Cement-Wood Composites: Effects of Wood Species, Particle Treatments and Mix Proportion

The aim of this research was to investigate the effects of wood species, particle treatments and mix proportion on the physical (density) and mechanical (compressive strength and dynamicmodulus of elasticity) properties of cement-wood composites. Different mix proportions were investigated, based on the cement: wood ratio of 0.3:0.7, in volume, with Pinus elliottii and Eucalyptus grandis sawdust percentages of 0-100, 25-75, 50-50, 75-25 or 100-0. Sawdust particles were pre-treated with either lime or cement coating to improve cement and wood compatibility. Results show that wood species, particle treatments and mix proportions may influence the physical and mechanical properties of cement-wood composites. In general, results confirm that Eucalyptus sawdust and cement are naturally compatible and do not require any previous particle treatment to avoid compatibility problems.


Introduction
Brazil is one of the five largest producers of industrial round wood with USA, Russian Federation, China and Canada. These countries produced in 2013 54% of total global production [1]. As consequence, huge quantities of wood waste are produced annually by sawmills, whose improper disposal can lead to environmental damage and economic concerns for wood companies.
Plant biomass has already being used to produce engineering materials, encompassing technological and scientific aspects as well as economic, environmental and social issues [2]. The use of wood in cement matrix, in particular, has been investigated for more than one hundred years, whereas its industrial utilization has started later. Initially, small wood particles were embedded in cement matrixto produce low-density boards mainly used for insulating purposes [3]. However, recent studies have also investigated the use of wood waste as raw material for manufacture of many other composites such as wood-based panels, cement slabs and plastic-wood [4].
Wood waste in form of fibers, particles or strands have the potential to be used as reinforcing agent or filler in cement composites [3,[5][6][7][8][9].Cement-wood composites have many advantages over other conventional wood materials, including better insulation and fire performance, better resistance to water soaking, better bactericidal properties and higher stiffness [5].
Wood fibers, as natural fibers, are composed by three major components: cellulose, hemicellulose and lignin. The relative portion of these components may vary based on the types of fibers and the plant source [2]. For some specific applications, certain chemical elements may have to be removed [10].
Due to the chemical morphology of cement and cellulose, hydrogen bonding and/or hydroxyl bridges may play a major role in the bonding of these components [3]. On the other hand, hemicelluloses, starches, sugars, phenols and acids existing in wood fibers, at different levels for each wood specie, tend to inhibit cement hydration, not only leading to a longer setting time but also limiting the strength of the cementitious composite due to micro-fracturing of the matrix during cement hydration [3,6].The complex system of cement hydration reactions itself is influenced by the concentration of Ca(OH) 2 , pH and temperature. Different conditions determine different yields for the reactions and different stabilities for the final products [11]. Inhibition of cement hydration occurs when the calcium silicate hydrate (C-S-H) nucleation sites, which are originally positively charged surfaces, are poisoned by the sugar-acid anions. The wood-fiber surface is probably the most likely place to find the inhibitory effects, related cement hydration [7].
Variability in properties of natural fibers, in general, depends on the growing environment (temperature, humidity, soil composition, air and age) but also depends on the way the plants are harvested and processed [2]. Therefore, physical and mechanical properties of cement-wood composites are strongly influenced by the wood species used in the manufacture, as well as by wood content [3], wood particle size [3,4] and particle treatments [3,5,7,8].
The main purpose of this study was to investigate the effects of two wood species, three particle treatments and fifteen different mix proportions, on the physical and mechanical behavior of cement-wood composites, through the analysis of density, compressive strength and dynamic modulus of elasticity.

Materials
Brazilian cement CPV-ARI [12] similar to Portland cement Type III -HESC -High Early Strength Cement (ASTM C150-12) was used as binder.
Manufacture process consisted in mixing cement, water and sawdust contents (Table 3) in a planetary mortar mixer, placing five samples of each composite in cylindrical φ20mmx40mmmetallic molds. After seven curing days, samples were demolded and conditioned in a laboratory room, protected from air currents and direct insulation, for 21 days.

Physical and Mechanical Properties
Cement-wood composites performance, regarding wood species, particle treatments and mix proportions, were evaluated through physical and mechanical properties.

