curbside collection, with recycled material expected to double by 1992 (Nulty, 1990). Revenues in the post-consumer business rose to about $6 billion in 1989 compared to $4.8 billion in 1988. Revenue is expected to grow 4 between 25% and 30% annually over the next five years (Nulty, 1990). The recyclable material of choice these days is plastic. Throughout 1990 industry's giants aggressively promoted plastics recycling and made capital investments to bring reprocessing capacity on line throughout the country (Recycling Times, December 1990). Leidner (1981) describes the action of recycling plastics from both post consumer-and industrial waste as secondary recycling. Redefined by Selke (1990) , secondary recycling is the manufacture of new products from recycled materials which possess less stringent specifications than the original. According to the U.S. Environmental Protection Agency, the recycling rate for plastic is around 1% vs. 20% for paper, 27% for aluminum beverage cans, and 7% for glass (Thayer, 1989). In 1988, the volume of discarded plastics in the MSW stream was 19.9%, although, by weight, plastics represented only 9.2%. Of the 14.3 million tons of plastics discarded, 1.1% was recovered (Recycling Times, July 1990) . The largest single source of plastic waste in 1985 was packaging, representing 40-weight percent (4,921,000 metric tons) of resin use (Resource Integration Systems Ltd, 1987). Plastic packaging is highly visible as a waste management problem due to its overall volume percent and short life span. Also, it 5 is predicted that 50% of the total packaging material will be plastic by the year 2000, up from approximately 25% in 1985 (Resource Integrated Systems Ltd, 1987) . Currently, the most recycled plastic packaging is the polyethylene terephthalate (PET) beverage container. More than 175 million pounds of PET was recovered and recycled in 1988. The second most recovered plastic packaging was high-density polyethylene (HDPE) , the plastic used to make one-gallon milk and water jugs (American Metal Market, September 1990) . Recovered high- density polyethylene bottles were purchased at an average of 9 to 14 cents a pound in 1988 (American Metal Market, April 1990) . Current buying prices for baled plastic containers are about 5 cents a pound, while baled PET consisting of only clear materials is worth 9 cents a pound (American Metal Market, September 1990). The market share of plastic food containers has increased quite rapidly in recent years and is projected to grow at over 10% per year for the next 10 years. The fastest growing segments of plastic packaging for food will be coextruded, high barrier, multi-layer bottles, jars, and cans, and thermoplastic dual-ovenable trays (Lindsay, 1988). Multi- layer, high barrier products have come under attack in environmental circles recently because of alleged recycling problems. Yet, multi-layer packaging, in itself, is a form of source reduction. To match the properties achieved with a 6 multi-layer material, single plastic packages may need up to four times the amount of material (Thayer, 1990) . One polyolefin which is being used extensively in multi-layer plastic packaging is polypropylene. After three decades in commercial production, polypropylene (PP) has a world market growth rate of about 10 percent (Potter, 1990). Yet, Kaushick (1990) pegs U.S. growth at a 5.7 percent rate. Polypropylene has captured some of the demand for high-density polyethylene or polystyrene because of its pricing (Kaushick, 1990) . In 1987, 12.6 billion pounds of plastics were used for packaging in the U.S.: roughly 10 percent (1.3 billion pounds) were of the PP resin type (Selke, 1990) . The price for virgin PP homopolymer in 1990 was at 40 to 41 cents a pound in October, up from 35 cents in March (Chemical Marketing Reporter, October 1990a) . Although PP is .in its infant stage as a recycled medium, Du Pont and American National Can have developed a program to collect, sort, and market recycled PP, as well as PVC, PS, and LDPE bottles. They plan to develop markets for their output and intend to pioneer automated sorting technology to separate the different plastics (Chemical Marketing Reporter, October 1990b). Addition of recycled material from multi-layer barrier juice, ketchup bottles, and retort containers improved some mechanical properties of polypropylene (PP) homopolymers, 7 primarily in the areas of tensile strength, elongation, flexural modulus, and impact strength (Plastics World, 1990). In addition, a study which was conducted.by the Plastic Bottle Institute (PBI) concluded that.multi-layer bottle reclaim can be used in a mixed monolayer polypropylene stream (Plastics World, 1990). Since the mechanical properties of recycled multi-layer polypropylene containers have demonstrated some improvement over homopolymers, this study will focus on the use of ketchup bottles as a low-cost matrix with wood fibers to form a composite. Composite materials, as defined, are composed of a reinforcing structure, surrounded by a matrix (Richardson, 1987). Creating a PP/wood fiber composite material provides a way to extract some value from recycled, multi-layer plastic, while acquiring the stiffness properties of the wood fibers. The wood fibers chosen to be used as the reinforcement are aspen hardwood fibers in the form of thermomechanical pulp (TMP). When used as a reinforcement in composite materials, wood pulp fibers possess strength and modulus properties which compare favorably with glass fibers when one considers the density of the fibers (Woodhams et al., 1984). Wood fibers also have a number of advantages such as lower cost, light weight, and resistance to damage during processing (Raj et al, 1988). A strong interface between the reinforcement and the matrix is 8 extremely important to develop composites with improved physical and mechanical properties. Unfortunately, good bonding between the wood fibers and polypropylene is difficult because wood fibers are hydrophilic and polar while polypropylene is hydrophobic and nonpolar. Moreover, fibers have a high degree of intermolecular hydrogen bonding and during the mixing of fibers and thermoplastics the fibers tend to agglomerate, unless fibers are wetted to reduce fiber-to- fiber’bonding (Kokta.et al., 1990). Yet, wood fibers produced from mechanical pulping still retain most of their lignin and natural waxes, materials which can aid fiber dispersion in nonpolar hydrocarbon polymers (Woodhams et al., 1984). In 1985, reinforced thermoplastics represented a 15% share of the composite market. Nylon and polypropylene were the most commonly used resins representing 50% of total production (Vu- Khanh, 1987). The PP/wood fiber composite has potential for high volume processing into a myriad of products because of the inherent ease of processing and the increase in stiffness and creep resistance. Bigg and Preston (1989) examined the potential for solid-state stamped parts from thermoplastic matrix sheet composites. Recycled aseptic packaging containers (juice boxes) made primarily of softwood fibers and polyethylene, are currently being processed into recycled products, eg. , plastic lumber (Recycling Times, August 1990b) . Additionally, the composite has the potential to replace 9 material such as wood and concrete for products such as mailbox posts, picnic tables, speed bumps, highway markers, parking stops, fences, park benches, etc. The primary objectives of this study were: (i) to investigate the use of a recycled multi-layer polypropylene container (ketchup bottle) as the matrix with wood fiber to form a composite: (ii) to determine the effect fiber content and fiber orientation has on the mechanical properties of the composite structure: (iii) to examine the effect barrier materials in the PP reclaim.have on the dimensional stability of the composite structure under extreme environmental conditions: (iv) to develop a secondary use for discarded multi-layer packaging made from predominately polypropylene resin: and (v) to compare actual mechanical properties with theoretical mechanical properties, where appropriate. 10 LITERATURE REVIEW QQHBQEII§_£LIEBIAL§ A composite is formed when two or more materials are combined, with the intent of accomplishing better results than can be obtained by a single, homogeneous material. Composites are separated into two basic forms: (i) composite materials and (ii) composite structures. Composite materials are composed of a reinforcing structure surrounded by a continuous matrix, where as composite structures display a discontinuous matrix. The matrix usually has a lower strength than the reinforcement and is the material that holds the reinforcement together. Composite materials can be classified as either fibrous or particulate, depending on the geometry of the reinforcement. There are three basic components in a fiber-reinforced composite: matrix, fiber, and the fiber-matrix interface (Katz and Milewski, 1978). The main functions of the matrix are to transfer and apportion stresses onto the fiber, and to maintain desired orientation and separation of the fibers. The matrix also provides protection against fiber exposure to environmental conditions as well as fiber abrasion. The matrix will additionally cause the fibers to act as an aggregation in resisting deformation 11 or failure under load, along with limiting the maximum temperature to which the composite can be exposed. The fiber gives the composite high tensile strength and modulus, and provides resistance to bending and breaking under the applied load or stress. The fiber-matrix interface determines the potential properties of the composite. The stresses acting on the matrix are transmitted to the fiber across the interface. Bonding at the interface is due to adhesion between the matrix and the fiber. Fibers are often coated with a coupling agent which forms a bond between the fiber and the matrix providing improvement in interfacial conditions. There are two types of fibrous reinforced composites: continuous and discontinuous. Continuous-filament (unidirectional) composites have greater strength and modulus in the fiber axis direction and generally lack physical strength in the transverse direction. In a discontinuous fiber composite, the stress along the fiber is not uniform. The length (l) to diameter (d) ratio of the fiber, or aspect ratio (l/d) determines the level of strength the composite will achieve. If the fiber is shorter than the critical length, the composite will fail at a low strength level. Therefore, it is important that the properties of a composite 12 be predictable from a knowledge of the component matrix, fiber, fiber volume, and fiber orientation. RBEDIQIIQE_QI_RBQEEBII§§ The mechanical properties of a composite are related to the properties and distribution of its components, and their chemical and physical interactions. Many analytical models and failure theories have been used in the analysis of the mechanical properties of different types of composites. The modulus is the bulk property of a composite that depends primarily on the geometry, modulus, particle size distribution and concentration of the filler (Bigg, 1987). The Rule of Mixtures as specified.by.Jindal (1986) can be used to estimate the strength of a composite, by: ac =ofo ”‘0me (1) where: a, = Ultimate strength of the composite of ,¢%, = Ultimate strength of the fibers, matrix V,,.