Phase-Defined van der Waals Schottky Junctions ... - ACS Publications

Jun 8, 2017 - preamplifier (virtual ground), the positive sign of the photocurrent represents the flow of the photocurrent from β to α. As shown in ...
0 downloads 0 Views 5MB Size
Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Letter

Phase-Defined Van der Waals Schottky Junctions with Significantly Enhanced Thermoelectric Properties Qiaoming Wang, Liangliang Yang, Shengwen Zhou, Xianjun Ye, Zhe Wang, Wenguang Zhu, Matthew D. McCluskey, and Yi Gu J. Phys. Chem. Lett., Just Accepted Manuscript • Publication Date (Web): 08 Jun 2017 Downloaded from http://pubs.acs.org on June 8, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry Letters is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

Phase-Defined Van der Waals Schottky Junctions with Significantly Enhanced Thermoelectric Properties

Qiaoming Wang,*1 Liangliang Yang,*1 Shengwen Zhou,1 Xianjun Ye,1 Zhe Wang,2 Wenguang Zhu,2 Matthew D. McCluskey,1 and Yi Gu§1

1 2

Department of Physics and Astronomy, Washington State University, Pullman, WA 99164, USA

Department of Physics, University of Science and Technology of China, Hefei, Anhui 230026, China

*

Equal contribution

§

Corresponding Author: [email protected] 1 ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abstract: We demonstrate a van der Waals Schottky junction defined by crystalline phases of multilayer In2Se3. Besides ideal diode behaviors and the gate-tunable current rectification, the thermoelectric power is significantly enhanced in these junctions, by more than 3 orders of magnitude, compared to single-phase multilayer In2Se3, with the thermoelectric figure-of-merit approaching ~ 1 at room temperature. Our results suggest that these significantly improved thermoelectric properties are not due to the two-dimensional (2D) quantum confinement effects, but instead are a consequence of the Schottky barrier at the junction interface, which leads to hot carrier transport and shift the balance between thermally and field-driven currents. This “bulk” effect extends the advantages of van der Waals materials beyond the few-layer limit. Adopting such an approach of using energy barriers between van der Waals materials, where the interface states are minimal, is expected to enhance the thermoelectric performance in other 2D materials as well.

TOC Graphic:

2 ACS Paragon Plus Environment

Page 2 of 31

Page 3 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

Metal-semiconductor junctions (MSJs) are a fundamental component in electronic circuits, and the capability to control the behaviors of MSJs (Schottky vs. Ohmic) is central to realizing various device functions. Such a control, in principle, can be achieved by varying the junction energy barrier height, via using metals with different work functions. However, in conventional MSJs, the Fermi level is often pinned to a certain position within the semiconductor bandgap, rendering the energy barrier height insensitive to the metal work function. While the exact mechanisms of this Fermi level pinning are complex,1 the nature of the chemical bonds at the metal-semiconductor interface plays a critical role. Recently, two-dimensional (2D) layered semiconductors, particularly the transition metal dichalcogenides (TMDs), have been extensively explored as the material of choice for applications in electronics and optics. However, the MSJs based on conventional metals and TMD layers suffer from the formation of gap states, as a result of strong interactions between metal and TMD surface atoms.2-6 This has motivated efforts in using 2D metals as the contact material that bond to 2D semiconductors via the van der Waals interaction, since this weak bonding is expected to minimize the gap states formation.7 Most of the experimental efforts have centered on achieving Ohmic-type van der Waals MSJs. For example, by using metallic-phase MoS2 and WSe2 as the electrodes, low contact resistance and enhanced performance were achieved in 2D MoS2 and WSe2 transistors.8,9 However, stabilizing these metallic phases remains a challenge.10 Recently, a stable metallic-MoTe2 has been demonstrated that leads to lowresistance Ohmic contacts.11 On the other hand, 2D Schottky MSJs, with promising applications in novel field-effect Schottky barrier transistors and potentially in high-frequency electronics, have received less attention and so far are mostly limited to graphene/2D semiconductor junctions.12-14 The presence of interface carrier traps in such junctions has been suggested.15 3 ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

In this work, we demonstrate a van der Waals Schottky diode defined by the phase junction of multilayer In2Se3. Particularly, with the layered α (β) crystalline phase of In2Se3 showing semiconducting (metallic) characteristics, the vertical phase hetetrostucture (α – β) exhibits ideal diode behaviors with tunable current rectification. Most interestingly, an enhanced thermoelectric power (Seebeck coefficient) was observed in this phase junction, which is more than 3 orders of magnitude higher than that in single-phase multilayer and bulk In2Se3, with the thermoelectric figure-of-merit (ZT) at room temperature approaches ~ 1. Our findings suggest that the Schottky energy barrier, formed between van der Waals materials with the same composition and similar bonding nature, plays an important role in enhancing the thermoelectric performance. This “bulk” effect extends the advantages of 2D materials beyond the few-layer limit, and opens up new opportunities in thermoelectrics. Both α- and β-In2Se3 have a layered crystal structure with each quintuple layer (Se-In-Se-In-Se) bonded to each other through van der Waals interactions, as shown schematically in the insets to Figs. 1 (a) and 2 (a), respectively. The structural difference between α- and β-In2Se3 involves relative shifts between atomic layers inside each quintuple layer,16 with the difference between in-plane (vertical) lattice constants to be ~ 0.9% (~ 1.8%).17 Recent studies also suggested that both phases belong to the same space group (R-3m).18 These two phases can be identified by their phonon modes via Raman spectroscopy;17 particularly, the dominant A1 phonon mode is ~ 103 cm-1 and ~ 110 cm-1 for α- and β-In2Se3 layers, respectively [Figs. 1 (a) and 2 (a), respectively]. The similarity of these phonon mode frequencies is consistent with the structural similarity between these two phases. In bulk single crystals, the β-In2Se3 is unstable at room temperature. However, when in 2D form, the β phase can persist in ambient conditions.17 The phase transition (α → β) in multilayer In2Se3 can be driven by a simple thermal annealing at 4 ACS Paragon Plus Environment

