Photocurrent Switching of Monolayer MoS2 Using ... - ACS Publications

Dec 28, 2016 - Jin Hee Lee†‡, Hamza Zad Gul†‡, Hyun Kim†‡, Byoung Hee Moon‡, Subash Adhikari†‡, Jung Ho Kim†‡, Homin Choi†‡,...
0 downloads 0 Views 873KB Size
Subscriber access provided by Fudan University

Communication 2

Photocurrent Switching of Monolayer MoS using Metal–Insulator Transition Jin Hee Lee, Hamza Zad Gul, Hyun Kim, Byoung Hee Moon, Subash Adhikari, Jung Ho Kim, Homin Choi, Young Hee Lee, and Seong Chu Lim Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.6b03689 • Publication Date (Web): 28 Dec 2016 Downloaded from http://pubs.acs.org on December 28, 2016

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.

Nano 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 19

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

Nano Letters

Photocurrent Switching of Monolayer MoS2 using Metal–Insulator Transition Jin H. Lee,1,2,€ Hamza Z. Gul,1,2,€ Hyun Kim,1,2 Byoung H. Moon,2 Subash Adhikari,1,2 Jung H. Kim,1,2 Homin Choi,1,2 Young H. Lee1,2,* and Seong C. Lim1,2,* 1

Department of Energy Science, Sungkyunkwan University, Suwon 16419, Republic of Korea.

2

Center for Integrated Nanostructure Physics, Institute for Basic Science (IBS), Suwon 16419,

Republic of Korea. €

These two authors contributed to this work equally.

*

Correspondence should be addressed to: [email protected], [email protected]

ACS Paragon Plus Environment

Nano 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 achieve switching on/off the photocurrent of monolayer molybdenum disulfide (MoS2) by controlling the metal–insulator transition (MIT). N-type semiconducting MoS2 under a large negative gate bias generates a photocurrent attributed to the increase of excess carriers in the conduction band by optical excitation. However, under a large positive gate bias, a phase shift from semiconducting to metallic MoS2 is caused, and the photocurrent by excess carriers in the conduction band induced by the laser disappears due to enhanced electron–electron scattering. Thus, no photocurrent is detected in metallic MoS2. Our results indicate that the photocurrent of MoS2 can be switched on/off by appropriately controlling the MIT transition by means of gate bias.

Keywords: MoS2, Metal–insulator transition, Photocurrent switching, Optical phonon scattering, Electron–electron scattering

ACS Paragon Plus Environment

Page 2 of 19

Page 3 of 19

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

Nano Letters

Semiconducting transition metal dichalcogenides (s-TMDs), including MoS2, WSe2, and WS21-3, with an electronic band gap of 1–2 eVs, have attracted worldwide attention as alternative future electronic materials. The single atomic thickness and the absence of dangling bonds of s-TMDs is expected to benefit future sub-microelectronics since the surface scattering of electrons in the channel materials—an important limiting factor of Si devices as the device size scales down—is suppressed with s-TMDs. Furthermore, unlike graphene, the optical band gap of s-TMDs broadens their application area to optoelectronic devices such as photodetectors and solar cells4-8. The modulation of the electrical properties of MoS2 has recently been achieved by controlling the gate bias. Depending on the carrier concentration n, MoS2 changes from an insulator to a metal. This constitutes the so-called metal–insulator transition (MIT)9, 10. The dramatic change in the electrical behavior of MoS2 significantly affects its optical properties. For instance, at low carrier concentrations, MoS2 illuminated by a laser with Eλ > Eg, forms electron–hole pairs, excitons, owing to a strong Coulomb interaction such that luminescence is dominated by exciton states. However, at high carrier concentrations, excitons bind with electrons, forming negative trions11-13. Then, the luminescence peak shifts to a lower energy than the exciton binding energy. Generally, the photocurrent of MoS2 can be generated due to a few different causes. In addition, its modulation is demonstrated by controlling the barrier height or band alignment using the gate bias14 and by exposing it to gas adsorbates15. In the channel of MoS2 FET, a strong photovoltaic effect is observed14,16. In addition, a large and tunable thermoelectric effect is observed from the layered

materials17,18.