\Q = Volume fraction of the fibers, matrix (Jindal, 1986) This equation assumes a continuous-unidirectional composite with stress direction parallel to the fibers and perfect bonding of all fibers to the matrix (Katz and.Mi1ewski, 1978). Yet, the stress is not uniform along the fiber for a discontinuous fiber composite. The properties of a l3 discontinuous (short) fiber composite are a function of the fiber length and fiber ends (Agarwal and Broutman, 1980). A fiber can be defined as any material that has a high ratio of length to cross sectional area, with the minimum ranging from 10:1 to 100:1, a maximum cross sectional area of 7.85 x 107 ixn? and a maximum transverse dimension of 0.010 in. (Katz and Milewski, 1978). In comparison, Richardson (1977) defined a short-fiber composite as having a fiber length-to-diameter ratio between 10 and 1000. The mechanism of stress transfer can allow interpretation of the performance of a discontinuous fiber composite. The distribution of stress along the length of a fiber can be explained by examining the force equilibrium of an infinitesimal length, dz, of a short-fiber composite: (wr2)of-+(2nrdz)1==(nr2)(a,+dof) (2) do, = 21 TE? where: (n = the fiber stress in the axial direction 1 = shear stress on the cylindrical interface r = the fiber radius dz = element of length in axial direction (Agarwal and Broutman, 1980) This indicates that the rate of increase of fiber stress is proportional to the shear stress at the interface, provided the fiber is of uniform radius. The equation (2) can be 14 integrated to obtain the fiber stress on a cross section a distance 2 away from the fiber end: 2 2 a ==a +-_. rdz: 3 1‘ f0 rIo ( ) where: am = the stress on the fiber end (Agarwal and Broutman, 1980) At equal strain, the average stress will be lower in a short fiber than in a continuous composite fiber, because the ends of short fibers are not loaded to the same level as the center (Richardson, 1977) . When the fiber has an aspect ratio (length-to-diameter ratio) that equals or exceeds the critical aspect ratio, the stress in the center will be equivalent to that of a continuous length fiber (Katz and Milewski, 1978). The critical aspect ratio (l/d)c can be expressed as: 1) Sf = __ (4) (a C 23: where: l,d. = the length and diameter of the fiber (l/d)c = the critical aspect ratio S.f = the tensile stress of the fiber Y = the yield strength of the matrix in shear or the fiber-matrix interfacial shear strength, whichever is the lowest (Katz and Milewski, 1978) 15 The critical aspect ratio would result in fiber fracture at its midpoint. The stressed fiber will de-bond from the matrix and the composite will fail at a low level of strength when the fiber is shorter than the critical length (Katz and Milewski, 1978). Fiber length is an important parameter in the determination of stress to be transferred from the matrix to the fiber. This may be expressed as: L: =DSTf (5) where: I“ = critical fiber length D = diameter of the fiber S = strength of the matrix bond to fiber (approx. equal to shear strength of matrix) I; = tensile strength of fiber (Richardson, 1987) As the angle between the fiber axis and loading direction increases, the strength of an aligned short-fiber composite decreases. For a discontinuous fiber composite, Katz and Milewski (1978) provide an estimate for tensile strength using the rule of mixtures expressed as: S, = v, s,[1-71i.]+vIn sm (6) 16 where: Sc == tensile strength of the composite \h = volume fraction of the fiber V; = volume fraction of the matrix 8. == tensile strength of the fiber 8,,I := tensile strength of the matrix 1.6 = critical length of the fiber H II length of the fiber (Katz and Milewski, 1978) When applying the rule of mixtures to plastic matrix filled composites, the Vf must be greater than 10% and less than 70%. A fiber volume fraction lower than 10% will yield a strength similar to that of the matrix and anything higher than 70% will show a decrease in properties (Katz and.Milewski, 1978). The interfacial strength between the matrix and the filler is difficult to determine. JMany bonding mechanisms are possible between a polymer and filler: these include ionic, covalent, electrostatic, and van der Waals (Bigg, 1987). Hull (1981) describes five main mechanisms for adhesion at the interface (either in isolation or in combination) to produce the bond, (1) adsorption and wetting, (2) interdiffusion, (3) electrostatic attraction, (4) chemical bonding, and (5) mechanical adhesion. 17 When two solids are brought together, the surface roughness at the micro level prevents the surfaces from coming into contact except at isolated points. For effective wetting of a fiber surface the liquid resin.must cover every part of the surface to displace all the air. Yet, this bonding mechanism is usually not achieved because of the contamination of the fiber surface, entrapped air at the solid surface, and because of displacements at the surface due to shrinkage stresses during the cure process. Interdiffusion (also referred to as autohesion in relation to adhesives), as a bonding mechanism, is achieved by molecular entanglements of two polymer surfaces promoted by the presence of plasticizing agents and solvents. Interdiffusion is accomplished by pre-coating the fibers with polymer prior to mixing with the matrix. Ionic bonding or acid-base interactions are forces of attraction which allow surfaces with dissimilar charges (+,-) to bond. This electrostatic attraction plays a more important role as a coupling agent and is not likely to contribute extensively to the bond strength of the fiber- matrix composite. Coupling agents, which are applied to fiber surfaces prior to mixing of composites, allow for chemical bonding to form 18 between themselves and chemical groups in the matrix. Interface failure then must involve breaking of the bond(s), adding to the strength of the composite. Mechanical bonding occurs by the interlocking of two surfaces mechanically. Unless a large number of re-entrant angles appear on the fiber surface, a high strength in tension is unlikelyu Roughness of the fiber surface may provide strength in shear by supplying a large surface area for increased chemical bonding with the use of surface wetting. The need to understand the role of the fiber/matrix interface has led to the introduction of additional experimental techniques particularly designed to measure the interfacial shear strength. A fiber pull-out test will allow for the direct measurement of the interfacial shear strength of a fiber/polymer’ composite (Mader' and Freitag, 1990). The interfacial shear strength can be obtained by the equation: Fd 1 =.______ d any L (7) where: 13 = interfacial shear strength of the composite F} = the debonding load I} = the fiber radius L = the embedded length (Mader and Freitag, 1990) 19 The shear-lag analysis is restricted to the case of an elastic fiber and a plastic interface yielding at constant stress. Compression of fibrous materials will cause fibers to align perpendicular to the orifice of a mold resulting in a random orientation. Randomly oriented short-fiber composites are produced to obtain composites that are isotropic in a plane (Agarwal and Broutman, 1980). Prediction of properties of randomly' oriented short-fiber' composites. is complex. Zadorecki et al. (1986) demonstrated how Tsai and Pagano's equation derived from orthotropic elastic theory can be used to predict the modulus of composites containing fibers that are randomly oriented in a plane: 3 5 Erandom = g EL + 5 ET (8) where: EL := longitudinal modulus of the composite ET == transverse modulus of the composite (Zadorecki et al., 1986) Both moduli must be of an aligned short-fiber composite having the same fiber aspect ratio and fiber volume fractions as the composite under consideration (Zadorecki et al., 1986). Randomly oriented short-fiber composites are quasiisotropic, that is, having similar mechanical properties in all 20 directions. Therefore, a laminate analysis procedure is utilized to predict the strength of randomly oriented short- fiber composites (Agarwal and Broutman, 1980). There are no theoretical developments capable of predicting the impact strength enhancement of composites (Bigg et al, 1988) . Although impact strength is an important property, factors such as microscale morphological changes in the polymer, caused by the filler, affect the strength of fiber composites (Bigg, 1987). In addition, fiber orientation, fiber aspect ratio, and interfacial adhesion, affect the impact strength of fiber composites (Bigg et al., 1988). The packing fraction of a filler'has a functional influence on the mechanical properties of a composite. Large volumes are occupied by poorly packing fillers, contributing to a reduction in continuity of the matrix in a composite. The packing of random, rigid fibers is extremely poor. Discontinuous short fibers do not pack as well as continuous fibers and their reinforcing efficiencies are only 50-70% of continuous fibers. Packing fractions will remain lower for discontinuous fibers unless the fibers are perfectly parallel and butted end to end (Parratt, 1972). The material packing of rigid fibers depends on their aspect ratio. For three- dimensional random fibers, volume concentration is proportional to (L/d)'2, compared with random fibers in a plane 21 where packing is proportional to (L/d)". The packing of parallel fibers is independent of the aspect ratio (Parratt, 1972). Since the matrix absorbs the shock of impact, fillers having high packing fractions will tend to decrease impact strength much less at the same filler volume (Katz and Milewski, 1978). Packing fraction is also a determining factor in creep and stress relaxation. Relative packing fraction determines the proportion of free matrix. Therefore, maximum creep resistance can be expected at the maximum packing fraction, Pf. Creep for composites can be estimated from relative modulus data: e=e1(t)%‘ (9) where: 6 = the creep elongation at any time (t) e, = the corresponding creep of the matrix E = the elastic modulus of the composite Eh = the elastic modulus of the matrix (Katz and Milewski, 1978) Calculations are valid only up to a point when the filler begins to debond from the matrix; .At'this point, catastrophic failure most likely will occur. 22 EBIQB_E§§§ABQ§ Wood fiber as a reinforcement for thermoplastic composites is attracting an increasing amount of attention in research laboratories, primarily with the inclusion of dispersants and coupling agents. Hardwood pulp, in the form of highly bleached cellulose, was added to polypropylene at several concentrations by Bataille, Ricard and Sapieha (1989) . They studied the effects of surface pre-treatment of cellulosic fibers and the processing time and temperature on the cellulose-containing polypropylene. Little adhesion was found at high elongation levels between the untreated fibers and polypropylene. Interfacial adhesion improved significantly with the treatment of fibers with coupling agents. Noted was the improvement of adhesion and dispersion of fibers in the matrix with the presence of maleic anhydride modified polypropylene. Woodhams, Thomas and Rodgers (1984) compared stiffness characteristics of talc, glass, and softwood pulp fiber-filled polyolefins. Softwood pulp fibers in the form of Kraft (bleached and unbleached) , mechanical and chemical-mechanical pulps, waste pulps, and reclaimed newspapers were dispersed into high density polyethylene (HDPE) and isotactic polypropylene (PP). Carboxylic dispersion agents were added to aid dispersion. They concluded that the addition of 23 carboxylic waxes aids dispersion and permits inclusion of 40 to 50 weight-percent wood fibers. Flexural strengths of 70 MPa and flexural moduli of 5 to 6 GPa were obtained for both isotactic PP and HDPE when filled with 40 to 50 weight-percent wood fibers. Raj, Kokta, Maldas and Daneault (1988) studied the reinforcement of thermoplastics with wood fibers. Aspen chemithermomechanical pulp (CTMP) was dispersed in linear low density (LLDPE) and high density (HDPE) polyethylenes using four different isocyanates as bonding agents. The tensile properties of aspen fiber composites compared favorably with glass and mica fiber reinforced composites. The isocyanate provided significant improvement in mechanical properties (i.e., stress and modulus) for'both HDPE and LLDPE composites. Cellulose in the form of highly bleached hardwood pulp, used as a filler in linear low density polyethylene (LLDPE) and high density polyethylene (HDPE), was examined by Bataille, Allard, Cousin and Sapieha (1990) to determine the mechanical properties. Benzoyl peroxide (BPO) was applied. to 'the cellulose fibers both prior to ("MS" method) and during processing ("DM" method). No significant effect of the various treatments was gained except for the benzoyl peroxide treated composite using method DM. Higher yield strengths were indicated for LLDPE than HDPE, probably due to easier 24 interaction between the polymer chain and other constituents. Kokta, Maldas, Daneault and Beland (1990) studied the mechanical properties of 'treated. hardwood. aspen (chemithermomechanical pulp (CTMP) and sawdust) incorporated into poly(viny1 chloride). The fibers were latex coated or grafted with vinyl monomers, in addition to treatment with coupling agents (e,g., maleic anhydride, obietic acid, and linoleic acid) and various