Controlled Doping of Vacancy-Containing Few-Layer MoS2 via Highly

Oct 27, 2015 - Although vacancies present in 2D materials are generally regarded as ..... Contact-Engineered Electrical Properties of MoS 2 Field-Effe...
0 downloads 0 Views 5MB Size
Download PDF
Dong Min Sim,† Mincheol Kim,‡ Soonmin Yim,† Min-Jae Choi,† Jaesuk Choi,† Seunghyup Yoo,‡ and Yeon Sik Jung*,†

ARTICLE

Controlled Doping of Vacancy-Containing Few-Layer MoS2 via Highly Stable Thiol-Based Molecular Chemisorption †

Department of Materials Science and Engineering, Korea Advanced Institute of Science and Technology (KAIST), 291 Daehak-ro, Yuseong-gu, Daejeon, 305-701, Republic of Korea and ‡Department of Electrical Engineering, Korea Advanced Institute of Science and Technology (KAIST), 291 Daehak-ro, Yuseong-gu, Daejeon 305-701, Republic of Korea

ABSTRACT MoS2 is considered a promising two-dimensional active channel

material for future nanoelectronics. However, the development of a facile, reliable, and controllable doping methodology is still critical for extending the applicability of MoS2. Here, we report surface charge transfer doping via thiol-based binding chemistry for modulating the electrical properties of vacancy-containing MoS2 (v-MoS2). Although vacancies present in 2D materials are generally regarded as undesirable components, we show that the electrical properties of MoS2 can be systematically engineered by exploiting the tight binding between the thiol group and sulfur vacancies and by choosing different functional groups. For example, we demonstrate that NH2-containing thiol molecules with lone electron pairs can serve as an n-dopant and achieve a substantial increase of electron density (Δn = 3.7  1012 cm2). On the other hand, fluorine-rich molecules can provide a p-doping effect (Δn = 7.0  1011 cm2) due to its high electronegativity. Moreover, the n- and p-doping effects were systematically evaluated by photoluminescence (PL), X-ray photoelectron spectroscopy (XPS), and electrical measurement results. The excellent binding stability of thiol molecules and recovery properties by thermal annealing will enable broader applicability of ultrathin MoS2 to various devices. KEYWORDS: molybdenum disulfide . surface charge transfer doping . chemisorption . sulfur vacancy . thiol chemistry

T

he discovery of graphene has promoted significant interest in other two-dimensional (2D) materials, especially transition metal dichalcogenides (TMDs), due to their unique electrical, chemical, and mechanical properties. Although graphene possesses a high mobility of up to 106 cm2/ V 3 s,1,2 the absence of a band gap seriously limits its application as an active channel material in electronic or optoelectronic devices such as photodetection or switching device applications. Alternatively, molybdenum disulfide (MoS2), one of the most widely studied semiconducting TMDs, has a thickness-dependent energy band gap and band structure: an indirect band gap of 1.2 eV in a bulk state3 and a direct band gap of 1.8 eV in a monolayer form.4 Because of these reasons, MoS2 has gained much attention as a promising channel material in next-generation nanoelectronics.510 In general, a mechanical exfoliation approach using bulk MoS2 can produce SIM ET AL.

high-quality MoS2 sheets with minimal defects.11 Field-effect transistors (FET) based on mechanically exfoliated MoS2 show a high on/off current ratio of up to 108, a steep subthreshold swing of 70 mV/dec, and a large in-plane carrier mobility of 200500 cm2/ V 3 s.12,13 On the other hand, liquid and chemical exfoliation,14,15 physical vapor deposition (PVD) such as vaporsolid growth,16,17 and chemical vapor deposition (CVD)1820 of MoS2 are more suitable for large-scale integration of nanoelectronic devices. However, chemically exfoliated and CVD or PVD-grown MoS2 accompanies abundant disordered structural defects. Intrinsic and extrinsic defects such as vacancies, point defects, and grain boundaries on MoS2 are inevitable during the synthesis process.18,21 Moreover, these defects are the major traps or scattering centers of charge carriers, leading to a relatively low mobility and current level in FETs.22 On the other hand, defects can provide new opportunities to tailor the physical and VOL. 9

’

NO. 12

’

* Address correspondence to [email protected]. Received for review August 18, 2015 and accepted October 26, 2015. Published online October 27, 2015 10.1021/acsnano.5b05173 C 2015 American Chemical Society

12115–12123

’

2015

12115 www.acsnano.org

RESULTS AND DISCUSSION To investigate the doping effect of thiol-terminated molecules, we prepared 4-layer MoS2 samples after SIM ET AL.

