Electron Excess Doping and Effective Schottky ... - ACS Publications

Sep 20, 2016 - ABSTRACT: Layered hexagonal boron nitride (h-BN) thin film is a dielectric that surpasses carrier mobility by reducing charge scatterin...
0 downloads 0 Views 1MB Size
Subscriber access provided by - Access paid by the | UCSB Libraries

Communication

Electron excess doping and effective Schottky barrier reduction on MoS/h-BN heterostructure 2

Min-Kyu Joo, Byoung Hee Moon, Hyunjin Ji, Gang Hee Han, Hyun Kim, Gwan Mu Lee, Seong Chu Lim, Dongseok Suh, and Young Hee Lee Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.6b02788 • Publication Date (Web): 20 Sep 2016 Downloaded from http://pubs.acs.org on September 21, 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 24

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

Electron excess doping and effective Schottky barrier reduction on MoS2/h-BN heterostructure Min-Kyu Joo§,†, Byoung Hee Moon§,†, Hyunjin Ji||, Gang Hee Han†, Hyun Kim†,||, Gwan Mu Lee†,||, Seong Chu Lim†,||, Dongseok Suh†,||, and Young Hee Lee*,†,||



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

Republic of Korea. ||

Department of Energy Science, Sungkyunkwan University, Suwon, 440-746, Republic of

Korea.

1

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 Layered hexagonal boron nitride (h-BN) thin film is a dielectric that surpasses carrier mobility by reducing charge scattering with silicon oxide in diverse electronics formed with graphene and transition metal dichalcogenides. However, the h-BN effect on electron doping concentration and Schottky barrier is little known. Here, we report that use of h-BN thin film as a substrate for monolayer MoS2 can induce ∼6.5 × 1011 cm−2 electron doping at room temperature which was determined using theoretical flat band model and interface trap density. The saturated excess electron concentration of MoS2 on h-BN was found to be ∼5 × 1013 cm−2 at high temperature and was significantly reduced at low temperature. Further, the inserted h-BN enables to reduce the Coulombic charge scattering in MoS2/h-BN and lower the effective Schottky barrier height by a factor of three, which gives rise to four times enhanced the field−effect carrier mobility and an emergence of metal−insulator transition at a much lower charge density of ∼1.0 × 1012 cm–2 (T = 25K). The reduced effective Schottky barrier height in MoS2/h-BN is attributed to the decreased effective work function of MoS2 arisen from h-BN induced n-doping, and the reduced effective metal work function due to dipole moments originated from fixed charges in SiO2.

KEYWORDS: MoS2, h-BN, Substrate doping, Interface trap density, Coulomb scattering, Schottky barrier height, Dipole alignment

2

ACS Paragon Plus Environment

Page 2 of 24

Page 3 of 24

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

3

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

Atomically thin layered materials have all of their atoms on the material surface, making them susceptible to dielectric environment effects. One important feature of this configuration is the enhanced Coulomb scattering effects on a carrier transport due to diminished screening effects, which can be further modulated with adjacent dielectric substrates.1-6 This situation often leads to perturbed intrinsic properties of these layered materials. For example, when two-dimensional (2D) materials such as graphene and transition metal dichalcogenides (TMDs) are transferred onto oxide substrates to fabricate the field-effect transistors (FETs), three main issues are related to the oxide: i) excess doping in the channel, ii) carrier scattering from located traps at the oxide, and iii) surface roughness that provokes straining in the atomically thin TMD layer.6-9 In addition, the substrate induced effects as mentioned above are hardly distinguishable in the conventional device structure prepared on SiO2 dielectrics. A simple but efficient approach to preserve the intrinsic properties of channel materials is the use of a hexagonal boron nitride (hBN) thin film as a substrate, which has been successfully demonstrated for use with graphene transistors.7, 10 Its inert and flat surface that does not create the strain on the host materials provides an ideal platform for substrates as well as tunnel barriers or flexible insulators, which are critical components in composite 2D optoelectronic devices.8, 11, 12 Besides, the much cleaner (i.e., less charged impurities) flat surface significantly enhances their field−effect mobility.5, 7, 8, 11, 13, 14

There have been a number of published reports investigating the carrier transport of molybdenum disulfide (MoS2) on h-BN substrate implicitly evidencing the beneficial impact of h-BN substrate.14, 15 For example, the Shubnikov-de Haas oscillation was observed in MoS2 thin 4

