The precursors selected for our study contain more or less water. The amount of water depends on several factors, among which we cite the very nature of the precursor, its origin, the duration of its exposure in a humid environment and the duration of drying. This study was carried out without prior treatment of 5.0 g of the Sawdust, PET and Sawdust + PET mixture sample. These materials were baked at 80˚C, 90˚C, 100˚C and 110˚C for 24 hours. The final mass made it possible to eliminate the loss of water from the three samples, according to relationship (2) and we obtained a profile of variation in the humidity rate given in Figure 1.

There are several ways to follow the evolution of the mass, among many others, we can mention the differential thermal analysis [5] . However, in the context of our study, we based ourselves on the variations in the mass of the precursors, from 25.0 to 75.0 g depending on the temperature, following a constant heating rate (10˚C/min) and 75 µm in diameter. The results of this study are shown in Figure 2.

w ( % ) = m 0 − m f i n a l e m 0 × 100 (2)

In order to study the influence of this physical factor on the loss of mass, this study was carried out for sawdust. This study concerned sizes varying from 75 to 630 µm in diameter at 700˚C and 10˚C/min. The results are reported in Figure 3.

Four speeds have been demonstrated, 5˚C/min, 10˚C/min, 15˚C/min and 20˚C/min for the different temperatures, 500˚C, 600˚C and 700˚C. The mass of the precursor retained for this work is 50 g and the results are reported in Figure 4.

In order to study the influence of time on thermal reactivity, 50 g of the three 75 µm precursors were subjected to 700˚C in the reactor. This study was carried out in a time interval varying from 30 to 180 min. Burn-off and density were determined, the results have been reported in Figures 5-7.

Figure 1 represents the variation of the humidity rate as a function of the drying temperature, relating to the pyrolysis of sawdust, PET and CPS (mixture PET + Sawdust). We note that the water loss of SC and CPS increases with temperature, unlike that of PET whose variation is substantially zero. Without heat treatment, sawdust alone contains about 7% water. It is therefore sawdust from wood that is poor in water compared to other species such as: green wood and sapwood which can contain up to 25% water. This is a consequence of the fact that the humpback is a tropical tree that has a great ability to adapt to arid climate and drought.

Furthermore, it can be noted that PET has a substantially zero humidity rate, which can be justified by its crystalline structure, preventing the attachment of the water molecule. SC, PET and CPS baked at 90˚C still contain respectively; 4.34%, 0.0015% and 2.03%, of water. According to Saastamoinen [11] , a sample is only completely dry if the boiling point of water is exceeded. Thus, around 110˚C, anhydrous materials are obtained which quickly regain a certain humidity as soon as it is exposed to the open air. In order to avoid problems related to the humidity level in the pyrolysis ovens and for a better interpretation of the results, we processed the data under dry bases and the baking temperature was stopped at 110˚C for a duration of 24 hours, because this temperature does not degrade the nature of the materials even less its structure. In short, it does not engage a complex thermal reactivity, but the only reaction involved is dehydration.

One can observe in Figure 2, a variation of the loss of mass in relation to the initial mass of the three precursors at 500˚C, 600˚C, 700˚C. This shows the greatest mass losses for the PET samples, between 87.19% and 96.05%, followed by that of the mixture, between 47.33% and 64.37%. And the lowest are observed, with those of sawdust (from 24.02% to 62.6%). However, here we can say that the influence of the mass is not great, given the slight difference between the losses by temperature. Indeed, according to Delmon [12] , the path of heat energy inside a pulverulent mass would be represented by a complicated network thus creating a heap effect. On the other hand, among the gaseous molecules resulting from the pyrolysis of the three precursors, there are polymerizable products. These could form large molecules that are likely to be trapped between the grains of the sample when the pile grows in size.

Taking this result into account, all subsequently processed samples will have an initial mass in the order of 5 mg to reduce the relative error due to sensitivity and to control the thermal diffusion phenomenon due to the heap effect.

