Fermi Level Pinning at Electrical Metal Article Contacts ...dasan.skku.edu/~ndpl/2017/download/42.pdf · Fermi Level Pinning at Electrical Metal Contacts of Monolayer Molybdenum Dichalcogenides

Fermi Level Pinning at Electrical MetalContacts of Monolayer MolybdenumDichalcogenidesChangsik Kim,†,⊥ Inyong Moon,†,⊥ Daeyeong Lee,† Min Sup Choi,† Faisal Ahmed,†,‡ Seunggeol Nam,§

Yeonchoo Cho,§ Hyeon-Jin Shin,§ Seongjun Park,§ and Won Jong Yoo*,†

†Samsung-SKKU Graphene Center (SSGC), SKKU Advanced Institute of Nano-Technology (SAINT), and ‡School of MechanicalEngineering, Sungkyunkwan University, 2066, Seobu-ro, Jangan-gu, Suwon, Gyeonggi-do 16419, Korea§Device & System Research Center, Samsung Advanced Institute of Technology (SAIT), 130 Samsung-ro, Yeongtong-gu, Suwon,Gyeonggi-do 16676, Korea

*S Supporting Information

ABSTRACT: Electrical metal contacts to two-dimensional (2D) semiconducting transition metal dichalcogenides(TMDCs) are found to be the key bottleneck to the realization of high device performance due to strong Fermi levelpinning and high contact resistances (Rc). Until now, Fermi level pinning of monolayer TMDCs has been reported onlytheoretically, although that of bulk TMDCs has been reported experimentally. Here, we report the experimental study onFermi level pinning of monolayer MoS2 and MoTe2 by interpreting the thermionic emission results. We also quantitativelycompared our results with the theoretical simulation results of the monolayer structure as well as the experimental resultsof the bulk structure. We measured the pinning factor S to be 0.11 and −0.07 for monolayer MoS2 and MoTe2, respectively,suggesting a much stronger Fermi level pinning effect, a Schottky barrier height (SBH) lower than that by theoreticalprediction, and interestingly similar pinning energy levels between monolayer and bulk MoS2. Our results further implythat metal work functions have very little influence on contact properties of 2D-material-based devices. Moreover, wefound that Rc is exponentially proportional to SBH, and these processing parameters can be controlled sensitively uponchemical doping into the 2D materials. These findings provide a practical guideline for depinning Fermi level at the 2Dinterfaces so that polarity control of TMDC-based semiconductors can be achieved efficiently.

KEYWORDS: molybdenum disulfide, molybdenum ditelluride, transition metal dichalcogenides, Schottky barrier height,Fermi level pinning, contact resistance

Ultrathin two-dimensional (2D) semiconducting tran-sition metal dichalcogenides (TMDCs) have receivedsignificant attention for their potential utility in the

development of future low-power device applications, whichappear infeasible in the current silicon-based semiconductorregime.1−3 Among the various TMDCs examined to date, MoS2is considered to be the most promising material due to itsabundance in natural reserves and its feasibility in large-scalesynthesis.4 MoS2 is currently being intensively investigated in

various semiconductor applications because it exhibits a sizableband gap in the range of 1.2−1.8 eV and a high on/off ratio of upto 108.5 The absence of dangling bonds and a useful band gaphave enabled MoS2 applications in heterostructure deviceprototypes, such as pn junctions, optoelectronic devices,

Received: October 24, 2016Accepted: January 15, 2017Published: January 15, 2017

Artic

lewww.acsnano.org

© 2017 American Chemical Society 1588 DOI: 10.1021/acsnano.6b07159ACS Nano 2017, 11, 1588−1596

memories, invertors, tunneling devices, as well as in emergingflexible and transparent electronics.6−12MoTe2 has recently beenexamined as an alternative promising 2D TMDC. The onlydifference between MoS2 and MoTe2 is a chalcogen atom, buttheir electrical structures are totally different from each other.MoTe2 has a narrow band gap of 1.28 eV, and its phase isreported to be transited by laser.13,14 Unlike MoS2, whichdisplays inherent n-type semiconductor behavior, MoTe2 can beused in ambipolar field-effect transistors (FETs) that display anon/off ratio of >106, a subthreshold swing of 140 mV/dec, andcomparative field effect and Hall mobilities of 10 cm2/V·s.15,16

Although MoS2 and MoTe2 have promising properties forelectronics, it is reported that charge transport in devices basedon 2Dmaterials is largely dominated by the contact properties ofthese materials due to large surface ratio and the lack of chemicalbonding.17−19 Thus, understanding interfacial properties of 2Dmaterials and the bulk metal electrode is very important. In thisregard, Schottky barrier height (SBH) and contact resistance(Rc) are important parameters for examining contact properties.According to the Schottky−Mott rule, SBH is determined by thework function of contact metal. However, the SBH between bulkMoS2 and metal very weakly depend on the metal work function,indicating a strong Fermi level pinning effect which induces highcontact resistance and obstructs realization of ohmic contact.20

The experimental studies on Fermi level pinning have beencarried out on multilayer MoS2. Theoretical calculations haverevealed a strong Fermi level pinning effect between monolayerMoS2 and various metals, whereas an experimental study onmonolayer MoS2 and MoTe2 has not been examined yet.In this work, we report the experimental study about Fermi

level pinning of monolayer MoS2 andMoTe2 by investigating the

SBH in FETs. Based on our results, we quantitatively comparedifferences between theoretical and experimental results ofmonolayer MoS2 andMoTe2. Furthermore, the relation betweenSBH and electrical Rc is investigated with temperaturedependence. Finally, the surface charge transfer-based chemicaldoping technique is employed to depin the Fermi level, and itenabled suppression of Rc and polarity control by reducing theSBH between a 2D material and a metal.

RESULTS AND DISCUSSIONPinning Factor and Charge Neutrality Level. In order to

describe the Fermi level pinning effect, we adopted the pinningfactor and charge neutrality level.21 Figure 1 shows a schematicimage of a device structure and a conceptual band diagram of a2D semiconductor−metal interface of monolayer TMDC andmetal. According to the Schottky−Mott rule, SBH for theelectron (ϕBn) is given by the difference between the workfunction of a metal (ϕm) and the electron affinity of thesemiconductor (χ).

