Correction: Applied Physical Sciences
Correction: Applied Physical Sciences
Correction: Applied Physical Sciences
CORRECTION
University, Shenzhen, 518060, China. The corrected author
and affiliation lines appear below. The online version has been
corrected.
www.pnas.org/cgi/doi/10.1073/pnas.2012778117
Downloaded at DKFZ-HGF on August 16, 2020
Edited by Eva Y. Andrei, Rutgers, The State University of New Jersey, Piscataway, NJ, and approved May 11, 2020 (received for review April 22, 2020)
Local impurity states arising from atomic vacancies in two-dimensional impurity states, especially magnetic impurities, not only strongly
APPLIED PHYSICAL
(2D) nanosheets are predicted to have a profound effect on charge couple with itinerant charge carriers but can also significantly
transport due to resonant scattering and can be used to manipulate affect the band structure and even reverse charge transport behavior
thermoelectric properties. However, the effects of these impurities (6–8). Few-atoms-thick layered materials are highly sensitive
SCIENCES
are often masked by external fluctuations and turbostratic inter- to such extrinsic effects, and therefore, they serve as a good
faces; therefore, it is challenging to probe the correlation between platform for exploiting thermoelectric properties by intro-
vacancy impurities and thermoelectric parameters experimentally. ducing such impurities. Among these materials, MoS2 has
In this work, we demonstrate that n-type molybdenum disulfide attracted special attention because of its well-defined spin-
(MoS2) supported on hexagonal boron nitride (h-BN) substrate re- splitting under light illumination and/or applied magnetic
veals a large anomalous positive Seebeck coefficient with strong fields (9–11). Extrinsic magnetic impurities in MoS2 can be
band hybridization. The presence of vacancies on MoS2 with a large introduced by vacancies, dislocations, edges, strain, or doping
conduction subband splitting of 50.0 ± 5.0 meV may contribute to by magnetic ions (12–16). However, due to its extreme sen-
Kondo insulator-like properties. Furthermore, by tuning the chem- sitivity to external fluctuations and turbostratic interfaces,
ical potential, the thermoelectric power factor can be enhanced by
up to two orders of magnitude to 50 mW m−1 K−2. Our work shows
Significance
that defect engineering in 2D materials provides an effective strat-
egy for controlling band structure and tuning thermoelectric
transport. The study of correlated phenomena in 2D semiconductors
opens up new pathways toward understanding and engi-
thermoelectric | Kondo | 2D materials | Seebeck | phonon drag neering material functionalities (such as thermoelectrics) in
easily accessible van der Waals solids. Local structural defects
such as vacancies inevitably exist in natural as well as synthetic
A B 4 D
3
Energy (eV)
1
6 Layer
0 MoS2
-1
15 nm
-2
Γ M K Γ
C 0.6
4
0.5 C4 E
Band 40
Sample Bias (eV)
0.4 Hybridization
0
C3b C3 C3 VSul
DOS (states/eV)
0.3 20
C3a
0.2 C2b C2 0
C2
C2a
0.1 C1b C1 C1
-20 -1
VSul
C1a
0.0
-40
-3 -2 -1 0 1 2
-0.1
Energy (eV)
dI/dV (a.u.)
Fig. 1. Electronic properties and band structure of six-layer MoS2 with sulfur vacancy. (A) Large-area STM image of the bare MoS2 on h-BN substrate (VS =
1.0 V, I = 0.7 nA). The dark topographic contrast, which shows the single sulfur vacancy features, can be found over the whole scanned area. (B) DFT calculation of
six-layer MoS2. The zoomed-in image shows the conduction subbands (Cn) near the CBM at point of six-layer MoS2. (C) dI/dV spectrum (VS = −0.7 V, I = 1.3 nA)
at the vicinity of a sulfur vacancy (Left side) (scale bar, 0.6 nm) and pristine region (Right) (scale bar, 0.8 nm) of MoS2 near CBM. Conduction subbands C1, C2,
and C3 can be revealed from the resonance peaks in STS curve of pristine MoS2. However, each of those subbands split into two bands (Cna and Cnb) due to
band hybridization in the vicinity of a sulfur vacancy. (D) Top and side views of spin density associated with single sulfur vacancy in monolayer MoS2. (Purple
larger spheres represent molybdenum atoms, and smaller yellow spheres represent sulfur atoms.) The spin density associated with VSul−1, with a significant
component localized at the three exposed Mo atoms. (E) Electronic DOS of the sulfur vacancy in neutral (VSul0) and −1 charged state (VSul−1) (for intuitive
understanding, the VBMs for the two cases were aligned).The charging of the vacancy leads to the lifting of the degeneracy of defective level indicated by the
splitting of curves of spin-up (dashed line) and spin-down (solid line) states for VSul−1.