Density
Density was determined by measuring the mass and volume of each sample, according toEquation1, where M 0 is the weight (g), V 0 is the volume (cm 3 ) and ρ is the density (g/cm 3 ). Results were reported as the average of five measurements.
Compressive strength and Dynamic Modulus of Elasticity Compressive strength tests were carried in a universal testing machine, following a procedure adapted from the Brazilian Technical Standard for Compressive Strength of Cement [16], considering a loading speed of 0,05MPa/s and reported as the average of five samples.
Dynamic modulus of elasticity was determined [17] through the measurement of ultrasonic pulse velocity using aTICO Proceq testing device with transducers of 54 Hz. Results were reported as the average of five samples, based on Equation 2, where V is the ultrasonic pulse velocity (mm/µs), ρ is the density (kg/m³) and µ is the Poisson ratio.

Statistical Analysis
Analysis of variance (ANOVA) was performed using the Statgraphics commercial software. Fisher´s Least Significant Difference (LSD) test was used to compare the difference among the mean values for the properties at the level of 0.05.

Results and Discussion
Trends in the average properties of cement-wood composites values are difficult to ascertain, mainly because wood fibers are a biological material and had inherent variability in fiber length and properties [7]. Apart from that, treatments of each wood fiber may not be uniform over the fiber surface. However, wood specimens, treatments and mix proportions, proposed in this experimental program, resulted in a data set that allowed an examination of apparent trends in the average physical and mechanical properties of cement-wood composites, as discussed in this section. Figure 3 shows results of density for cement-wood composites, whose values are presented in Table 3. Results agree with others reported by literature [3,4,8,11,19].

Density
The importance of studying physical properties of cement-wood composites, in particular density, refers to the fact that, in general, the wood incorporation leads to a composite with lower density and, consequently, lower compressive strength [11,18]. Cement-wood composites with higher densities also present higher values of modulus of elasticity and modulus of rupture [19]. On the other hand, reduction of density results on lighter elements that could be an advantage regarding handling and transporting.
In this study, density of cement-wood composites was significantly influenced by particle treatments. Composites with sawdust treated by cement coating showed higher density, followed by the ones with in natura sawdust and the ones in which sawdust was immersed in alkaline solution.
Regarding P100E0 that contains only Pinus elliottii sawdust, test results indicate that values of density are higher for cement coating, decreasing when immersion in alkaline solution or in natura sawdust are applied.  On the other hand, composites with Eucaliptus grandis, P0E100, P25E75 P50E50 e P75E25, show a different behavior: density is higher for cement coating, decreasing when in natura sawdust and immersion in alkaline solution particle treatments are applied.
Mix proportions in which blends of both wood species were used (P25E75 P50E50 e P75E25) and sawdust particles were immersed in alkaline solution showed the lowest values of density and no statistic difference among means. Figure 4 shows results of compressive strength for cement-wood composites, whose values are presented in Table 4. Results agree with others reported by literature [3,4,7,8].

Compressive Strength
Results showed that compressive strength was influenced by wood species. Cement-wood composites with 100% of Pinus elliottii sawdust showed better performance regarding compressive strength when sawdust particles were pre-treated by cement coating. Samples with blends of Pinus elliottii and Eucaliptus grandis follow the same behavior: higher percentages of Pinus elliottii and cement coating particle treatment tend to result higher values of compressive strength. Considering samples with 100% of Eucaliptus grandis, higher values of compressive strength were reached with in natura sawdust. A previous study of cement-wood composites with in natura Pinnus sp. sawdust and cement: wood: water ratio of 1:0.21:0.6 resulted in a compressive strength mean of 5MPa [8], which is compatible to the mean value of 4.16MPa (Table 4), for P100E0 nat with cement: wood: water ratio of 1: 00:48:0.60.
As a general trend, this research shows that the highest values of compressive strength are reached in composites with higher percentage of in natura Eucaliptus grandis sawdust. On the other hand, high percentages of in natura Pinus elliottii decreased the compressive strength.
Experimental results [20] of compressive strength tests performed in cement wood-composites show that Eucalyptus sawdust and cement are naturally compatible and do not require any previous particle treatment to avoid compatibility problems. Those results are aligned to the results found in the present study. These results, regarding composites with Eucaliptus sawdust are very important for industrial purposes, since the Eucaliptus wood is the most important raw material for cellulose and paper industry in Brazil [21] and the wood processing leads to a great amount of waste.  Eucaliptus grandis sawdust pre-treated by immersion in alkaline solution resulted lower compressive strength values compared to in natura sawdust. Mix proportions in which blends of both wood species were used (P25E75 P50E50 e P75E25) and sawdust particles were immersed in alkaline solution showed the lowest values of compressive strength and no statistic difference among means.
The bonding between the wood fiber and the cement may be chemical, physical, or a combination of both. At the interface between the wood fiber and the cement matrix, stress is transferred between the wood fiber and the cement [7]. A strong or weak interfacial bonding influences the mechanical behavior of the composite. If a strong bonding exists, the result is a brittle material with high strength, as shown in Figure 5(a). On the other hand, a weak bonding results in a tougher material, lacking high strength ( Figure  5(b)). Figure 6 shows Stress versus Displacement behaviour of cement-wood composites with in natura Eucaliptus grandis sawdust, during compressive strength tests, which results are presented in Table 4.