ϕ ϕ χ= −Bn m (1)

However, interface states formed between metal and asemiconductor deviate the SBH from the Schottky−Mott rule,resulting in Fermi level pinning. In that case, the electron SBH ischaracterized quantitatively by introducing a pinning factor (S)and charge neutrality level (CNL, ϕCNL).

ϕ ϕ ϕ ϕ χ ϕ= − + − = +S S b( ) ( )Bn m CNL CNL m (2)

These can be used as figures of merits calculated from the slopeof the linear fitted line and the SBH. That is, S was defined as theslope S = dϕBn/dϕm, which varied from S = 1 for an unpinned

Figure 1. (a) Schematic illustration of fabricated device. (b) Schematic illustration of pinning factor (S) and charge neutrality energy level (ϕCNL).(c) Band diagram at the contact with interface states, including tunnel barrier (van der Waals gap), orbital overlapped TMDC under metal, anddefect states. These can modify the SBH and induce Fermi level pinning. (d) Optical image of a device structure used for the transmission linemethod measurement. (e) Transmission electron microscopy (TEM) image of a monolayer MoS2 transistor on a 285 nm thick SiO2-covered Siwafer. (f) TEM image of the FET channel region. Between van derWaals gaps, monolayerMoS2 is observed. TEM images of Pd contact (g) andTicontact (h). Unlike the FET channel region, monolayer TMDCs under metal are distorted.

ACS Nano Article

DOI: 10.1021/acsnano.6b07159ACS Nano 2017, 11, 1588−1596

1589

interface to S = 0 for a strongly pinned interface. The y-intercept(b) of the plotted line is related to the CNL, defined as the energyabove which the states of a neutral surface are empty. In otherwords, the CNL is the energy at which the Fermi level of a metalis pinned. The CNL could be estimated using the followingequation:21

ϕ χ= +−

bS1CNL (3)

Here, ϕCNL is estimated from the vacuum energy level. For S =1, which indicates no pinning, SBH can be obtained from ϕBn =ϕm − χ. By contrast, for S = 0, which indicates strong pinning,SBH can be obtained from ϕBn = ϕCNL − χ, as shown in Figure1b. Note that the pinning factor of conventional semiconductingmaterials such as Si, GaAs, and Ge are known to be ∼0.3, 0.1,22and 0.05.23 For the hole SBH, the methodology is described inSupporting Information 1.According to the theoretical calculations for monolayer MoS2,

tunnel barrier and orbital overlap are suggested to be the intrinsicorigins for Fermi level pinning.24−26 Theoretical pinning factorsof monolayer MoS2 and MoTe2 are suggested to be 0.31 and−0.17, respectively. The band diagram of the interface is shownin Figure 1c. Besides, the defect states including atomic vacanciesand structural defects in monolayer MoS2 may be responsible forinterfacial properties.27−29 The tunnel barrier, orbital overlap,and defect states are hard to be precisely controlled andmeasured but result in a weak dependence of SBH on the workfunction of contact metals. As a result, most MoS2 FETs haveshown the stubborn n-type behavior with high Rc.Fermi level pinning of MoS2 and MoTe2 can be characterized

experimentally by measuring SBH, associated with electricalcontacts between the 2D materials and metals. That is, in thisstudy, we measured SBH between TMDCs of MoS2 and MoTe2

and metals of titanium (Ti), chromium (Cr), gold (Au), andpalladium (Pd) by analyzing temperature dependency of theelectrical current. These metals are chosen to study Fermi levelpinning as low and high work function metals from 4.3 to 5.6eV.30 Fabrication process steps and material details are inSupporting Information 2 and 3.

Schottky Barrier Heights of Monolayer MoS2 andMoTe2. Carrier transport at Schottky contacts can be describedby using the following thermionic emission eqs 4 and 5.31

ϕ

π

= * −

* =*

⎜ ⎟⎛⎝⎜

⎞⎠⎟

⎛⎝

⎞⎠I WA T

q

kTqVDkT

Aq m k

h

exp exp ,

8

2D 2D3/2 Bn

2D

3

2 (4)

ϕ =Δ

Δ −

⎡⎣⎢⎢

⎤⎦⎥⎥

kq

I TT

ln( / )Bn

D3/2

1(5)

Here, W is the channel width, A2D* is the modified Richardsonconstant, ϕBn is the Schottky barrier height, VD is the drainvoltage, and m* is the effective mass. The derivation of thisequation is described in Supporting Information 4. Thisframework assumed that the application of a high drain voltage(VD) reduces the influence of the Schottky barrier at the drain tonegligible levels, and the current flow is dominated by the SBH atthe source side, as shown in Figure S3a. So we applied VD = 1 V toall of the tested devices for fair comparison. The low-temperatureI−V measurement method is employed to extract the SBHs ofmultilayer MoS2 structures by using the thermionic emissionequations.20

Under this assumption, we measured the current as a functionof the temperature and extracted the SBH from the slope of a

Figure 2. Transfer curve and SBH obtained frommonolayer MoS2 FETs. Transfer curves obtained fromMoS2 with Ti (a), Cr (b), Au (c), and Pd(d) at VD = 1 V, L = 1 μm,W = 2 μm. SBH VG of MoS2 for the metals of Ti (e), Cr (f), Au (g), and Pd (h). The blue and red dots indicate low-temperature (173−293 K) and high-temperature (323−473 K) ranges from each transfer curve, respectively.