Page 4 of 31

Page 5 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

moderate temperatures (~ 260 oC – 390 oC).17 The single crystallinity of the lattice is maintained during this phase transition, confirmed by our previous transmission electron microscopy results.17

Figure 1. (a) Raman spectrum of α-In2Se3 with schematic crystal structure shown in the inset; (b) output (with a representative optical microscopy image of a device shown in the inset), (c) transfer characteristics, (d) electrical resistance as a function of temperature, and (e) photocurrent spectroscopy of α-In2Se3 layers.

Here we first show that α- and β-In2Se3 layers have distinctive electrical properties, despite their similar lattice structures. For this, the α layers were obtained by exfoliating In2Se3 powders (Alfa Aesar) on SiO2 (90 nm)/Si substrates; a thermal annealing at 350 oC in Ar transformed the α phase into the β phase. For electrical measurements, In/Au electrodes, which form Ohmic contacts to In2Se3,17 were defined by e-beam lithography followed by metallization, with the Si substrate as the backgate. Figures 1 (b) and (c) show the output and the transfer characteristics of 5 ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

a representative α-In2Se3 layer device (thickness ~ 147 nm) respectively. The drain current, ID, increases by ~ 3 orders of magnitude, with the backgate voltage, VG, varying from -40 to 40 V. This suggests that α-In2Se3 is an n-type semiconductor. The field-effect mobility estimated from the transfer curves is up to ~ 55 cm2/Vs among our devices, consistent with other studies.19 In addition, the electrical resistance of α-In2Se3 decreases with increasing temperature [Fig. 1 (d)]; such a behavior is typically observed in semiconductors due to the thermal excitation of carriers from donor/acceptor states. Consistent with its semiconducting nature, photocurrent spectroscopy results [Fig. 1 (e)] show a bandgap energy at ~ 1.46 eV, in agreement with theoretical results16 and experimental observations.20

Figure 2. (a) Raman spectrum of β-In2Se3 with schematic crystal structure shown in the inset; (b) output (with a representative optical microscopy image of a device shown in the inset), (c) transfer characteristics, (d) electrical resistance as a function of temperature, and (e) photocurrent spectroscopy of β-In2Se3 layers.

6 ACS Paragon Plus Environment

Page 6 of 31

Page 7 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

In contrast, β-In2Se3 layers exhibits substantially lower electrical resistances [Fig. 2 (b); layer thickness ~ 86 nm]. While ID increases with the increasing VG [Fig. 2 (c)], indicating the n-type behavior that is consistent with our Hall-effect measurement results (not shown here), the relative variations of ID are much less significant than those of α-In2Se3. Furthermore, the electrical resistance of β-In2Se3 increases with increasing temperature [Fig. 2 (d)], demonstrating the metallic characteristic that can be attributed to the stronger carrier-phonon interactions at higher temperatures. Interestingly, the β-In2Se3 also shows a bandgap at ~ 1.38 eV from the photocurrent spectroscopy measurements [Fig. 2 (e)].a This suggests that the β-In2Se3 is a degenerate semiconductor with the Fermi level inside the conduction band, and thus behaves electrically as a metal. The origin of this metal-insulator transition during the α → β phase transformation is unclear and requires further studies, although the superlattice structure observed in β-In2Se3, which has been attributed to vacancy ordering21 or charge density waves,17 might play a role. Two types of phase junction devices, α on β (α/β) and β on α (β/α), were fabricated on SiO2 (90nm)/Si substrates. For the former, the bottom β-In2Se3 layers were obtained by annealing asexfoliated α-In2Se3 layers; for the latter, the top β-In2Se3 layers were exfoliated from powders purchased from 2D Semiconductors.b Before the exfoliation of the top layer, the bottom layer surface was cleaned by oxygen plasma and buffered HF etching. All devices were then annealed at 180 oC in Ar to improve the interface quality.

a

As the metallic β-In2Se3 layers are very electrically conductive, it was necessary to apply a small bias (20 mV) for the photocurrent spectroscopy measurements, so that the dark current did not overload the lock-in amplifier. This small bias resulted in weak photocurrent signals, which are quite noisy. In contrast, the semiconducting nature of the α-In2Se3 layers allowed us to apply a larger bias (0.5 V), which led to stronger photocurrent signals and less noises. b We note that not all samples as received are of the beta phase, consistent with the metastable nature of the beta phase in bulk In2Se3. 7 ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 3 (a) Experimental current-voltage relation (open circles) of an α/β junction device, with the dashed line representing the simulated result and the inset showing the optical image of an α/β junction device; (b) CPD map of an α/β junction with representative CPD profiles shown in (c); (d) schematic band alignments of α- and β-In2Se3; (e1) schematic and (e2) simulated band diagrams of an α/β Schottky junction; (f) current-voltage relations of a β/α junction under various backgate voltages.