additive dispersants (e.g. , stearic acid or anhydrides). Generally, the mechanical properties improved, compared to untreated composites. Grafting of the fibers was most effective. Coupling agents performed better than dispersants, of which linoleic acid was most promising. Maldas and.Kokta (1990) studied.the potential of the recycling of polystyrene-hardwood aspen fiber (CTMP) composites. They evaluated the dimensional stability and.mechanical properties of the recycled composites and the original polymer. The influence of a coupling agent and various treatments on the properties of the polystyrene-hardwood fiber composite were also studied. They concluded that treated wood-fiber-filled thermoplastic composites offer excellent mechanical properties and dimensional stability under extreme conditions (e.g., exposure to boiling water). Moreover, in comparison to the original composites, the mechanical properties and dimensional stability of the recycled composites did not change 25 significantly even after exposure to extreme conditions. Henequen fibers as a reinforcing element with waste polyethylene (PE) were formed into a panel and examined by Ramirez and. Solis (1984) to. determine the ‘physical and mechanical properties for comparison with properties of commercial panels. The waste PE was recovered from waste films from packing. River sand in the form of feldspar and quartz was added to the matrix to increase environmental resistance (e.g., ultraviolet (UV) light). Ramirez and Solis concluded that the composite material had good mechanical and physical properties and an increase in environmental resistance by about 1000% with the incorporation of sand. In addition, the fibers did not degrade in the molding process, sustaining their inherent mechanical properties. Zadorecki and Michell (1989) examined the future use of cellulose wood fibers as reinforcements in synthetic organic polymers to form composites. Wood cellulose proposed as reinforcement was in the form of wood fibers, cellulose fibers, microfibrillar, and microcrystalline cellulose. Zadorecki and, Michell concluded that more sophisticated processing will be required bringing together the separation of fibers and.the formation of the polymers, if full potential of the reinforcement is to be achieved. Furthermore, in the commodity field it is expected that cellulose fiber- 26 thermoplastics with coupling agents will be introduced commercially, to compete with mineral filled polymers. Composites prepared. by' Hua, Zadorecki and Flodin (1987) involved combining unsaturated polyester and surface treated cellulose as a reinforcing material. Formaldehyde and di- methylolmelamine (DMM) were used as surface treatments for cellulose fibers in the form of bleached kraft paper. The experimental design utilized five treatments. Tensile strength and the elongation of the cellulose fibers were determined in dry and wet conditions along with the tensile strength and modulus of the cellulose-polyester composites. DMM was found to be an effective surface modifying agent for cellulose fibers. An improvement by more than 50% was achieved in the wet strength of the composite, along with a reduction of water uptake (46 to 52 percent) by the composite. Lightsey, Short and Sinha (1977) tested pulp mill residue as a filler for low density polyethylene (LDPE) and polystyrene (PS) to form composites using a Brabender Rhomex extruder. Kraft pulp mills in Georgia, Florida, and Louisiana supplied the wood pulp residue, the major species being southern yellow pine (loblolly and slash pine). They concluded that the tensile strength and modulus of the composites containing wood residue are only slightly greater than composites filled with wood flour. Bonding between the polystyrene matrix and the 27 filler was weak. The effect of alkali treatment on the surface adhesion of sisal fibers to polyester resin was examined by Navin and Rohatgi (1986). Retted sisal fibers from Bhopal, India were soaked in an aqueous solution of NaOH ( 5 wt%) for various time periods. Alkali treatment significantly increased wettability of the surface of sisal fibers with polyester resin. In addition, Navin and Rohatgi found that alkali treatment of sisal fibers for 90 hours resulted in adhesion with polyester and an increase in tensile strength. Aspen hardwood fibers in the form of TMP, used as a filler in high density polyethylene (HDPE) , was examined by Gogoi (1989) to determine the mechanical properties. Gogoi also studied the effect of fiber pretreatment, screw configuration, and compounding temperature on the mechanical properties of the composite. Gogoi found that the screw configuration which has the longest mixing time imparts the best overall strength, although it produces maximum damage to the fibers. Gogoi also concluded better adhesion to acetylated and untreated fibers than to heat-treated fibers at 30 percent fiber concentration. Keal (1990) studied the effect of combining two additives with high density polyethylene (HDPE) and 30 weight-percent aspen wood fibers. The additives studied were stearic acid, a 28 dispersing agent; ionomer modified polypropylene: and maleic anhydride modified polypropylene, both to increase interfacial bonding; INo significant. improvement 'was gained by' the combination of additives as compared to the use of one additive, except for modulus of elasticityu In addition, Keal concluded that the use of additives improved properties, except impact strength. 29 EXPERIMENTAL Materials Thamoplasti cs The material used for the matrix was primarily a polypropylene random copolymer (Fortilene’4104) in the form of regrind from multi-layer squeezable ketchup bottles. Regrind was chosen as a representation of post consumer waste. The regrind was in granulated form, made from containers recycled in the closed loop manufacturing operation. The container has four primary layers bound with two adhesive layers to ensure product quality. The layers consist of polypropylene, ethylene vinyl alcohol, adhesive, and regrind (PP/tie/EVOH/tie/regrind/PP). PolyprOpylene random copolymers are a type of polypropylene in which the basic structure of the polymer chain has been changed by the incorporation of a different monomer molecule, in this case, an ethylene comonomer is used. The physical properties of the PP are changed providing an improvement in optical properties, improved impact resistance, increased flexibility, and a decreased melting point (Walsh, 1990). Yet, the same chemical resistance, water barrier properties, and organoleptic properties as PP homopolymers are preserved. 30 Random copolymer PP typically contains between 99 to 93 wt.-% of propylene molecules and 1 to 7 wt.-% of ethylene molecules (Walsh, 1990) . Ethylene/propylene random copolymers are produced in the same reactors used to produce homopolymer PP by the simultaneous polymerization of propylene and ethylene molecules. Random copolymers are usually more flexible than homopolymer PP with flexural modulus values (secant at 1% strain) ranging from 70,000 to 150,000 psi, compared to 150,000 to 200,000 psi for homopolymers. Table 1 provides a summary of mechanical properties of the Fortilene'PP used in this experiment. Table 1. Properties of Fortilene'PP (Solvay, 1990). Density (g/cc) 0.898 Melt Flow Index 1 (g/10 min) 1.7 Tensile Strength at Break (psi) 3900 Elongation at Break (%) 550 Tensile Modulus 2 (psi) 97,000 Flexural Modulus 2 (psi) 130,000 Notched Izod Impact Strength (ft-lbf/in) 0.8 ‘ 230°C/2160 g 2 Secant at 1% strain Ethylene Vinyl Alcohol (EVOH) is a semi-crystalline polymer formed by hydrolyzed copolymers of vinyl acetate and ethylenes. The resulting EVOI-I copolymer is an atactic polymer with the following molecular structure (Fig. 1). 31 ---(CH2 - CH2)m - (CH2 - CH)n --- / OH Ethylene Unit Vinyl Alcohol Unit Fig. 1. Molecular structure of EVOE (Foster, 1987). EVOH is hydrophilic and will absorb moisture due to the presence of hydroxyl groups (-OH) in the molecular structure. Moisture absorbed material becomes plasticized and hydrogen bonding is reduced, thereby increasing chain mobility in the amorphous regions. The EVOH used in the multi-layer ketchup bottle is Grade EP-F 101 (EVAL’Solarnol DC, EVALCA) with a density of 1.19 g/cc and a melting point of 181%: (Foster, 1987). EVOH has high mechanical strength, elasticity, surface hardness, good abrasion resistance, and excellent weatherability. Specialty designed adhesive resins (tie resins) are used due to the poor adhesion between the EVOH and PP resins. The trade name of the adhesive used in the production of the multi-layer ketchup bottle is Admer' (Mitsui Monoply MT38) manufactured by Mitsui Petrochemical Industries, Ltd. EVOH resin contained in the ketchup bottle is recovered and reused as a layer of regrind due to economic reasons. The percent of regrind.used varies depending on the amount of trim 32 and scrap produced. The coextruded ketchup bottle contains almost 95 weight-percent PP with the remainder of the structure containing EVOH and adhesive (Table 2). Table 2. Ketchup bottle multi-layar structure (weight %). Resin Material Weight Percent Random Copolymer PP (Fortilene‘4101, Solvay) 94.50 EVOH (EVAL'Solarnol DC, EVALCA) 3.75 Adhesive (Admer', Mitsui Monoply MT38) 1.75 F er Hardwood aspen (Populus Tremuloides, Michx), in the form of thermomechanical pulp (TMP) was supplied by Lionite Hardboard, Phillips, Wisconsin. The hardwood fiber was produced using the pressurized refiner process. In this process, wood chips are softened by a digester at about 100 p.s.i. of steam pressure for 3 to 5 minutes, to help equalize chip moisture content. The wood chips are then ground into fiber using a Bauer 418 refiner. In the refiner, the chips are forced between two oppositely rotating grinding blades. Exiting the blades, the wood pulp is then blown continuously through a Heil flash tube dryer (40 in. diam. x 150 ft.) at an entrance temperature of 475°F. In the dryer, the fiber is suspended for " 20 seconds and exits at temperature of 175°F, with a moisture content of " 7%. The pulp yield is about 95% with insignificant modification and removal of the lignin. The 33 majority of the fibers are separated with slight damage (i.e., exposure of secondary wall surface). Wood is a composite material containing three major polymers: 1) cellulose, 2) hemicellulose, and 3) lignin, which serve as the skeletal, matrix, and encrusting substances, respectively (Schniewind, 1989). Cellulose is a polydisperse linear syndiotactic organic polymer with the basic monomeric unit D-glucose (Fig. 2). D-glucose links through a glycosidic bond in the beta configuration between carbon 1 and carbon 4 of adjacent units forming long-chain 1,4-B-glucans. Each monomeric unit.within the cellulose chain has three hydroxyl groups, specifically two secondary and one primary hydroxyl group (Schniewind, 1989). The cellulose fiber is made up of 55-75% crystalline and 25-45% amorphous regions. In the amorphous regions, the hydroxyl groups are highly accessible and readily reactive in all chemical reactions. The crystalline regions are not readily accessible to reactant molecules, and it is the crystalline nature of cellulose that provides strength and stiffness. Cellulose is a hydrophilic polymer (it absorbs water readily and.swells), although swelling is limited.to the amorphous regions of the fiber. 34 H CH20H CH20H —0 0H OH H H OH H OH H H CH20H CH20H Cellulose H OH H H OH H H H OH H H OH H H H OH H H 0 OH H H o 0 H H 0 H H H OH H H OH Xylen Fig. 2. Chain formulas of cellulose and xylan (Browning, 1963). The hemicellulose is linear polysaccharides that are displayed continuously with cellulose and lignin in plant cell walls. Xylan is the predominant hemicellulose in hardwood comprising of 20-30% of the wood substance. Hardwood xylans consist of a main chain with random side chains along the backbone (Fig. 2). The main chain is made up of 1,4-linked B-D-xylopyranose residues, some transporting a single terminal 4-O-mythyl-a-D- glucuronic acid unit attached to C-2 (Schniewind, 1989) . Additionally, an average of seven acetyl groups per ten xylose units are attached to either C-2 or C-3. Hardwood xylans are amorphous in their natural state. In addition to xylan, hardwoods also contain less than 5% of a glucomannan, composed of 1,4-linked B-D-glucopyranose and B-D-mannopyranose residues (Schniewind, 1989). 