confirming the number of layers by a Raman scattering analysis and an atomic force microscopy (AFM) analysis, as shown in Figure S1 in the Supporting Information. The frequency difference of 24.5 cm1 between the E12g and A1g modes in the Raman spectra and the sample thickness of ∼2.7 nm (thickness of monolayer MoS2 = ∼0.65 nm) from the AFM measurement support the successful acquisition of 4-layer MoS2 via mechanical exfoliation.38 Figure 1a and 1b depict optical microscopy (OM) and AFM images of an asexfoliated 4-layer MoS2 sample with residual contaminants from adhesive tape. To remove the residue and also to create binding sites for molecular attachment (this will be discussed in detail later), the MoS2 samples were annealed at 250 °C in air for 1h, and thus a clean surface was obtained (Figure 1c and d). Moreover, because thermal annealing was performed at a relatively low temperature of below 300 °C, triangular etch pits were not formed.39,40 Meanwhile, the AFM image obtained for the sample annealed at 310 °C in air for 1 h reveals the formation of surface etch pits with an average size of 113 nm, as shown in Figure S2 in the Supporting Information. During the thermal treatment, sulfur vacancies can be artificially created because the sulfur vacancy formation energy is lower than the formation energies of other vacancies.22 Thus, we investigated the samples before and after thermal annealing using X-ray photoelectron spectroscopy (XPS) to identify whether sulfur vacancies were generated by the thermal treatment. All XPS spectra were calibrated using the C 1s peak at 284.5 eV as a reference. As shown in Figure 1e, the XPS spectra of Mo 3d consisted of three sets of peaks that can be respectively assigned to MoO3 (yellow), intrinsic MoS2 (i-MoS2), and vacancy-containing MoS2 (v-MoS2). The deconvoluted Mo4þ 3d5/2 and Mo4þ 3d3/2 peaks depict the contributions of i-MoS2 (doublets located at 232.3 and 229.1 eV) and v-MoS2 (peaks at 231.8 and 228.8 eV).41 After thermal annealing, the intensity of the i-MoS2 peak decreased, while the intensity of the v-MoS2 peak significantly increased. As a result, the peaks of Mo4þ 3d3/2 and Mo 3d5/2 were shifted to the lower binding energy side. This reveals that the density of sulfur vacancies was increased by the thermal annealing at 250 °C, because the v-MoS2 peak is directly associated with sulfur vacancies.41 The generation of sulfur vacancies was further confirmed by the atomic ratio of S/Mo acquired from the XPS spectra. The S/Mo ratio decreased from 1.88 (as exfoliated, Figure 1f, top) to 1.39 (thermal-annealed, Figure 1f, bottom). The magnitude of the peaks corresponding to MoO3, which can be deconvoluted into Mo6þ doublets of 3d5/2 and 3d3/2, located at 232.8 and 235.4 eV, also slightly increased after thermal annealing, suggesting that MoS2 was partially oxidized. This partial oxidation is related to the compensation of carrier density, which will be discussed in the next part. VOL. 9

’

NO. 12

’

12115–12123

’

ARTICLE

chemical properties of MoS2.22,23 In general, perfect MoS2 sheets are free of dangling bonds, and therefore it is difficult to attach functional molecules on their surfaces. However, sulfur-containing groups can be bonded with unsaturated Mo edges or sulfur vacancies in MoS2.24,25 On this background, surface functionalization of MoS2 using sulfur vacancies and thiol molecules has been reported recently.14,26,27 In addition, defects can also serve as active sites to improve the hydrogen evolution reaction (HER) with a large cathodic current density and small Tafel slope and to realize sensors with excellent adsorption ability.2830 To further enhance the applicability of ultrathin TMDs including few-layer MoS2 as a 2D active channel material, the development of a more facile and universal method for modulating their electrical properties is essential. However, few-layer MoS2 cannot be doped via traditional techniques such as ion implantation due to its ultrathin nature. Chemical and molecular doping via surface charge transfer thus were previously demonstrated using various types of molecules such as potassium,31 benzyl viologen (BV),32 polyethylenimine (PEI),33 chloride molecules,34 cesium carbonate,6 and self-assembled monolayers.35,36 However, these approaches still suffer from stability issues or mainly focus on reducing the contact resistance to improve the performance of FETs. Therefore, it is highly desired to develop a more reliably and systematic way to adjust the doping level of few-layer MoS2 to facilitate wider application to future nanoelectronic devices. In this work, we report a facile and effective chemical doping method for modulating the carrier concentration of few-layer MoS2 based on the attachment of thiol molecules on vacancy sites existing on MoS2 surfaces. Because sulfur-containing functional groups can form covalent bonds with unsaturated Mo edges or sulfur vacancies in MoS2, thiol molecules with various functional groups can easily be absorbed and tightly bound to sulfur vacancies of MoS2 and can act as donors or acceptors via surface charge transfer. To demonstrate the doping effect, we chose two thiol-terminated molecules containing NH2 or F-containing groups as n-doping and p-doping sources, respectively. NH2 groups have lone electron pairs, and thus can donate electrons to MoS2,36 while F-containing groups withdraw electrons from MoS2 due to its high electronegativity.37 The proposed doping method was tested in FETs, and the observed electrical characteristics were shown to be consistent with the other experimental data. The excellent binding stability of thiol molecules and the reversibility of their doping effect can be exploited to realize more robust 2D devices such as sensors with better sensitivity and reversibility.