Figure 3 (a) shows a representative current-voltage (ID vs. VDS) relation of an α/β device, which shows the rectifying characteristics with the β-In2Se3 biased. This current rectification was consistently observed in all devices. To understand this behavior, a knowledge of the band

8 ACS Paragon Plus Environment

Page 8 of 31

Page 9 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

diagram across the junction is required. For this, we performed scanning Kelvin probe microscopy (SKPM) on an α/β device with the drain/source electrodes grounded, as shown in Fig. 3 (b). The contact potential difference (CPD) signal from SKPM represents the work function difference between the sample surface and the SKPM probe tip. Therefore, the difference in the CPD signals from the α-In2Se3 and the β-In2Se3 regions corresponds to the work function difference between these two phases. Particularly, the β-In2Se3 has a larger work function than the α-In2Se3, and the difference is 0.25 ± 0.05 eV [Fig. 3 (c)]. With this, as shown schematically in Figs. 3 (d) and (e1), the electrons transfer from the α-In2Se3 to the β-In2Se3 at the interface, leading to the formation of the Schottky barrier and thus the rectifying ID vs. VDS relation in Fig. 3 (a). These results are consistent with density functional theory calculations (Supplementary Note 1). We further conducted numerical electrical simulations of the α – β junction (Supplementary Note 2). The simulated ID vs. VDS relation, with the α/β work function difference defined at 0.2 eV corresponding to the experimental value, is plotted in Fig. 3 (a). As the simulations assume ideal interface conditions (i.e. without interface states), the agreement between experimental and simulated results suggests that the α – β junctions demonstrate ideal Schottky diode behaviors. Fig. 3 (e1) shows a simulated band diagram, exhibiting the characteristic band bending at the interface that is consistent with the Schottky barrier. Moreover, the β/α devices, where the αIn2Se3 is at the bottom, allow for controlling the junction electrical properties via VG.c As shown in Fig. 3 (f), both forward and reverse currents can be modulated by about two orders of magnitudes within ± 40 V of VG. We note that, while the magnitude of the forward current, especially at large drain-source voltages, is largely controlled by the resistance of the α-In2Se3 c

For α/β devices, the bottom β-In2Se3 layer screens the backgate electric field, so the junction electrical properties cannot be controlled effectively via VG. 9 ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 31

(see also below) outside the junction area, the reverse current is a function of the Schottky barrier height. Therefore, the modulation of the reverse current represents a gate-tunable Schottky barrier height, indicating minimal Fermi level pinning at the interface and consistent with the ideal diode behaviors.

Figure 4 (a) Optical reflection and photocurrent maps of an α/β junction device under (b) 1 V and (c) 0 V biases; (d) equivalent circuit model for the α/β junction under the local laser excitation.

To probe the optoelectronic properties of the junctions, we used scanning photocurrent microscopy with the 532 nm laser excitation (see Methods). The laser power used was ~ 200 nW; such a low optical power was needed to avoid damaging the devices during extended periods of scanning. Figure 4 (b) shows the photocurrent map from an α/β device at VDS = 1 V, and the junction area can be identified from the optical reflection image [Fig. 4 (a)]. The thicknesses of the α-In2Se3 and β-In2Se3 layers are 64 nm and 50 nm, respectively. With the β-In2Se3 biased and the α-In2Se3 connected to a current pre-amplifier (virtual ground), the positive sign of the photocurrent represents the flow of the photocurrent from β to α. As shown in Fig. 4 (b), the photocurrent appears inside the α-In2Se3 region VDS = 1 V, indicating that this region has the largest electrical resistance and that the junction resistance is negligible. With the β-In2Se3 grounded (VDS = 0 V), the photocurrent occurs inside the junction region with a rather uniform spatial distribution [Fig. 4 (c)]. Most importantly, the photocurrent remains positive, i.e. the photocurrent still flows from β to α. To exclude the possibility of the “ground 10 ACS Paragon Plus Environment

Page 11 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

loop” effect, which can arise due to the different grounds (earth ground vs. virtual ground) connected to the α- and β-In2Se3, we exchanged the connections between these two terminals, and we observed the same photocurrent flow direction. We note that, given the band diagram of the junction [Figs. 3 (e1) and (e2)], the photogenerated electrons, from either the band-to-band excitation or the internal electron emission (from β-In2Se3), would move towards the α side due to the photovoltaic (PV) effect, resulting in a negative photocurrent. We thus believe that the photothermoelectric (PTE) effect, which has been observed in conventional metal-2D material junctions,22-24 is the dominant mechanism in this case. Particularly, the laser illumination can lead to a temperature gradient along the vertical direction of the junction, with the temperature higher in the top layer (α-In2Se3). This temperature gradient drives the electrons (which are the majority carriers) towards the β side, leading to a positive photocurrent as observed.d We note that, in contrast to previous studies, our semiconductor (top)/metal (bottom) configuration allows us to unambiguously differentiate the PV and the PTE effects, as they result in photocurrents with opposite signs. To evaluate the contributions from PV and PTE effects, we consider the equivalent circuit model shown in Fig. 4 (d): Rα, Rβ, and Rj are the resistances of the α/β-In2Se3 layers and the junction, respectively; the PTE effect is represented by a constant voltage source,25 with the PV effect represented by a constant current source;26 Iph is the experimentally measured photocurrent, and IPV is the photocurrent due to the PV effect. Simple circuit analysis (Supplemental Note 3) shows that, under the zero-bias condition (VDS = 0), Iph ≈ IPTE. In addition, under the zero-current

d

We note that the local optical illumination also leads to temperature gradients along the lateral direction. However, these lateral temperature gradients are rather symmetric around the laser beam, especially inside the junction area, as shown in Fig. S11 (c), so they do not produce a net thermoelectric current. 11 ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

condition (Iph = 0),

, where

Page 12 of 31

is the negative x-intercept in the Iph vs. VDS plot

(see also below).