ACS Paragon Plus Environment

Page 4 of 24

Page 5 of 24

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

film encapsulated by h-BN substrates at temperatures as low as 0.3K.14 A metal-insulator transition (MIT) mechanism in MoS2/h-BN heterostructure interpreted by a percolation theory was also reported based on capacitance measurements.15 In addition to substrate-induced charge scattering effects, Schottky barrier (SB) and channel access resistance can be inevitably formed at the contact, which often govern the transport properties. However, the effect of h-BN as a substrate to such the effective Schottky barrier height and/or contact resistance reduction are not reported yet. Therefore, improving just the channel quality in semiconductor devices is not sufficient, since their performance is considerably obscured by the contact property. Low contact resistance, highly crystalline channels, and better interface quality with gate oxides are required to improve device performance. Here, we demonstrate that the use of h-BN substrate can comply with all these requirements over the conventional MoS2/SiO2 in terms of i) excess n-doping effects in MoS2 channel, ii) suppressed interfacial trap density, iii) less channel activation energy (Ea), iv) alleviated the Coulombic charge scattering effect, and v) reduced effective Schottky barrier height (ΦSB) as well as contact resistance (RCT) at metal-MoS2 junction. The h-BN substrate effectively n-dopes MoS2 compared with SiO2 and lowers ΦSB as well as RCT, enabling to the four times enhanced field−effect carrier mobility and the early emergence of MIT. Figure 1a illustrates the device architecture under test with corresponding optical images. The h-BN thin film (15 nm) was first exfoliated onto a 300 nm SiO2/Si substrate. Monolayer MoS2 flakes synthesized by chemical vapor deposition (CVD) were transferred on the prepared h-

5

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

BN/SiO2/Si substrate. The detailed CVD growth procedures are described in Experimental Section. After identifying two MoS2 flakes, one properly located on h-BN and the other directly on top of SiO2 without h-BN, standard electron-beam lithography and a Cr/Au metallization process with a pre-determined design layout were conducted to fabricate contact electrodes. The height profile across MoS2 on SiO2 from atomic force microscope (AFM) confirms a single layer MoS2 (tch ∼ 0.8 nm) (Figure 1b). The dimensions of both samples defined by electrodes are ∼12 µm in length (L) and ∼10.4 µm in width (W). Detailed AFM images of MoS2/SiO2 and MoS2/hBN and the thickness profile of h-BN thick film are presented in Supplementary Information (SI) Figure S1. Raman and photoluminescence (PL) measurements were performed on MoS2 flakes deposited on SiO2 and h-BN. Figure 1c shows Raman spectra of these samples; no obvious deviations were ଵ observed between the two at the ‫ܧ‬ଶ௚ peak position, while the ‫ܣ‬ଵ௚ peak was blue-shifted by 2.4 ଵ cm–1, resulting in a frequency difference (∆f) between ‫ܧ‬ଶ௚ and ‫ܣ‬ଵ௚ of about 17.2 cm–1 on SiO2

and 19.6 cm–1 for h-BN. These upshifted Raman peaks indicate that the inserted h-BN layers can effectively alleviate hole doping effects that are caused by charged impurities located in SiO2.13, 16, 17

In addition, the PL emission spectra (SI Figures S2a and S2b) show that MoS2 on h-BN

exhibits a blue-shift in peak energy positions and an increased trion peak intensity relative to the neutral exciton peak, compared to MoS2 on SiO2. It also clearly demonstrates the role of relative n-doping by h-BN as explained by Raman analysis.13, 16, 18, 19 Detailed discussion on the n-doping effect arising from h-BN will be made in the next section.

6

ACS Paragon Plus Environment

Page 6 of 24

Page 7 of 24

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

We here discuss the fundamental temperature-dependent transport properties for two devices and extract the excess doping concentration from them. Figure 2a presents ID-VBG transfer curves for MoS2 deposited on h-BN and SiO2, measured at several temperatures as low as 25K. Currents are normalized by the channel width. We note that MoS2/h-BN exhibits higher on-currents than MoS2/SiO2. The subthreshold swing (SS) was significantly improved to 2.17 V/dec by using the h-BN substrate, from 9.75 V/dec on the SiO2 substrate at room temperature. The flat band voltage (VFB) was downshifted in the case of MoS2/h-BN compared to MoS2/SiO2 substrate, which indicates electron doping in the channel. Furthermore, a signature of MIT was observed near VBG ≈ 40 V in MoS2/h-BN, which was not visible in MoS2/SiO2. Figure 2b displays the temperature dependence of the interfacial trap density (DIT) extracted from SS as described in below Equation (1). Approximately 5-fold decrease of DIT in MoS2/h-BN for the entire temperature range corroborates much cleaner interface provided by the h-BN substrate. Another important parameter VFB is related with the doping level in these systems (Figure 2c). The smaller VFB for MoS2/h-BN indicates more n-doped, which results from the smaller interface trap density. In addition, to demonstrate MoS2 sample variability issue, we fabricated another two devices (MoS2/h-BN and MoS2/SiO2) within one MoS2 flake at the same time and obtained the consistent results with Figure 2a. The further discussions are described in SI Figure S3. We next derive the temperature-dependent doping concentration for two devices by applying the flat band voltage model. The excess electron doping concentration can be determined using the following Equations. In general, SS is defined by gate oxide capacitance and trap density as shown in Equation (1),20 7