35 The third important component of the cell wall is lignin. Wood lignins are predominantly aromatic and almost totally insoluble in solvents, not hydrolyzable to monomeric units, and devoid of the highly regular structure so characteristic of other' natural polymers (Browning, 1963). Lignin is composed of phenylpropane units that encrust the intercellular space of the cell wall. Most hardwoods contain 22-28% of lignin. Hardwoods have a guiacylsyringyl lignin with one or two methoxyl groups (Schniewind, 1989). Of the total amount of lignin present, 20-25% occurs mainly in the intercellular region and primary wall, while the cell wall contains 75-80%. Lignin concentration varies from 50-100% in the middle lamella-primary wall to 20-25% in the secondary wall (Schniewind, 1989). Lignin is completely' amorphous and softens at temperatures of 165-175°C under normal conditions (Bodig and.Jayne, 1982). Browning (1963) lists the components of Trembling aspen (Populus tremuloides) in percent of extractive-free wood (Table 3). Table 3. Composition of Trembling aspen (percent of extractive-free wood (Browning, 1963). Uronic Species Glues: Mm Select-I Xylm Ardfinm “yd-ids Acetyl timin Ash Trdlim aspen 57.3 2.3 0.8 16.0 0.4 3.3 3.4 16.3 0.2 (Palm tmloidu) 36 The structure of a cellulose fiber consists of (1) a primary wall (”0.1g thick) containing mostly noncellulosic substances (i.e., waxes and pectin), (2) a secondary wall (‘4u thick) which contains almost all the cellulose present in the fiber along with a considerable amount of lignin, and ( 3) the lumen, which.is the hallow center of the fiber cell (Browning, 1963). Three distinct layers (S1, Sat 53) are almost always present in the secondary wall with the $2 layer being the most extensive of the three (Bodig and Jayne, 1982). Hardwoods are composed of fibers which are threadlike cells accompanied by larger-diameter vessels and short parenchyma cells (rays) . The fibers provide mechanical support while the vessels are the main channel of fluid transportation, and the parenchyma cells are for food.storage (Bodig and.Jayne, 1982). In aspen hardwood, like other hardwoods, fibers and vessels are the most abundant type. IBrowning (1963) provides the cell type proportions and fiber dimensions of Quaking aspen (Populus tremuloides Michx) (Table 4). 37 Table 4. Cell type proportions and fiber dimensions (Browning, 1963). Cell Mansion Portionof Cal ls, peremt volt” Vessel (and. Fiber tenth Fiber din-ate.- Vessels Fibers lays Ava., - Avg" - rsme h min. aspen 11.11 55.1 11.1 0.67 1.01 10 - 27 (Paula tmloides Iliohx) 38 eth 9229221219.; This experiment is comprised of eight treatments in terms of material composition, two of which were pure polymer material (100%). Aspen wood fiber was incorporated into the remaining six matrix materials at concentration levels of 30, 40, and 50 percent by weight. The exact wood fiber-matrix content is presented in Table 5. (See Table 17, Appendix A for actual material used by weight.) The wood fibers were air-dried in ambient conditions (23°C, 50% RH) equilibrium. Table 5. Actual composite contents by weight-percent. until reaching moisture Treatments Composite Wood Fiber (%) Matrix (%) 30% Fiber Level 1 - Virgin PP 30.80 69.20 2 - PP Reclaim 24.66 75.34 40% Fiber Level 3 - Virgin PP 40.41 59.59 4 - PP Reclaim 38.66 61.34 50% Fiber Level 5 - Virgin PP 52.44 47.56 6 - PP Reclaim 51.44 48.56 . 7 100% Virgin PP 100.00 8 100% PP Reclaim 100.00 Note: For actual material used by weight refer to Appendix.A. 39 The virgin PP resin was made into plates approximately 0.0625 in. thick and granulated with a BTP Granulator Model 68 SPL (Polymer Machinery, Berlin, CT) to achieve an equal density and friction as the PP Regrind. Like densities were important to produce similar fiber to polymer ratios, since feed rates could not be altered sufficiently to allow for the dissimilar frictions. Compounding .A Baker-Perkins Model MPC/V-30 DE, 38mm, 13:1 intermeshing self-wiping corotating twin-screw extruder (Baker-Perkins, Saginaw, MI) was used to mix the polymer and wood fibers. The temperature of the feeder, transition, and metering zones of the extruder was 185°C. The compounder speed was 100 rpm. The polymer feed rate and subsequent average load percent varied for each of the fiber concentrations. The polymer was added at the feeder zone while the wood fibers were added at an open port in the transition zone. The extruded material was allowed to cool to room temperature and then was compression molded into plates approximately 0.125 in thick using a Carver Model M 25 Ton laboratory press (Fred S. Carver, Inc., Menomonee Falls, MI). Plates were made using three lengths of material. The mold was heated at 185°C platen temperature for 15 min under pressure (30,000 psi), and then cooled down under pressure to a temperature of “50°C by circulating cold. water for' about 10 min in the press. 40 Specimens for tensile, impact, flexural modulus, creep, and water adsorption were made according to ASTM standards (ASTM, 1988). Business Molded plates were cut into tensile and creep specimens (Type I dumbbell shape with a "2 in. gauge length) using a Tensilkut Model 10-13 specimen cutter (Tensilkut Engineering, Danbury, CT). Creep specimens were cut parallel to the direction of the extrudate (lengthwise) . Flexural modulus samples were cut into 6.0 in. x 0.5 in. x 0.125 in. bars using a band saw. Impact specimens were cut into 2.5 in. x 0.5 in. x 0.125 in. bars and notched using a TMI Notching Cutter Model TMI 2205 Tensile Machines, Inc., Amityville, NY). The specimens for tensile, impact, and flexural modulus were made in both lengthwise and crosswise direction to the extrudate to study the effects. Specimens used for the water absorption test were cut.by a circular drill bit" Vernier callipers were used for measurements of composite specimen dimensions. Conditioning The specimens were conditioned at standard laboratory atmosphere (23°C, 50% RH) for 40 h using Procedure A of ASTM D618-61: Conditioning Plastics and Electrical Insulating Materials for Testing (ASTM, 1988), prior to testing. Air circulation was provided on all sides of each specimen. 41 Testing Tensile strength, tensile modulus, and elongation at break were measured using an Instron Tester Model 4201 (Instron, Canton, MA), following ASTM D 638-87b: Tensile Properties of Plastics (ASTM, 1988) at ambient conditions (23W3, 50% RH). The rate of elongation was 2 in/min., gauge length.was 3.5 in, and full scale load.was 500 lbs. Sandpaper was lodged between specimen and grips to deter slippage. The average of five measurements was used to report mechanical properties. Flexural modulus was tested using Method I, Procedure A of ASTM D790-86: Flexural Properties of Unreinforced and Reinforced Plastics and Electrical Insulating Materials (ASTM, 1988) on a electromechanical test frame fitted with a 20 lb. load cell (United Testing System, UTS). Crosshead speed was 1.00 in/min, support span length was 4.0 in, and a 16:1 span- to-depth ratio was used" 'The average of five measurements was used to report the tangent modulus of elasticity. Impact strength was tested using Method A (Izod Type) of ASTM D256-87: Impact Resistance of Plastics and Electrical Insulating Materials (ASTM, 1988). Fracture energies were determined using a TMI 43-1 Izod Impact Tester (Testing Machines, Inc., Amityville, NY) with a 5 ft-lb. pendulum load. The average of eight measurements was used to report Izod impact strength. 42 Water sorption was determined using a 2-h boiling water procedure of ASTM D570-81: Water Absorption of Plastics (ASTM, 1988). Moisture gain was reported as an average of three measurements. For creep extension ASTM D 2990-77: Tensile, Compressive, and Flexural Creep and Creep-Rupture of Plastics (ASTM, 1988) was used. Weights (SO-lb) were attached to the bottom of the end grips. iMeasurements were made at set time intervals up to 500 hours. Creep extension was measured by grip separation and was tested in ambient (23°C, 37% RH) and extreme (37°C, 92% RH) conditions. Extension was reported as an average of two samples and is suggestive rather than conclusive. MSTAT statistical program (version 5.0, Michigan State University, 1988) was used to perform statistical analyses for tensile strength, elongation at break, tensile modulus, flexural modulus, and Izod impact strength. The following statistical analyses were performed: . A two-way analysis of variance to specify any statistically significant variable effects. - A one-way analysis of variance and orthagonal contrasts to determine if the compared variables are statistically significant. 43 RESULTS A two-way analysis of variance was performed for the tensile test, Izod impact test, and flexural modulus to determine any significance between the means, at an alpha level of .05. Also, a one-way analysis of variance with accompanying orthagonal comparisons was performed to determine any significance between variable means (Table 37, Appendix C). The results from these statistics can be found in Appendix C. en e Test Tensile strength results are located in Table 6. The fiber content for each matrix material was tested parallel (lengthwise) and perpendicular (crosswise) to the extrudate, providing a total of 12 variables. In addition, the matrix materials were ‘tested. parallel to the extrudate. IEach variable had five replications for a total of 70 samples. Tensile strength data are located in Table 19 of Appendix B. The samples in the lengthwise direction for each matrix and fiber content, exhibited significantly higher tensile strength compared to the crosswise fiber direction of the extrudate. Results for percent elongation are located in Table 7. Data for percent elongation are located in Appendix B, Table 20. 44 Table 6: Tensile strength at break (lb/i112) . Var. Fiber No. Material Direction Mean SD 1. 30% Wood fiber-PP lengthwise 3100.8 307.9 2. 30% Wood fiber-PP Reclaim lengthwise 3469.5 332.2 3. 30% Wood fiber-PP crosswise 2059.1 77.7 4. 30% Wood fiber-PP Reclaim crosswise 2522.8 187.5 5. 40% Wood fiber-PP lengthwise 2848.7 628.7 6. 40% Wood fiber-PP Reclaim lengthwise 2835.6 684.6 7. 40% Wood fiber-PP crosswise 1803.2 187.5 8. 40% Wood fiber-PP Reclaim crosswise 1619.4 58.0 9. 50% Wood fiber-PP lengthwise 2259.8 225.1 10. 50% Wood fiber-PP Reclaim lengthwise 2375.9 453.6 11. 50% Wood fiber-PP crosswise 1355.7 105.3 12. 50% Wood fiber-PP Reclaim crosswise 1238.7 190.4 13. 100% PP 2619.2 173.3 14. 100% PP Reclaim 2397.7 102.6 Table 7: Percent elongation at break (%). Var. Fiber No. Material Direction Mean SD 1. 30% Wood fiber-PP lengthwise 7.39 0.87 2. 30% Wood fiber-PP Reclaim lengthwise 11.44 1.11 3. 30% Wood fiber-PP crosswise 5.36 0.38 4. 30% Wood fiber-PP Reclaim crosswise 7.61 1.23 5. 40% Wood fiber-PP lengthwise 7.39 0.82 6. 40% Wood fiber-PP Reclaim lengthwise 5.88 0.82 7. 40% Wood fiber-PP crosswise 4.14 0.82 8. 40% Wood fiber-PP Reclaim crosswise 3.52 0.31 9. 50% Wood fiber-PP lengthwise 3.78 0.61 10. 50% Wood fiber-PP Reclaim lengthwise 3.93 0.96 11. 50% Wood fiber-PP crosswise 2.76 0.56 12. 50% Wood fiber-PP Reclaim crosswise 2.34 0.25 13. 100% PP 689.40 269.50 14. 100% PP Reclaim 214.60 43.29 Samples 'with low fiber’ content. exhibited longer' percent elongation at break. Overall, samples in the lengthwise fiber 45 direction showed significantly higher elongation when compared to the crosswise direction, for both matrices. Young's modulus of elasticity for each of the variables are located in Table 8. Tensile modulus of elasticity data are located in Table 21 of Appendix B. An increase in fiber content increased the modulus in both directions. Overall, the samples iNL the lengthwise fiber direction, displayed significantly higher tensile modulus when compared to the crosswise direction of the extrudate. Also, tensile modulus was significantly higher for each fiber content of the PP-wood fiber composite in both directions, compared. to the PP Reclaim-wood fiber composite. Table 8: Young's modulus of elasticity (lb/in?). Var. Fiber No. Material Direction Mean SD 1. 30% Wood fiber-PP lengthwise 98054 3088 2. 30% Wood fiber-PP Reclaim lengthwise 63186 4513 3. 30% Wood fiber-PP crosswise 85284 4114 4. 30% Wood fiber-PP Reclaim crosswise 59495 5998 5. 40% Wood fiber-PP lengthwise 101226 2514 6. 40% Wood fiber-PP Reclaim lengthwise 77269 7993 7. 40% Wood fiber-PP crosswise 93732 11525 8. 40% Wood fiber-PP Reclaim crosswise 63583 7822 9. 