Displacement (mm)
The tough material behavior of the cement-wood composites that shows the efficiency of load transfer between matrix and wood fiber can be verified in Figure 6, as well as the increase in the compressive strength with the higher percentage of in natura Eucaliptus grandis sawdust.

Dynamic Modulus of Elasticity
The importance of determining the modulus of elasticity refers to the fact that this mechanical property is directly related to stiffness, deformability and cracking control of cement-wood composites [22]. The modulus of elasticity of a cement paste depends basically on its porosity and water: cement ratio. However, in a cement-wood composite, the proportion between paste and aggregate, the wood specie and wood particle treatments may also be relevant. Figure 7 shows results of dynamic modulus of elasticity for cement-wood composites, whose values are presented in Table 5. Results agree with others reported by literature [3,4,8,9]  The modulus of elasticity of Pinus and Eucaliptus may range from 7GPa-13GPa and 12GPa-19GPa, respectively. As a result of that, the mechanical behavior of cement-wood-composites, especially regarding stiffness and deformability, may vary considerably if blends of woods species are used.
Experimental results (Table 5) showed the influence of the wood species in the modulus of elasticity. Although composites with high amount of Eucaliptus grandis sawdust (P25E75 and P0E100) showed the highest values of compressive strength (Figure 4), such composites did not present the same behavior regarding modulus of elasticity.
The high coefficient of variation (Table 5) may result from the heterogeneity of the cement-wood samples but also may occur because several factors such as path length, moisture content, temperature, shape and size of specimen may affect ultrasonic measurements. The influence of the aggregate in the modulus of elasticity sometimes may be more relevant than strength or age. For composites with different types of aggregate this influence may be even more complex [23]. In this research, it was found that different composites with same range of strength may present very different modulus of elasticity and it is assumed that these results come from aggregate influence. Figure 8 shows dynamic modulus of elasticity of cement-wood composites with statistical data of comparison between means, where same letters imply treatments with equivalent means. Composite P75E25 with particles pre-treated by cement coating showed the highest value of modulus of elasticity. Cement-wood composites with 100%, 75% and 50% of Pinus elliottii sawdust and particles pre-treated by cement coating showed better performance regarding modulus of elasticity. Composites P100E0 and P50E50 showed no statistic difference among means for particles pre-treated by immersion in alkaline solution.
Composites with 100% of Eucaliptus grandis sawdust showed best results regarding modulus of elasticity for in natura sawdust, followed by cement coating and immersion in alkaline solution treatments.
The interfacial bond strength between the fiber and the cement matrix is influenced by the moisture content due to the reduced bending strength of wet fiber, making it more flexible and less likely to inhibit cracking in the cement matrix [3]. Thus, considering the hygroscopic behavior of wood, the modulus of elasticity of a cement-wood composite is lower than the one of the cement paste itself, and tends to decrease the higher the percentage of wood.

Conclusions
Results showed that, wood species, particle treatments and mix proportions might influence the physical and mechanical properties of cement-wood composites.
Blends In general, results confirm that Eucalyptus sawdust and cement are naturally compatible and do not require any previous particle treatment to avoid compatibility problems.