APPLIED PHYSICAL
dependent transport measurements were carried out. Fig. 2A shows However, for the MoS2/h-BN device, instead of saturating at a
the top and section view of the MoS2/h-BN heterostructure, and a residual value as T→0, a conductance peak (temperature at this
representative optical image is shown in Fig. 2B. To be consistent, a critical point is defined as Tmax; red dashed line in Fig. 3A) in the
SCIENCES
MoS2 flake with thickness 4.2 ± 0.3 nm was selected (six layers; Fig. metallic region is observed. This observed anomalous resistance
1C) (more transport results of various thicknesses are provided in SI minimum at low temperature is another signature of a Kondo
Appendix, Note S15). Four-probe measurements were performed to insulator behavior.
exclude the effects of electrical contact resistance, and the nano- This effect is suppressed in the MoS2/SiO2 device due to the
fabricated heater allows thermoelectric measurements. The linear imperfect interface between MoS2 and SiO2 but clearly detect-
Isd–Vsd curves collected at room temperature (Fig. 2C) indicate able for the six-layer MoS2/h-BN devices (more data on six-layer
A B
5
1
2
3
4
Thickness
4
Ti / Au 2
MoS2
0
h-BN
0.0 0.5 1.0 1.5 2.0 2.5
Distance ( μm )
C D
0.04 70V 60V 200 T (K) 20
50V 40V 50
100
0.02 150 150
200
Vsd (V)
G ( μS)
0.00 250
100
300
-0.04 10V 0V 0
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 -60 -40 -20 0 20 40 60 80
I sd (PA) V (V)
g
Fig. 2. Structural and electronic properties of MoS2/h-BN heterostructure. (A) Schematic diagram of the device. Electrode 1 acts as an electrical heater.
Electrodes 2 and 5 act as a current source for four-probe electrical measurements, while electrodes 3 and 4 act as thermometers. (Bottom) The section view of
the heterostructure. (B) Optical image of a complete device. The dotted red dashed box outlines the MoS2 flake. (Scale bar, 10 μm.) (Bottom) AFM scan gives
the thickness of MoS2 flakes as 4.2 ± 0.3 nm. (C) Four-probe Isd-Vsd curves at different Vg values at 300 K. (D) Four-probe FET characterization at different
temperatures. Clear crossover can be found at Vg ∼ 20V for T > 100 K, which is defined as the MIT. For Vg > VMIT, MoS2 shows metallic behavior, and con-
ductance decreases with increasing T. For Vg < VMIT, conductance increases with increasing T, which is a typical insulating behavior. When T < 100 K, the
conductance (in the metallic region) drops anomalously as T decreases.