ACS Nano Article

DOI: 10.1021/acsnano.6b07159ACS Nano 2017, 11, 1588−1596

1590

linear fit to ln(I/T3/2) as a function of 1/T, using eq 5. Figures 2and 3 plot the transfer curves and SBH obtained frommonolayerMoS2 and MoTe2 devices as a function of temperature.Regardless of the metal work function, all measured MoS2devices showed n-type behavior, whereas the MoTe2 devicesshowed ambipolar behavior with strong p-branch due to thesmall band gap. The transfer curves were used to plot ln(I/T3/2)as a function of 1/T, according to eq 5. The slope of a linear fityielded the effective SBH for the particular VG, as shown inFigures 2e−h and 3e−h. The SBH for MoS2 in the low-temperature range is indicated by the blue line in Figure 2e−h.The SBH was underestimated due to the weak thermionicemission at low temperature. In our experiments, the SBH ofmonolayer MoS2 could not readily be obtained at a low-temperature range of 173−293 K, due to the low thermal energyand large band gap of the monolayer MoS2. Extracted SBH at thelow-temperature range can be misinterpreted whether it is lowSchottky contact or ohmic contact due to temperature-independent characteristics arising from tunneling transportthrough SBH. However, when the sample is at a hightemperature, the current significantly increases based onactivated thermionic emission. At the high temperature rangeof 323−473 K, the thermionic emission equation was moreuseful for extracting the SBH. If carrier transport is dominated bythermionic emission, the application of VG changes the effectiveSBH in a linear fashion. Applying VG from −40 V to a flat bandvoltage (VFB), however, induces the carriers to tunnel throughthe Schottky barrier, and their transport is no longer described bythermionic emission alone. Thus, the SBH obtained at the end-point of linearity indicates the true SBH (VG = VFB) between a2D semiconductor and a metal. The temperature and VG

dependence of SBH are plotted in Figure S3b. The exact SBHvalues and information are shown in Tables S1 and S2.

In Figure 3, transfer curves depending on temperature and theeffective SBH are shown. MoTe2 has a narrow band gap andshows ambipolar behavior, thus it needed a different approachfrom n-type MoS2. The hole current increased significantly withincreasing temperature at the low-temperature range, but theelectron current seemed to be independent of temperature. Thismeans that hole transport is dominated by thermionic emission,but electron transport is dominated by tunneling through SBrather than thermionic emission. As we mentioned in the aboveparagraph and Figure S3b, electron SBH of MoTe2 also can beunderestimated. At the high-temperature range, electron currentseems to increase but is still not enough to extract electron SBH.For this reason, the sum of the hole and electron SBHs is notequal to the band gap (1.28 eV) of MoTe2. Additionally, the offstates of MoTe2 made it difficult to determine the true SBH atVFB. That is, the hole and electron currents overlap each otheraround off states. The exact SBH values and energy bandinformation are shown in Tables S3 and S4. In this regard, wefocused the hole SBH of MoTe2. Based on all measured SBHs ofmonolayer MoS2 and MoTe2, Fermi level pinning is discussedbelow.

Fermi Level Pinning of Monolayer MoS2 and MoTe2.The SBH for monolayer MoS2 and MoTe2, extracted from thisstudy, the reported SBH for bulk MoS2,

20,32−39 and thetheoretical values for monolayer MoS2

24−26,40 are plottedtogether as a function of the metal work function in Figure 4.Detailed references to each point are summarized in Tables S5−S7. We extracted the pinning factor and the CNL, based on eqs 2and 3. The electron affinity of the monolayer is 4.28 eV13,25 witha band gap of 1.88 eV.25 The results are compared with the banddiagram, as shown in Figure 4d−f. The experimental SBHobtained from the monolayer MoS2 is lower than the theoreticalvalues, and the obtained pinning factor (S = 0.11) is lower thanthe theoretical values (S = 0.31). Figure 4d shows the pinning

Figure 3. Transfer curve and SBH obtained frommonolayer MoTe2 FETs. Transfer curves obtained fromMoTe2 with Ti (a), Cr (b), Au (c), andPd (d) at VD = 1 V, L = 1 μm,W = 2 μm. SBH VG of MoTe2 for Ti (e), Cr (f), Au (g), and Pd (h). The blue and red dots indicate low-temperature(173−273 K) and high-temperature (323−423 K) ranges from each transfer curve, respectively.

ACS Nano Article

DOI: 10.1021/acsnano.6b07159ACS Nano 2017, 11, 1588−1596

1591

energy levels for the employed metals obtained by theoreticalexpectation and our experimental result. The calculated SBHs(green dots) are widely dispersed, whereas the experimentalSBHs (red dots) are converged below the conduction band. TheCNL of experimental and theoretical SBHs are 4.48 and 4.76 eV.The extracted value of S = 0.11 indicates that the SBH is nearlyindependent from the metal work function close to the strongFermi level pinning around the CNL of 4.48 eV. Theexperimental CNL close to the conduction band (<0.20 eVfrom the conduction band) displays a SBH lower than that fromthe calculation. The experimental and theoretical SBH40 valuesobtained from MoTe2 are plotted in Figure 4b,e. Our MoTe2devices display strong p-type and weak n-type properties, so weonly plot the hole SBH. An S of −0.07 is obtained frommonolayer MoTe2, which is lower than the theoreticallyextracted value of −0.17,40 indicating very strong Fermi levelpinning at the metal−MoTe2 interface. The experimental andtheoretical CNLs of monolayer MoTe2 are 4.19 and 4.77 eV.MoTe2 shows an ambipolar behavior with the strong p-branch,and the hole SBHs are pinned near the valence band.Difference in Theoretical and Experimental Results of

Fermi Level Pinning. The discrepancies between experimentaland theoretical values can be caused by the presence of defectstates, which are attributed to atomic vacancies introducedduring synthesis or process-induced defects. The previoustheoretical values were obtained by considering only the tunnelbarrier and orbital overlap without defects.24,25,40 The pristinesurfaces of 2D MoS2 and MoTe2 without dangling bondscontribute to form van der Waals gap that acts as a tunnel barrier.In the monolayer structure, the d-orbital of Mo atoms can beoverlapped by adjacent metal with the weakened intralayer Mo−S bonding. Thus, dangling-bond-free surface and ultrathin body

are regarded as the primary reason for a weak dependence of themetal work function on SBH. Meanwhile, it has been reportedthat the defects affected contact properties.41,42 Atomic vacanciesin 2D materials are considered as defects that induce interfacestates. The S vacancies and Mo vacancies in MoS2 can act asdonors and acceptors, respectively, at different spatial points ofthe same sample.43 The formation energy of S vacancies is lowerthan that of Mo vacancies, so S vacancies can be formed easily.41

The pinning energy level induced by S vacancy is near theconduction band.28,42 This means that rich S vacancies result inFermi level pinning stronger than that in theoretical calculations.In the case of Ti contact, theoretical SBH decreased from 0.28 to0.15 eV due to intrinsic vacancies in monolayer MoS2.