Therefore,

the

metal-MoS2

junction

can

exhibit

intense

photothermoelectric effects18,19 owing to the different Seebeck coefficients of the layers. MoS2 has shown various ways of optical response to the light, which are correlated with different physical aspects of MoS2. Similarly, the photocurrent of MoS2 is expected to be affected by the

ACS Paragon Plus Environment

Nano 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

metallicity of the MoS2 channel, the latter being controlled by the gate bias. However, an MITbased photocurrent modulation has not been reported yet. In this study, we characterize the photocurrent as a function of the back gate bias, which transforms insulating MoS2 to metallic MoS2. At the insulator state, the photocurrent significantly increases in response to optical input, resulting from an increase in the amount of excess carriers. However, at the metallic state, the photocurrent is not generated under illumination. The absence of a photocurrent in the metallic state is attributed to increased carrier– carrier scattering that nullifies the generation of excess carriers in the MoS2 conduction band. Our results indicate that the photocurrent can be switched on/off using the MIT transition. A monolayer MoS2 was synthesized using chemical vapor deposition20. Monolayer flakes grew into triangles with each side length of about 30–50 µm. They were randomly located on a substrate, as shown in Fig. 1(a) [see Supporting Information 1 for the growth conditions]. Then, the monolayer MoS2 flakes were coated using PMMA and transferred onto a substrate for ebeam lithography. After patterning the electrodes, Cr (5 nm)/Au (50 nm) were deposited using an e-beam evaporator in order to construct a MoS2 field-effect transistor (FET) on an approximately 300-nm-thick SiO2 substrate, as shown in Fig. 1(a). To understand the effect of the Schottky barrier height (SBH) on the photocurrent, the monolayer MoS2 FET was additionally fabricated on an exfoliated, approximately 28-nm-thick multilayer h-BN, as shown in Fig. 1(b). MoS2 FETs were transferred into a closed cycle refrigerator at a base pressure of 1 × 10−7 Torr. In the manuscript, the discussion is regarding the device on SiO2 substrate unless specified otherwise. Then, the transport properties of the MoS2 FETs on different substrates were studied in the temperature range of 80–280 K. We modulate the gate voltage from −60 V to +80 V at 1 V increments. Ids–Vgs curves of the MoS2 FET on the SiO2 substrate shown in Fig. 2(a) demonstrate

ACS Paragon Plus Environment

Page 4 of 19

Page 5 of 19

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

Nano Letters

the n-type behavior of MoS2, manifested by an increase in Ids at the positive gate bias. The n-type behavior of MoS2 is reportedly attributed to S orbital vacancies21. As shown in Fig. 2(a), the temperature-dependent transport behavior of MoS2 reveals that the Ids increases with temperature when Vgs < ≈60 V, whereas, when Vgs > ≈60 V, it decreases. The transition of MoS2 from semiconductor to metal is assumed to occur when the Fermi wavelength

κ F of electrons is longer than the electron mean free path l22. The transition is theoretically predicted to occur at the critical charge concentration of nc > 1 × 1013/cm2. Therefore, the MIT occurs at approximately VMIT = 60 V from a device on the SiO2 substrate, as shown in Fig. 2(a). Both the n-type behavior and MIT are also reproduced in the MoS2 FET on the h-BN substrate, as shown in Fig. 2(b), although VMIT for the device on h-BN is about 45 V [supporting information 2 for the carrier concentration]. Figure 2(c) shows the SBHs obtained only from Au-MoS2 junctions, both on SiO2 and h-BN substrates at 300 K. The SBH gradually reduces with increasing Vgs–Vth. A higher SBH is observed when the device is on the SiO2 substrate rather than on the h-BN substrate. In Fig. 2(c), the SBHs of MoS2 with 2 terminals are approximately 68 meV and 40 meV at Vgs–Vth = 0 V for SiO2 and h-BN, respectively. For a comparison, MoS2 devices with 4-electrodes were prepared on SiO2 and h-BN substrates as well. The theoretical model and procedure to evaluate SBHs and their values from 4-electrode devices are described in Supporting Information 3. In addition to SBH, the field-effect mobility µFE of each device was determined. µFE is expressed as