μ ( cm V S )
-2.1
~T
-1
Vg 5 200
Vs (e 2 /h)
Vs (e 2 /h)
4 Vg
-1
4 70V
3 70V
2
2 3
50 MoS2/h-BN -1.7
2
~T
1 35 MoS2/SiO2
0 10V 1 10V
20
50 100 150 200 250 300 50 100 150 200 250 300 10 25 40 70 100 250 400
T (K) T (K) T (K)
D E F
90 1.0
15 Vg =70V NRG Calculation
80 0.8
Kondo Fit
5
70
Vg =50V 0.6 Vg (V)
R ( KΩ )
RK / R K 0
25
T K (K)
60 10
15 ~ Vcritical 0.4 20
Kondo Fit 30
50
300 40
Vg=30V 50
0.2
40 60
100 Kondo Fit 70
30 0.0
0 100 200 300 10 20 30 40 50 60 70 0.1 1 10
T (K) Vg (V) T / TK
Fig. 3. Carrier density and temperature-dependent properties of defective MoS2 on h-BN substrate. (A) Four-probe sheet conductivity of MoS2/h-BN devices
as a function of T and Vg. MIT can be observed when σs ∼ e2/h. The anomalous peaks Tmax at low temperatures are marked out by the red dashed line. (B)
Four-probe sheet conductivity of MoS2/SiO2 devices as a function of T and Vg. (C) Temperature-dependent field effect mobility. The solid line shows the
phonon-limited power law μph ∼ T −γ. γ = 1.7 and μ ∼ 200 cm2 V−1 s−1 for MOS2/SiO2 sample. For MoS2 on h-BN, γ = 2.1 and μ reaches 405 cm2 V−1 s−1. For T < 100 K,
dμFE =dT > 0 is observed for MoS2/h-BN sample indicating an anomalous scattering mechanism. (D) Temperature dependence of four-probe resistance at Vg =
70, 50, and 30 V for MoS2/h-BN device, with the resistance minima at 70, 89, and 135 K, respectively. (E) Gate tuning of the Kondo temperature Tk. (F)
Normalized Kondo resistance versus normalized temperature for the data from Vg = 10 V to Vg = 70 V. The red dashed line describes the universal Kondo
behavior from numerical renormalization group calculations (35).
MoS2/SiO2 and MoS2/h-BN as well as monolayer and bilayer TK), magnetic impurities quantum-mechanically exchange spin
MoS2/h-BN are in SI Appendix, Note S15). At low temperatures, with conduction electrons of many wave vectors (momenta)
the hybridization of the conduction band with the localized impurity (30), thereby creating a resonant scattering state at the Fermi
state manifests in a large band splitting. Kondo resonant scattering
level with a width ∼ kB TK [kB TK ∼ Δ exp DðE−1F ÞJ0 , where D(EF)
can dominate charge transport in this temperature range and give
rise to the anomalous component of the resistance. Such a picture is is the electron DOS at the Fermi level, Δ is the bandwidth, and
also complemented by the change in four-probe mobility ðμÞ with J0 is a constant representing a Kondo scattering exchange en-
temperature (Fig. 3C). For MoS2 on SiO2, μ is mainly limited by ergy] (31, 32). As a result, a dip in the resistance vs. temper-
optical phonon scattering when T > 100 K, and its temperature ature curve appears. Fig. 3E summarizes the TK variation as a
dependence follows a power law of μ ∼ T −γ (27). Once T < 100 K, function of Vg for MoS2/h-BN sample. At Vg > VMIT, TK re-
most optical phonons are deactivated, and saturation is observed, mains nearly unchanged because of the constant 2D DOS.
resulting in dμ=dT ∼ 0 as T→ 0 (27). On the contrary, MoS2/h-BN When Vg < VMIT, the Fermi level shifts toward the gap region;
device deviates from this trend where its mobility decreases as T < therefore, TK starts to increase with lower doping, similar to the
100 K and shows dμ=dT > 0. behavior observed in heavily doped fermion bulk systems such as
Fig. 3D shows the four-probe resistance plots of MoS2/h-BN CePd3 (33). We also scale the R(T) curves at each Vg and observe a
device at three different back gates. For each curve, a clear re- universal Kondo behavior (6, 34), in which the normalized
sistance minimum can be observed at 70, 89, and 135 K. For a Kondo resistance R k /R k0 vs. T/Tk at all gate voltages collapses
better understanding of this data, apart from 2D Block–Gruneisen onto a single universal Kondo behavior curve (Fig. 3F). This is in
resistance ð ∝ T 4 Þ (28, 29) and electron–phonon high temperature accordance with numerical renormalization group theory (35),
resistance ð ∝ TÞ, we employ a Kondo resistance model [RK ðTK , TÞ, which attests to the Kondo-driven transport in MoS2/h-BN devices.