Figure 5 (a) PL spectra from the α-In2Se3 layer and the junction area; (b) PL energy map of the junction device; (c) corresponding temperature map of the junction device under the local optical illuminations.

To further explore this PTE effect, we used the finite-element simulations to obtain the temperature variations along the vertical direction of the junction area. While there might be a presence of hot electrons under the local optical excitation, we used the lattice temperature here as the contribution from the hot electrons is minimal in our case, as discussed in Supplemental Note 4. As a calibration for our simulation model, we first obtained the temperatures of α-In2Se3 layers from micro-photoluminescence (PL) measurements; these experimental values were then used to calibrate the simulation model. Specifically, the α-In2Se3 layer exhibits a near-band-edge PL peak at ~ 1.36 eV, as shown in Fig. 5 (a), under the 532 nm laser excitation with the power of ~ 525 µW (note the much higher optical power used here than for the photocurrent measurements in order to detect clear PL energy shifts; see also below). In contrast, we did not observe any PL from the β-In2Se3, which can be attributed to the efficient non-radiative Auger process due to the large number of free electrons and is consistent with its indirect band gap (Supplementary Note 1). The PL spectrum of the junction area [Fig. 5 (a)] shows a lower peak 12 ACS Paragon Plus Environment

Page 13 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

energy, by ~ 34 meV, than that of the α-In2Se3; such a difference is also captured by the PL energy map shown in Fig. 5 (b). While this energy shift can be due to several possible mechanisms, including the change in the dielectric environment,27 charge transfer between materials,28 and the formation of interlayer type-II excitons,29 our additional PL measurements point to the higher temperature in the junction area under the local optical excitation as the origin of this PL energy shift (Supplementary Note 5). The connection between the PL and the temperature allows us to obtain the device temperature via the change in the PL peak energy (Supplementary Note 6), as shown in the temperature map of the device [Fig. 5 (c)]. This serves as a calibration of our thermal simulation model, where the thermal conductivity was measured in suspended α- and β-In2Se3 layers (Supplementary Note 7), by allowing us to match the simulated temperatures to the experimental values for the α-In2Se3 layer both inside and outside the junction area, by limited tuning of thermal parameters (Supplementary Note 8). With the calibrated simulation model, we calculated the average temperature across the α-β junction area under the photocurrent measurement conditions, ∆Tα–β, which is in the range of ~ 4 – 9 mK (Table S3 in Supplementary Note 8).e We note that ∆Tα–β is in line with simulated values obtained in SnS layers under optical excitations.f The PTE effect can then be described by the relation

e

∆Tα–β is significantly larger (with an average factor of ~ 5 across devices) than the temperature difference between the top and the bottom of the α-In2Se3 layer. This can be attributed to the relative low temperatures in the bottom β-In2Se3 layers that are in direct contact with the substrate (which acts as a heat sink). f In Ref. 23, the temperature rise in SnS layers under 45 µW optical excitation is 1.667 K. The thermal conductivity of SnS layers is close to that of In2Se3 layers. Assuming that the thermal conductivity is constant within a small temperature range, the temperature rise is proportional to the optical power. Therefore, the temperature rise in SnS layers under the 0.2 µW optical excitation would be ~ 7.4 mK, in good agreement with our simulated values.

13 ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 31

(1)

where ∆VPTE was obtained as the x-intercept from the Iph vs VDS curve [Fig. 6 (a)]. Using this relation, we obtained |Sjunction| ~ 105 – 106 µV/K [Fig. 6 (b)] for junctions with various thicknesses.

Figure 6 (a) Photocurrent-voltage relation of an α/β junction with another example shown in the inset; (b) -Sjunction as a function of the α-In2Se3 layer thickness (solid circles) and as a function of the total junction thickness (open circles), with the solid square indicating -Sα estimated from the Mott relation; (c) schematic demonstration of hot electron emission into the β-In2Se3 layer.

The value of S determined here is about more than 3 orders of magnitude higher than that of bulk In2Se3 (~225 µV/K),30 and also compares favorably to the value reported in SnS and MoS2 thin layers.23,24 It has been argued22 that the large value of S might be an artifact from the underestimation of ∆T: specifically, the photogenerated electrons can have higher temperatures than the lattice temperature. However, the presence of these “hot electrons” is typically observed under strong optical excitations, as opposed to the very weak laser power used in our study (Supplementary Note 4). Furthermore, our temperature simulation models were calibrated against the experimental results from the PL measurements, which are sensitive to the hot electron effects. Therefore, in our case, the optical excitation itself does not generate hot carriers. 14 ACS Paragon Plus Environment