µ FE =

Lg m , where W is the width of the device, Cox is the gate oxide capacitance per cm2, WCoxVds

and gm is the transconductance defined by g m =

∂I ds ∂Vgs

23

. For our devices on the SiO2 substrate,

Vds

ACS Paragon Plus Environment

Nano 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

Cox and gm become 11.5 ×10−9 nF/cm2 and 3.4 ×10−8 (A/V), respectively. µ FE ranges approximately 61.0 cm2/V·s. The same procedure was applied for the device on the h-BN layer. In such a case, µ FE shows 99.2 cm2/V·s. Since Vds from 2-terminal devices was incorporated into µFE for both devices, our µFE showed a lower limit. On the h-BN substrate, the lower SBH along with the higher carrier mobility for MoS2 can be attributed to reduced coulomb carrier scattering on the h-BN layer24, 25. The optical properties of a CVD-grown monolayer of MoS2 were studied in a vacuum system at a base pressure of 5 × 10−5 Torr. The sample was illuminated using laser light with a wavelength of 532 nm. The incident laser light was focused and positioned at the center of the MoS2 channel. Since the channel was about 30 µm wide and the beam diameter of the incident laser is only on the order of 1 µm, no excitation at the metal-MoS2 junction occurs. Thus, our narrow beam diameter prevents the simultaneous occurrence of both photothermal and photovoltatic effects at the junction. This is due to the different Fermi levels and Seebeck coefficients of the Au electrode and the MoS2 contacts. The laser is absorbed not only by MoS2, but mostly by the Si substrate after passing through the SiO2 or h-BN gate oxide. It can ionize trap sites at the SiO2/MoS2 interface resulting in an electric field at the channel in addition to Vgs. The photoinduced gating effect is different from the electrostatic Si back gate Vgs26-28. In addition to Ids caused by Vgs, photoinduced gating increased the current that originated from the metal electrode, and not due to interband transition from the valence band of MoS2. In order to separate the photoinduced gating current Ids-pg from the photocurrent by excess carriers Ids-pc, we compiled a characterization system outlined in Supporting Information 4. Since the photoinduced gating effect is caused by trapped charges, it

ACS Paragon Plus Environment

Page 6 of 19

Page 7 of 19

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

Nano Letters

has a considerably long time constant26, 28 [Supporting Information 5]. Therefore, using chopped light with a frequency of 100–300 Hz, Ids-pg can be selectively filtered out from the photocurrent measurement26. The inset of Fig. 3(a) is the optical image of the MoS2 channel patterned using the reactive-ion etching (RIE) method using SF6 gases. The optical response of MoS2 on the SiO2 substrate in the labeled area of the inset was investigated using different Vgs and laser powers. The areal photocurrent distribution is shown in Figs. 3(a) and 3(b). Under Vgs < Vth < VMIT, power of 1 µW, and wavelength of 532 nm, we observed a high Ids-pc through the MoS2 FET on a SiO2 substrate, as shown in Fig. 3(a). The channel demonstrates considerably higher Ids-pc than that at the MoS2– metal junction. However, under Vgs > VMIT > Vth and a power of 10 µW, the Ids-pc from the channel significantly drops, as shown in Fig. 3(b). These results indicate that the photocurrent switching using Vgs results from the optical excitation inside the MoS2 channel, rather than from the SBH variation. Figures 4(a–f) depict the photocurrent density Jds-pc of MoS2 on SiO2 dielectrics as a function of Vgs. Under a large negative gate bias, MoS2 exhibits semiconducting behavior, under which a significant generation of Jds-pc is observed, as shown in Fig. 4(a). As the gate bias is swept towards the positive side, the photocurrent of MoS2 fluctuates, but gradually decreases. At Vgs > 45 V, Jds-pc becomes zero. It is also possible that the laser heats up and reduces the resistance of MoS2, resulting in an increase in the photocurrent. However, the absorption of 532 nm by MoS2 is not affected by Vgs29. Therefore, no enhanced photoconduction is caused by the laser from the photothermal heating of MoS2. Thus, the absence of Jds-pc above VMIT stems from other mechanisms, rather than absorption changes. In addition to the absence of Jds-pc in metallic MoS2, the photocurrent fluctuates during the gate bias sweep in Fig. 4(a).

