The raw material used for this work consists of wood chips, expanded polystyrene and clay powder.

The wood chips used are passing of 5 mm mesh sieve. They are lamellar in shape and 90% of particles are larger than 2 mm in size (Figure 1(a)). They were collected in carpentry units in Abidjan (Ivory Coast).

· Expanded polystyrene (EPS) (Figure 1(b)) comes from the packaging of household appliances. It was collected in pre-collection bins near shops in Abidjan. Once stripped of any labels, the polystyrene is dissolved into acetone to obtain a resin. This dissolution reduces its volume by 98% [14] and thus makes its collection profitable.

This resin will be used as a binder to make composites. The dissolution of EPS is done in a ratio of approximately 0.7 kg of EPS per liter of solvent.

The mineral filler used is a gray to slightly brown colored clay. This clay has been the subject of many characterization and valuation studies which show that it is essentially made of kaolinite and illite [7] [15] [16]. It was dried, and crushed. The passing of the 63 μm sieve was used for this work (Figure 1(c)).

The production method is summarized in Figure 2. The base material is made of wood chips (75% by volume) bound by polystyrene resin (25% by volume). To this composition, clay additions were made in proportions of 5%, 10% and 15%. After drying, some of the samples were subjected to thermoforming. To do this, the sample is put between two metal plates; the whole is heated at 150˚C for 50 minutes and hot compacted using a manual static press.

Finally, two categories of samples are produced: non-thermoformed samples and thermoformed samples.

The results of the fire persistence test are listed in Table 2 below. This table shows that samples without mineral addition (0% clay) present combustion with flame. In fact, it is a continuous combustion until their total consumption. However, tastes do not form during their combustion.

After adding mineral filler (5% to 15%), the samples present combustion with a flame for variable duration. The combustion time decreases when mineral filler content increases.

The ignition time is approximately 65 seconds for samples at 5% of filler, for samples at 15% non-thermoformed and thermoformed is respectively about 15 seconds and 11 seconds. Overall, there is no incandescence that persists after the flame is extinguished.

The samples without mineral filler have low fire resistance due to the presence of cellulose (a very flammable material) in the wood and the fact that EPS is a hydrocarbon and therefore an oxidizer by nature. This explains the continuous combustion with flame until the total consumption of these composites.

When the mineral filler is added to the EPS-Wood chips mixture, the clay particles become intercalated into the matrix and hinder the propagation of the flame because clay is not combustible. This discomfort becomes more and more important when clay content increases. This could explain the fact that composites become less flammable. The mineral filler behaves like a flame retardant. Indeed clay particles cover the wood chips by constituting a barrier between the flame and the chips.

Similar results have been reported by [20] [21] on the incorporation of glass silicon carbide and basalt fiber into composites. They explain that these mineral filler reduce the ignition time and the rate of flame propagation considerably. These fillers behave like a barrier layer.

Likewise [22] during their studies on polypropylene composites filled with natural fibers (Kenaf) made the same observations.

For these authors, a carbonized layer is formed at the surface of the polymer during thermal degradation. As an insulating barrier, this carbonized layer reduces; the transfer of the flame to the polymer as well as the diffusion of oxygen into the material.

According to the M classification of construction and arrangement materials (standard NF P 92-504), samples without mineral addition are of class M3, i.e. moderately flammable combustible materials while those with mineral filler are of class M2: flame retardant materials such as particle board, wall carpets, etc. Class M2 composites are conform with (French) fire safety legislation [23].

The flexural (three-point) and compressive strength of non-thermoformed and thermoformed composites are shown in Figure 4.

When the mineral filler is added to the EPS-Wood shavings mixture, the flexural strength increases up to 10% of filler, beyond that the strength drops (Figure 4(a)), both for samples which have sustained thermoforming or not. The flexural strength for thermoformed samples range from 11.73 MPa at 0% of filler to 12.34 MPa at 10% then drop to 9.2 MPa with 15% filler.

For non-thermoformed samples, their flexural strength drop from 5.4 MPa (0% filler) to 7.45 MPa (10%) then drop to 6.28 MPa (15%).

The compressive strength of non-thermoformed samples increases from 14.24 MPa to 15.94 MPa and then drop to 13.32 MPa when filler content goes from 0% to 10% and then to 15%. After thermoforming, the strength range from 18.67 MPa to 21.72 MPa before dropping to 18.13 MPa (Figure 4(b)).

This variation can be explained by the fact that when mineral filler is added, during the stress in flexural or in compression, the force which was previously supported only by the EPS matrix and the wood chips, is now supported by these two constituents and by clay. Hence the strength increases as the mineral filler increases.

The drop in strength for more than 10% mineral filler would be due to the fact that because of high quantity of mineral filler in the mixture, the action of the binder becomes ineffective. It promotes existence of contacts between clay particles. These contacts weaken the composites, and leading to the drop in mechanical strength.

The same observation was made by [4] in his work on sand-plastic composites filled with clay and by [24] on the incorporation of mineral filler into a biocomposite reinforced with rice husks and sawdust.

Incorporating fibers into a polymer matrix also gives similar results [24] [25]. These authors explain the drop in mechanical strengths after optimal filler rate, by the fact that from this rate, the bringing together of the fibers seem to hinder fluidity of the matrix, and causing weakening of fiber-matrix bonds. This behaviour can reduce mechanical strength of composite materials.

It can be notice that mechanical performance of thermoformed samples is greater than those of non-thermoformed samples. This is explained by the fact that thermoforming ensures a good distribution of the EPS matrix into the composites.

The combined effect of heat and pressure induces reorganization of matrix and wood chips, reducing voids in the mixture while strengthening the bonds between the EPS matrix and the wood chips.

Therefore the optimal rate of mineral filler would be around 10% for these composites.

For some authors [26] [27], this rate depends on kind of polymer in the mixture. For example [25] shows that the optimal filler rate is 5% for polypropylene while [26] place this rate at 20% for corn starch. For the same polymer, this rate may vary when its structure is modified, then the optimal filler rate drops from 5% to 3% for modified propylene [28] [29].