Page 15 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

The dramatic increase of S in 2D layers compared to bulk materials has been attributed to the enhanced local electron density due to quantum confinement.23,31 In that case, S is expected to increase as the layer thickness decreases.23 However, as shown in Fig. 6 (b), |Sjunction| shows no obvious dependence on the α-In2Se3 layer thickness or the total junction thickness. Furthermore, most previous studies of the thin layer thermoelectric properties are based on the photothermoelectric current observed close to the metal electrodes; however, in most of our devices, we did not observe such a photocurrent at the metal electrodes (which are Ohmic contacts). This indicates a relatively small S in single-phase In2Se3 layers. To verify this, we estimated the Seebeck coefficient of the α-In2Se3, |Sα|, to be in the range of ~ 110 – 470 µV/K [Fig. 6 (b)], using the Mott relation, with parameters determined by experimental measurements (Supplemental Note 9). These values agree well with that determined for bulk α-In2Se3 (~225 µV/K).30 We expect the Seebeck coefficient of the β-In2Se3 to be even smaller than |Sα| because of its metallic nature. We therefore propose that the Schottky barrier at the α – β junction plays a critical role in enhancing Sjunction. Particularly, we believe that there are at least two possible mechanisms. In homogeneous materials, the transport of electrons induced by the temperature gradient leads to an accumulation of excess electrons at the cold end, setting up an electric field that opposes the thermally driven electron motion. In α – β junctions, the excess electrons are on the β side (cold end); however, the corresponding electric field is screened by the positive space-charge region associated with the Schottky barrier, shifting the balance towards the thermally driven current. Alternatively, one can consider that the counterbalancing field-driven electron transport, with the direction from β to α, is impeded by the barrier, thereby enhancing the thermally driven current and the Seebeck coefficient. 15 ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 31

In addition, as shown schematically in Fig. 6 (c), once the electrons, which are in thermal equilibrium with the α phase lattice, move across the junction interface into the β side, they become hot electrons due to the conduction band offsets between α- and β-In2Se3. The subsequent transport of these hot electrons enhances the Seebeck coefficient, which can be understood as the following. The Seebeck coefficient, S, can be written as in eq. (2),32

(2)

where e is the electric charge, σ is the electrical conductivity, ε is the electron energy, εF is the Fermi energy, and f (ε) is the Fermi-Dirac distribution, with σ (ε) as the differential conductivity that is proportional to the density of states. The term

is a bell-shaped symmetric function,33

with the peak at the Fermi energy and the width of ~ kBT, where kB is the Boltzmann constant. From eq. (2), it is clear that only those electrons with energy close to the Fermi level can contribute to S. Critically, the contributions from electrons with energies below and above the Fermi level are of the opposite signs; this reduces the magnitude of S. Therefore, by increasing the number of electrons above the Fermi level (i.e. hot electrons), S can be enhanced. This approach has been discussed extensively in previous theoretical studies,32-34 with one of them35 showing orders-of-magnitude enhancement of S. We acknowledge that further studies, which are beyond the scope of this work, are necessary to identify the exact origins for the enhanced thermoelectric properties. In both scenarios, we suggest that the minimal density of interface states in the Schottky junction, expected between van der Waals materials and demonstrated here, is required to minimize carrier trapping for enhancing S.

16 ACS Paragon Plus Environment

Page 17 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

Knowing Sjunction allows us to estimate the thermoelectric figure-of-merit (ZT) of the junction with the following relation:

, where

and

are the electrical conductivity

and the thermal conductivity in the direction perpendicular to the layers (out of plane), respectively, and T is the ambient temperature. We note that

and

thermoelectric transport is across the vertical junction. While simulations (Supplementary Note 8), it is difficult to estimate

are used here because the was determined from our

, although one would expect

to be smaller than the in-plane conductivity. However, as mentioned above (Fig. 4), the junction resistance under the forward bias, involving the electron transport along the vertical direction from α-In2Se3 to β-In2Se3 (which is the same as the electron thermoelectric transport direction), appears to be negligible compared to the in-plane resistance of the α-In2Se3 layer ( ). We note that, in In3Se4, which also has a layered crystal structure, ZT with ~ 0.15 for

is ~ 4.36 We therefore calculated

= 1 and 4. The values of ZT range from ~ 0.2 to ~ 0.6 for

= 1, and ~ 0.06 to

= 4 at room temperature. These values compare favorably to the recently

reported largest ZT (0.13) of thin SnS layers,23 and also approach the threshold for industrial applications.36,37 More importantly, the extent of the enhancement in the thermoelectric performance by the Schottky barrier is significant. In summary, we have demonstrated van der Waals Schottky diodes based on In2Se3 phase junctions. In addition to the ideal diode behaviors and gate-tunable current rectification, these Schottky junctions have enhanced the Seebeck coefficient, and the corresponding ZT values approach ~ 1 at room temperature. We expect that such an approach of using energy barriers between van der Waals materials, where the interface states are expected to be minimal, to enhance thermoelectric performance, can be adopted in other 2D materials as well. 17 ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 31