ACS Paragon Plus Environment

Nano 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

The photocurrent Ids-pc of MoS2 is expressed as Ids-pc = (W/L)·Vds·σpc, where W and L are the width and length of the channel, respectively, and σpc is the electrical conductivity by excess charges. The only parameter depending on Vgs in the above expression of Ids-pc is the electrical conductivity. The change in the electrical conductivity can be further described by σ pc = q ( µ n + µ p ) ∆ nex , where µn (µp) is the electron (hole) mobility and ∆nex (= ∆pex ) is the

number of excess electrons (holes) that is expressed with the number of absorbed photons η and non-radiative carrier lifetime τ r , ∆n ex = ητ r . Since some amount of excess carriers recombine before they are attracted to the electrodes and are trapped in the defect states, the net photoconductivity σ pc of electron excess carriers can be expressed as σ pc = q µ n ( ∆ nex − pt ) , where pt is the trapped electron concentration26. MoS2 contains numerous trap sites with different densities and energy levels. Therefore, the fluctuating Jds-pc shown in Fig. 4(a) is expected to reflect the trap site density NT at different energy levels inside MoS2. However, as Vgs increases to a positive value, more trap sites are occupied by channel electrons. In such a case, Jds-pc should not reduce at high positive Vgs. In order to reduce the effect of trap sites on Jds-pc, our device in Fig. 4(a) was annealed at 130 oC for 2 h in a vacuum chamber under a base pressure of 5 × 10−6 Torr. After annealing, Jds-pc in Fig. 4(b) increased by an order of magnitude compared to that shown in Fig. 4(a). In addition, the fluctuations of Jds-pc were reduced. The enhancement in Jds-pc is attributed to the removal of trapped impurities, pt, of MoS2. After annealing, the correlation between Jds-pc and Vgs clearly manifested, i.e., a complete elimination of Jds-pc was achieved when the metallicity of MoS2 shifted from semiconductor to metal. In the expression for the net photoconductivity σ pc = q µ n ( ∆ nex − pt ) , even after the annealing, we still have some trap sites and pt depends on

ACS Paragon Plus Environment

Page 8 of 19

Page 9 of 19

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

Nano Letters

Vgs because the number of active trap sites can change depending on EF. From the Shockley– Read–Hall recombination process, τ r is inversely proportional to NT, τ r ≈ 1/ NT 30, 31. Thus, the number of trap sites should decrease as Vgs is increased, resulting in an increase of τ r . Then,

∆σ pc should increase since ∆nex is proportional to τ r . Unfortunately, we did not observe such behavior from the gate-dependent photocurrent. In σ pc = q µ n ( ∆ nex − pt ) , we assume that µn remains constant regardless of metallicity of MoS2. However, our results lead us to speculate on the variation of µn of excess carriers depending on the gate bias. One of the possible scenarios is the enhanced carrier scattering rate when MoS2 switches from semiconductor to metal using the gate bias. Therefore, the increase in Jds-pc by excess carrier ∆ nex can be suppressed by the increase of carrier scattering in the MoS2 channel. This will reduce the mobility of excess carriers since µn depends on the scattering time τ,

µ n = qτ / me , where τ is the carrier scattering time and me is the electron mass. In such a case, the photocurrent caused by excess carriers Jds-pc can drop to zero. For instance, when n > nc, the carrier scattering is dominantly mediated by electron–electron scattering32. The scattering time of electrons scales with carrier concentration as n−1/3. A similar behavior of the photocurrent has been observed for graphene. When graphene was heavily doped to resemble a metal, negative photoconductivity was observed owing to increased carrier scattering. However, lightly doped graphene resembling a semiconductor showed positive photoconductivity from the increase in the number of excess carriers33. Another possibility is that, with the increase in the number of excess carriers > 1013/cm2, pre-existing excitons at low n start to form trions with the same charge, and mass three times larger than the electron mass12. In order to emphasize on the photocurrent switching mechanism using MIT, schematics shown in