44

Experimentally, the concentration of S vacancies in monolayerMoS2 is N = (1.2 ± 0.4) × 1013 cm−2 by STEM.29 However, it ishard to measure defects between metal and 2D semiconductors.The TEM images of the contact region in Figure 1g,h show thatdistorted MoS2 under contact metal are likely introduced duringthe fabrication process. The deposition pressure45,46 and evenlow electron doses47 used for electron beam lithography canmodify thin MoS2 flakes and form an additional interlayerbetween TMDCs and metal. While fabricating all the devices, weapplied the same electron dose of 400 μC/cm2 and a depositionpressure of ∼10−5 Torr. These experimental conditions areexpected to affect monolayer MoS2 under metal contact due tothe sensitive surface of monolayer TMDCs. For this reason, wesupposed that the defects and experimental conditions were theorigins of the discrepancies.MoS2 and MoTe2 differ with respect to their chalcogen atoms,

which provide distinct band structures. The pinning factorobtained from MoTe2 was much lower than that obtained fromMoS2 because the band gap of MoTe2 is relatively small, and the

Figure 4. Fermi level pinning of MoS2 and MoTe2. (a,b) SBH of MoS2 and MoTe2 for the various metal work functions. The pinning factors Swere obtained from fits using the Schottky−Mott rule (dotted line), our extracted monolayer SBH (red line), and the calculated monolayer SBH(black line and dot). (d,e) Fermi level pinning with respect to theoretical SBH and extracted SBH for monolayer MoS2 and MoTe2. (c,f)Comparison of the Fermi levels in the bulk MoS2 and monolayer MoS2 (orange line, conduction band edge; green line, valence band edge; bluebar, charge neutrality level; blue S, the pinning factor).

ACS Nano Article

DOI: 10.1021/acsnano.6b07159ACS Nano 2017, 11, 1588−1596

1592

material shows distinct chalcogen vacancy levels and a differentoptical dielectric constant.41 In terms of defect states, the Svacancy energy levels of MoS2 are located below the conductionband,41 supporting the stubborn n-type performance and similarlocation with our experimental CNL ofmonolayerMoS2. The Tevacancy energy levels of MoTe2 are located in the middle of theband gap,41 supporting ambipolar behavior of MoTe2. The CNLof MoTe2 obtained from our experiment was hard to comparewith chalcogen vacancy levels due to electron tunneling transportoverriding thermionic emission. Unlike monolayer MoS2,monolayer MoTe2 cannot be sustained under ambient air. Thisis because telluride compounds are acutely sensitive to oxygenexposure, causing the energy level to fall to deep level states.48

We tried to avoid air exposure, as shown in experimentalmethods. It is understood that our MoTe2 was degraded due tothe sensitive surface of the monolayer. Unstable telluride maycarry many Te vacancies, which are expected to intensify Fermilevel pinning.Thickness Dependence of Fermi Level Pinning. Figure

4c,f shows the SBH of monolayer and bulk MoS2. The effect ofthe 2D material thickness on the SBH is expected to depend onits band gap and the electron affinity. The band gap of MoS2increased as the thickness decreased from ∼1.2 eV (indirect) to∼1.9 eV (direct).13 The electron affinity of the monolayer is 4.28eV with a band gap of 1.88 eV.25 Several studies have reportedthat the electron affinity of the bulk material is ∼4 eV.20,49 Thiselectron affinity suggested that the SBH should decrease as thethickness decreased; however, the extracted SBH50 values did notfollow this expected trend. The theoretically predicted SBH forbilayer MoS2 was also lower than that predicted for themonolayer26 because interlayer coupling enhanced the electroninjection efficiency. These results suggest that the electronaffinity of the bulk exceeded that of the monolayer. Here, weselected an electron affinity value of 4.45 eV, based on DFTsimulations of bulk MoS2.

51 Figure 4c,f plots the Fermi levelpinning factors measured from bulk MoS2 and monolayer MoS2.Interestingly, despite having distinct band structures, the bulkCNL (4.57 eV) and the monolayer CNL (4.48 eV) were close toone another (less than∼0.1 eV). In other words, regardless of thelayer thickness, the energy level at which the metal was pinnedtended to remain unchanged. The pinning factor of themonolayer, which was 0.11, was closer to zero than thecorresponding value of the bulk, which was 0.15. This is becausethe monolayer MoS2 may have a more sensitive surface due tohigh surface area to volume ratio.25 The defect states and

weakened intralayer Mo−S bonding strengthen the pinningeffect of monolayer MoS2 more than that of the bulk MoS2.

Relation between Rc and SBH. It is reasonable to expectthat Rc and SBH of 2Dmaterials are closely related to each other.Here, we explored the relationship between SBH and Rc. SBHwas determined by measuring current−voltage (I−V) character-istics at various temperatures, whereas Rc was measured via thetransmission line method (TLM). The total resistance measuredbetween two terminals (Rtotal) in the FET devices could bedecomposed into two components: 2Rc and Rch: Rtotal = 2Rc +Rch, where Rch is the channel resistance. Rc of 2D semiconductorsdepends mainly on the number of layers and VG; Rc increasesexponentially with decreasing thickness. The reported Rc ofmonolayer MoS2 is 10−103 kΩ·μm.19 Figure 5a showstemperature-dependent resistances of MoS2/Cr from 170 to470 K. The range of 170−293 K corresponded to the region inwhich Rc dominated carrier transport, weakly depending ontemperature. This is because electron transport across the energybarrier is dominated by tunneling.17 However, from 323 K, the Rcdecreases abruptly with decreasing Rtotal due to activatedthermionic emission. This means that the carrier transportchanged from tunneling to thermionic emission between 323and 373 K. In contrast, the Rch increases slightly based on thermalscattering, and then the ratio of Rc became less than that of Rch at423 K. The transition in the carrier transport mechanism alsoaffected the extraction of SBH. In order to describe therelationship between Rc and effective SBH, Rc dependent ontemperature was plotted as a comparison with SBH, as shown inFigure 5b. The SBH and the logarithmically scaled Rc decreasedaccording to a common trend. Other metal contact results areshown in Figure S4 and show a similar relationship. Theserelationships were confirmed by plotting the logarithm of Rc as afunction of SBH. Figure 5b clearly reveals an exponentialrelationship: log(Rc) ∝ SBH. The slope of Rc against SBHdecreases with increasing temperature due to dominantthermionic emission.In order to depin the Fermi level and further explore SBH-