Experimental Methods Sample Preparations and Device Fabrications Devices were fabricated on SiO2(90 nm)/Si substrates (University Wafers), which were first sonicated in acetone and isopropanol. The α-In2Se3 layers were mechanically exfoliated from In2Se3 powders (Alfa Aesar) onto the PDMS film (Gel-Pak, PF-30/17-X4). Under an optical microscope, layers with desired and uniform thickness were selected. The PDMS film was then attached to a glass slide mounted on an XYZ mechanical manipulator. The pre-selected layers were transferred from the PDMS film onto SiO2/Si substrates using the all-dry viscoelastic stamping method.38 To remove the residue from the PDMS film, layers were rinsed in methanol, acetone and isopropanol. To obtain the β-In2Se3 layers, α-In2Se3 layers on SiO2/Si substrates were annealed in a hot stage (INSTEC, HCS302) at 350 oC for 10 mins, under a continuous flow of ultra-pure Argon. For vertical junction fabrications, the bottom layers were first cleaned in O2 plasma, followed by HF etching and DI water rinse. The top layers were then exfoliated. The junction devices were annealed in ultra-pure Ar at 180 oC to improve the interface quality. The electrodes on the devices were defined by e-beam lithography followed by metallization via either sputtering or thermal evaporations. For fabrication of suspended layer structures, SiO2(285 nm)/Si and Si3N4(200nm)/Si substrates (University Wafers) were sonicated in acetone and isopropanol. These substrates were then spincoated with MMA and PMMA. Circular windows (~ 5 µm in diameter) were defined by e-beam lithography. After the patterns were developed, the substrates were then etched by deep reactiveion etching, with the MMA/PMMA coating as the protection layer, leading to the formation of hole arrays. After the lift-off of the MMA/PMMA coating, α-In2Se3 layers were exfoliated over these holes per the procedures mentioned above. After the completion of thermal measurements 18 ACS Paragon Plus Environment

Page 19 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

on α-In2Se3 layers, these layers were transformed into β-phase layers by annealing at 350 oC for 10 minutes, followed by the same thermal measurements on these layers. Scanning Kelvin Probe Microscopy (SKPM) and Scanning Photocurrent Microscopy (SPCM) SKPM was performed using an atomic force microscopy (Nanonics) with Cr/Pt coated ATEC tips. Topographical and CPD signals were measured at different tip frequencies within a single pass. For SPCM, a dual-axis scanning galvo mirror system was used to raster scan the 532 nm laser beam, which was focused by a microscope objective (50X, NA = 0.75). A mechanical chopper was used to modulate the optical excitation with a typical frequency of 1 kHz; the photocurrent and the reflected laser beam intensity were measured simultaneously at each pixel using the lock-in detection. All measurements were conducted at room temperature in ambient air. Photocurrent Spectroscopy A Xenon 91192, 1000W Solar Simulator along with AM 1.5 global filter, which was coupled to a monochromator, was used as the excitation source. The responsivity was obtained by dividing the photocurrent by the optical power on the layer at a specific wavelength. These measurements were conducted at room temperature in N2. Micro-Photoluminescence (PL) Spectroscopy and Imaging

PL spectroscopy and imaging were performed using a PL microscope (Klar Scientific). An excitation laser beam of 532 nm wavelength was directed by a dichroic mirror (Thorlabs, DMLP650R) onto the sample surface through a microscope objective (Zeiss, EC Epiplan 50 X / 0.75). The PL emission was collected by a 200 µm diameter multimode fiber, which was connected to a spectrometer (Ocean Optics, Maya2000 Pro-NIR). A piezoelectric

19 ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 31

nanopositioning stage (Physik Instrumente, P-611.2S) was used to scan the sample. Motion control and data acquisition were implemented by a custom-written C++ program. The PL mapping data were analyzed using Python with the modified Savitzky Golay filtering algorithm. Variable-temperature PL was conducted by using a hot/cool chamber (Instec), which was purged by high-purity Ar during the measurements.

Acknowledgements:

This work was supported by the US National Science Foundation under grants DMR-1506480 and DMR-1561419. Z.W. and W.G.Z. acknowledge support from the National Natural Science Foundation of China (Grant Nos. 11374273, 11674299, 11634011) and the Fundamental Research Funds for the Central Universities (Grant No. WK2340000063). The authors thank Prof. Brian Collins’ group for the assistance in the photocurrent spectroscopy measurements.

Supporting Information:

Density functional theory calculations, numerical electrical simulations, equivalent circuit model for junction devices, discussions of hot electrons and origin of the photoluminescence (PL) peak shift, determination of the temperature from the PL peak shift, thermal conductivity measurements, finite-element thermal simulations, and calculation of the Seebeck coefficient from the Mott relation.

References: (1)

Tung, R. T. The Physics and Chemistry of the Schottky Barrier Height. Appl. Phys. Rev. 2014, 1, 011304.

20 ACS Paragon Plus Environment

Page 21 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

(2)

Xu, Y.; Cheng, C.; Du, S; Yang, J.; Yu, B.; Luo, J.; Yin, W.; Li, E.; Dong, S.; Ye, P. et al. Contacts between Two- and Three-Dimensional Materials: Ohmic, Schottky, and p-n Heterojunctions. ACS Nano 2016, 10, 4895-4919.

(3)

Das, S.; Chen, H. Y.; Penumatcha, A. V.; Appenzeller, J. High Performance Multilayer MoS2 Transistors with Scandium Contacts. Nano Lett. 2013, 13, 100-105.

(4)

Chen, W.; Santos, E. J. G.; Zhu, W. G.; Kaxiras, E.; Zhang, Z. Y. Tuning the Electronic and Chemical Properties of Monolayer MoS2 Adsorbed on Transition Metal Substrates. Nano Lett. 2013, 13, 509-514.

(5)

Gong, C.; Colombo, L.; Wallace, R. M.; Cho, K. The Unusual Mechanism of Partial Fermi Level Pinning at Metal-MoS2 Interfaces. Nano Lett. 2014, 14, 1714-1720.