50% Wood fiber-PP lengthwise 119044 16053 10. 50% Wood fiber-PP Reclaim lengthwise 84300 8438 ll. 50% Wood fiber-PP crosswise 110116 6463 12. 50% Wood fiber-PP Reclaim crosswise 72884 6130 13. 100% PP 48842 6044 14. 100% PP Reclaim 51608 2432 46 Flexural Test Flexural modulus results are shown in Table 9. Five replications of the tangent modulus in bending were calculated to find the variable mean, for a total of 70 samples. Data from the flexural modulus test are located in Table 22 of Appendix B. The samples in the lengthwise direction showed significantly higher flexural modulus compared to the crosswise direction of the extrudate. .Additionally, flexural modulus was significantly higher in the lengthwise direction for the PP Reclaim-wood fiber composite at 40% and 50% fiber content, when compared to PP-wood fiber composite. Table 9: Flexural modulus (ft/inz). Var. Fiber No. Material Direction Mean SD 1. 30% Wood fiber-PP lengthwise 381566 70145 2. 30% Wood fiber-PP Reclaim lengthwise 302778 18038 3. 30% Wood fiber-PP crosswise 249337 26751 4. 30% Wood fiber-PP Reclaim crosswise 283459 17988 5. 40% Wood fiber-PP lengthwise 349872 40062 6. 40% Wood fiber-PP Reclaim lengthwise 449102 48934 7. 40% Wood fiber-PP crosswise 334162 43930 8. 40% Wood fiber-PP Reclaim crosswise 319691 30121 9. 50% Wood fiber-PP lengthwise 409282 24709 10. 50% Wood fiber—PP Reclaim lengthwise 459342 58922 11. 50% Wood fiber-PP crosswise 341608 29396 12. 50% Wood fiber-PP Reclaim crosswise 358206 9043 13. 100% PP 326106 23180 14. 100% PP Reclaim ' 195648 18566 47 l!£§££.22§£ Eight replications of the Izod impact testing were calculated to find the variable mean for a total of 112 samples. Izod impact testing results are located in.Table 10. Data from.the Izod impact test are located in .Appendix B, Table 23. Overall, the samples in the lengthwise direction displayed significantly higher impact strength compared to the crosswise direction of the extrudate, excluding the 30% wood fiber-PP Reclaim composite. In addition, impact strength was significantly higher in the lengthwise direction of the 30% wood fiber-PP composite compared to the 30% wood fiber-PP Reclaim composite in the lengthwise direction. Table 10: Izod impact strength (ft.lb/in). Var. Fiber No. Material Direction Mean SD 1. 30% Wood fiber-PP lengthwise 1.005 0.13 2. 30% Wood fiber-PP Reclaim lengthwise 0.788 0.11 3. 30% Wood fiber-PP crosswise 0.748 0.08 4. 30% Wood fiber-PP Reclaim crosswise 0.799 0.08 5. 40% Wood fiber-PP lengthwise 1.049 0.11 6. 40% Wood fiber-PP Reclaim lengthwise 1.145 0.07 7. 40% Wood fiber-PP crosswise 0.831 0.18 8. 40% Wood fiber-PP Reclaim crosswise 0.870 0.07 9. 50% Wood fiber-PP lengthwise 1.007 0.10 10. 50% Wood fiber-PP Reclaim lengthwise 1.043 0.04 11. 50% Wood fiber-PP crosswise 0.930 0.28 12. 50% Wood fiber-PP Reclaim crosswise 0.964 0.07 13. 100% PP 0.525 0.12 14. 100% PP Reclaim 0.641 0.12 48 ree e t Two replications were averaged for each variable in the parallel (lengthwise) direction of the extrudate, for each of the environmental conditions, for a total of 32 samples. Creep extension was reported as an average of two samples. Results after 500 hours are located in Table 11. Creep extension data for both ambient and extreme conditions are located in Table 24 and Table 25 of Appendix B, respectively. Creep extension results are displayed graphically for each environmental condition (Figures 3 and 4). Table 11. Effect of fiber content on creep extension.(soo h). Increasegin Length (in) Matrix No 30% 40% 50% and Condition Fiber Fiber Fiber Fiber Ambient PP 0.056 0.024 0.027 0.026 PP Reclaim 0.041 0.018 0.014 0.018 Extreme PP 0.126 0.094 0.181 ----- PP Reclaim 0.121 0.106 0.076 0.076 NOTE: 50% wood-PP samples failed in extreme cond. after 20 h. Creep extension did not level off after 500 h for any one of the variables. 49 .2.“ *NM .UoMNv ”SCH—.2500 Fauna-.1 2H IOHmzuth Bung-U .M H8303..— AIV m2: 000—. cow or _. ...0 rod 2: a_ _ :__:__ n :______ a :2.:. . 1.2:. _ no 1 rod can. so? -.m: [mod an. $09 Imr man—:3 $60 -6: 4 no.0 can; ion IXI m mean; 1.0% 1 v0.0 dais $o¢ IVT 00.0 AZ: ZO_wa._.xw 8:056:00 EmEE< _. 86538 820 50 .02 «a .083 982.328 2.956 an zonmzmtfi auugu .e mason“. 2.: m2: coo. o9 o. F to 8.6 11... . _ 1...... . 1...... _ 11.... . 1...... . 10 1 00.0 man. $00.. \\ ‘\ \ -.m: \+ ............ M. an. .89 [my . .. . :81586 -8: J to 1..-; .86 IX. can-..) .84 -m: 8-2. .84 1.: mails .88 -+ mnT>> $00 1 09.0 N0 AZ: ZO_me.—.xw .mcoEccoo mEotxm .. 55:93 39.0 51 Water Absorption Water sorption results for 8 variables are located in Table 12. Weight gain data from three replications were averaged for each variable, for a total of 24 samples. Data from the test are located in Appendix B, Table 26. Water sorption increased linearly with an increase in wood fiber content. Table 12. Water absorption test (%). Var. No. Composite Material Mean SD 1. 30% Wood Fiber-PP 1.65 0.05 2. 30% Wood Fiber-PPR 1.41 0.07 3. 40% Wood Fiber-PP 2.79 0.25 4. 40% Wood Fiber-PPR 2.36 0.15 5. 50% Wood Fiber-PP 3.76 0.18 6. 50% Wood Fiber-PPR 3.88 0.78 7. 100% PP 0.10 0.01 8. 100% PPR 0.24 0.01 ASTM D570 - 81, 2-hr Boiling Water Immersion Test was used. Results Summary Overall, both PP-wood fiber composites tested parallel (lengthwise) to the extrudate displayed significantly higher mechanical test results compared to the samples tested perpendicular (crosswise) to the extrudate. The virgin PP- wood fiber composite showed significantly higher results in both percent elongation at break and tensile modulus. The PP Reclaim-wood fiber composite displayed a significantly higher Izod impact strength and flexural modulus. Creep extension was more pronounced in the virgin PP-wood fiber composite, 52 particularly in extreme environmental conditions. Water absorption increased with an increase in fiber content. Lingar Regression A linear regression was performed for each mechanical test to obtain the slope of the line (excluding creep and water absorption). This could be used to predict a resulting property value, given an appropriate wood fiber concentration (i.e., less than 50% wood fiber). The mechanical tests demonstrating a good line fit (i.e., an R value ”1.0) are located in Table 13, with the corresponding equations located in Appendix A, Table 18. Table 13. Linear Regression. Mechanical Test R Value Izod Impact Strength - PP Crosswise 0.998 Izod Impact Strength - PPR Crosswise 0.997 Tensile Strength - PP Crosswise 0.967 Tensile Modulus - PP Crosswise 0.998 Tensile Modulus - PPR Crosswise 0.954 Tensile Modulus - PP Lengthwise 0.979 Tensile Modulus - PPR Lengthwise ‘ 0.976 Flexural Modulus - PPR Crosswise 0.995 Flexural Modulus - PPR Lengthwise 0.941 NOTE: Equations for the slope of each line above can be found in Appendix A, Table 18. 53 DISCUSSION Tens 1e Stren th Tensile strength as a function of wood fiber content and fiber orientation is shown in Figure 5. The tensile strength of both composite structures increased at 30% fiber content. Tensile strength for the PP Reclaim-wood fiber composite is comparable in value to a composite comprised of 30% CTMP and PP‘With addition of PP-maleated propylene wax (3% by weight of polymer) as a coupling agent (Raj et al., 1989). Although strength decreased steadily after 30% fiber content with the increase in filler concentrations for both composite structures in both.directions, the.EVOH and adhesive contained in the PP Reclaim may be contributing as coupling agents. Although the tensile strength of a composite is strongly dependent on the degree of adhesion between the fibers and the matrix, it is difficult to predict. Bigg (1987) described an upper and lower' bound response to tensile strength and empirical formulas associated with each bound for predicting the tensile strength of the composite. The lower bound response assumes weak or no adhesion between the polymer and filler; while the upper bound response assumes strong adhesion between the two materials. The formulas (not shown) can be 54 on . 20Hh> on cm 2 o ID 57 . _ . O 1 N 430$: u £5300: .3 £00.— «SVéS .1... .3 *2: .396. 5.38m ...E 1mT 1 4 2.93.: .5203. an. 1T .820. an. IT 1 o .596... an. IT 131. .2382 x322 1 m NF 35 ZO_._. .h muaunm 33 Emmi 110 ...Imu.w>> on o4 on cm 2 o . . . . O 1 com 1 84 9 1 000 5 Amway—0v chm—00m an lml .1 com 2851.... 5.33m 8.. 1.? .329 n... 1T 1 82 1 .584... a... ll _ 1 com. .3335. 5:22 684. 30.1w .ma. wDJDDOE w..:w2w... Stozmmfi .0 3:622 m.mc:o> 60 containing fibers randomly oriented in a plane: 3 5 Ema," - §EL + aET (11) 1+(2%)01Vf where EL = Em 1 ‘ "L Vf 1 + 2 :7T V,1 ET = III -——-_———— 1 17T V, E [-‘J -1 where - E'“ "L ‘ E E [-‘1-1 Elm 771’ = [E‘] _ +2 Em where E, = Modulus of the fiber E... = Modulus of the matrix material Vf = Volume fraction of fiber (.11 = Length and diameter of the fiber (Agarwal and Broutman, 1980) Calculating ER using the aforementioned empirical formulas results in the following tensile moduli (Table 15). 61 Table 15. Theoretical tensile moduli for the tested composite structures. Composite Structure Theoretical Value (PSI e+03) 30% PP 330 - 411 30% PP Reclaim 340 - 419 40% PP 451 - 552 40% PP Reclaim 463 - 562 50% PP 592 - 708 50% PP Reclaim 607 - 720 Although the theoretical tensile moduli represent composites containing randomly oriented fibers, the resulting values are more in-line with.those obtained by Woodhams et al. (1984) and Raj et a1. (1989). The percentage difference in flexural and tensile moduli is not usually as large as the difference in strength values, although a true correlation between flexural properties and other mechanical properties has never been established (Katz and Milewski, 1978). F exural Modulus The flexural modulus for the PP Reclaim-wood fiber composite increased with increasing fiber content and the PP-wood fiber composite showed varying results (Figure 8). During the application of load, one face of the sample is under compression and the other face is in tension, and the failure of fiber-reinforced composites normally occurs on the tension side (Katz and Milewski, 1978). 62 .20HF¢P2UH20 02¢ h2mh200 fluflflk 0003 E0 20Hh02=h ¢ m¢ 0040002 J¢¢axmdm .fl unawuu 33 mme “.0 ...Imu_w>> om o4 on on o. o . . . . 0 .396. 5.4.8.. an. 1m1 .595... 5.4.8.... n... 1.1 1 ooo. .320. an. IT .598... A... ll \ ..ooow \« 1 oooo .._.\ . $\ . 1 83 oooo ANO+m .wn: 031.000.). 4> o4 oo on o. o v.0 oEozxm1mdd 1m! . othoxmudd 1T1 mfio 2038(1de |.| 283.57%. l..| N0 AZ: ZO_me._.xm ammmo .5 com ..wtm 8.9.2.6. 390 68 The predicted values (Table 16) come extremely close to the actual data in the PP-wood fiber composite in ambient conditions as well as the PP Reclaim-wood fiber composite in extreme conditions. Yet, the remaining predicted values are not as accurate (Figure 11). Use of this formula to predict creep extension of a composite is not as reliable as testing. Table 16. iEffect of fiber’content on creep extension (500 h): Actual vs. predicted. Increase in Length (in) 30% 40% 50% Matrix and No Condition Fiber Actual Fred. Actual Pred. Actual Pred. Ambient (23°C, 37% RH) PP 0.056 0.024 0.028 0.027 0.027 0.026 0.023 PP Reclaim 0.041 0.018 0.033 0.014 0.027 0.018 0.025 Extreme (37°C, 94% RH) PP 0.126 0.094 0.063 0.181 0.061 ----- 0.052 PP Reclaim 0.121 0.106 0.099 0.076 0.081 0.076 0.074 NOTE: The 50% wood fiber-PP composite samples failed after 20 h. Ethylene vinyl alcohol copolymer (EVOH) , one of the structural materials in the PP Reclaim, is extremely affected by water. EVOH, above relative humidities of 80%, is plasticized to the point where its glass-transition temperature drops below room temperature and water absorption and water vapor diffusion both rise abruptly (Wachtel et al. , 1985) . The plasticization of the EVOI-I reduces hydrogen bonding, allowing segmental motion of the chains which may promote an increase in 69 .20Hh¢¢h2m0200 GNMHK h0 20Hh0202 ¢ m¢ 20Hm2mhxm ENMGU J¢0h0¢ .m> 0uhUHOmlm .HH ”Gaunt $3 mum... ....O ...Iw.m>> ._ moo 109.0 «.0 AZ: ZO_me._.Xm dmmmo 352:0. 51mm... 352:0. <15... .2335... 51m"... c.3555 <12... 352:0. 31.... 352:0. <1“... .2335... 51.... .5255 <1“... 32:28. x53: +++MM+ 830.591 .m> 2330.... 1 50.2.9.0. 