Note that during sawdust pyrolysis, the 75 and 300 µm curves have the same appearance and that the differences are insignificant, even if the diameter of the grains is varied by a factor of three. The curve for particles smaller than or equal to 630 µm deviates from the other two, thus reducing the loss of mass. The curve displays a different thermal behavior from those of 75 and 300 µm. We can respectively see the increase in weight loss from 26% to 36% in the interval from 500˚C to 600˚C followed by the decrease from 36% to 27% from 600˚C to 700˚C. These results show that the speed of heat propagation in the lignocellulosic material depends on the particle size. We observe a low thermal diffusivity of 600˚C to 700˚C, reflecting the slowness of the heat front to pass through the thickness of the material. The weight drop between 36% to 27% in the interval of 600˚C to 700˚C can be attributed to the phenomenon of thermal inertia. On the basis of these results and to better minimize the physical constraints such as the intragranular diffusibility of the volatile matter and the homogeneity of the temperature in the grains, all the experiments, the continuation, will be carried out on samples of the same order of mass, and with a particle size of 75 μm in diameter.

From Figure 4, we notice that when we vary the heating rate and keeping all the other factors constant, the mass loss at the end of the reaction is approximately the same depending on each precursor material. Nevertheless, it has been demonstrated that the heating rate has a strong influence on the nature of the reaction products, both for the volatile matter and for the solid residue, as well as on the kinetic parameters of the chemical reaction [5] [6] .

Indeed, according to the temperature of the reaction medium and according to the time spent in the reactor, the gaseous products which are released can degrade into small molecules of smaller sizes, polymerize or condense, then attach themselves to the solid residue [6] .

These fixed products are released when the temperature rises further. On the other hand, we notice a delay in the thermograms towards low temperatures as the heating rate decreases. This phenomenon is very often observed in the literature [3] [7] [8] .

This delay is caused by the energy supplied to the system which gains importance when the heating rate decreases, because even if the heat flow is low in the case of low heating rates, the residence time is too high to that the sample stores enough energy to react quickly and break the chemical bonds which break when the energy of their formation is reached.

In addition, one can note the slight shift in the loss on ignition between the speeds. We notice that when we increase the heating rate, the mass loss increases slightly, reflecting the significant thermal change. This leads us to say that the pyrolysis of PET, SC and CPS is a very sensitive reaction. This during the speed of 10˚C/min has been shown to be optimal.

Retention time is a very important factor in the carbonization process. Indeed, this factor considerably increases the porous volume of materials during carbonization [3] . In this sense, the influence of the heating time on the yield of carbonization and on the apparent density of polyethylene terephthalate (PET), sawdust (SC) of wood and mixture (CPS) was studied.

Figures 5-7 represent the variation of the carbonization yield and of the apparent density as a function of the heating time. We notice that the yield decreases with the residence time in the reactor. This means that the longer the carbonization time, the more the yield of solid residue decreases. Indeed, this phenomenon is accompanied by a degassing which is at the origin of this decrease. It can be noted that the yield of the three precursors stabilizes from 60 minutes and remains almost constant after a reaction time of between 60 and 180 minutes for PET, SC and CPS. This time can be considered as the end time of thermal reactivity of the materials where the heating time seems to have no more effects on the materials. These results corroborate those obtained by Elabed [13] , Tchakala [14] and Abo [15] during the activation of bagasse and corn cobs with phosphoric acid.

In addition, the variation of the apparent density shows a decrease as a function of the increase in the residence time of the materials in the reactor. As the carbonization time increases, the bulk density decreases. It is noted, for the lignocellulosic material that the apparent density stabilizes from 60 minutes. It varies very slightly if we further increase the retention time in the reactor. This could be explained by the fact that beyond one hour of treatment, the decrease in the thermal reactivity of the precursor. A thermal stability begins to manifest itself in the structure of the carbonaceous solid and on its surface. This promotes the crystallinity of the material if the residence time is prolonged leading to a decrease in porosity and consequently an increase in the apparent density of the activated carbon [11] . The same phenomenon is observed for the plastic material as well as the mixture from 150 minutes.

These results show a thermal insatiability of polyethylene terephthalate which gives it a good thermal reactivity, unlike sawdust whose stability is faster (1H). A residence time of less than one hour is insufficient for the reaction limit of the three precursors and the temperature. One hour seems to be the ideal treatment time since, beyond the latter, the porosity of our materials decreases.