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

SS = ln(10 )

k BT q

 qD IT + C D + C S  1 + C OX 

 k T  ≈ ln(10 ) B q 

Page 8 of 24

 qD IT  1 + C OX 

  

(1),

where q, COX, kB, CD, and CS denote electron unit charge, oxide capacitance per unit area (= 11.5 nF/cm2 for SiO2 and = 11.0 nF/cm2 for h-BN), the Boltzmann constant, depletion capacitance, and semiconductor capacitance, respectively (CD and CS are negligible in monolayer structure).15, 21

The flat band voltage VFB (the bias condition where the surface potential of a semiconductor is

zero) can be expressed as follows,

qV FB = ψ Si − ( χ MoS 2

N − N IN / EX + E C − E F ) + q IT C OX

= ψ Si − ( χ MoS 2 + EC − E F ) − q 

ψ Si = χ Si +  E g _ Si (0) − 

N D _ Eff C OX

4.73 × 10 − 4 ⋅ T 2 T + 636

 k BT  N A _ Si  + ln  n q   i _ Si

           

(2),

where ߖௌ௜ is the work function of the gate metal, ߯ெ௢ௌమ and ߯ௌ௜ are electron affinity of MoS2 (4.1 eV)22 and silicon (4.15 eV)20, respectively, and EC, EF, Eg_Si (≈ 1.17 eV)20, NIT, NIN/EX, NA_Si (≈ 1019 cm-3) and ni_Si (≈ 1.5 × 1010 cm-3)23 are the conduction band edge, Fermi level, band gap energy of silicon, interface trap concentration, intrinsic and/or extrinsic doping density, boron doping concentration for p-type silicon substrate, and intrinsic silicon doping concentration, respectively. Therefore, the last term of the first line in Equation (2) describes the band bending caused by the internal gate field induced by the net charge density −(NIT − NIN/EX), which is interpreted as an excess electron doping concentration, ND_Eff as expressed in Equation (2). Using 8

ACS Paragon Plus Environment

Page 9 of 24

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

VFB linearly extrapolated from ID-VBG (Figure 2c) and DIT (Figure 2b), we can extract information of excess electron doping concentration ND_Eff as illustrated in Figure 2d. Positive ND_Eff indicates excess free (untrapped) electron concentration, while negative ND_Eff indicates that the positive interface trapped charges exceed the carriers in number. As a result, most carriers are trapped, so that the charge transport by thermal excitations or hopping becomes much less. In the case of SiO2 substrate, all the carriers are already trapped within the observed temperature range. However, excess free carrier density in MoS2/h-BN is sustained down to T ∼ 165K, which means that NIN/EX is greater than the trap concentration NIT (DIT in different unit) making the device operate in a depletion mode. Therefore, the h-BN substrate causes a relative n-doping effect compared to SiO2, which is consistent with previous optical analysis. Taking into account the dependence of VFB on T with regards to material properties of monolayer MoS2 such as bandgap energy (Eg(T) = Eg(0) − AT2/(B + T), where Eg(0) = 1.874 eV, A = 5.9 × 10–4 eV/K and B = 430K),3 effective in-plain mass m*,24, 25 2D density of states g2D (gV⋅m*/π ℏଶ , where gV and ℏ are valley degeneracy around K point and reduced Plank constant),1, 3, 19-21, 25-28 and DIT(T), values of ND_Eff range −8.5 × 1011 cm–2 (T = 25K) ∼ 6.5 × 1011 cm–2 (T = 294K) for MoS2/h-BN device and −16.9 × 1011 cm–2 (T = 25K) ∼ −8.3 × 1011 cm–2 (T = 294K) for MoS2/SiO2, as displayed in Figure 2d. The difference between ND_Eff values in MoS2 on h-BN and SiO2 indicates the carrier number difference between the two devices at T = 294K (≈ 1.5 × 1012 cm–2). This value is almost identical to the 1.7 × 1012 cm–2 using a simple approximation (COX⋅∆VFB/q, ∆VFB ∼ 14.4 V at T = 294K, where ∆VFB is the difference in VFB between the two devices), rationalizing the validity of our approach. To provide an accurate 2D 9

