Thermionic Energy Conversion Based on Graphene van der Waals Heterostructures

www.nature.com/scientificreports OPEN received: 16 September 2016 accepted: 10 March 2017 Published: 07 April 2017 Thermionic Energy Conversion Based on Graphene van der Waals Heterostructures Shi-Jun Liang1, Bo Liu2, Wei Hu3, Kun Zhou2 & L. K. Ang1 Seeking for thermoelectric (TE) materials with high figure of merit (or ZT), which can directly converts low-grade wasted heat (400 to 500 K) into electricity, has been a big challenge. Inspired by the concept of multilayer thermionic devices, we propose and design a solid-state thermionic devices (as a power generator or a refrigerator) in using van der Waals (vdW) heterostructure sandwiched between two graphene electrodes, to achieve high energy conversion efficiency in the temperature range of 400 to 500 K. The vdW heterostructure is composed of suitable multiple layers of transition metal dichalcogenides (TMDs), such as MoS2, MoSe2, WS2 and WSe2. From our calculations, WSe2 and MoSe2 are identified as two ideal TMDs (using the reported experimental material’s properties), which can harvest waste heat at 400 K with efficiencies about 7% to 8%. To our best knowledge, this design is the first in combining the advantages of graphene electrodes and TMDs to function as a thermionic-based device. The most common approach to harvest the waste heat to generate electricity is thermoelectrics (TE), which is based on the Seebeck effect (see Table 1). The performance of TE-based devices is characterized by the figure of merit (ZT), given by ref. 1 ZT = α2 T , κl + LT µne (1) where α, T, κl, μ, n, and e are, respectively, the Seebeck coefficient, absolute mean temperature, lattice thermal conductivity, carrier mobility, carrier density and electron charge. Here, L is defined as the Lorenz number equal to 2.44 ×  10−8 WΩ K−2. This formula has recently been redefined to solve the inconsistence between theoretical predication and experimental measurement2. Before the 1990s, the progress of improving ZT had been slow and the best TE material was Bi2Te3 alloys with ZT ≈ 1.0 at 300 K3. To increase ZT, many new approaches have been proposed4,5, such as fabricating low-dimensional thermoelectric structures to increase large density of state, engineering the interface of materials to reduce the lattice thermal conductivity, and modulating dopants to increase carrier mobility. Subsequently, further improvements include ZT = 2.4 at 300 K for p-type Bi2Te3/Sb2Te3 superlattice6, and ZT = 3 at 550 K for Bi-doped n-type PbSeTe/PbTe quantum-dot superlattice7. A prospective of nanostructured TE materials can be found in a review paper8. For practical applications, other issue such as size, maintenance and fast response time must also be considered even if high-efficiency TE materials are found ref. 1. Recent interests in using two-dimensional (2D) transition metal dichalcogenides (TMDs) as new TE materials have attracted extensive attention9,10 due to high Seebeck coefficient offered by these 2D TMDs: a bilayer MoS2 gives α2σ = 8.5 mW/m/K2 9. If this MoS2-based TE material is able to realize the calculated thermal conductivity of κ ≈ 1.55 W/m/K11, we will have ZT = 1.6, which corresponds to an efficiency of about 6.5% in harvesting waste heat at Th (hot side) = 400 K and Tc (cold side) = 300 K [according to Eq. (1)]. For high temperature range, the more viable approach is based on thermionic energy convertor (TIC), which was first proposed by G. N. Hatsopoulos12. Due to the high work function of the metallic electrode, however TIC is limited to high-temperature operation above 1500 K. A potential method to harvest waste energy at 900 K was 1 SUTD-MIT International Design Center (IDC), Singapore University of Technology and Design (SUTD), 8 Somapah road, 487372, Singapore. 2School of Mechanical and Aerospace Engineering, Nanyang Technological University, 50 Nanyang Avenue, 639798, Singapore. 3Computational Research Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA. Correspondence and requests for materials should be addressed to L.K.A. (email: ricky_ ) or S.J.L. (email: ) Scientific Reports | 7:46211 | DOI: 10.1038/srep46211 1 www.nature.com/scientificreports/ System d [nm] κ for [W/m/K] Experiment SBH [V] TR [K] Optimal SBH [V] Max ηg [%] G/MoS2 50 0.3 0 to 0.11 847 0 3.15 G/WS2 50 0.2125 0 to 0.37 758 0 3.93 G/MoSe2 70 0.089 0 to 0.4 501 0.005 7.28 G/WSe2 62 0.048 0 to 0.44 428 0.02 8.56 Table 1. The designed parameters: TR, optimal effective barrier height φ′ and the calculated efficiency ηg of the proposed power generator operating at Th = 400 K and Tc = 300 K based on the experimentallymeasured cross-plane thermal conductivity of different TMDs materials at the temperature of 300 K [See S2: Supplementary materials]: (a) 50-nm of Gr/MoS2/Gr28, (b) 50-nm of Gr/WS2/Gr28, (c) 70-nm of Gr/ MoSe2/Gr27,39 and (d) 62-nm of Gr/WSe2Gr/27,39. recently proposed by using a suspended monolayer graphene as cathode to provide an efficiency of higher than 40%13. This improvement is attributed to the new thermionic law given by J(φ, T, EF) =  A* ×  T3 ×  exp[− (eφ −  EF)/kBT], where A⁎ = ekB3/π 3v 2f = 0.01158 A/cm2/K3, vf is the Fermi velocity, EF is the Fermi level, and φ is the barrier height at zero bias. Note that the new scaling has been compared well with a recent experiment14. For the wasted heat generated in the industrial or domestic process, low-grade heat (around 400 K to 500 K) is distributed more everywhere. developing an efficient approach remains a great challenge so far. In this paper, we propose a high-efficiency solid-state thermionic device by using van der Waals (vdW) heterostructure15 composed of 2D TMDs (MoS2, MoSe2, WS2, and WSe2) and graphene electrodes. By taking the advantage of the ultralow cross-plane thermal conductance of the 2D materials and the new thermionic emission over the Schottky barrier (SB) contact between the graphene and 2D materials (tunable via gate voltage or chemical doping), we predict that it is possible to realize high-efficiency power generation and refrigeration at the temperature of 300 K to 500 K, which may be better than (or at least comparable to) the traditional TE devices. Note that the concept of using multi-layers or superlattices in the thermionic devices was first suggested by two groups (Shakouri and Mahan) in late 1990s16,17. The performance of their proposed single-junction thermionic device was predicted to be better than the TE device using the same lnGaAs/lnAlAs material18. For simplicity, we will ignore the effect of non-conservation of lateral momentum in the thermionic emission18,19 in this paper. This is justified by the facts that Schottky barrier height is planar and homogenous at the interface between graphene and Transition metal dichalcogenide20. With the current advances in growing graphene and TMDs, the proposed vdW heterostructures such as Gr/ TMDs/Gr (Gr is the monolayer graphene) can be assembled experimentally15 (...truncated)