dependent Rc, we prepared n- or p-type doped MoS2 and MoTe2devices in which the depinning with polarity change can bedemonstrated, by using chemical doping techniques involvingbenzyl viologen52 or gold chloride (AuCl3),

53 respectively. TheSBH and Rc as a function of the gate bias are presented in FigureS5. These results imply that chemical doping even onto thechannel can control both SBH and Rc with polarity changes.However, Rc values of monolayer MoS2 and MoTe2 afteropposite type chemical doping are lower than that of the same

Figure 5. Relationship between Rc, SBH, and temperature (a) Plots of Rc vs temperature for MoS2 (VG = 40 V, VD = 1 V), and (b) Rc vs SBH for aMoS2/Cr FET at various temperatures. Rc decreased with increasing temperature and was exponentially proportional to SBH.

ACS Nano Article

DOI: 10.1021/acsnano.6b07159ACS Nano 2017, 11, 1588−1596

1593

type chemical doping. This is because the pristine polarities ofTMDCs are still maintained against opposite type chemicaldoping due to the stubborn Fermi level pinning effect. Theexponential relationship between SBH andRc was no longer validbecause the SBH decreased significantly to near 0 eV. This meansthat, although SBH determines Rc sensitively, controlling SBHhas a limitation to obtain very low Rc.

CONCLUSIONThe Fermi level pinning and contact resistance of FETs formedusing 2D molybdenum dichalcogenides were investigatedsystematically by fabricating monolayer MoS2 and MoTe2transistors using various metal contacts, including Ti, Cr, Au,and Pd. From temperature-dependent I−V, we obtained theexperimental Schottky barrier heights and the pinning factor Sbetween the metallic electrodes and the monolayer MoS2 andMoTe2. The pinning factors of monolayer MoS2 and MoTe2 are0.11 and −0.07, respectively. The CNL of monolayer MoS2 is4.48 eV less than 0.2 eV from the conduction band, whereas theCNL of monolayer MoTe2 is 4.48 eV higher than 0.22 eV fromthe valence band. A comparison of our results with previousanalytical reports shows much stronger Fermi level pinning andlower SBH than theoretical values. These results imply theinfluence of defect states caused by atomic vacancies andexperimental artifacts. Interestingly, the CNL of the monolayer(4.48 eV) and bulk (4.57 eV) MoS2 were found to be similar,suggesting that the pinning energy level was not affected by theband structure, despite the change in the thickness of the 2Dmaterials. Additionally, Rc obtained in this work is exponentiallyproportional to SBH, suggesting that Rc can be widely varied bycontrolling SBH. We also achieved depinning via chemicaldoping of MoS2 using benzyl viologen and AuCl3.

METHODSDevice Fabrication. The SBH and Rc were measured by fabricating

many electrodes separated by various channel lengths. These electrodeswere used for the TLM measurements. Fabrication process steps are inSupporting Information 1. A p-Si wafer covered with a 285 nm thickSiO2 layer was used as the substrate and was ultrasonically cleaned inacetone and isopropyl alcohol prior to use. Before exfoliation, thesubstrates were subjected to oxygen plasma treatment to removeambient adsorbents. For our studies, MoS2 was supplied by 2DSemiconductors andMoTe2 by HQGraphene. The details onMoS2 andMoTe2 samples are described in Supporting Information 3. After usingScotch-tape to exfoliate the MoS2 and MoTe2 flakes, we lightly rubbedthe flakes and annealed them for 10 min at 100 °C. These steps yieldedlarge thin flakes,54 as shown in Figure S1a. The channel region wasdefined by applying plasma etching using a SF6/O2 gas mixture. Auniform channel of width of 2 μmwas prepared, as shown in Figure S1b.All the flakes are rinsed in acetone and isopropyl alcohol to removeadsorbents. For patterning, we used double PMMA layers (495 PMMAA4 and 950 PMMA A6), which enable facile development. Electrodeswere patterned via electron beam lithography (JEOL JSM-7001F forSEM; Raith ELPHY Plus and Quantum for lithography; dose = 400 μC/cm2, beam step size = 0.05 μm). Metals were deposited via electronbeam evaporation (Korea Vac, KVE-E2000) at 10−5 Torr. The contactmetals of Ti, Cr, Au, and Pd with different work functions (Ti, 4.3 eV;Cr, 4.5 eV; Au, 5.2 eV; Pd, 5.6 eV)30 provided by iTASCO (purity, 4N,99.99%) were used. Each of these deposited metals are 5 nm thick, andthey are subsequently subjected to 50 nm thick Au deposition, as shownin Figures 1 and S1c. All of the devices were annealed at 423 K informing gas (Ar/H2) environment for 3 h to make the surface clean. Toprevent reaction with ambient air, all the devices were kept in a vacuumdesiccator. All the electrical measurements were conducted using asemiconductor parameter analyzer (Agilent 4155C) under vacuum(MSTECH, vacuum chamber M6VC). For low-temperature electrical

measurement using liquid nitrogen, a probe machine (MSTECHMST-1000B) was used. The details on MoS2 and MoTe2 samples aredescribed in Supporting Information 2. Figure 1e−h shows cross-sectional transmission electron microscopy images of MoS2/Ti andMoS2/Pd transistors.

