Program

P.13 -- Poster, IWTCP-3

Date: Wednesday, 29 July 2015

Time: 08h30 - 10h00

Improved thermoelectric properties of graphene devices by strain and doping engineering

M. Chung Nguyen (1,2), V. Hung Nguyen (1,2), H. Viet Nguyen (1), J. Saint-Martin (2) and P. Dollfus (2)

(1) Center for Computational Physics, Institute of Physics, VAST, Hanoi, Vietnam; (2) Institut d’Electronique Fondamentale, Université Paris Sud, Orsay, France

The thermoelectric effect enables direct conversion of a temperature difference into anelectric voltage and vice versa, and provides a viable route for electrical power generation from waste heat. It has been shown in the literature that low dimensional and/or nanostructured materials have better thermoelectric properties than that in bulk materials [1,2]. In this regard, graphene, a mono-layer of carbon atoms, is a truly 2D material and hence could be a promising channel for thermoelectric devices. However, graphene still have a drawback due to its gapless character, which makes it difficult to separate the contribution of electrons and holes and leads to a weak Seebeck effect (S < 100 µK/V in pristine graphene [3]). Hence, several energy-gap nanoengineerings have been suggested to be solve this issue [4]. Recently, we found that strain engineering is a promising technique to generate a finite energy-gap in graphene strain heterochannels [5]. In this work, we propose to exploit this effect to enhance the thermoelectric properties (particularly, Seebeck coefficient) of graphene devices. We demonstrate that due to the strain-induced energy-gap, the Seebeck coefficient in graphene strained heterochannels can be significantly enlarged, i.e., ~ 17 times higher than that in pristine graphene when a strain of ~ 10% is applied. However, to avoid the requirement of a large strain, we propose to use this type of heterochannel in graphene p-n devices. In the p-n devices, the displacement of electronic structure in two highly doped sections can further enlarge the energy-gap of transmission and hence can be helpful for further enhancing the Seebeck effect. Indeed, we demonstrated that in such graphene p-n devices with a local strain, a similarly high Seebeck coefficient can be achieved with a small strain of only ~ 5 %. The dependence of these phenomena on the lengths of strain section and transition region between highly doped parts has been also clarified. Thus, we demonstrated that besides its use in strain sensors [5], this design strategy is very promising to achieve good performance in graphene devices based on the Seebeck effect, as thermal sensors [6]. References: [1] M. S. Dresselhaus et al., Adv. Mater. 19, 1043 (2007); [2] G. Snyder and E. Toberer, Nat. Mater. 7, 105 (2008); [3] Y. Zuev et al., Phys. Rev. Lett. 102, 096807 (2009); [4] P. Dollfus et al., J. Phys.: Condens. Matter 27, 133204 (2015); [5] M. C. Nguyen et al., Semicond. Sci. Technol. 29 115024 (2014); [6] M. C. Nguyen et al., Physica E (2015) in press; DOI 10.1016/j.physe.2015.05.020

Presenter: Mai Chung Nguyen