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 24

carrier density, the Hall measurement was carried out for both devices and again confirmed the similar range of carrier density difference as plotted in Figure S4f. In detail discussions are located in SI Figure S4. Finally, the simulated temperature dependence of 2D surface carrier concentration as a function of doping concentration (ND) of MoS2 on h-BN is calculated from Equation (2) and plotted in Figure 2e. The curve for ND ≈ 5.0 × 1013 cm–2 properly describes the h-BN data in Figure 2e (blue dotted arrow). Therefore, we conclude that ND ≈ 5.0 × 1013 cm–2 is the total doping concentration of MoS2 on h-BN, much higher than the intrinsic doping density of ≈ 1010 cm–2 and the room temperature value n2D ≈ 6.5 × 1011 cm–2 yields ∼10 nm depletion width of SB, and this value increases as the temperature decreases. Figures 3a and 3b displays the temperature-dependent conductance G(T) curves in the insulating regime for several values of VBG for MoS2 on SiO2 and h-BN, respectively. The conductance slope in the inset of Figure 3b is positive at a negative gate bias and crosses over to a negative value at positive gate bias above VBG ∼ 30 V, which is clear evidence of MIT and in good contrast with SiO2 with no cross-over behavior (Figure 3a). The measured conductivity (σ) where the gate bias shows MIT is ≈ 0.5 (e2/h) at T = 294K. This value is smaller than those of monolayer MoS2 on any oxide substrates or h-BN substrate.1, 15, 29-33 These data are further analyzed at high T regime (100K ≤ T ≤ 260K) using a thermally activated charge transport mechanism expressed as G(T) = G0⋅݁ ିாೌ /௞ಳ ் where Ea is the activation energy specifying the energy difference between the Fermi level and the mobility edge, and G0 is the fitting parameter.1, 30, 31 Ea is extracted from the linear fit of this Equation to the data set for each 10

ACS Paragon Plus Environment

Page 11 of 24

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

value of VBG (Figures 3a and 3b). At temperatures below 80K, the large deviation from the fitting lines is attributed to the existing Coulomb impurities, which strongly localize the carriers at low temperature.30 Figure 3c shows the VBG dependence of Ea for two devices. It is at least a factor of three smaller for h-BN device than for SiO2. Since G(T) was measured using a 4-probe method, the obtained Ea indicates the activation barrier within the channel, not the contact-related ΦSB. At a zero gate bias, Ea is 10 meV for h-BN compared with 43 meV for SiO2 and approaches zero near VBG ∼ 20 V, which can be identified as a MIT point. On the other hand, Ea for SiO2 device decreases with VBG but still finite within our experimentally observed VBG range. This earlier emergence of MIT in h-BN device is ascribed to not only the higher n-doping density but also less Coulomb scattering in the channel. The carrier density corresponding to MIT is low, n2D ≈ 1.0 × 1012 cm–2 (T = 25K) and ≈ 2.6 × 1012 cm–2 (T = 294K), compared to ≈ 1013 cm–2 reported previously,1, 15, 29-32 implying the good quality of our h-BN device. Although the exact carrier density should be obtained via Hall measurement for direct comparison, such a trend of earlier emergence of MIT in h-BN device is still valid. For better insight into the charge scattering mechanism, the T dependence of field−effect mobility (µFE = Cox−1⋅(∂G/∂VBG)⋅(∂L12/∂W)) is analyzed using a simplified Matthiessen’s rule for Coulomb impurity scattering and phonon scattering (µFE(T) = (1/µC(T)α + 1/µPh(T)β)−1 where µC,

µPh, α and β denote the Coulomb-limited mobility, optical-phonon-limited mobility, and their exponents, respectively.). We assume that the surface roughness effect is reflected in µC, and other scattering sources produce negligible effects such as electron-electron scattering and strain-

11

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 24

induced scattering. In addition, long-range scatterers are also excluded in this analysis owing to their negligible contributions to µC as reported publications.4,

8, 14, 21

This is true since the

reported µC range for long-range scatterer is above ∼1,000 cm–2⋅V–1⋅s–1 which is much higher than our measured Coulomb mobility.14 As seen in Figure 4a, a distinct T dependence of µFE for both devices is well described by Matthiessen’s rule, yielding the fitting parameters for the selected VBG range as displayed in Figures 4b and 4c. In the MoS2 on h-BN device, Coulombic contributions dominate µFE at ∼20K and phonon contributes even at ∼200K. Meanwhile, in the SiO2 device, the Coulombic effects govern µFE over the observed temperature range. The crossover from phonon scattering to Coulomb scattering is present near T ∼ 170K in the h-BN device, indicating the change of dominant scattering source from phonon to Coulomb scatter as T decreases. This critical temperature could be linked to the temperature at which ND_Eff changes sign as shown in Figure 2d, although the origin of this behavior is not clear. We obtained an exponent α ≈ 0 in the h-BN device, an indication of minimal Coulombic scattering influence. However, α of SiO2 is larger than zero and inversely proportional to VBG, implying strong Coulombic scattering. The exponent β is almost constant (−1.5) for both devices, similar to a theoretical value of −1.69.4, 29 Now we turn to evaluate the contact quality. Since the large contact resistance in metalsemiconductor junctions arises mainly from the Schottky barrier, we estimated ΦSB from the thermionic emission current using Equation (3),14, 22