ACS Paragon Plus Environment

Nano 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 19

Figs. 4(c) and 4(d) were drawn. When Vgs < VMIT in Fig. 4(c), the photocurrent generation is attributed to the increase in the number of excess carriers by optical excitation. However, when Vgs > VMIT in Fig. 4(d), the excess carriers are severely scattered by the increased channel current, and no photocurrent is observed. In the case of MoS2 FET on the h-BN layer, the dependence of the photocurrent on the back gate bias is observed in a considerably narrower voltage window, from Vgs = −60 V to Vgs = −40 V, as shown in Fig. 4(e). The power dependence of the photocurrent in this Vgs range indicates that the measured current is attributed to excess carriers. When Vgs > −40 V, no photocurrent is detected, regardless of the laser power. The peculiar phenomenon of the MoS2 photocurrent on the h-BN substrate is associated with a lower SBH, which, via the photoinduced gating effect, enables the carrier concentration of MoS2 to easily reach above nc . Because of a relatively high SBH on SiO2, the back gate bias for MIT is slightly shifted by the photoinduced gating effect. In order to understand the transport behavior of excess carriers of MoS2, the photocurrent is characterized as a function of the temperature. Figure 4(f) shows the Jds-pc versus T curve measured at Vgs = −75 V, at which MoS2 behaves as the semiconductor. As the temperature decreases, the photocurrent increases. Since we probe only the temperature dependence of excess carriers, it should not follow the intrinsic behavior of a semiconductor, J~Jo·exp(−Eg/kBT), where Jo is the current density at a given temperature and Eg is the bandgap of MoS2. The Jds-pc versus T curve should reflect the dependence of excess carrier scattering on temperature. The photocurrent decreasing with temperature, as shown in Fig. 4(f), is obtained under constant light power. Thus, decreasing Jds-pc with temperature indicates an increase in phonon scattering, when the temperature becomes higher. The power law fit of Jds-pc as a function of T reveals a fitting exponent of 1.67. Since during the experiment, ∆ nex , q, τ r , and electric field E do not change,

ACS Paragon Plus Environment

Page 11 of 19

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

Nano Letters

the dependence of Jds-pc on T is dominated by µn. Thus, the dependence of Jds-pc on T can be explained by the scattering of excess carriers by optical phonons34. Unlike graphene, which has a large optical phonon energy of 190 meV35, MoS2 has a rather low energy of optical phonons, 48 A

meV ( ωLO1 g = 400 cm−1), which play as a dominate scattering centers in our temperature range. Therefore, excess carriers are interfered by optical phonons at around 300 K, rather than acoustic phonons36. The contribution of acoustic phonons to the transport of excess carriers is expected to appear below 100 K after the suppression of optical phonons. In contrast to the temperature dependence of Jds-pc at Vgs = −75 V, the photocurrent density Jds-pc obtained at Vgs = 60 V does not change with temperature. In this metallic region, electron– electron interaction is the dominant scattering source. Usually, electron–electron scattering does not exhibit temperature dependence. Below 10 K, electron–electron scattering is expected to show dependence on T2 in highly crystalline samples. Since electron–phonon scattering and electron–impurity scattering37 are much stronger in our temperature range than electron–electron scattering, no temperature dependence is postulated to occur under our experimental conditions. In summary, we studied the photoelectric behavior of MoS2 and its electrical properties were modulated using the back gate bias. The photoconductance detected was attributed to excess carriers at the semiconducting state. In the metallic state, however, increased scattering of excess carriers completely eliminated the photocurrent. The photoconductivity switching can form the basis for establishing logic gates under light illumination.

