One of the major concerns of materials researchers today is to reduce the thermal loads of buildings. Reducing this load is one of the solutions for minimising energy consumption in buildings and combating the effects of climate change. There is an intrinsic relationship between the energy consumption of materials and the environmental impact of a building’s life cycle [1] . However, to reduce energy consumption in homes, it is essential to promote alternative methods from the construction stage onwards, through the use of materials that provide thermal comfort and reduce conventional energy requirements. Raw earth is one of the building materials recognised for its high thermal inertia and abundance, but it suffers from problems of sensitivity to water, systematic cracking and hygrothermal performance measurement. To remedy these problems, several types of soil stabilisation have been used to improve the characteristics of the soil. In addition, measuring its hygrothermal performance remains another key issue, when stabilisers such as cement, lime, ash, fibres, etc. have been incorporated into the clay soil [2] - [5] . In this study, diatomite is used as an additive in clay matrices because of its interesting structure and thermal properties [6] . The aim of this research is to develop and determine the hygrothermal behaviour of composites made from two clay soils and diatomite. Particular emphasis will be placed on the capacity of these composites to absorb water by capillary action, resistance to erosion, thermal conductivity and thermal effusivity as a function of diatomite percentages. These parameters are particularly useful for characterising a material from a hygrothermal point of view.

Figure 4 shows the variations in absorption of blocks from two mixtures as a function of diatomite percentages. Samples without the addition of diatomite have an average absorption coefficient of 49.01%, whereas other samples amended with different levels of diatomite have an average water absorption coefficient of 60.13%, i.e. an increase of 11.12%. With regard to BYD composites, the absorption coefficients are distributed as follows: 50.11% of mortar without diatomite and 56.29% of composites, i.e. an increase in absorption coefficients of

around 6.18%. Absorption coefficients increase as the percentage of diatomite in the clay matrix increases. This capillary increase in composites depends in particular on the open porosity of the material [9] . Diatomite incorporated into the clay matrix favours total porosity, these results are in agreement with the conclusion of [8] on compressed earth block CEB amended with 20% diatomite waste and of [10] [11] on CEB or adobes amended with plant fibres or aggregates and ash. These latter conclusions are contrary to those of [12] [13] , using pumice, fly ash and peanut shell powders. Certain values of the absorption coefficients show that the addition of a few levels of diatomite cannot have a beneficial effect on the hydrous behaviour of composites.

The results of the mass loss of the composites after the rain erosion test presented in Figure 5 show that the two composites BTD and BYD have different erosion resistances. The weight loss of BTD composites increases with the percentage of diatomite (37.12% to 53.92%) to around 10.49% compared to reference mortars. On the other hand, the loss of mass decreased with the diatomite content (23.41% to 14.79%) compared to the untreated specimens. A reduction in mass of around 17.72%. Compared with BTD composites, BYD composites have good resistance to rain erosion, but even so, the mass loss values for both composites are higher than the maximum permitted [14] . The mass loss of the BYD₅₀ composite is close to the recommended 10% for building bricks.

The various thermal conductivity results are shown in Figure 6 . The curves obtained show variations in thermal conductivity as a function of the percentage of diatomite in the BTD and BYD composites. Adding diatomite content lowers the thermal conductivity values of BTD composites from 0.72 to 0.29 W/m∙K and BYD composites from 0.52 to 0.21 W/m∙K respectively.

The drop in thermal conductivity values can be explained by the increase in pores due to the intrinsic characteristics of diatomite and voids due to the evaporation of water in the specimens after drying. According to [15] , steady-state heat transfer is by conduction, so pores represent gaps for heat flow transfer. According to the results obtained, the thermal conductivity values measured decrease considerably as the diatomite content increases. This phenomenon is consistent with the intrinsic properties of diatomite, which is a thermal insulator. However, for BTD and BYD composites, the thermal conductivities decrease by an average of around 40% for the different formulations, compared with specimens without the addition of diatomite. These results are better than those reported by [16] , using diatomite in the clay matrix, but in line with those reported by [9] [17] . The thermal conductivity results of this work corroborate the recommended values (0.5 to 1.7 W/m∙K) for stabilised or unstabilised CEB [8] [17] [18] . The low thermal conductivity of these materials helps to reduce heat conduction from the outside to the inside, and regulates the temperature inside buildings. This provides good thermal comfort inside buildings, while reducing electrical energy costs.

The influence of curing time on thermal conductivities shown in Figure 7 and Figure 8 respectively is not noticeable in either composite. This phenomenon can be interpreted by the departure of water, leaving voids and thus intrinsic diatomite characteristics in the mixtures. The results of this work corroborate the conclusions of [19] [20] on CEB and [13] on peanut-amended adobes. As the specimens age, the water in them evaporates, leaving voids.

The thermal effusivity values of the composites are shown in Figure 9 . The variation in the thermal effusivity curves of composites is similar to that of thermal conductivity. A decrease in this property with increasing percentages of diatomite is also observed in all cases. These results corroborate the conclusions of work by authors [16] [21] using diatomite.

Figure 10 shows the thermal conductivities as a function of composite density. Thermal conductivity decreases with density as the diatomite content increases, leading to a reduction in thermal properties. For all composites (BTD and BYD), when thermal conductivity values increase with specimen density.

The various thermal conductivity results in this work are compared with the mechanical and thermal requirements of construction materials such as structural and insulating concretes in accordance with the RILEM recommendation [22] . According to the RILEM functional classification, the properties of composites are comparable to those of lightweight aggregate concretes as structural and thermal insulation materials [22] . The determination of crucial parameters such as the density, mechanical strength and thermal resistance of materials in the field of building construction makes it possible to choose the optimum point and

type of materials used in construction. A sufficient knowledge of these three important parameters for composite use is therefore necessary to make optimum use of the qualities of strength, lightness and insulation.

The hydric and thermal behaviour of various composites based on clayey soils and diatomite presented and discussed in this article gave satisfactory results, particularly with regard to thermal properties. We note that the absorption of water by capillary action of the composites increases as the percentage of diatomite increases. This absorption is much more noticeable in BYD composites than in BTD composites, but is perfectly acceptable for construction purposes. The rain erosion test showed divergent results between BTD and BYD composites. The severe deterioration of BTD blocks by water is due to the high quantity of the active element (clay) in the T soil. They are qualified as unsuitable for the construction of buildings in regions with high rainfall. Thermal parameters generally decrease considerably as the diatomite content increases in clay matrices. However, the incorporation of diatomite reduces thermal conductivity values by more than 40%. In fact, for the different levels of diatomite used in clay soils, the measurements of the thermal parameters of the composites are very close to the recommended thermal requirements for construction materials. These composites can be used as new bio-composite materials to reduce energy loads in buildings.

In the future, the authors plan to carry out a study of the durability and aging of these composites by simulation, in order to predict the service life and maintenance requirements of the proposed materials.

The authors would like to thank respectively: Mr Mahamat Saleh Abdel-khadir, head of the Energy and Materials Research Laboratory (LREM) at the Institut National Supérieur des Sciences et Techniques d’Abéché (INSTA) and Dr Abakar Ali, head of the Laboratory at the Ecole Nationale Supérieure des Travaux Publics (LENSTP) in N’Djamena, Chad.