full expression in SI Appendix, Note S11]:
Thermoelectric Performance Driven by Band Hybridization. To in-
R = AT 4 + BT + RK ðTK , TÞ + R0 , [1] vestigate further if the thermoelectric values show a similar ef-
fect, a DC current is introduced via a nanofabricated heater to
where R0 is the temperature-independent term arising from a create a temperature gradient along the devices (Fig. 2A). Fig.
residual zero temperature resistance (Fig. 3D). In the Kondo pic- 4A shows the Seebeck coefficient (S) as a function of tempera-
ture, close to a characteristic temperature (Kondo temperature, ture at Vg = 70, 50, and 30 V for both MoS2/h-BN and SiO2
S (mV/K)
MoS2 to the Seebeck coefficient via the phonon drag effect, Sph 1.0
(40). Both MoS2/SiO2 and MoS2/h-BN samples show similar
values of negative S and behavior for T > 120 K. However, when 0.5
T < 120 K, an anomalous sign change (from negative to positive) MoS2 / SiO2
in S is observed in MoS2/h-BN sample even though electrons
remain as the dominant charge carriers. The positive S reaches a 0.0
magnitude as high as 2 mV/K for Vg = 30 V at 60 K, which is
almost one order of magnitude enhancement compared to -0.5
MoS2/SiO2 sample (−270 μV/K) under the same experimental
conditions. This is a remarkably large value for a material in the 0 50 100 150 200 250 300 350
metallic state. To better illustrate the sign change of S, a 2D map
of S as a function of both T and Vg is plotted in Fig. 4B. Clearly, S
T (K)
changes from negative to positive values at all gate voltages. It is B
also found that the critical temperature (defined as the tem- 70
-0.6
perature at which S = 0) and the maximum positive S (highest
intensity; red in Fig. 4B) increases as Vg decreases, suggesting 60 0.0
APPLIED PHYSICAL
that the Seebeck coefficient can be tuned by electrostatic gating.
Importantly, this critical temperature across all temperatures for
S = 0 is identical to the conduction maximum line Tmax in Fig. 3A, V g (V) 50 0.6
SCIENCES
further evidencing that this anomalous thermoelectric transport
has the same origins as the electronic transport. 40 1.2
Since the transfer curve shows unambiguous n-type behavior
and the quasiparticle bandgap of six-layer MoS2 is relatively large
(41), the valence band states of MoS2 are inaccessible and hence
30 1.8
cannot result in a positive S, especially at low temperatures. When
electrons are the majority carriers, the single-particle Boltzmann 20 S
transport equation (BTE) can be employed to describe thermo- (mV/K)
electric transport and is defined as SBTE. In the degenerate state 10
(highly conducting on-state), the BTE reduces to the Mott formula 50 100 150 200 250 300
described below (42):
T (K)
π 2 k2B T ∂ln τ ∂ln g ∂ln v2 C
SBTE = − + + = Sτ + SN , [2]
3 jej ∂E ∂E ∂E E=EF
0.0
B ∂lnτ π 2 kB T ∂ln g ∂ln v2
2 k2 T 2
with Sτ = −π3 jej ∂E and S = − + .