ASSOCIATED CONTENT Supporting Information. The Supporting Information contains characterization of CVD-grown monolayer MoS2 using Raman spectroscopy, extraction of SBH, experimental setup, and the

ACS Paragon Plus Environment

Nano 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 12 of 19

shift of threshold voltage due to the photogating effect. This material is available free of charge via the Internet at http://pubs.acs.org. AUTHOR INFORMATION J. H. and H. Z. set up the experiment and acquired the data. H. K. grew the MoS2 monolayer. B. H. analyzed the MoS2 FET on the h-BN substrate. J. H. and Homin fabricated the MoS2 FETs. Y. H. and S. C. organized the data and prepared the manuscript.

Corresponding Author *E-mail address: [email protected], [email protected].

Notes The authors declare no competing financial interest.

ACKNOWLEDGMENT This work was supported by IBS-R011-D1.

ACS Paragon Plus Environment

Page 13 of 19

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

Nano Letters

REFERENCES 1. Radisavljevic, B.; Radenovic, A.; Brivio, J.; Giacometti, i. V.; Kis, A. Nat. Nanotech. 2011, 6, (3), 147-150. 2. Movva, H. C.; Rai, A.; Kang, S.; Kim, K.; Fallahazad, B.; Taniguchi, T.; Watanabe, K.; Tutuc, E.; Banerjee, S. K. ACS nano 2015, 9, (10), 10402-10410. 3. Cui, Y.; Xin, R.; Yu, Z.; Pan, Y.; Ong, Z. Y.; Wei, X.; Wang, J.; Nan, H.; Ni, Z.; Wu, Y. Adv. Mater. 2015, 27, (35), 5230-5234. 4. Mak, K. F.; Lee, C.; Hone, J.; Shan, J.; Heinz, T. F. Phys. Rev. Lett. 2010, 105, (13), 136805. 5. Chang, Y.-H.; Zhang, W.; Zhu, Y.; Han, Y.; Pu, J.; Chang, J.-K.; Hsu, W.-T.; Huang, J.K.; Hsu, C.-L.; Chiu, M.-H. ACS nano 2014, 8, (8), 8582-8590. 6. Tsai, M.-L.; Su, S.-H.; Chang, J.-K.; Tsai, D.-S.; Chen, C.-H.; Wu, C.-I.; Li, L.-J.; Chen, L.-J.; He, J.-H. ACS nano 2014, 8, (8), 8317-8322. 7. Choi, W.; Cho, M. Y.; Konar, A.; Lee, J. H.; Cha, G. B.; Hong, S. C.; Kim, S.; Kim, J.; Jena, D.; Joo, J. Adv. Mater. 2012, 24, (43), 5832-5836. 8. Yin, Z.; Li, H.; Li, H.; Jiang, L.; Shi, Y.; Sun, Y.; Lu, G.; Zhang, Q.; Chen, X.; Zhang, H. ACS nano 2012, 6, (1), 74-80. 9. Radisavljevic, B.; Kis, A. Nat. Mater. 2013, 12, (9), 815-820. 10. Schmidt, H.; Wang, S.; Chu, L.; Toh, M.; Kumar, R.; Zhao, W.; Castro Neto, A.; Martin, J.; Adam, S.; Özyilmaz, B. Nano Lett. 2014, 14, (4), 1909-1913. 11. Mak, K. F.; He, K.; Lee, C.; Lee, G. H.; Hone, J.; Heinz, T. F.; Shan, J. Nat. Mater. 2013, 12, (3), 207-211. 12. Lui, C.; Frenzel, A.; Pilon, D.; Lee, Y.-H.; Ling, X.; Akselrod, G.; Kong, J.; Gedik, N. Phys. Rev. Lett. 2014, 113, (16), 166801. 13. Mouri, S.; Miyauchi, Y.; Matsuda, K. Nano Lett. 2013, 13, (12), 5944-5948. 14. Wu, C.-C.; Jariwala, D.; Sangwan, V. K.; Marks, T. J.; Hersam, M. C.; Lauhon, L. J. J. Phys. Chem. Lett. 2013, 4, (15), 2508-2513. 15. Huo, N.; Yang, S.; Wei, Z.; Li, S.-S.; Xia, J.-B.; Li, J. Sci. Rep. 2014, 4. 16. Lopez-Sanchez, O.; Lembke, D.; Kayci, M.; Radenovic, A.; Kis, A. Nat. Nanotech. 2013, 8, (7), 497-501. 17. Echtermeyer, T. J.; Nene, P.; Trushin, M.; Gorbachev, R. V.; Eiden, A. L.; Milana, S.; Sun, Z.; Schliemann, J.; Lidorikis, E.; Novoselov, K. S. Nano Lett. 2014, 14, (7), 3733-3742. 18. Buscema, M.; Barkelid, M.; Zwiller, V.; van der Zant, H. S.; Steele, G. A.; CastellanosGomez, A. Nano Lett. 2013, 13, (2), 358-363. 19. Zhang, Y.; Li, H.; Wang, L.; Wang, H.; Xie, X.; Zhang, S.-L.; Liu, R.; Qiu, Z.-J. Sci. Rep. 2015, 5. 20. Han, G. H.; Kybert, N. J.; Naylor, C. H.; Lee, B. S.; Ping, J.; Park, J. H.; Kang, J.; Lee, S. Y.; Lee, Y. H.; Agarwal, R. Nat. Comm. 2015, 6. 21. Qiu, H.; Xu, T.; Wang, Z.; Ren, W.; Nan, H.; Ni, Z.; Chen, Q.; Yuan, S.; Miao, F.; Song, F. Nat. Comm. 2013, 4. 22. Rosenbaum, T.; Milligan, R.; Paalanen, M.; Thomas, G.; Bhatt, R.; Lin, W. Phys. Rev. B. 1983, 27, (12), 7509. 23. Schroder, D. K., Semiconductor materials and device characterization. 3rd ed.; John Wiley & Sons: New York, 2005; p 500. 24. Dean, C. R.; Young, A. F.; Meric, I.; Lee, C.; Wang, L.; Sorgenfrei, S.; Watanabe, K.; Taniguchi, T.; Kim, P.; Shepard, K. Nat. Nanotech. 2010, 5, (10), 722-726.