30.0 70 interfacial adhesion with an increase in reaction with the hydroxyl groups (-OH) found on the cellulosic fibers. An important aspect of fatigue is that local failures in the matrix and at the weak interface can ruin the integrity of the composite.even though.the fibers remain unchanged (Agarwal and Broutman, 1980). Water Absorption Samples of the composites were immersed in boiling water and Figure 12 shows the water uptake that occurred during a 2-hour period. Water sorption was estimated using the formula: w -w w = "w ° x 100% (13) O W = Increase in weight (%) W11 = Weight of sample after removal from water Wo = Weight of dry sample where Water uptake increased with an increase in fiber content, which is what could be expected from the hydrophilic nature of cellulose fibers (Zadorecki and Flodin, 1986). Discussion Summan Wood fiber incorporated into both composites provide an improvement in tensile strength and stiffness, while the matrix provides environmental protection (e.g., moisture). Yet, the PP Reclaim-wood fiber composite has been proven to be :more useful in structural applications that will be exposed to extreme environmental conditions (i.e., temperature and RH). 71 . F2MF200 .v _ «menu coo: no zonhozou < at A». zonhaeom amh<3 .«H museum $3 ...Imu_m>> Z. m0> .4... 8:208... .255 72 CONCLUSION The PP Reclaim-wood fiber composite exhibited improved mechanical properties compared to the PP-wood fiber composite. Increase in content of wood fibers improved mechanical properties for both composites, except for tensile strength and percent elongation at break which decreased after 30% fiber content. Overall, orientation of wood fibers displayed significantly improved mechanical properties for both composites, at each fiber concentration. The highest flexural modulus (459,342 PSI) was achieved in the 50% wood fiber-PP Reclaim composite (lengthwise), 18% higher than the PP-wood fiber composite. Izod impact strength for both composites generally increased with an increase in fiber content. This suggests an increase in interfacial adhesion. The PP Reclaim—wood fiber composite also displayed longer retention of interfacial strength under stress in both ambient and extreme environmental conditions. 73 RECOMMENDATIONS Since the PP Reclaim matrix offers excellent dimensional stability and improvements in mechanical properties, further research should be carried out in the areas of 1) improving fiber-matrix bonding at the interface, 2) the effect of the mixing apparatus, and 3) ultimate use for an injection molding composite. The addition of Epolene wax (maleated propylene wax) to the pulp appears to improve bonding of the fiber-matrix interface when used as a coupling agent for 40% CTMP aspen fiber-PP composite. IResults indicated an improvement in 'tensile strength by 35% (Raj et al., 1989). The similarity of the isotactic structures permits segmental crystallization to occur, whereas the carboxyl groups provide polar or chemical attachments to the cellulosic fibers (Woodhams et al., 1984). It.has also been shown that the average fiber length and fiber matrix bond are affected by the conditions under which the fibers are mixed into the polymer using a co-rotating twin screw extruder (Wall, 1989). The length of the mixing section is the most significant parameter. 74 Additionally, compounds must be in granulated form to be capable of injection molding; Therefore, composite materials could be granulated prior to injection molding into tensile and impact bars to improve the accuracy of the mechanical properties and provide more precise end-use data. APPENDIX A 75 APPENDIX A Table 17. Actual composite composition by weight. Treatments Composite Wood Fiber (9) Matrix (9) 30% Wood Fiber 1 - Virgin PP 143.43 322.32 2 - PP Reclaim 206.18 630.00 40% Wood Fiber 3 - Virgin PP 173.94 256.47 4 - PP Reclaim 340.28 540.00 50% Wood Fiber 5 - Virgin PP 248.00 224.90 6 - PP Reclaim 450.00 424.84 7 100% Virgin PP 900.00 8 100% PP Reclaim 900.00 Table 18. Equations for Linear Regression. Mechanical Test Equation Izod Impact Strength - PP Crosswise y = 0.5209 + 7.686e03x Izod Impact Strength - PPR Crosswise y = 0.6413 + 6.176e03x Tensile Strength - PP Crosswise y = 26.748 - 0.23115x Tensile Modulus - PP Crosswise y = 48.827 + 1.1538x Tensile Modulus - PPR Crosswise y = 50.631 + 0.39252x Tensile Modulus - PP Lengthwise y = 50.760 + 1.3273x Tensile Modulus - PPR Lengthwise y = 50.318 + 0.65434x Flexural Modulus - PPR Crosswise y = 19.900 + 0.31456x Flexural Modulus - PPR Lengthwise y = 19.226 + 0.55580x Y x Mechanical Property Percent Wood Fiber APPENDIX B 76 APPENDIX B Table 19. Tensile strength data from tensile test (lb/inz) . Replications Var e No. Composite 1 2 3 4 s 1. 30% W-PP (LW) 2938.7 2971.9 3472.6 3372.6 2748.4 2. 30% W-PPR (LU) 4024.0 3200.0 3506.4 3245.0 3371.9 3. 30% 9.22 (cw) 1975.9 2050.8 2016.0 2070.4 2182.3 4. 30% w-rrn (09) 2703.1 2271.3 2503.7 2425.6 2710.1 5. 40% 9.22 (19) 3093.9 2172.7 2480.0 3800.0 2696.9 6. 40% w-rrn (Lu) 3131.8 2338.5 3861.1 2695.4 2151.1 7. 40% w-rr (cw) 1728.0 1938.2 1998.3 1827.6 1523.9 8. 4o: W-PPR (cw) 1643.7 1524.8 1681.6 1622.2 1624.8 9. so: w-rr (1w) 2292.2 2593.8 2062.0 2310.9 2040.0 10. so: w-rrn (cw) 2186.4 2250.4 1856.5 3070.2 2516.0 11. so: w-rr (cw) 1343.2 1412.4 1430.7 1414.9 1177.3 12. so: w-rrn (CW) 1320.8 1308.5 1150.5 958.3 1455.6 13. 100% PP 2741.9 2766.7 2451.6 2725.8 2409.8 14. 100% PPR 2292.9 2469.7 2523.1 2297.1 2405.8 W - wood Fiber PP - Polypropylene PPR - Polypropylene Reclaim LW - Lengthwise direction of the extrudate CW - Crosswise direction of the extrudate 77 Table 20. Percent elongation data from tensile test (%). Replications Var. Mo. Composite 1 2 3 4 5 l 30% W-PP (EU) 8.15 6.30 8.30 6.75 7.45 2. 30% W-PPR (LU) 12.50 9.60 11.90 11.90 11.30 3. 30% W-PP (CW) 5.35 5.35 4.75 5.75 5.60 4. 30% W-PPR (CW) 8.85 5.95 7.25 7.20 8.80 5. 40% W-PP (LN) 6.50 8.35 7.50 8.00 6.60 6 40% W-PPR (LU) 6.85 5.00 6.65 5.55 5.35 7. 40% W—PP (CW) 4.55 4.75 4.70 3.90 2.80 8. 40% W-PPR (CW) 3.20 3.50 3.95 3.25 3.70 9. 50% W-PP (LW) 4.50 4.15 3.40 3.90 2.95 10. 50% W-PPR (LU) 3.35 2.85 3.70 5.30 4.45 11. 50% W-PP (CW) 2.55 2.80 3.60 2.80 2.05 12. 50% W-PPR (CW) 2.35 2.25 2.40 2.00 2.70 13. 100% PP 1055.50 911.50 699.50 369.40 411.35 14. 100% PPR 242.35 268.35 140.60 220.90 200.65 W - 900d Fiber PP - Polypropylene PPR - Polypropylene Reclaim EU - Lengthwise direction of the extrudate CW - Crosswise direction of the extrudate 78 Table 21. Young's modulus of elasticity data (lb/inz) . Replications Var. Mo. Composite 1 2 3 4 5 l. 30% W-PP (EN) 98387 101562 100000 93548 96774 2. 30% W—PPR (LW) 64800 67460 66935 58140 58594 3. 30% W-PP (CW) 85714 90476 86400 84800 79032 4. 30% WkPPR (CW) 68750 52713 56296 60800 58915 5. 40% W-PP (LW) 100000 104545 98461 100000 103125 6. 40% W-PPR (LW) 71212 68461 84733 86364 75573 7. 40% W-PP (CW) 81600 97561 104132 81301 104065 8. 40% W-PPR (CW) 74219 56391 68000 63492 55814 9. 50% W-PP (LW) 93750 112308 128682 131250 129231 10. 50% W-PPR (LW) 83333 91603 70992 91603 83969 11. 50% W-PP (CW) 113600 119355 108800 105600 103226 12. 50% W-PPR (CW) 75591 76613 69841 63636 78740 13. 100% PP 54839 50000 40984 44355 54032 14. 100% PPR 49643 55344 52692 50725 49635 W - W00d Fiber PP - Polypropylene PPR - Polypropylene Reclaim LW - Lengthwise direction of the extrudate CW - Crosswise direction of the extrudate 79 Table 22. Flexural modulus of elasticity data (lb/inf). Replications Var. No. Composite 1 2 3 4 5 1. 30% W—PP (EW) 392336 408158 337443 293054 476838 2. 30% W-PPR (LW) 294851 310233 294346 330386 284076 3. 30% W—PP (CW) 260841 225714 223966 288401 247765 4. 30% W-PPR (CW) 272638 312989 266140 281309 284221 5. 40% W-PP (LW) 386381 324082 371219 375385 292293 6. 40% W-PPR (LW) 470836 494770 406469 387240 486193 7. 40% W-PP (CW) 386552 287697 360454 346385 289723 8. 40% W-PPR (CW) 337490 301665 275321 338510 345471 9. 50% W-PP (LW) 434393 411127 391764 431551 377575 10. 50% W-PPR (LW) 475059 552302 448470 417412 403466 11. 50% W-PP (CW) 296619 375117 358543 336900 340860 12. 50% W-PPR (CW) 355598 364777 344568 367753 358334 13. 100% PP 317153 346737 329668 290966 346005 14. 100% PPR 215038 200573 203028 165277 194322 W - W0od Fiber PP - Polypropylene PPR - Polypropylene Reclaim EW - Lengthwise direction of the extrudate CW - Crosswise direction of the extrudate 80 Table 23. Data from Izod impact test (ft.lb/in). Replications Var. No. Composite 1 2 3 4 5 6 7 8 l. 30% W-PP (LW) 0.877 1.192 1.183 0.886 1.110 0.974 .118 .101 2. 30% W-PPR (LW) 0.730 0.816 1.002 0.883 0.667 0.749 .742 .716 3. 30% W-PP (CW) 0.766 0.741 0.803 0.656 0.735 0.791 .877 .619 4. 30% W-PPR (CW) 0.645 0.766 0.779 0.779 0.864 0.914 .830 .816 5. 40% W-PP (LW) 1.004 1.045 1.192 0.870 0.929 1.136 .126 .090 6. 40% W-PPR (LW) 1.099 1.084 1.154 1.209 1.146 1.285 .088 .097 7. 40% W-PP (CW) 0.797 0.943 0.592 1.118 0.667 0.735 .772 .025 8. 40% W-PPR (CW) 0.978 0.891 0.829 0.859 0.936 0.863 .747 .854 9. 50% W-PP (LW) 0.929 1.065 1.082 0.943 1.045 1.084 .825 .082 10. 50% W-PPR (LW) 1.042 1.082 1.038 1.071 0.982 1.095 .001 .033 11. 50% W-PP (CW) 0.760 0.779 0.816 1.609 0.882 0.741 .877 .974 12. 50% W-PPR (CW) 1.030 0.914 0.893 1.091 0.944 1.001 .936 .907 13. 100% PP 0.513 0.622 0.602 0.640 0.349 0.443 .636 .397 14. 100% PPR 0.727 0.784 0.498 0.684 0.518 0.501 .651 .766 W - Wood Fiber PP - Polypropylene PPR - Polypropylene Reclaim LW - Lengthwise direction of the extrudate CW - Crosswise direction of the extrudate 81 Table 24. Creep elongation data (in.) in ambient conditions. Time N0. 0.017 0.1 0.2 0.5 1 2 5 20 50 100 200 500 A-l .003 .005 .007 .009 .009 .013 .013 .015 .018 .023 .026 .033 -2 .001 .002 .003 .003 .006 .006 .006 .007 .009 .009 .011 .014 Avg .002 .004 .005 .006 .008 .010 .010 .011 .014 .016 .019 .024 8-1 .002 .007 .010 .010 .010 .012 .015 .015 .018 .019 .022 .025 -2 .002 .003 .004 .004 .005 .006 .006 .006 .008 .008 .010 .010 Avg .002 .005 .007 .007 .008 .009 .011 .011 .013 .014 .016 .018 C-1 .009 .011 .012 .013 .014 .017 .017 .019 .022 .026 .028 .034 -2 .007 .008 .008 .009 .010 .010 .011 .011 .012 .014 .016 .019 Avg .008 .010 .010 .011 .012 .014 .014 .015 .017 .020 .022 .027 D-l .004 .006 .008 .008 .008 .010 .010 .012 .013 .014 .016 .018 -2 .002 .003 .004 .005 .005 .005 .005 .006 .007 .007 .008 .009 Avg .003 .005 .006 .007 .007 .008 .008 .009 .010 .011 .012 .014 E-l .012 .013 .013 .015 .015 .016 .018 .019 .021 .025 .026 .032 -2 .010 .011 .011 .011 .012 .013 .013 .013 .013 .016 .017 .020 Avg .011 .012 .012 .013 .014 .015 .016 .016 .017 .021 .022 .026 F-l .011 .012 .013 .013 .013 .013 .013 .014 .014 .015 .016 .018 -2 .005 .006 .008 .009 .009 .009 .010 .011 .011 .011 .013 .017 Avg .008 .009 .011 .011 .011 .011 .012 .013 .013 .013 .015 .018 C-1 .019 .021 .022 .025 .027 .031 .034 .038 .044 .048 .052 .062 -2 .014 .019 .021 .022 .022 .023 .029 .032 .037 .042 .046 .049 Avg .017 .020 .022 .024 .025 .027 .032 .035 .041 .045 .048 .056 H—l .017 .018 .019 .021 .023 .024 .027 .030 .031 .038 .040 .047 -2 .006 .009 .012 .013 .015 .018 .020 .020 .023 .026 .031 .034 Avg .012 .014 .016 .017 .019 .021 .024 .025 .027 .032 .036 .041 30% Wood Fiber-PP 30% Wood Fiber-PPR 40% Wood Fiber-PP 40% Wood Fiber-PPR 50% Wood Fiber-PP 50% Wood Fiber-PPR 100% PP 100% PPR 2350, 37% Relative Humidity. IOWMUOG> I 82 Table 25. Creep elongation data (in.) in extreme conditions. Time N0. 0.017 0.1 0.2 0.5 1 2 5 20 50 100 200 500 A-l .019 .020 .022 .024 .026 .030 .034 .039 .041 .045 .050 .064 -2 .043 .045 .047 .053 .055 .059 .065 .078 .087 .096 .106 .124 Avg .031 .033 .035 .040 .041 .045 .050 .059 .064 .071 .078 .094 B-l .031 .037 .038 .040 .044 .048 .054 .063 .072 .083 .090 .105 -2 .036 .039 .041 .042 .046 .051 .056 .069 .077 .085 .094 .107 Avg .034 .038 .040 .041 .045 .050 .055 .066 .075 .084 .092 .106 C-1 .018 .024 .028 .031 .032 .038 .048 .066 .083 .102 .133 .208 -2 .022 .027 .028 .032 .036 .039 .048 .062 .078 .093 .116 .154 Avg .020 .026 .028 .032 .034 .039 .048 .064 .081 .098 .125 .181 D-l .020 .022 .023 .024 .026 .028 .032 .041 .049 .056 .066 .083 -2 .012 .013 .017 .018 .019 .020 .024 .028 .039 .045 .055 .068 Avg .016 .018 .020 .021 .023 .024 .028 .035 .044 .051 .061 .076 E-l .026 .029 .033 .035 --- --- --- --- --- --- --- --- -2 .015 .017 .018 .019 .022 .024 .030 .055 --- --- --- --- Avg .021 .023 .026 .027 .022 .024 .030 .055 --- --- --- --- F-l .001 .013 .014 .015 .024 .026 .029 .043 --- --- --- --- -2 .019 .022 .024 .026 .027 .029 .032 .039 .047 .051 .061 .076 Avg .010 .018 .019 .021 .026 .028 .031 .041 .047 .051 .061 .076 G-l .042 .052 .057 .066 .076 .084 .097 .107 .116 .120 .123 .129 -2 .033 .045 .050 .057 .067 .078 .089 .101 .110 .116 .119 .123 Avg .038 .049 .054 .062 .072 .081 .093 .104 .113 .118 .121 .126 H-l .050 .055 .060 .064 .071 .078 .090 .099 .111 .115 .118 .129 -2 .038 .042 .047 .056 .061 .068 .077 .092 .098 .103 .107 .112 Avg .044 .049 .054 .060 .066 .073 .084 .096 .105 .109 .113 .121 A - 30% Wood 816652? 