12

ACS Paragon Plus Environment

Page 13 of 24

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

I D = A ∗T

3 2

 q Φ SB   q (V D − I D R Ch )   exp  exp  − − 1  nk B T  k BT   

(3),

where A*, RCh, and n are the effective Richardson constant, channel resistance, and ideality factor, respectively. From the linear relation of ln(T3/2/ID) versus q/kBT, ΦSB can be easily determined with a properly chosen range of n (= 1 ∼ 2) as a boundary condition. Here ID is measured using a 2-probe method to include RCT, and RCh was obtained using a 4-probe method. As illustrated in Figure 4d, ΦSB of MoS2 on h-BN is approximately three times smaller than that of MoS2 on SiO2, which corroborates the lower RCT in MoS2/h-BN device shown in Figure 4e. Thus, the 4-probe method helps to observe the MIT. The reduced ΦSB and RCT are good indications for stronger band bending effect at metal−semiconductor interface because of the relatively higher n-doping effect at MoS2/h-BN compared to the case of MoS2/SiO2. More specifically, the inserted h-BN substrate makes the Fermi level of MoS2 upshift and reduces the depletion width, which will increase tunneling at the junction. It has been recently reported that the oxygen molecules can lead to extensive p-doping effect, resulting in the large increase of the work function of MoS2.34 Since h-BN film in this study is inert and clean substrate compared to SiO2 in terms of oxygen molecules and dangling bonds, the reduced effective Schottky barrier height of MoS2/h-BN can be understood by the reduced p-doping from SiO2. However, the difference of the doping level for two devices, which corresponds to ∆VFB ∼ 14.4 V, seems insufficient to explain the difference of ΦSB, as shown in Figure 4d.

13

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

In addition to the substrate doping effect, we suggest the dipole alignment effect at the interface between MoS2 and dielectrics for further adjustment since this provides the strong effect to suppress ΦSB and RCT as discussed in silicon according to bond polarization theory35 as discussed in silicon transistors.36 As we described earlier, SiO2 contains numerous positively charged trap sites near the surface. These charges induce the image charges in the metal, composing dipoles which modulate the metal work function, and cause the band bending in MoS2 as illustrated in Figure 5. Without h-BN, SiO2 directly contacts with MoS2 and draw the electrons to be neutralized. For this reason, dipole effects can be stronger in MoS2 on h-BN, resulting in lower

ΦSB. Although this is a rough sketch, we believe that our suggestion provides a perceptional view on our new finding, Schottky barrier lowering in MoS2/h-BN heterostructure on SiO2. In summary, we have demonstrated the h-BN effect on electron doping concentration and Schottky barrier lowering based on the temperature-dependent carrier transport of MoS2/h-BN heterostructure. As an optimal substrate, the use of h-BN reveals several advantages over SiO2 substrate; i) excess n-doping concentration of ∼ 5 × 1013 cm−2 at high temperature, ii) suppressed interface trap density by a factor of 5, iii) 3-fold less Ea and alleviated Coulomb scattering for the charge transport in the channel, and iv) 3 times reduced ΦSB in metal-MoS2 junction due to a higher energy band bending and dipole alignment effect compared to SiO2. Consequently, the field−effect mobility of MoS2/h-BN at T = 25K was enhanced four times and allowed the operation of MoS2 transistor in the depletion mode, which results in early emergence of MIT. These results will shed light to further enhance the performance of 2D electronic systems.