p-Type Dopant Preparation.Gold(III) chloride (AuCl3; >99.99%trace metal basis; Sigma-Aldrich) and acetonitrile (CH3CN) wereprepared. Next, 0.034 g of AuCl3 (20 mM) and 5 mL of CH3CN wereinjected into each vial. All of the air inside the vial was replaced bynitrogen through the degassing process. Then, 5 mL of CH3CN wasinjected into a vial including AuCl3 using an injector. After being stirredfor a few minutes, when all of the solute was dissolved in solvent, theinside of the vial became a vacuum state. Finally, the prepared dopantsolution was placed into a glovebox to be applied to our target sample.

n-Type Dopant Preparation. Three milliliters of DI water wasinjected into NaBH4. After being stirred for 3−4 min, NaBH4 was welldissolved in the DI water. Then, 10 mL of DI water, 10 mL of toluene,and 0.151 g of NaBH4 were injected in order inside a vial containing0.0409 g of benzyl viologen (equivalent to 20 mM solution). Afterward,all of the substances were mixed well. The stirring process was used for10 h. After the stirring process was completed, we could check that twolayers were clearly formed. A Schlenk line was used to make a vacuum,and the fabricated dopant solution was placed into a glovebox. The toplayer of solvent was applied to our target sample.

ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acsnano.6b07159.

Pinning factor and charge neutrality level for hole SBH(Supporting Information 1), fabrication process steps(Figure S1), material details (Figure S2), extraction ofSBH (Supporting Information 4 and Figure S3), SBH andenergy band values of MoS2 and MoTe2 (Tables S1−S4),detailed SBH values from references (Tables S5−S7),SBH and Rc against Vg before and after chemical doping(Figures S4 and S5 and Table S8) (PDF)

AUTHOR INFORMATIONCorresponding Author*E-mail: [email protected] Lee: 0000-0002-1723-4749Won Jong Yoo: 0000-0002-3767-7969Author Contributions⊥C.K. and I.M. contributed equally to this work.NotesThe authors declare no competing financial interest.

ACKNOWLEDGMENTSThis work was supported by the Global Research Laboratory(GRL) Program (2016K1A1A2912707) and Global FrontierR&D Program (2013M3A6B1078873) at the Center for HybridInterface Materials (HIM), both funded by the Ministry ofScience, ICT & Future Planning via the National ResearchFoundation of Korea (NRF).

REFERENCES(1) Wang, Q. H.; Kalantar-Zadeh, K.; Kis, A.; Coleman, J. N.; Strano,M. S. Electronics and Optoelectronics of Two-Dimensional TransitionMetal Dichalcogenides. Nat. Nanotechnol. 2012, 7, 699−712.(2) Chhowalla, M.; Shin, H. S.; Eda, G.; Li, L.-J.; Loh, K. P.; Zhang, H.The Chemistry of Two-Dimensional Layered Transition MetalDichalcogenide Nanosheets. Nat. Chem. 2013, 5, 263−275.