ACS Paragon Plus Environment

Nano 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 19

25. Decker, R.; Wang, Y.; Brar, V. W.; Regan, W.; Tsai, H.-Z.; Wu, Q.; Gannett, W.; Zettl, A.; Crommie, M. F. Nano Lett. 2011, 11, (6), 2291-2295. 26. Furchi, M. M.; Polyushkin, D. K.; Pospischil, A.; Mueller, T. Nano Lett. 2014, 14, (11), 6165-6170. 27. Miller, B.; Parzinger, E.; Vernickel, A.; Holleitner, A. W.; Wurstbauer, U. Appl. Phys. Lett. 2015, 106, (12), 122103. 28. Wu, Y.-C.; Liu, C.-H.; Chen, S.-Y.; Shih, F.-Y.; Ho, P.-H.; Chen, C.-W.; Liang, C.-T.; Wang, W.-H. Sci. Rep. 2015, 5. 29. Newaz, A.; Prasai, D.; Ziegler, J.; Caudel, D.; Robinson, S.; Haglund Jr, R.; Bolotin, K. Solid State Comm. 2013, 155, 49-52. 30. Sze, S. M., Physics of Semiconductor Devices. 2nd ed.; John Wiley & Sons: New York, 1981; p 37. 31. Macdonald, D.; Cuevas, A. Phys. Rev. B. 2003, 67, (7), 075203. 32. Nie, Z.; Long, R.; Sun, L.; Huang, C.-C.; Zhang, J.; Xiong, Q.; Hewak, D. W.; Shen, Z.; Prezhdo, O. V.; Loh, Z.-H. ACS nano 2014, 8, (10), 10931-10940. 33. Shi, S.-F.; Tang, T.-T.; Zeng, B.; Ju, L.; Zhou, Q.; Zettl, A.; Wang, F. Nano Lett. 2014, 14, (3), 1578-1582. 34. Kaasbjerg, K.; Thygesen, K. S.; Jacobsen, K. W. Phys. Rev. B. 2012, 85, (11), 115317. 35. Nika, D. L.; Balandin, A. A. J. Phys.: Condens. Matter 2012, 24, (23), 233203. 36. Kaasbjerg, K.; Bhargavi, K.; Kubakaddi, S. Phys. Rev. B. 2014, 90, (16), 165436. 37. Nie, Z.; Long, R.; Teguh, J. S.; Huang, C.-C.; Hewak, D. W.; Yeow, E. K.; Shen, Z.; Prezhdo, O. V.; Loh, Z.-H. J. Phys. Chem. C 2015, 119, (35), 20698-20708.

