*3.5. Module Design*

Commercial modules, such as those from Table 6, generally have a pre-designed configuration from the supplier, which means that for some applications, they are not suitable or do not show their optimal performance. On the other hand, the design is an important factor since by means of the design parameters, the efficiency, the cost and the useful life of the modules can be improved. Likewise, most of the studies related to TEG components or systems are carried out using simulation tools [15], due to the complexity of their manufacture, as well as the cost that these would generate for their development if they were carried out exclusively by experimental means. Some of the most relevant studies in this regard are listed below.

Segmented modules represent one of the most viable alternatives from the view point of design. In these, various TE materials are used to manufacture the module legs, seeking to increase the working temperature, minimize the thermal effects, and increase the efficiency of the TEG modules. Recent work related to the manufacture of segmented modules includes the work of Hung et al. [46], who implemented commercial TE materials in cold areas of the cell and oxides in hot areas, in order to increase the working temperature of the modules. The viability of these TEG modules was analyzed using numerical modeling, which found that the oxide-segmented modules have an efficiency of around 10%. Ouyang et al. [36] carried out a study in which high-ZT TE materials were evaluated by finite element analysis. A systematic model was achieved for the segmented modules, finding a cost-performance ratio of ~0.86 \$/W with an efficiency of 17.8%. Similarly, Jiang et al. [27] evaluated a TEG module composed of np Bi2Te3 and np PbS/PbTe, which exhibited an efficiency of 11.2% with a ∆T = 317 ◦C. In addition, they optimized the ratio of the legs at low and high temperatures, determining the optimal ratio to be 7:17. The maximum power obtained in this TEG module was 0.53 W for a ∆T = 312 ◦C.

During the design of TEG modules, the optimization of parameters such as the length of the legs or the number of thermocouples of the TE materials helps to reduce the amount of material used, without compromising the efficiency of the modules. According to Rezania et al. [49], the temperature differences at the n- and p-type junctions of TE elements are not identical. Such temperature differences are lower in n-type TE elements, compared with those in p-type TE elements, due to the higher thermal conductivity in the n-type material. Consequently, the footprint size of the n-type element must be larger than the footprint of the p-type TE element, due to the higher thermal conductivity in the n-type material. Therefore, the optimal ratio of footprint areas to achieve the maximum generation and the best cost-efficiency ratio in thermoelectric modules must satisfy that An/Ap < 1, where An and Ap are the footprint areas of the the n-type and p-type junctions, respectively. Brito et al. [41], found that when the thickness of the TE elements is smaller, the electrical resistance is reduced, but this will impact the ∆T of the TE module, because a lower

thermal resistance will increase the thermal output and attenuate the temperature difference between the hot and cold sources. However, this will only occur if the usable hot source is low and the other thermal resistances are high enough to significantly affect the ∆T of the TE module. In their study, Dongxu et al. [31] found that the thickness of the TE legs can be reduced to 1.1 mm, which is 4 mm less than commercial modules, while still achieving the same efficiency. In all these works, simulation tools were successfully used to find the relationships between the different parts of the modules and their respective powers.