(6)

Guo, Y. Z.; Liu, D. M.; Robertsont, J. 3D Behavior of Schottky Barriers of 2D Transition-Metal Dichalcogenides. ACS Appl. Mater. Interfaces 2015, 7, 25709-25715.

(7)

Liu, Y.; Stradins, P.; Wei, S.-H. Van der Waals Metal-Semiconductor Junction: Weak Fermi level Pinning Enables Effective Tuning of Schottky Barrier. Sci. Adv. 2016, 2, e1600069-6.

(8)

Kappera, R.; Voiry, D.; Yalcin, S. E.; Branch, B.; Gupta, G.; Mohite, A. D.; Chhowalla, M. Phase-Engineered Low-Resistance Contacts for Ultrathin MoS2 Transistors. Nat. Mater. 2014, 13, 1128-1134.

(9)

Ma, Y. G.; Liu, B. L.; Zhang, A. Y.; Chen, L.; Fathi, M.; Shen, C. F.; Abbas, A. N.; Ge, M. Y.; Mecklenburg, M.; Zhou, C. W. Reversible Semiconducting-to-Metallic Phase Transition in Chemical Vapor Deposition Grown Mono layer WSe2 and Applications for Devices. ACS Nano 2015, 9, 7383-7391.

21 ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(10)

Page 22 of 31

Guo, Y. S.; Sun, D. Z.; Ouyang, B.; Raja, A.; Song, J.; Heinz, T. F.; Brus, L. E. Probing the Dynamics of the Metallic-to-Semiconducting Structural Phase Transformation in MoS2 Crystals. Nano Lett. 2015, 15, 5081-5088.

(11)

Cho, S.; Kim, S.; Kim, J. H.; Zhao, J.; Seok, J.; Keum, D. H.; Baik, J.; Choe, D. H.; Chang, K. J.; Suenaga, K. et al. Phase Patterning for Ohmic Homojunction Contact in MoTe2. Science 2015, 349, 625-628.

(12)

Di Bartolomeo, A. Graphene Schottky diodes: An Experimental Review of the Rectifying Graphene/Semiconductor Heterojunction. Phys. Rep. 2016, 606, 1-58.

(13)

Tian, H.; Tan, Z.; Wu, C.; Wang, X. M.; Mohammad, M. A.; Xie, D.; Yang, Y.; Wang, J.; Li, L. J.; Xu, J. Novel Field-Effect Schottky Barrier Transistors Based on GrapheneMoS2 Heterojunctions. Sci Rep-UK 2014, 4, 5951.

(14)

Gao, A. Y.; Liu, E. F.; Long, M. S.; Zhou, W.; Wang, Y. Y.; Xia, T. L.; Hu, W. D.; Wang, B. G.; Miao, F. Gate-Tunable Rectification Inversion and Photovoltaic Detection in Graphene/WSe2 Heterostructures. Appl. Phys. Lett. 2016, 108, 223501.

(15)

Kim, W.; Li, C. F.; Chaves, F. A.; Jimenez, D.; Rodriguez, R. D.; Susoma, J.; Fenner, M. A.; Lipsanen, H.; Riikonen, J. Tunable Graphene-GaSe Dual Heterojunction Device. Adv. Mater. 2016, 28, 1845-1852.

(16)

Ding, W. Z., J.; Wang, Z.; Gao, Y.; Xiao, D.; Gu, Y.; Zhang, Z.; Zhu, W. Predictive Design of Intrinsic Two-Dimensional Ferroelectrics in In2Se3 and Other III2-VI3 Van der Waals Materials. Nat. Commun. 2017, 8, 14956.

(17)

Tao, X.; Gu, Y. Crystalline-Crystalline Phase Transformation in Two-Dimensional In2Se3 Thin Layers. Nano Lett. 2013, 13, 3501-3505.

22 ACS Paragon Plus Environment

Page 23 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

(18)

Rasmussen, A. M.; Teklemichael, S. T.; Mafi, E.; Gu, Y.; McCluskey, M. D. PressureInduced Phase Transformation of In2Se3. Appl. Phys. Lett. 2013, 102, 062105.

(19)

Island, J. O.; Blanter, S. I.; Buscema, M.; van der Zant, H. S. J.; Castellanos-Gomez, A. Gate

Controlled

Photocurrent

Generation

Mechanisms

in

High-Gain

In2Se3

Phototransistors. Nano Lett. 2015, 15, 7853-7858.

(20)

Jacobs-Gedrim, R. B.; Shanmugam, M.; Jain, N.; Durcan, C. A.; Murphy, M. T.; Murray, T. M.; Matyi, R. J.; Moore, R. L.; Yu, B. Extraordinary Photoresponse in TwoDimensional In2Se3 Nanosheets. ACS Nano 2014, 8, 514-521.

(21)

Peng, H. L.; Xie, C.; Schoen, D. T.; Cui, Y. Large Anisotropy of Electrical Properties in Layer-Structured In2Se3 Nanowires. Nano Lett. 2008, 8, 1511-1516.

(22)

Zhang, Y. W.; Li, H., Wang, L.; Wang, H. M.; Xie, X. M.; Zhang, S. L.; Liu, R.; Qiu, Z. J. Photothermoelectric and Photovoltaic Effects Both Present in MoS2. Sci Rep-Uk 2015, 5, 7938.