ACS Paragon Plus Environment

Page 15 of 19

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

Nano Letters

FIGURE LEGENDS:

Figure 1. Optical images of MoS2 FETs fabricated on (a) a SiO2 substrate and (b) an h-BN substrate.

ACS Paragon Plus Environment

Nano Letters

9

0.7

0.0 -60

Vds = 0.1 V -30

0

Vgs (V)

30

60

(c)

280 K 240 K 200 K 160 K 120 K 80 K

6 3

Vds = 0.5 V

0 -40

-20

SiO2/MoS2 h-BN/MoS2

120

Φb (meV)

1.4

(b) 12

280 K 240 K 200 K 160 K 120 K 80 K

Ids (µA)

(a) 2.1 Ids (µA)

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 19

0

20

40

Vgs (V)

80

40 -20

-10

0

10

Vgs - Vth (V)

Figure 2. Ids–Vgs characteristics of the MoS2 FETs fabricated on (a) SiO2 and (b) h-BN substrates. (c) The SBHs of the MoS2 FETs on the SiO2 and h-BN substrates.

ACS Paragon Plus Environment

Page 17 of 19

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

Nano Letters

Figure 3. Photocurrent mapping images of the MoS2 FET under (a) Vgs < Vth < VMIT and (b) Vgs > VMIT > Vth. The inset shows the optical image of the MoS2 FET. The dashed lines indicate the edge of the MoS2 channel. The unit of the color scale is ampere (A).

ACS Paragon Plus Environment

Nano Letters

Dark 2 2.5 pW/µm

(b)

0.2

0.1 Metal Semiconductor

0.0

(c)

Dark 5 pW/µm2 150 pW/µm2 2 500 pW/µm

Vds = 10 mV

7.5 2

Vds = 5 mV

Jds-pc (pA/µm )

(a) 0.3

݄ߥ

5.0

Metal

0

30

60

(d)

(e) 16

݄ߥ

Au

MoS2

Au Vgs > VMIT

-60

-30

0

30

60

Vgs (V)

Vgs (V)

Vds = 10 mV

12

(f)

Dark 5 pW/µm2 2 150 pW/µm 500 pW/µm2

0.30

Jds-pc (pA/µm2)

-30

Au

Vgs < VMIT

0.0 -60

MoS2

Au

Semiconductor

2.5

Jds-pc (pA/µm2)

Jds-pc (pA/µm2)

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 19

8

4

-60

-40

-20

0

Vgs (V)

Vds = 5 mV

Vgs = 60 volt Vgs = -75 volt α I~T

0.25 0.20

α = -1.69

0.15 0.01 0.00 160

200

240

280

Temperature (K)

Figure 4. Jds-pc versus Vgs of the MoS2 FET on the SiO2 substrate (a) before annealing and (b) after annealing at 130 oC for 2 h in a vacuum system. The schematic diagrams of scattering of excess carriers when (c) Vgs < VMIT and (d) Vgs > VMIT. (e) Jds-pc versus Vgs of the MoS2 FET on the h-BN substrate without annealing. (f) Jds-pc versus T of the MoS2 FET on the SiO2 substrate measured at Vgs = 60 V (black circles) and Vgs = −75 V (red circles).

ACS Paragon Plus Environment

Page 19 of 19

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

Nano Letters

Table of Contents Graphic

ACS Paragon Plus Environment