37W, 4 Relative fiumidity. B - 30% Wood Fiber-PPR C - 40% Wood Fiber-PP D - 40% Wood Fiber-PPR E - 50% Wood Fiber-PP F - 50% Wood Fiber-PPR C - 100% PP H - 100% PPR 83 Table 26. Water absorption data (%). Replications Composite 1 3 5 Average SD 30% Wood Fiber-PP 1.62 1.71 .61 1.65 0.05 30% W00d Fiber—PPR 1.49 1.39 .36 1.41 0.07 40% Wead Fiber-PP 2.89 2.51 .98 2.79 0.25 40% Wood Fiber-PPR 2.19 2.39 .50 2.36 0.15 50% Wood Fiber-PP 3.88 3.84 .55 3.76 0.18 50% Wood Fiber-PPR 3.10 4.65 .83 3.88 0.78 100% PP 0.09 0.11 .11 0.10 0.01 100% PPR 0.25 0.24 .24 0.24 0.01 APPENDIX C 84 APPENDIX C Table 27. Two-way Analysis of Variance over variable 1 (Treatments) with values from 1 to 14 and variable 2 (Replications) with values from 1 to 5. Variable 3: Tensile strength. ANALYSIS OF VARIANCE TABLE Degrees of Sum of Mean Source Freedom Squares Square F-value Probability Treatments 13 27723199.81 2132553.832 20.16 0.0000 Replications 4 470792.75 117698.187 1.11 0.3607 Error 52 5501681.31 105801.564 Non-additivity l 544904.98 544904.978 5.61 Residual 51 4956776.33 97191.693 Total 69 33695673 87 Grand Mean - 2321.860 Grand Sum - 162530.210 Total Count - 70 Coefficient of Variation - 14.01% Means for Tensile Strength Treatments Treatments Treatments (Var. No.) Mean (Var. No.) Mean (Var. No.) Mean 1 3100.84 6 2835.58 11 1355.70 2 3469.46 7 1803.20 12 1238.74 3 2059.08 8 1619.42 13 2619.16 4 2522.76 9 2259.78 14 2397.72 5 2848.70 10 2375.90 85 Table 28. Two-way Analysis of Variance over variable 1 (Treatments) with values from 1 to 12 and variable 2 (Replications) with values from 1 to 5. Variable 3: Percent elongation at break. ANALYSIS OF VARIANCE TABLE Degrees of Sum of Mean Source Freedom Squares Square F-value Probability Treatments 11 378.56 34.415 57.66 0.0000 Replications 4 3.53 0.882 1.48 0.2254 Error 44 26.26 0.597 Non-additivity l 3.00 2.999 5 54 Residual 43 23.26 0.541 Total 59 408.35 Grand Mean - 5.462 Grand Sum - 327.700 Total Count - 60 Coefficient of Variation - 14.15% Means for Percent Elongation Treatments Treatments Treatments (Var. No.) Mean (Var. No.) Mean (Var. No.) Mean 1 7.390 6 5.880 11 2.760 2 11.440 7 4.140 12 2.340 3 5.360 8 3 520 13 not used* 4 7.610 9 3.780 14 not used* 5 7.390 10 3.930 * Means for PP and PPR were not used because their percentages were greater than 100% (much higher). 86 Table 29. Two-way Analysis of Variance over variable 1 (Treatments) with values from 1 to 14 and variable 2 (Replications) with values from 1 to 5. Variable 3: Tensile modulus. ANALYSIS OE VARIANCE TABLE Degrees of Sum of Mean Source Freedom Squares Square F—value Probability Treatments 13 31226656706.47 2402050515.882 39.74 0.0000 Replications 4 47533121.09 11883280.27l 0.20 0.9391 Error 52 3143425265.31 60450485.87l Non-additivity 1 164435095.14 l64435095.138 2.82 Residual 51 2978990170.18 58411571.964 Total 69 34417615092.87 Grand Mean - 80615.957 Grand Sum - 5643117.000 Total Count - 70 Coefficient of Variation - 9.64% Means for Tensile Modulus Treatments Treatments Treatments (Var. No.) Mean (Var. No.) Mean (Var. No.) Mean 1 98054.2 6 77268.6 11 110116.2 2 63185.8 7 93731.8 12 72884.2 3 85284.4 8 63583.2 13 48842.0 4 59494.8 9 119044.2 14 51607.8 5 101226.2 10 84300.0 87 3 : Flexural modulus . ANALYSIS OF VARIANCE TABLE 33692006.96 Table 30. Degrees of Source Freedom Treatments l3 Replications 4 Error 52 Non-additivity 1 Residual 51 Total 69 4 Grand Mean - 3400.114 Grand Sum - 238007.959 Total Count - 70 Sum of Squares 556434. 7056235. 568562. 6487672. Coefficient of Variation - 10.83% Means (e+002) for Flexural Modulus Treatments (Var. No.) Mean 1 3815.658 2 3027.784 3 2493.374 4 2834.594 5 3498.720 Treatments (Var. No.) Mean Square F-value Probability 259l692.843 60 139108.649 01 135696.827 39 568562.395 62 127209.267 4491.016 3341.622 3196.914 4092.820 4593.418 39.74 1.03 4.47 Treatments (Var. No.) 11 12 13 14 Two-way Analysis of Variance over variable 1 (Treatments) with values from 1 to 14 and variable 2 (Replications) with values from 1 to 5. Variable 3416.078 3582.060 3261.058 1956.476 88 Table 31. Two-way Analysis of Variance over variable 1 (Treatments) with values from 1 to 14 and variable 2 (Replications) with values from 1 to 8. Variable 3: Izod impact strength. ANALYSIS OF VARIANCE TABLE Degrees of Sum of Mean Source Freedom Squares Square F-value Probability Treatments 13 3.19 0 246 15.66 0 0000 Replications 7 0.11 0.016 1.04 0 4069 Error 91 1.43 0.016 Non-additivity 1 0.01 0 015 0 95 Residual 90 1.41 0.016 Total 111 4 74 Grand Mean - 0.885 Grand Sum - 99.171 Total Count - 112 Coefficient of Variation - 14.15% Means for Impact Strength Treatments Treatments Treatments (Var. No.) Mean (Var. No.) Mean (Var. No.) Mean 89 Table 32. One way.hnalysis of Variance grouped over variable 1 (Treatments) with values from 1 to 14. Variable 3: Tensile strength. ANALYSIS 0? VARIANCE TABLE Degrees of Sum of Mean Source Freedom Squares Square F-value Probability Between 13 27723199.811 2132553.832 19.996 0.0000 Within 56 5972474.058 106651.322 Total 69 33695673.869 Coefficient of Variation - 14.07% Var. Tensile Strength (1b./in2) 1 Number Sum Average SD SE 1 5 15504 200 3100.840 307 89 146 05 2 5 17347 300 3469.460 332 19 146 05 3 5 10295 400 2059.080 77 71 146 05 4 5 12613 800 2522.760 187 52 146 05 5 5 14243 500 2848.700 628 70 146 05 6 5 14177 900 2835 580 684 60 146 05 7 5 9016 000 1803 200 187 45 146 05 8 5 8097 100 1619.420 57 98 146 05 9 5 11298 900 2259 780 225 09 146 05 10 5 11879 500 2375 900 453 63 146 05 11 5 6778 500 1355 700 105 27 146 05 12 5 6193 700 1238 740 190 43 146 05 13 5 13095 810 2619 162 173 29 146 05 14 5 11988 600 2397 720 102 56 146 05 Total 70 162530.210 2321.860 698.82 83.52 Within 326.58 Bartlett's test Chi-square - 47.208 Number of Degrees of Freedom - 13 Approximate significance - 0.000 90 Table 33. One way Analysis of Variance grouped over variable 1 (Treatments) with values from.1 to 12. Variable 3: Percent elongation at break. ANALYSIS OF VARIANCE TABLE Degrees of Sum of Mean Source Freedom Squares Square F-value Probability Between 11 378.564 34.415 55.456 0.0000 Within 48 29.788 0.621 Total 59 408.352 Coefficient of Variation - 14.42% Var. Elongation at Break (%) 1 Number Sum Average SD SE 1 5 36.950 7.390 0.87 0.35 2 5 57.200 11.440 1.11 0.35 3 5 26.800 5.360 0.38 0.35 4 5 38.050 7.610 1.23 0.35 5 5 36.950 7.390 0.82 0.35 6 5 29.400 5.880 0.82 0.35 7 5 20.700 4 140 0.82 0.35 8 5 17.600 3.520 0.31 0.35 9 5 18.900 3.780 0.61 0.35 10 5 19.650 3.930 0.96 0.35 11 5 13.800 2.760 0.56 0.35 12 5 11.700 2.340 0.25 0.35 Total 60 327.700 5.462 2.63 0.34 Within 0.79 Bartlett's test Chi-square - 15.833 Number of Degrees of Freedom - 11 Approximate significance - 0.000 91 Table 34. One way Analysis of Variance grouped over variable 1 (Treatments) with values from 1 to 14. Variable 3: Tensile modulus. ANALYSIS OF VARIANCE TABLE Degrees of Sum of Mean Source Freedom Squares Square F-value Probability Between 13 31226656706.471 2402050515.882 42.155 0.0000 Within 56 3190958386.400 56981399.757 Total 69 34417615092.871 Coefficient of Variation - 9.36% Var Tensile Modulus (1b./in2) 1 Number Sum Average SD SE 1 5 490271.000 98054.200 3088 13 3375.84 2 5 315929.000 63185.800 4513 19 3375.84 3 5 426422.000 85284.400 4114.27 3375.84 4 5 297474.000 59494.800 5998.41 3375.84 5 5 506131.000 101226.200 2513 63 3375 84 6 5 386343.000 77268.600 7993.35 3375 84 7 5 468659.000 93731.800 11525.05 3375 84 8 5 317916.000 63583.200 7821 84 3375 84 9 5 595221.000 119044.200 16053 14 3375 84 10 5 421500.000 84300.000 8438.23 3375 84 11 5 550581.000 110116.200 6463.27 3375 84 12 5 364421.000 72884.200 6130.08 3375 84 13 5 244210.000 48842.000 6044.06 3375 84 14 5 258039.000 51607.800 2432.41 3375 84 Total 70 5643117.000 80615 957 22333 97 2669.42 Within 7548 60 Bartlett's test Chi-square - 27.033 Number of Degrees of Freedom - 13 Approximate significance - 0.000 92 Table 35. One way Analysis of Variance grouped over variable 1 (Treatments) with values from 1 to 14. Variable 3: Flexural modulus. ANALYSIS OF VARIANCE TABLE Degrees of Sum of Mean Source Freedom Squares Square F-value Probability Between 13 33692006.963 259l692.843 19.065 0.0000 Within 56 7612669.607 135940.529 Total 69 41304676.570 Coefficient of Variation - 10.84% Var. Flexural Modulus (lb. /in2 e+002) 1 Number Sum Average SD SE 1 5 19078 290 3815.658 701 45 164 89 2 5 15138 920 3027 784 180 38 164 89 3 5 12466 870 2493 374 267 51 164 89 4 5 14172 970 2834.594 179 88 164 89 5 5 17493 600 3498 720 400 62 164 89 6 5 22455 080 4491.016 489 34 164 89 7 5 16708 110 3341.622 439 30 164 89 8 5 15984 570 3196 914 301 21 164 89 9 5 20464 100 4092.820 247 09 164 89 10 5 22967 090 4593.418 589 22 164 89 11 5 17080 390 3416.078 293 96 164 89 12 5 17910 300 3582.060 90 43 164 89 13 5 16305 290 3261.058 231 80 164 89 14 5 9782 380 1956 476 185 66 164 89 Total 70 238007.959 3400.114 773.70 92.48 Within 368.70 Bartlett's test Chi-square - 25.592 Number of Degrees of Freedom - 13 Approximate significance - 0.000 93 Table 36. One way Analysis of Variance grouped over variable 1 (Treatments) with values from 1 to 14. Variable 3: Izod Impact strength. ANALYSIS OF VARIANCE TABLE Degrees of Sum of Mean Source Freedom Squares Square F-value Probability Between 13 3.193 0.246 15.607 0.0000 Within 98 1.542 0.016 Total 111 4 735 Coefficient of Variation - 14.17% Var. Impact Strength (ft.lb./in.) 1 Number Sum Average SD SE 1 8 8.441 1.055 0.13 0.04 2 8 6.305 0.788 0.11 0.04 3 8 5.988 0.748 0.08 0.04 4 8 6.393 0.799 0.08 0.04 5 8 8.392 1.049 0.11 0.04 6 8 9.162 1.145 0.07 0.04 7 8 6.649 0.831 0.18 0.04 8 8 6 957 0.870 0.07 0.04 9 8 8.055 1.007 0.10 0.04 10 8 8.344 1.043 0.04 0.04 11 8 7.438 0.930 0.28 0.04 12 8 7.716 0.964 0.07 0.04 13 8 4.202 0.525 0.12 0.04 14 8 5.129 0.641 0.12 0.04 Total 112 99.171 0.885 0.21 0.02 Within 0.13 Bartlett's test Chi-square - 43.554 Number of Degrees of Freedom - 13 Approximate significance - 0.000 94 'Table 37. Coefficients for the partitioning of the sum of squares among fourteen treatments into fifteen independent (orthagonal) comparisons. VTreatments 1 Response to W -1 -1 -1 -1 -1 -1 -1 -1 2 30% W vs. All -5 -5 -5 -5 2 2 2 2 3 40% W vs. All 2 2 2 2 -5 -5 -5 -5 4. 50% W vs. All 2 2 2 2 2 2 2 2 5. LW vs. CW -1 -1 1 1 -1 -1 1 1 6 30% LW vs. CW -1 -1 1 1 0 0 0 0 7 40% LW vs. CW 0 0 0 0 -1 -1 l 1 8 50% LW vs. CW 0 0 0 0 0 0 0 O 9. 3PL vs. 3RL -1 1 0 0 0 0 0 0 10. 3P6 vs. 3R0 0 0 -1 1 0 0 0 0 ll. 4PL vs. 4RL 0 0 0 0 -1 1 0 0 12. 4PC vs. 4R6 0 0 O 0 O 0 -1 1 l3. SPL vs. SRL 0 0 O 0 O 0 O 0 l4. 5P0 vs. 5R0 0 0 0 0 0 0 0 0 15. PP vs. PPR 0 0 O 0 0 0 0 0 OOHOOCO l -1 -1 -1 2 2 2 2 2 2 -5 -5 -5 -l 1 l 0 0 0 0 0 0 -l l 1 0 0 0 0 0 0 0 0 0 0 0 0 l 0 0 0 -1 1 0 0 0 HOODOCOOOOONMNO‘ HOOOOOOOOOOMNNCh W - Wood Fiber, DW - Lengthwise, CW - Crosswise, and PL, RL, PC, RC - Polypropylene (PP) lengthwise, PP Reclaim (PPR) lengthwise, PP crosswise, and PPR crosswise, respectfully. Numbers 3, 4, 5, and 10 - 30%, 40%, and 50% wood fiber content, and 100% polymer, respectfully. NOTE: Percent elongation for comparison number 1 was not tabulated due to exceedingly high values (>> 100%). ORTEAGONAL COMPARISONS 1. Response to wood fiber. Tens. Strength Elong. Tens. Mod. Sum of Squares 406144.938 10775524133.038 Effect 31.097 -5065.176 Error 15.935 368.334 F value 3.808 189.106 Probability 0.056 0.000 Flex. Mod. Izod 7306012.264 -131.891 17.991 53.744 0.000 0.005 108.369 0.000 95 12” 30% wood fiber content vs. the remaining treatments. Tens. Strength Elong. Tens. Mod. Flex. Mod. Izod Sum of Squares 6084932.107 185.754- 473245165.491 3573796.3951 0.064 Effect -186.470 -1.244 1644.463 142.904 0.015 Error 24.687 0.072 570.621 27.871 0.007 F value 57.054 299.322 8.305 26.289 4.054 Probability 0.000 0.000 0.006 0.000 0.047 3. 40% wood fiber content vs. the remaining treatments. Tens. Strength Elong. Tens. Med. Flex. Mod. Izod Sum.of Squares 57041.079 1.576 311701168.401 1506478.306 0.349 Effect 18.054 0.115 -1334.597 -92.782 -0.035 Error 24.687 0.072 570.621 27.871 0.007 F value 0.535 2.539 5.470 11.082 22.192 Probability 0.118 0.023 0.002 0.000 4. 