ACS Nano Article

DOI: 10.1021/acsnano.6b07159ACS Nano 2017, 11, 1588−1596

1594

(3) Jariwala, D.; Sangwan, V. K.; Lauhon, L. J.; Marks, T. J.; Hersam,M.C. Emerging Device Applications for Semiconducting Two-Dimen-sional Transition Metal Dichalcogenides. ACS Nano 2014, 8, 1102−1120.(4) Ahn, C.; Lee, J.; Kim, H.-U.; Bark, H.; Jeon, M.; Ryu, G. H.; Lee, Z.;Yeom, G. Y.; Kim, K.; Jung, J.; Kim, Y.; Lee, C.; Kim, T. Low-Temperature Synthesis of Large-Scale Molybdenum Disulfide ThinFilms Directly on a Plastic Substrate Using Plasma-Enhanced ChemicalVapor Deposition. Adv. Mater. 2015, 27, 5223−5229.(5) Radisavljevic, B.; Radenovic, A.; Brivio, J.; Giacometti, V.; Kis, A.Single-Layer MoS2 Transistors. Nat. Nanotechnol. 2011, 6, 147−150.(6) Choi, M. S.; Qu, D.; Lee, D.; Liu, X.; Watanabe, K.; Taniguchi, T.;Yoo,W. J. Lateral MoS 2 P−n Junction Formed by Chemical Doping forUse in High-Performance Optoelectronics. ACS Nano 2014, 8, 9332−9340.(7) Zhang, W.; Huang, J.-K.; Chen, C.-H.; Chang, Y.-H.; Cheng, Y.-J.;Li, L.-J. High-Gain Phototransistors Based on a CVDMoS 2Monolayer.Adv. Mater. 2013, 25, 3456−3461.(8) Choi, M. S.; Lee, G.-H.; Yu, Y.-J.; Lee, D.-Y.; Hwan Lee, S.; Kim, P.;Hone, J.; Yoo, W. J. Controlled Charge Trapping by MolybdenumDisulphide and Graphene in Ultrathin Heterostructured MemoryDevices. Nat. Commun. 2013, 4, 1624.(9) Yu, W. J.; Li, Z.; Zhou, H.; Chen, Y.; Wang, Y.; Huang, Y.; Duan, X.Vertically Stacked Multi-Heterostructures of Layered Materials forLogic Transistors and Complementary Inverters. Nat. Mater. 2012, 12,246−252.(10) Britnell, L.; Gorbachev, R. V.; Jalil, R.; Belle, B. D.; Schedin, F.;Mishchenko, A.; Georgiou, T.; Katsnelson, M. I.; Eaves, L.; Morozov, S.V.; Peres, N. M. R.; Leist, J.; Geim, A. K.; Novoselov, K. S.;Ponomarenko, L. A. Field-Effect Tunneling Transistor Based onVertical Graphene Heterostructures. Science (Washington, DC, U. S.)2012, 335, 947−950.(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. Flexible and Transparent MoS 2 Field-EffectTransistors on Hexagonal Boron Nitride-Graphene Heterostructures.ACS Nano 2013, 7, 7931−7936.(12) Li, H.-M.; Lee, D.; Qu, D.; Liu, X.; Ryu, J.; Seabaugh, A.; Yoo,W. J.Ultimate Thin Vertical P-N Junction Composed of Two-DimensionalLayered Molybdenum Disulfide. Nat. Commun. 2015, 6, 6564.(13) Kang, J.; Tongay, S.; Zhou, J.; Li, J.; Wu, J. Band Offsets andHeterostructures of Two-Dimensional Semiconductors. Appl. Phys. Lett.2013, 102, 012111.(14) Cho, S.; Kim, S.; Kim, J. H.; Zhao, J.; Seok, J.; Keum, D. H.; Baik,J.; Choe, D.-H.; Chang, K. J.; Suenaga, K.; Kim, S. W.; Lee, Y. H.; Yang,H. Phase Patterning for Ohmic Homojunction Contact in MoTe2.Science (Washington, DC, U. S.) 2015, 349, 625−628.(15) Pradhan, N. R.; Rhodes, D.; Feng, S.; Xin, Y.; Memaran, S.; Moon,B.-H.; Terrones, H.; Terrones, M.; Balicas, L. Field-Effect TransistorsBased on Few-Layered α-MoTe 2. ACS Nano 2014, 8, 5911−5920.(16) Lin, Y.-F.; Xu, Y.; Wang, S.-T.; Li, S.-L.; Yamamoto, M.;Aparecido-Ferreira, A.; Li, W.; Sun, H.; Nakaharai, S.; Jian, W.-B.; Ueno,K.; Tsukagoshi, K. Ambipolar MoTe 2 Transistors and TheirApplications in Logic Circuits. Adv. Mater. 2014, 26, 3263−3269.(17) Ahmed, F.; Choi, M. S.; Liu, X.; Yoo, W. J. Carrier Transport atthe metal−MoS 2 Interface. Nanoscale 2015, 7, 9222−9228.(18) Liu, H.; Si, M.; Deng, Y.; Neal, A. T.; Du, Y.; Najmaei, S.; Ajayan,P. M.; Lou, J.; Ye, P. D. Switching Mechanism in Single-LayerMolybdenum Disulfide Transistors: An Insight into Current Flowacross Schottky Barriers. ACS Nano 2014, 8, 1031−1038.(19) Allain, A.; Kang, J.; Banerjee, K.; Kis, A. Electrical Contacts toTwo-Dimensional Semiconductors. Nat. Mater. 2015, 14, 1195−1205.(20) Das, S.; Chen, H.-Y.; Penumatcha, A. V.; Appenzeller, J. HighPerformance Multilayer MoS 2 Transistors with Scandium Contacts.Nano Lett. 2013, 13, 100−105.(21) Robertson, J. Band Offsets, Schottky Barrier Heights, and TheirEffects on Electronic Devices. J. Vac. Sci. Technol., A 2013, 31, 050821.(22) Tung, R. T. The Physics and Chemistry of the Schottky BarrierHeight. Appl. Phys. Rev. 2014, 1, 011304.

(23) Dimoulas, A.; Tsipas, P.; Sotiropoulos, A.; Evangelou, E. K.Fermi-Level Pinning and Charge Neutrality Level in Germanium. Appl.Phys. Lett. 2006, 89, 252110.(24) Kang, J.; Liu, W.; Sarkar, D.; Jena, D.; Banerjee, K. ComputationalStudy of Metal Contacts to Monolayer Transition-Metal Dichalcoge-nide Semiconductors. Phys. Rev. X 2014, 4, 031005.(25) Gong, C.; Colombo, L.; Wallace, R. M.; Cho, K. The UnusualMechanism of Partial Fermi Level Pinning at Metal−MoS 2 Interfaces.Nano Lett. 2014, 14, 1714−1720.(26) Zhong, H.; Quhe, R.; Wang, Y.; Ni, Z.; Ye, M.; Song, Z.; Pan, Y.;Yang, J.; Yang, L.; Lei, M.; Shi, J.; Lu, J. Interfacial Properties ofMonolayer and BilayerMoS 2 Contacts withMetals: Beyond the EnergyBand Calculations. Sci. Rep. 2016, 6, 21786.(27) KC, S.; Longo, R. C.; Addou, R.;Wallace, R.M.; Cho, K. Impact ofIntrinsic Atomic Defects on the Electronic Structure of MoS 2Monolayers. Nanotechnology 2014, 25, 375703.(28) Zhou, W.; Zou, X.; Najmaei, S.; Liu, Z.; Shi, Y.; Kong, J.; Lou, J.;Ajayan, P. M.; Yakobson, B. I.; Idrobo, J.-C. Intrinsic Structural Defectsin Monolayer MolybdenumDisulfide.Nano Lett. 2013, 13, 2615−2622.(29) Hong, J.; Hu, Z.; Probert, M.; Li, K.; Lv, D.; Yang, X.; Gu, L.; Mao,N.; Feng, Q.; Xie, L.; Zhang, J.; Wu, D.; Zhang, Z.; Jin, C.; Ji, W.; Zhang,X.; Yuan, J.; Zhang, Z. Exploring Atomic Defects in MolybdenumDisulphide Monolayers. Nat. Commun. 2015, 6, 6293.(30) Lide, D. R. CRC Handbook of Chemistry and Physics, 76th ed.;CRC Press: Boca Raton, FL, 1995.(31) Anwar, A.; Nabet, B.; Culp, J.; Castro, F. Effects of ElectronConfinement on Thermionic Emission Current in a Modulation DopedHeterostructure. J. Appl. Phys. 1999, 85, 2663.(32) Kaushik, N.; Nipane, A.; Basheer, F.; Dubey, S.; Grover, S.;Deshmukh, M. M.; Lodha, S. Schottky Barrier Heights for Au and PdContacts to MoS2. Appl. Phys. Lett. 2014, 105, 113505.(33) Bhattacharjee, S.; Ganapathi, K. L.; Nath, D. N.; Bhat, N. IntrinsicLimit for Contact Resistance in Exfoliated Multilayered MoS2 FET.IEEE Electron Device Lett. 2016, 37, 119−122.(34) Yoo, G.; Lee, S.; Yoo, B.; Han, C.; Kim, S.; Oh, M. S. ElectricalContact Analysis of Multilayer MoS2 Transistor With MolybdenumSource/Drain Electrodes. IEEE Electron Device Lett. 2015, 36, 1215−1218.(35) Lee, S.; Tang, A.; Aloni, S.; Philip Wong, H.-S. Statistical Study onthe Schottky Barrier Reduction of Tunneling Contacts to CVDSynthesized MoS 2. Nano Lett. 2016, 16, 276−281.(36) Kim, S. Y.; Park, S.; Choi, W. Variability of Electrical ContactProperties in Multilayer MoS2 Thin-Film Transistors. Appl. Phys. A:Mater. Sci. Process. 2014, 117, 761−766.(37) Sano, K.; Takahashi, T.; Uchida, K. Large Variability of ContactResistance in Au/Cr/MoS 2 System and Its Suppression by CrThinning. Jpn. J. Appl. Phys. 2016, 55, 036501.(38) Qiu, D.; Kim, E. K. Electrically Tunable and Negative SchottkyBarriers in Multi-Layered Graphene/MoS2 Heterostructured Tran-sistors. Sci. Rep. 2015, 5, 13743.(39) Kaushik, N.; Karmakar, D.; Nipane, A.; Karande, S.; Lodha, S.Interfacial N-Doping Using an Ultrathin TiO 2 Layer for ContactResistance Reduction in MoS 2. ACS Appl. Mater. Interfaces 2016, 8,256−263.(40) Guo, Y.; Liu, D.; Robertson, J. 3D Behavior of Schottky Barriers of2D Transition-Metal Dichalcogenides. ACS Appl. Mater. Interfaces 2015,7, 25709−25715.(41) Guo, Y.; Liu, D.; Robertson, J. Chalcogen Vacancies inMonolayerTransition Metal Dichalcogenides and Fermi Level Pinning at Contacts.Appl. Phys. Lett. 2015, 106, 173106.(42) Liu, D.; Guo, Y.; Fang, L.; Robertson, J. Sulfur Vacancies inMonolayer MoS2 and Its Electrical Contacts. Appl. Phys. Lett. 2013, 103,183113.(43) McDonnell, S.; Addou, R.; Buie, C.; Wallace, R. M.; Hinkle, C. L.Defect-Dominated Doping and Contact Resistance inMoS 2.ACSNano2014, 8, 2880−2888.(44) Feng, L.-P.; Su, J.; Li, D.-P.; Liu, Z.-T. Tuning the ElectronicProperties of Ti-MoS2 Contacts through Introducing Vacancies inMonolayer MoS2. Phys. Chem. Chem. Phys. 2015, 17, 6700−6704.