14

ACS Paragon Plus Environment

Page 14 of 24

Page 15 of 24

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

Experimental Section CVD MoS2 growth process: Three types of solutions are firstly prepared; Solution A − OptiPrep density gradient medium (Sigma-Aldrich, D1556, 60% of (w/v) solution of iodixanol in water), Solution B − Sodium cholate (SC) hydrate (Sigma-Aldrich, C6445) in deionized (DI) water, and Solution C − Solution-phase ammonium heptamolybdate (AHM) precursor (Dissolving ammonium molybdate tetrahydrate (Sigma-Aldrich, 431346) into DI water). After mixing all solutions with the ratio of A:B:C = 0.55:3:1, the solution was dropped on SiO2/Si wafer and the spin-casting process was then carried out. A two-zone CVD system was then employed to respectively control sulfur (zone 1: 200mg of sulfur) and substrate (zone 2) at atmospheric pressure. The further specific growth conditions has been reported elsewhere.37, 38 Device fabrication: Mechanically exfoliated multi-layered h-BN flakes (purchased from 2D Semiconductors) were first transferred onto 300 nm SiO2/Si substrates. Single layer MoS2 flakes grown using CVD were then placed on SiO2/Si substrates with and without h-BN. After finding clean and properly located MoS2 monolayer flakes with and without h-BN using optical microscopy (Axio imager 2, CARL ZEISS), Cr/Au metal electrodes (of 2/60 nm thick) were selectively patterned using conventional electron-beam lithography. The geometrical factors of channel length, width, and the thickness of MoS2 flakes on SiO2 and h-BN nanosheets were confirmed by AFM (SPA 400, SEIKO). Optical and electrical characterizations: To obtain the optical properties of MoS2 on h-BN and SiO2, Raman (XperRam 200, Nano Base) and photoluminescence (NTEGRA-SPECTRA, NT15

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

MDT) techniques were employed. For electrical transport characterizations, a commercial semiconductor characterization system (4200-SCS, Keithley) and dual-channel system (2636A, Keithley) were used. The operating temperature was controlled using a cryogenic probe system (335, LakeShore).

16

ACS Paragon Plus Environment

Page 16 of 24

Page 17 of 24

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 1. (a) Optical images (top) and 3D architecture (bottom) of MoS2 devices on SiO2 and hBN. (b) AFM image of MoS2 on SiO2. The inset describes the thickness profile of monolayer MoS2 on SiO2 along the white line in the AFM image. (c) Raman spectroscopy of MoS2 on SiO2 (red) and h-BN (blue).

17

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 18 of 24

Figure 2. (a) The back gate bias dependence of ID as a function of T. The temperature dependence of (b) DIT and (c) VFB with respect to VBG. (d) The analytical calculation of ND_Eff as a function of T with respect to MoS2 on h-BN and SiO2 substrates. (e) The numerical simulation of T dependence of 2D surface carrier concentration with respect to ND.

18

ACS Paragon Plus Environment

Page 19 of 24

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. The temperature and gate-bias dependence of G(T) curves of MoS2 on (a) SiO2 and (b) h-BN substrates. The inset in (b) represents the sheet conductivity σ = G⋅L12/W, where L12 = 5.3 µm, which is the distance between voltage probes from the insulating to the metallic regime in the 4-probe measurement, as illustrated in Figure 1a. The crossover of sheet conductance occurs ∼0.5 (e2/h) at T = 294K. (c) Ea curves as a function of VBG for SiO2 and h-BN substrates.

19

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 20 of 24

Figure 4. (a) Temperature dependence of measured µFE(T) (symbols) obtained by 4-probe measurement at VBG = 50 V. The respective solid, long-dashed, and dashed-dot lines indicate calculated total mobility, Coulomb-limited mobility (µC), and optical-phonon-limited mobility (µPh). VBG dependence of (b) µC and µPh as well as (c) their exponents α and β, respectively. (d) ΦSB behaviour as a function of VBG for SiO2 and h-BN substrates. (e) Temperature dependence of RCT at VBG = 50 V obtained using 4-probe measurements.

20

ACS Paragon Plus Environment

Page 21 of 24

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 5. Energy-band diagrams with relevant schematics of monolayer MoS2 on h-BN and SiO2 under considering the different effect of i) substrate doping, ii) dipole alignment, and iii) substrate doping with dipole alignment on ΦSB. The Schottky barrier height is simply defined by the mismatch of metal work function (ߖெ ) and the electron affinity of material as illustrated in (a) ΦSB = ߖெ − ߯ெ௢ௌమ . The effective work function of MoS2 can be decreased by (b) the n-doping effect on MoS2/h-BN and increased by (c) the p-doping effect on MoS2/SiO2, resulting in lowering the effective Schottky barrier height (ΦSB_Eff) at MoS2/h-BN because of the enhanced carrier injection through the much thinner tunneling barrier. In addition to substrate doping effect, the dipole polarity of Au in front of MoS2 can be negatively aligned owing to the positive fixed charges at the surface of SiO2. This results in reduction of metal work function of (d) ΦSB_Eff − ∆ΦSB for MoS2/h-BN and (e) ΦSB_Eff − ∆ΦSB′ for MoS2/SiO2, respectively. Blue dotted box describes the vertical dipole potential profile, which can be a main reason for the modulation of ߖெ as described with red long dashed line in Figures 5(d) and 5(e). Consequently, the energy band diagram can be summarized by taking into account of both substrate doping and dipole alignment effects on ΦSB, which explains (f) the ΦSB_Eff reduction mechanism of MoS2/h-BN compared with (g) MoS2/SiO2.