50% wood fiber content vs. the remaining treatments. Tens. Strength Elong. Tens. Mod. Flex. Mod. Izod Sum.of Squares 7406994.267 153.115 7141317677.041 7599772.876 0.453 Effect 205.732 1.130 -6388.077 -208.392 -0.040 Error 24.687 0.072 570.621 27.871 0.007 F value 69.451 246.728 125.327 55.905 28.795 Probability 0.000 0.000 0.000 0.000 0.000 5. Treatments in the lengthwise direction vs. treatments in crosswise. Tens. Strength Elong. Tens. Mod. Flex. Mod. Izod Sum of Squares 16492170.067 82.603 l400912768.067 9027884.595 0.595 Effect -524.280 -1.l73 -4832.033 -387.898 -0.079 Error 42.161 0.102 974.520 47.599 0.013 F value 154.636 133.105 24.585 66.411 37.809 Probability 0.000 0.000 0.000 0.000 0.000 Table 37 (cont'd.) 96 6. Treatments with 30% wood fiber in lengthwise direction vs. 30% wood fiber content in the crosswise direction. Tens. Strength Elong. Tens. Mod. Flex. Med. Izod Sum of Squares 4942465 . 906 42 . 924 338697420 . 800 2870827 . 450 0 . 175 Effect —497.115 -1.465 -4115.200 -378.869 -0.074 Error 73.024 0.176 1687.919 82.444 0.022 F value 46.342 69.168 5.944 21.118 11.106 Probability 0.000 0.000 0.018 0.000 0.001 7. Treatments with 40% wood fiber in lengthwise direction vs. 40% wood fiber content in the crosswise direction. Tens. Strength Elong. Tens. Mod. Flex. Med. Izod Sum of Squares 6393882.362 39.340 560729910.050 2632476.977 0.487 Effect -565.415 -1.403 -5294.950 -362.800 -0.123 Error 73.024 0.176 1687.919 82.444 0.022 F value 59.951 63.392 9.841 19.365 30.950 Probability 0.000 0.000 0.003 0.000 0.000 8. Treatments with 50% wood fiber in lengthwise direction vs. 50% wood fiber content in the crosswise direction. Tens. Strength Elong. Tens. Mod. Flex. Mod. Izod Sum.of Squares 5208325.561 18.515 517337748.050 3562102.013 0.048 Effect -510.310 -0.653 -5085.950 -422.025 -0.039 Error 73.024 0.176 1687.919 82.444 0.022 F value 48.835 13.721 9.079 26.203 3.078 Probability 0.000 0.001 0.004 0.000 0.082 9. 30% wood fiber-PP in lengthwise vs. 30% wood fiber-PP Reclaim in lengthwise direction. Tens. Strength Elong. Tens. Mod. Flex. Med. Izod Sum.of Squares 339701.707 41.006 3039513296.400 1551863.500 0.285 Effect 184.310 2.025 -17434.200 -393.937 -0.133 Error 103.272 0.249 2387.078 116.594 0.031 P value 3.185 66.077 53.342 11.416 18.119 Probability 0.080 0.000 0.000 0.001 0.000 Table 37 (cont'd.) 97 110. 30% wood fiber-PP in crosswise vs. 30% wood fiber-PP Reclaim in the crosswise direction. Tens. Strength Elong. Tens. Mbd. Flex. Mod. Izod Sum of Squares 537497.981 12.656 1662758670.400 291077.754. 0.010 Effect 231.840 1.125 ~12894.800 170.610 0.025 Error 103.272 0.249 2387.078 116.594 0.031 F value 5.040 20.349 29.181 2.141 0.651 Probability 0.029 0.000 0.000 0.149 11. 40% wood fiber-PP in lengthwise vs. 40% wood fiber-PP Reclaim in the lengthwise direction. Tens. Strength Elong. Tens. Mod. Flex. Mod. Izod Sum of Squares 430.331 5.700 1434916494.400 2461628.360 0.037 Effect -6.560 -0.755 -11978.800 496.148 0.048 Error 103.272 0.249 2387.078 116.594 0.031 F value 0.004 9.185 25.182 18.108 2.355 Probability 0.004 0.000 0.000 0.128 12. 40% wood fiber-PP in crosswise vs. 40% wood fiber-PP Reclaim in the crosswise direction. Tens. Strength Elong. Tens. Med. Flex. Mod. Izod Sum of Squares 84437.725 0.961 2272345204.900 52351.054 0.006 Effect -91.890 -0.310 -15074.300 -72.354 0.019 Error 103.272 0.249 2387.078 116.594 0.031 F value 0.792 1.549 39.879 0.385 0.377 Probability 0.219 0.000 13. 50% wood fiber-PP in lengthwise vs. 50% wood fiber-PP Reclaim in the lengthwise direction. Tens. Strength Elong. Tens. Med. Flex. Med. Izod Sum of Squares 33709.619 0.056 3017898584.100 626496.011 0.005 Effect 58.060 0.075 -17372.100 250.299 0.018 Error 103.272 0.249 2387.078 116.594 0.031 P value 0.316 0.091 52.963 4.609 0.332 Probability 0.000 0.036 Table 37 (cont'd.) ‘14» 50% wood fiber-PP in crosswise vs. crosswise direction. Tens. Strength 98 Flex. Mod. 50% wood fiber-PP Reclaim in the Izod Sum of Squares 34199.103 Effect -58.480 Error 103.272 P value 0.321 Probability 15. 100% PP vs. Tens. Strength 100% PP Reclaim. Elong. Tens. Mod. 0.441 3465554560.000 -0.210 -18616.000 0.249 2387.078 0.711 60.819 0.000 Elong. Tens. Mod. 68875.087 82.991 116.594 0.507 Flex. Mod. Sum Of Squares 122591.390 563777.533 l9124l24.100 Effect -110.721 Error 103.272 F value 1.149 Probability 0.288 Table 37 (cont'd.) -237.440 1382.900 25.794 2387.078 84.740 0.336 0.000 4254835.691 -652.291 116.594 31.299 0.000 BIBLIOGRAPHY 99 BIBLIOGRAPHY Agarwal, Bhagwan D. and Lawrence J. Broutman. Agaly§i§_§gg 2erf2rmanse_2f_£iher_§2m22§ite§- 355 pp- New York: John Wiley & Sons, 1980. American_Me§al_Markg§. "Plastic Recycling to Pick Up: Report." 98, p. 7(1), April 3, 1990. t. "Plastic Seen Repeating Aluminum Recycling Success." 98(186), p. 9(1), September 24, 1990. S ° 8 ' 5. Philadelphia, PA: American Society for Testing and Materials, 1988. Bataille, P., P. Allard, P. Cousin, and S. Sapieha. "Interfacial Phenomena in Cellulose/Polyethylene Composites." Eglymg;_§gmpggite§, 11(5), 301-304, 1990. Bataille, P., L. Ricard, and S. Sapieha. "Effects of Cellulose Fibers in Polypropylene Composites." Polymer Cgmposites, 10(2), 103-108, 1989. Beck, Melinda, Mary Hager, Patricia King, Sue Hutchison, Kate Robins, and Jeanne Gordon. "Buried Alive." Newsweek, pp. 66-76, November 27, 1989. Bigg, D. M. "Mechanical Properties of Particulate Filled PolymerS-" Pglxmer_22m22§ite§. 8(2). 115-122. 1987- Bigg, D. M., D. F. Hiscock, J. R. Preston, and E. J. Bradbury. "Thermoplastic Matrix Sheet Composites." Polymer C O ° 5, 9(3), 222-228, 1988. Bigg, D. M. and J. R. Preston. "Stamping of Thermoplastic Matrix Composites." Eglymg; Composites, 10(4), 261-268, 1989. Bodig, Joseph and Benjamin A. Jayne. ' W nd EQQQ_QQme§iLe§. 712 pp. New York: Van Nostrand Reinhold CO., 1982. Brown, Lester R. , Christopher Flavin, and Sandr Postel. "World Without End." Natural_fli§§gry, pp. 89-94, May, 1990. 100 Browning. B. L. The_§hsmi§trx_2f_flcgd. 689 pp- New York: Interscience Publishers, 1963. Chand, Navin and P. K. Rohatgi. "Adhesion of Sisal Fibre - Polyester System-" Eelxmsr_99mmunicaticns. 27.157-160. 1986. c e , "Du Pont on Plastics Waste: Recycling is the Answer." p3(2), May 7, 1990. "Polypropylene 1990." 238(16), p35(l), October 15, 1990a. Chesisal_narketing_flsperter. "Recycling Agreement Set." 238(16), p35(l), October 15, 1990b. Cook, William J. "A Lot of Rubbish." . e s W d Beggrg, 107(25), 60-61, 1990. Crosby, J. M. and T. R. Drye. "Fracture Studies of Discontinuous Fiber Reinforced Thermoplastic Composites." ° 0 -e- ..s 0 1e :ue ' .1 _- - . ou-o- - - Technical.§2nferen2e. pp- 245-251- Lancaster. PA- Technomic Publishing Company, Inc., 1986. Fahey, Robert E. "Landfill Lodes." Natural_flis§9:y, pp. 58-60, May, 1990. Foster, Ron. "Ethylene Vinyl Alcohol Resins for Coextruded Packaging Applications." our a o P c ' e o , 1(1), 12-15, 1987. Gogoi, Binoy, K. "Processing-Morphology-Property Relationships for Compounding Wood Fibers with Recycled HDPE Using a Twin-Screw Extruder." M.S. Thesis, Michigan State University, 1989. Hua, Li, Pawel Zadorecki, and Per Flodin. "Cellulose Fiber- Polyester Composites With Reduced Water Sensitivity (1) - Chemical Treatment and Mechanical Properties." Eglymg; QQEEQ§1E§§. 8(3): 199-202, 1987- Hull, Derek. 0 10 o 05' ' . 246 pp. Cambridge, England: Cambridge University Press, 1981. Jindal, U. C. "Development and Testing of Bamboo-Fibres Reinforced Plastic Composites." J9u;ngl_gf_§2m29§i§g MQEEIIQLEI 20: 19'29: 1935- Katz, Harry S. and John V. Milewski. flangbggk_gfi_£illg;§ and_Bsinfersement§_fcr_£la§tis§. 652 pp- New York: Van Nostrand Reinhold Co., 1978. 101 Keal, Maria D. "The Effects of Dual Additive Systems on the Mechanical Properties of Aspen Fiber/Recycled High Density Polyethylene Composites." M.S. Thesis, Michigan State University, 1990. Kaushick, Sandeep. "PP Market Reacts to Strong Demand." Chemisal_narketing_aeeorter. 238(16). pp- 1(2). October 15, 1990. Kokta, B. V., D. Maldas, C. Daneault, and P. Beland. "Composites of Poly(Vinyl Chloride) and Wood Fibers. Part II: Effect of Chemical Treatment. " Pelymez_ggmngfii§§§. 11(2), 84-89,1990. Leidner, Jacob. ELQSLIQS Weste. 317 pp. New York, NY: Marcel Dekker, Inc., 1981. Lightsey, G. R., P. H. Short, and V. K. K. Sinha. "Low Cost Polyolefin Composites Containing Pulp Mill Wood Residue." 29lxmer_Enginesriag_and_§sisnse. 17(5). 305-310. 1977. Lindsay, Karen E. "Fast Growth, Tough Issues Face Packaging Industry." Packagiss_nnginesring pp- 21- 27 June 1988- Luoma, Jon R. "Trash Can Realities." beggeen, 92(2), pp. 86-97, March, 1990. Mader, E. and K-H. Freitag. "Interface Properties and their Influence on Short Fibre Composites." gemeeeieee, 21(5), 397-402, 1990. Maldas, D. and B. V. Kokta. "Effect of Recycling an the Mechanical Properties of Wood Fiber-Polystyrene Composites. Part I: Chemithermomechanical Pulp as a Reinforcing Filler." Polymer_gom29sifes. 11(2). 77-83. 1990- Mohan, R. and D. F. Adams. "Nonlinear Creep-Recovery Response of a Polymer Matrix and its Composites." Experimental_ueshanic§ pp- 262-271 1985- 9 ‘ 9° 0 0119- ‘ . 7 a _'°' 11'. ‘ soon 0. Michigan State University, 1988. Navin, Chand and P. K. Rohatgi. "Adhesion of Sisal Fibre- Polester System." Polyee; Coemenice;ions, 27, 157-160, 1986. Nulty, Peter. "Recycling Becomes a Big Business." EQILeQe, pp. 81-86, August 13, 1990. Parratt, N. J. 'b e- ' o 'a s . 180 pp. London, England: Van Nostrand Reinhold Co., 1972. 102 121§§§12§_E2£1Q- "Business Watch." 48(9), p.61, August, 1990. ‘Potter, David J. B. "Polypropylene is Popping Out All Over." Qhemieei_fleek, 146(13), pp. 20-23, April 4, 1990. Raj, R. G., B. V. Kokta, and C. Daneault. "Use of Wood Fiber as Filler in Common Thermoplastics: Studies on Mechanical Properties. " e a ' CW. 1(3). 80-98.1989. Raj, R. G., B. V. Kokta, D. Maldas, and C. Daneault. "Use of Wood Fibers in Thermoplastic Composites: VI. Isocyanate as a Bonding Agent for Polyethylene-Wood Fiber Composites." 29W. 9(6). 404-411. 1988- Ramirez, Amando Padilla and Antonio Sanchez Solis. "Development of a New Composite Material from Waste Polymers, Natural Fiber, and Mineral Fillers." lemmmel_ef Appiied Peiymeg Science, 29, 2405-2412, 1984. es, "Aseptic Packaging Recycled into Plastic Lumber." 2(17), p.15, August 14, 1990b. s, "EPA Considers 40% Goal by 1996, Stresses Markets." 2(18), p.1-3, August 28, 1990a. Beexeling_1imee, "EPA Estimates 13% Recovery; Updates Waste Characteristics." 2(14), p.3-7, July 3, 1990. Becyeiing Times, "Plastics Recycling Takes Off With New Facilities, Methods." 2(27), p.6-7, December 31, 1990. Resource Integration Systems, Ltd. "Market Study for Recyclable Plastics." 111 pp. State of Michigan Department of Natural Resources, 1987. Richardson. M- 0- W- W 568 pp. London, England: Applied Science Publishers, 1977. Richardson, Terry. Composimes; A peeigm guide. 343 pp. New York, NY: Industrial Press Inc., 1987. Schniewind, Arno P. 'se 0 ' W Wood- Beseg Meterials. 354 pp. Oxford, England: Pergamon Press: Cambridge, Mass.: MIT Press, 1989. Selke, Susan E. M. P a ' a E ' e . 179 pp. Lancaster, PA: Technomic Publishing Company, Inc., 1990. Solvay Polymers. "FortileneO PP." 1 pg. Soltex Polymer Corporation, 1990. 103 Spang, Aletha. "Recycling in Review: Year Two for New JerseY-" Qoxernmsat_£inanse_nexiew. 6(1). pp- 11-14. 1990- Thayer, Warren. "Solid Waste: The Problem You Can't Ignore." Progressixe_§recer. 68(3). pp-77-82. March. 1989. Vu-Khanh, T. "Survey of Current Status and Future Trends in Reinforced Plastics Composites." Eelymer_gempeeiree, 8(6), 363-370, 1987. Wachtel, James A., Bob C. Tsai, and Christopher J. Farrell. "Retortable Plastic Cans, Keep Air Out, Flavor In." £1ast1cs_Eng1aeering. 41(2). 41-45. 1935. Wall, David. "The Processing of Fiber Reinforced Thermoplastics Using Co-Rotating Twin Screw Extruders." Eoiymer Composites, 10(2), 98-102, 1989. Walsh, Thomas S. "PP Random Coplymers." od 5 'cs Encyeiepegia, 88-90, October, 1990. Woodhams, R. T., G. Thomas, and D. K. Rodgers. "Wood Fibers as Reinforcing Fillers for Polyolefins." Poiymer Engineering_and_§siense. 24(15). 1166-1171. 1984- Zadorecki, Pawel and Per Flodin. "Properties of Cellulose- Polyester Composites." Peiymer_§emmeeiree, 7(3), 170-175, 1986. Zadorecki, Pawel, Hakan Karnerfors, and Sven Lindenfors. "Cellulose Fibers as Reinforcement in Composites: Determination of the Stiffness of Cellulose Fibers." Qempoeites Scienee eme Technology, 27(4), 291-303, 1986. Zadorecki, Paul and Anthony J. Michell. "Future Prospects for Wood Cellulose as Reinforcement in Organic Polymer CompositeS-" Polymer_§9meosites. 10(2). 69-77. 1989. STRTE UNIV. LIB "‘jfiflfiflwmmmmmwm mm RRRIES lllHfllHHll 9300901 298