E=EF
N 3 jej ∂E ∂E
E=EF
-0.1
Here kB, τ, g, and v indicate the Boltzmann constant, re-
S (mV/K)
A 1.0
B
1.5 Experiment Data
0.5 S>0
S (mV/K)
S N + S ph
1.0 S Fitting
0.0
S total
S (mV/K)
-0.5 0.5
μ(cm V S )
-1
400
-1
0.0
2
200
Tcritical
-0.5
0
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300
T (K)
T (K)
C D 60
MoS2/h-BN
100 MoS2/h-BN Vg (V) 70
60 This work
50 G/h-BN
40
PF ( mWm-1 K -2)
PF ( mWm K )
40
-2
10
-1
20 MoS2/SiO2
G/SiO2
1
TiS2
0.4 WSe2
MoS2/SiO2
0.1
0.2 Black P
0.0
20 100 200 1 10 100 1000 10000
T (K) 2
μ ( cm V S )
-1 -1
Fig. 5. Thermolectric performance of MoS2/h-BN heterostructure. (A) Seebeck coefficient and mobility as function of temperature at Vg = 70 V for MoS2/h-BN
device. (B) Total S values of MoS2/h-BN sample at n = 2 × 1012 cm−2 are contributions from the energy-dependent diffusive part SN, phonon-drag part Sph, and the
Kondo scattering part Sτ . At a fixed n, the total S first exhibits a diffusive negative value at high temperatures from conducting electrons described by SN + Sph. As
the temperature decreases, the conventional diffusive contribution is weakened, and the Kondo scattering term Sτ starts to dominate and shows large positive
values. As temperature is decreased further, all of the physical interactions start to freeze, and the total S goes back to zero as expected (64). (C) Comparison of PF of
MoS2/h-BN and MoS2/SiO2 sample as a function of temperature at different gate voltages. Additional band hybridization-induced peaks as high as 50 mW m−1 K−2
can be observed at 30 K ∼ 50 K for MoS2/h-BN sample. (D) The PF value of our MoS2/h-BN heterostructure shows superior thermoelectric performance compared to
even the highest reported values in other 2D materials at room temperature. [G/SiO2 and h-BN (50), MoS2 (36, 37), WSe2 (60), TiS2 (61, 62), and BP (51).]
APPLIED PHYSICAL
current mode was employed to capture STM images, and external lock-in
PF of the MoS2/h-BN sample is comparable with that of the MoS2/ technique was employed to record STS data. Four-probe measurements
SiO2 sample. However, with decreasing temperature, the PF of the were employed to eliminate contact resistance and extract a precise elec-
SCIENCES
MoS2/h-BN sample drops significantly due to the positive Sτ con- trical conductivity of MoS2 flakes for different gate voltages. The
tribution which cancels the contribution from traditional negative temperature-dependent mobility is extracted from transfer characteristics,
SN. When temperature further decreases, Kondo scattering starts via μFE = C1ox dV
dσ
g
, where Cox = 12 × 10−9
F=cm is the capacitance for 285-nm-
2
to dominate the thermoelectric transport. A large enhancement in thick SiO2. The sheet conductivity σ s = WL G is converted from G by con-
PF up to two orders of magnitude is observed. Such a high PF sidering the length (L) and width (W) of the MoS2 flake, and electrical
value of 50 mW m−1 K−2 (T = 30 to 50 K), originating from band conductivity is calculated from σ = σds by considering its thickness (d). To
hybridization-enhanced S, is two to four orders of magnitude measure the Seebeck coefficient S = VTEP/ΔT, a temperature gradient is re-
larger than the PF of other 2D materials (36, 37, 50, 51, 60–62) alized by applying a DC current bias (I) through a metal heater using a
in the same temperature regime and exhibits the highest PF in 2D source meter (Keithley 6221); therefore, Joule heat and a temperature
materials at all reported temperatures (Fig. 5D). gradient along the sample are generated. The temperature gradient gen-
erates the thermoelectric voltage VTEP, which is measured by a voltmeter
Conclusion (Keithley 6430). As a series of DC current bias with increasing values are ap-
plied to the heater, the corresponding VTEP were recorded. The temperature
In summary, we discover that band hybridization due to sulfur
gradient is measured through the four-probe resistance of thermometers.
vacancies exerts a strong influence on the thermoelectric prop-
erties of few-layer MoS2 supported on h-BN substrate, whereas
Data and Materials Availability. All data supporting the findings of this study
these effects are suppressed on SiO2 substrate. The electric and are available within the article or SI Appendix.