21

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

ASSOCIATED CONTENT Supporting Information Supplementary figures and discussion contains atomic force microscope images, photoluminescence emission spectra, device variability, and Hall effect measurement data for MoS2/SiO2 and MoS2/h-BN, respectively. This material is available free of charge via the Internet at http://pubs.acs.org. AUTHOR INFORMATION Corresponding Author *

E-mail (Y. H. Lee): [email protected] Author Contribution

§

These authors (M-K. Joo and B. H. Moon) contributed equally to this work. Notes The authors declare no competing financial interest.

ACKNOWLEDGEMENTS This work was supported by the Institute for Basic Science (IBS-R011-D1) and also supported by the national project number (no. 2015069202) through the National Research Foundation of Korea (NRF), funded by the Ministry of Science, ICT & Future Planning, Republic of Korea.

22

ACS Paragon Plus Environment

Page 22 of 24

Page 23 of 24

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.; Kis, A. Nat Mater 2013, 12, (9), 815-820. 2. Liu, H.; Ye, P. D. Electron Device Letters, IEEE 2012, 33, (4), 546-548. 3. Wang, Z. M., MoS2: Materials, Physics, and Devices. Springer Science & Business Media: 2013; Vol. 21. 4. Kaasbjerg, K.; Thygesen, K. S.; Jacobsen, K. W. Physical Review B 2012, 85, (11), 115317. 5. Butler, S. Z.; Hollen, S. M.; Cao, L.; Cui, Y.; Gupta, J. A.; Gutiérrez, H. R.; Heinz, T. F.; Hong, S. S.; Huang, J.; Ismach, A. F.; Johnston-Halperin, E.; Kuno, M.; Plashnitsa, V. V.; Robinson, R. D.; Ruoff, R. S.; Salahuddin, S.; Shan, J.; Shi, L.; Spencer, M. G.; Terrones, M.; Windl, W.; Goldberger, J. E. ACS Nano 2013, 7, (4), 2898-2926. 6. Jang, C.; Adam, S.; Chen, J. H.; Williams, E. D.; Das Sarma, S.; Fuhrer, M. S. Physical Review Letters 2008, 101, (14), 146805. 7. Dean, C. R.; Young, A. F.; MericI; LeeC; WangL; SorgenfreiS; WatanabeK; TaniguchiT; KimP; Shepard, K. L.; HoneJ. Nat Nano 2010, 5, (10), 722-726. 8. Wang, Q. H.; Kalantar-Zadeh, K.; Kis, A.; Coleman, J. N.; Strano, M. S. Nat Nano 2012, 7, (11), 699-712. 9. Lafkioti, M.; Krauss, B.; Lohmann, T.; Zschieschang, U.; Klauk, H.; Klitzing, K. v.; Smet, J. H. Nano Letters 2010, 10, (4), 1149-1153. 10. Novoselov, K. S.; Falko, V. I.; Colombo, L.; Gellert, P. R.; Schwab, M. G.; Kim, K. Nature 2012, 490, (7419), 192-200. 11. Lee, G.-H.; Yu, Y.-J.; Cui, X.; Petrone, N.; Lee, C.-H.; Choi, M. S.; Lee, D.-Y.; Lee, C.; Yoo, W. J.; Watanabe, K.; Taniguchi, T.; Nuckolls, C.; Kim, P.; Hone, J. ACS Nano 2013, 7, (9), 7931-7936. 12. Fu, L.; Sun, Y.; Wu, N.; Mendes, R. G.; Chen, L.; Xu, Z.; Zhang, T.; Rümmeli, M. H.; Rellinghaus, B.; Pohl, D.; Zhuang, L.; Fu, L. ACS Nano 2016, 10, (2), 2063-2070. 13. Buscema, M.; Steele, G.; van der Zant, H. J.; Castellanos-Gomez, A. Nano Res. 2014, 7, (4), 561-571. 14. Cui, X.; Lee, G.-H.; Kim, Y. D.; Arefe, G.; Huang, P. Y.; Lee, C.-H.; Chenet, D. A.; Zhang, X.; Wang, L.; Ye, F.; Pizzocchero, F.; Jessen, B. S.; Watanabe, K.; Taniguchi, T.; Muller, D. A.; Low, T.; Kim, P.; Hone, J. Nat Nano 2015, 10, (6), 534-540. 15. Chen, X.; Wu, Z.; Xu, S.; Wang, L.; Huang, R.; Han, Y.; Ye, W.; Xiong, W.; Han, T.; Long, G.; Wang, Y.; He, Y.; Cai, Y.; Sheng, P.; Wang, N. Nat Commun 2015, 6, 6088. 16. Lijun, L.; Inyeal, L.; Dongsuk, L.; Moonshik, K.; Gil-Ho, K.; Nobuyuki, A.; Yuichi, O.; Kenji, W.; Takashi, T. Nanotechnology 2015, 26, (29), 295702. 17. Mao, N.; Chen, Y.; Liu, D.; Zhang, J.; Xie, L. Small 2013, 9, (8), 1312-1315. 18. Ross, J. S.; Wu, S.; Yu, H.; Ghimire, N. J.; Jones, A. M.; Aivazian, G.; Yan, J.; Mandrus, D. G.; Xiao, D.; Yao, W.; Xu, X. Nat Commun 2013, 4, 1474. 19. Mak, K. F.; He, K.; Lee, C.; Lee, G. H.; Hone, J.; Heinz, T. F.; Shan, J. Nat Mater 2013, 12, (3), 207-211. 20. Sze, S. M.; Ng, K. K., Physics of semiconductor devices. John Wiley & Sons: 2006. 21. Kim, S.; Konar, A.; Hwang, W.-S.; Lee, J. H.; Lee, J.; Yang, J.; Jung, C.; Kim, H.; Yoo, J.-B.; Choi, J.-Y.; Jin, Y. W.; Lee, S. Y.; Jena, D.; Choi, W.; Kim, K. Nat Commun 2012, 3, 1011. 22. Yu, L.; Lee, Y.-H.; Ling, X.; Santos, E. J.; Shin, Y. C.; Lin, Y.; Dubey, M.; Kaxiras, E.; Kong, J.; Wang, H. Nano Letters 2014, 14, (6), 3055-3063.