(23)

Lee, M. J.; Ahn, J. H.; Sung, J. H.; Heo, H.; Jeon, S. G.; Lee, W.; Song, J. Y.; Hong, K. H.; Choi, B.; Lee, S. H. et al. Thermoelectric Materials by Using Two-Dimensional Materials With Negative Correlation Between Electrical and Thermal Conductivity. Nat. Commun. 2016, 7, 12011.

(24)

Buscema, M.; Barkelid, M.; Zwiller, V.; van der Zant, H. S. J.; Steele, G. A.; CastellanosGomez, A. Large and Tunable Photothermoelectric Effect in Single-Layer MoS2. Nano Lett. 2013, 13, 358-363.

(25)

Siouane, S.; Jovanovic, S.; Poure, P. Equivalent Electrical Circuit of Thermoelectric Generators Under Constant Heat Flow. In 2016 IEEE 16th International Conference on

23 ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 31

Environment and Electrical Engineering (EEEIC), Florence, Italy, June 07-10, 2016; IEEE, New York, 2016; 1-6.

(26)

Sze, S. M. Physics of Semiconductor Devices; Jon Wiley & Sons: New York, 1981.

(27)

Lin, Y. X.; Ling, X.; Yu, L. L.; Huang, S. X.; Hsu, A. L.; Lee, Y. H.; Kong, J.; Dressehaus, M. S.; Palacios, T. Dielectric Screening of Excitons and Trions in SingleLayer MoS2. Nano Lett. 2014, 14, 5569-5576.

(28)

Kim, M. S.; Seo, C.; Kim, H.; Lee, J.; Luong, D. H.; Park, J. H.; Han, G. H.; Kim, J. Simultaneous Hosting of Positive and Negative Trions and the Enhanced Direct Band Emission in MoSe2/MoS2 Heterostacked Multilayers. ACS Nano 2016, 10, 6211-6219.

(29)

Rivera, P.; Schaibley, J. R.; Jones, A. M.; Ross, J. S.; Wu, S. F.; Aivazian, G.; Klement, P.; Seyler, K.; Clark, G.; Ghimire, N. J. et al. Observation of long-lived interlayer excitons in monolayer MoSe2-WSe2 heterostructures. Nat. Commun. 2015, 6, 6242.

(30)

Cui, J. L.; Wang, L.; Du, Z. L.; Ying, P. Z.; Deng, Y. High Thermoelectric Performance of a Defect in Alpha-In2Se3-Based Solid Solution Upon Substitution of Zn For In. J. Mater. Chem. C 2015, 3, 9069-9075.

(31)

Yoshida, M.; Iizuka, T.; Saito, Y.; Onga, M.; Suzuki, R.; Zhang, Y. J.; Iwasa, Y.; Shimizu, S. Gate-Optimized Thermoelectric Power Factor in Ultrathin WSe2 Single Crystals. Nano Lett. 2016, 16, 2061-2065.

(32)

Nishio, K.; Hirano, T. Improvement of the Efficiency of Thermoelectric Energy Conversion by Utilizing Potential Barriers. Jpn. J. Appl. Phys. 1 1997, 36, 170-174.

24 ACS Paragon Plus Environment

Page 25 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry Letters

(33)

Bahk, J. H.; Bian, Z. X.; Shakouri, A. Electron Energy Filtering by a Nonplanar Potential To Enhance the Thermoelectric Power Factor in Bulk Materials. Phys. Rev. B 2013, 87, 075204.

(34)

Berland, K.; Song, X.; Carvalho, P. A.; Persson, C.; Finstad, T. G.; Lovvik, O. M. Enhancement of Thermoelectric Properties by Energy filtering: Theoretical Potential and Experimental Reality in Nanostructured ZnSb. J. Appl. Phys. 2016, 119, 125103.

(35)

Shakouri, A.; LaBounty, C.; Abraham, P.; Piprek, J.; Bowers, J. E. Enhanced Thermionic Emission Cooling in High Barrier Superlattice Heterostructures. Mater. Res. Soc. Proc 1999, 545, 449.

(36)

Rhyee, J. S.; Lee, K. H.; Lee, S. M.; Cho, E.; Il Kim, S.; Lee, E.; Kwon, Y. S.; Shim, J. H.; Kotliar, G. Peierls Distortion As a Route To High Thermoelectric Performance in In4Se3-Delta Crystals. Nature 2009, 459, 965-968.

(37)

Zhao, L. D.; Lo, S. H.; Zhang, Y. S.; Sun, H.; Tan, G. J.; Uher, C.; Wolverton, C.; Dravid, V. P.; Kanatzidis, M. G. Ultralow Thermal Conductivity and High Thermoelectric Figure of Merit in SnSe Crystals. Nature 2014, 508, 373.

(38)

Castellanos-Gomez, A.; Buscema, M.; Molenaar, R.; Singh, V.; Janssen, L.; van der Zant, H. S. J.; Steele, G. A. Deterministic Transfer of Two-Dimensional Materials by All-Dry Viscoelastic Stamping. 2D Mater. 2014, 1, 011002.

25 ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Paragon Plus Environment

Page 26 of 31

Page 27 of 31

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Paragon Plus Environment

Page 28 of 31

Page 29 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

The Journal of Physical Chemistry Letters

ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

ACS Paragon Plus Environment

Page 30 of 31

Page 31 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38

The Journal of Physical Chemistry Letters

ACS Paragon Plus Environment