thermoelectric transport in few-layer MoS2 on h-BN match well
with a strongly correlated Kondo-like behavior, leading to a large ACKNOWLEDGMENTS. This research was supported by Agency for Science,
anomalous positive Seebeck coefficient of 2 mV/K in n-type MoS2 Technology and Research (A*STAR) Pharos Funding from the Science and
in the on state. Importantly, our work demonstrates that such an Engineering Research Council (SERC) (Grant 152 70 00015). K.P.L. acknowl-
effect can be electrostatically tuned to manipulate the Seebeck edges Ministry of Education (MOE) Tier 3 Grant “Two-dimensional crystal
quantum exciton photonic,” Project MOE2018-T3-1-005. C.-W.Q. acknowl-
coefficient, leading to two orders of magnitude enhancement in
edges the A*STAR SERC Pharos Grant 152 70 00014 with Project R-263-
thermoelectric PF. The ability to exhibit both negative Seebeck 000-B91-305. A.H.C.N. acknowledges the National Research Foundation
(diffusive) and positive Seebeck (band hybridized) coefficients in (NRF) Competitive Research Programme Award R-144-000-295-281. We also
n-type MoS2 suggests that a singly doped material could be used to thank Professor Zheng Yi for the helpful discussion and comments.
1. G. J. Snyder, E. S. Toberer, Complex thermoelectric materials. Nat. Mater. 7, 105–114 10. Z. Y. Zhu, Y. C. Cheng, U. Schwingenschlögl, Giant spin-orbit-induced spin splitting in
(2008). two-dimensional transition-metal dichalcogenide semiconductors. Phys. Rev. B Con-
2. J. Wu, Y. Chen, J. Wu, K. Hippalgaonkar, Perspectives on thermoelectricity in layered dens. Matter Mater. Phys. 84, 1–5 (2011).
and 2D materials. Adv. Electron. Mater. 14, 1800248 (2018). 11. K. Komider, J. W. González, J. Fernández-Rossier, Large spin splitting in the con-
3. C. Chang et al., 3D charge and 2D phonon transports leading to high out-of-plane ZT duction band of transition metal dichalcogenide monolayers. Phys. Rev. B Condens.
in n-type SnSe crystals. Science 360, 778–783 (2018). Matter Mater. Phys. 88, 1–7 (2013).
4. L. D. Zhao et al., Ultralow thermal conductivity and high thermoelectric figure of 12. L. Cai et al., Vacancy-induced ferromagnetism of MoS2 nanosheets. J. Am. Chem. Soc.
merit in SnSe crystals. Nature 508, 373–377 (2014). 137, 2622–2627 (2015).
5. F. Ghahari et al., Enhanced thermoelectric power in graphene: Violation of the Mott 13. A. S. Rodin, A. H. C. Neto, Localized magnetic states in two-dimensional semicon-
relation by inelastic scattering. Phys. Rev. Lett. 116, 136802 (2016). ductors. Phys. Rev. B 97, 235428 (2018).
6. J. H. Chen, L. Li, W. G. Cullen, E. D. Williams, M. S. Fuhrer, Tunable Kondo effect in 14. J. Zhang et al., Magnetic molybdenum disulfide nanosheet films. Nano Lett. 7, 2370–
graphene with defects. Nat. Phys. 7, 535–538 (2011). 2376 (2007).
7. A. Demion, A. D. Verga, Nonlinear electric transport in graphene with magnetic 15. H. Pan, Y. W. Zhang, Tuning the electronic and magnetic properties of
disorder. Phys. Rev. B Condens. Matter Mater. Phys. 90, 1–9 (2014). MoS 2 nanoribbons by strain engineering. J. Phys. Chem. C 116, 11752–11757
8. D. Xiao, G. Bin Liu, W. Feng, X. Xu, W. Yao, Coupled spin and valley physics in mono- (2012).
layers of MoS 2 and other group-VI dichalcogenides. Phys. Rev. Lett. 108, 1–5 (2012). 16. S. W. Han et al., Electron beam-formed ferromagnetic defects on MoS2 surface along
9. Q. H. Wang, K. Kalantar-Zadeh, A. Kis, J. N. Coleman, M. S. Strano, Electronics and 1 T phase transition. Sci. Rep. 6, 38730 (2016).
optoelectronics of two-dimensional transition metal dichalcogenides. Nat. Nano- 17. K. F. Mak, K. He, J. Shan, T. F. Heinz, Control of valley polarization in monolayer MoS2
technol. 7, 699–712 (2012). by optical helicity. Nat. Nanotechnol. 7, 494–498 (2012).