23

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 24 of 24

23. Kim, D. M., Introductory quantum mechanics for semiconductor nanotechnology. John Wiley & Sons: 2010. 24. Jin, W.; Yeh, P.-C.; Zaki, N.; Zhang, D.; Sadowski, J. T.; Al-Mahboob, A.; van der Zande, A. M.; Chenet, D. A.; Dadap, J. I.; Herman, I. P.; Sutter, P.; Hone, J.; Osgood, R. M. Physical Review Letters 2013, 111, (10), 106801. 25. Peelaers, H.; Van de Walle, C. G. Physical Review B 2012, 86, (24), 241401. 26. Mak, K. F.; He, K.; Shan, J.; Heinz, T. F. Nat Nano 2012, 7, (8), 494-498. 27. Carvalho, A.; Neto, A. H. C. Physical Review B 2014, 89, (8), 081406. 28. Yu, Z.; Pan, Y.; Shen, Y.; Wang, Z.; Ong, Z.-Y.; Xu, T.; Xin, R.; Pan, L.; Wang, B.; Sun, L.; Wang, J.; Zhang, G.; Zhang, Y. W.; Shi, Y.; Wang, X. Nat Commun 2014, 5, 5290. 29. Baugher, B. W. H.; Churchill, H. O. H.; Yang, Y.; Jarillo-Herrero, P. Nano Letters 2013, 13, (9), 4212-4216. 30. Ghatak, S.; Pal, A. N.; Ghosh, A. ACS Nano 2011, 5, (10), 7707-7712. 31. RadisavljevicB; RadenovicA; BrivioJ; GiacomettiV; KisA. Nat Nano 2011, 6, (3), 147-150. 32. Schmidt, H.; Wang, S.; Chu, L.; Toh, M.; Kumar, R.; Zhao, W.; Castro Neto, A. H.; Martin, J.; Adam, S.; Özyilmaz, B.; Eda, G. Nano Letters 2014, 14, (4), 1909-1913. 33. Licciardello, D. C.; Thouless, D. J. Journal of Physics C: Solid State Physics 1975, 8, (24), 4157. 34. Lee, S. Y.; Kim, U. J.; Chung, J.; Nam, H.; Jeong, H. Y.; Han, G. H.; Kim, H.; Oh, H. M.; Lee, H.; Kim, H.; Roh, Y.-G.; Kim, J.; Hwang, S. W.; Park, Y.; Lee, Y. H. ACS Nano 2016, 10, (6), 61006107. 35. Tung, R. T. Materials Science and Engineering: R: Reports 2001, 35, (1–3), 1-138. 36. Coss, B. E.; Sivasubramani, P.; Brennan, B.; Majhi, P.; Wallace, R. M.; Kim, J. Journal of Vacuum Science & Technology B 2013, 31, (2), 021202. 37. Ly, T. H.; Perello, D. J.; Zhao, J.; Deng, Q.; Kim, H.; Han, G. H.; Chae, S. H.; Jeong, H. Y.; Lee, Y. H. Nat Commun 2016, 7, 10426. 38. 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.; Johnson, A. T. C. Nat Commun 2015, 6, 6128.

24

ACS Paragon Plus Environment