ACS Nano Article

DOI: 10.1021/acsnano.6b07159ACS Nano 2017, 11, 1588−1596

1595

(45) McDonnell, S.; Smyth, C.; Hinkle, C. L.; Wallace, R. M. MoS 2 −TitaniumContact Interface Reactions. ACS Appl. Mater. Interfaces 2016,8, 8289−8294.(46) Smyth, C. M.; Addou, R.; McDonnell, S.; Hinkle, C. L.; Wallace,R. M. Contact Metal−MoS 2 Interfacial Reactions and PotentialImplications on MoS 2 -Based Device Performance. J. Phys. Chem. C2016, 120, 14719−14729.(47) Matsunaga, M.; Higuchi, A.; He, G.; Yamada, T.; Kruger, P.;Ochiai, Y.; Gong, Y.; Vajtai, R.; Ajayan, P. M.; Bird, J. P.; Aoki, N.Nanoscale-Barrier Formation Induced by Low-Dose Electron-BeamExposure in Ultrathin MoS 2 Transistors. ACS Nano 2016, 10, 9730−9737.(48) Chen, B.; Sahin, H.; Suslu, A.; Ding, L.; Bertoni, M. I.; Peeters, F.M.; Tongay, S. Environmental Changes in MoTe 2 Excitonic Dynamicsby Defects-Activated Molecular Interaction. ACS Nano 2015, 9, 5326−5332.(49) Schlaf, R.; Lang, O.; Pettenkofer, C.; Jaegermann, W. BandLineup of Layered Semiconductor Heterointerfaces Prepared by vanDer Waals Epitaxy: Charge Transfer Correction Term for the ElectronAffinity Rule. J. Appl. Phys. 1999, 85, 2732.(50) Li, S.-L.; Komatsu, K.; Nakaharai, S.; Lin, Y.-F.; Yamamoto, M.;Duan, X.; Tsukagoshi, K. Thickness Scaling Effect on Interfacial Barrierand Electrical Contact to Two-Dimensional MoS 2 Layers. ACS Nano2014, 8, 12836−12842.(51) Jiang, H. Electronic Band Structures of Molybdenum andTungsten Dichalcogenides by the GWApproach. J. Phys. Chem. C 2012,116, 7664−7671.(52) Kiriya, D.; Tosun, M.; Zhao, P.; Kang, J. S.; Javey, A. Air-StableSurface Charge Transfer Doping of MoS 2 by Benzyl Viologen. J. Am.Chem. Soc. 2014, 136, 7853−7856.(53) Liu, X.; Qu, D.; Ryu, J.; Ahmed, F.; Yang, Z.; Lee, D.; Yoo, W. J. P-Type Polar Transition of Chemically Doped Multilayer MoS 2Transistor. Adv. Mater. 2016, 28, 2345−2351.(54) Huang, Y.; Sutter, E.; Shi, N. N.; Zheng, J.; Yang, T.; Englund, D.;Gao, H.-J.; Sutter, P. Reliable Exfoliation of Large-Area High-QualityFlakes of Graphene and Other Two-Dimensional Materials. ACS Nano2015, 9, 10612−10620.

ACS Nano Article

DOI: 10.1021/acsnano.6b07159ACS Nano 2017, 11, 1588−1596

1596