12116

2015 www.acsnano.org

ARTICLE Figure 1. Optical microscopy (OM) and AFM images of pristine and vacancy-created few-layer MoS2 samples on SiO2/Si substrate. (a, b) As-exfoliated MoS2 and (c, d) v-MoS2 after 250 °C thermal annealing. AFM images correspond to the white circles in each OM image. (e) High-resolution XPS spectra for Mo 3d before (top) and after thermal annealing (bottom) of MoS2. Red and blue lines represent the intrinsic i-MoS2 and v-MoS2, respectively. (f) Schematic of pristine MoS2 (top) and v-MoS2 annealed at 250 °C (bottom).

To further evaluate the change of surface stoichiometry, we annealed the MoS2 sheets at 250 °C in air with increasing treatment time, as shown in Figure S3. With a longer treatment time, Mo4þ 3d3/2 and Mo4þ 3d5/2 peaks shifted to the lower binding energy side due to the increase of the v-MoS2 peak and the MoO3 peak, as shown in Figure S3a. We also observed that the ratio of i-MoS2 to v-MoS2 and S/Mo decreased with increasing annealing time, indicating an increase of sulfur vacancy concentration (Figure S3b). In particular, the change of the surface S/Mo ratio from 1.88 to 1.39 by annealing the sample for 1 h implies the formation of a substantial amount of sulfur vacancies. The evaluation of the surface stoichiometry provides a useful metric for quantitatively estimating the degree of vacancy creation by thermal annealing. To compare the quality of MoS2 samples with and without thermal annealing, we systematically performed electrical characterizations using back-gated FETs made of as-exfoliated MoS2 and MoS2 annealed at 250 and 310 °C (Figure S4a). The transfer curve of the FET device with 250 °C-annealed MoS2 is close to that of pristine MoS2, and its on-current only slightly decreased compared to the pristine sample (Figure S4b), despite the formation of additional sulfur vacancies serving as an n-dopant. The small change of the oncurrent is attributed to the effect of partial oxidation, causing a p-doping effect.42,43 Additionally, the SIM ET AL.

nonlinear output curve of the 250 °C-annealed MoS2 FET, as shown in the inset of Figure S4b, indicates the presence of a Schottky barrier between Ti and MoS2. On the other hand, the on-current level of the FET device with 310 °C-annealed MoS2 was greatly reduced (from 4.6  107 A to 4.6  1012 A at gatesource voltage (VGS) = 20 V), and the on/off ratio thus became lower than 10. This can be explained by a high defect density and the existence of many etch pits. Thus, we used the v-MoS2 samples treated at 250 °C to investigate the thiol-binding-based chemical doping of vacancy-containing MoS2. Figure 2a presents a schematic illustration of the thiolated ligand conjugation mechanism on v-MoS2. After the cleavage of SH bonds, remaining sulfur atoms can be chemically absorbed on the sulfur vacancies.14,26,27 We presumed that the different ligand molecules attached on MoS2 enable charge transfer between MoS2 and the ligands, and thus they can act as surface charge-transfer donors and acceptors. Two thiol-terminated molecules containing NH2 or fluorine-rich groups were used to modulate the electrical properties of v-MoS2 in this study, as depicted in Figure 2b. Because of the lone pair of electrons present in the NH2 group, it is expected that mercaptoethylamine (NH2-terminated thiol, MEA) can donate electrons to v-MoS2, consequently leading to an n-doping effect.36 In contrast, 1H,1H,2H,2H-perfluorodecanethiol VOL. 9

’

NO. 12

’

12115–12123

’

12117

2015 www.acsnano.org

ARTICLE Figure 2. XPS Characterization of functionalized MoS2 via thiol chemistry. (a) Schematic illustration of chemisorption of thiol molecules onto sulfur vacancies. (b) Molecular structure of two thiol-terminated molecules containing NH2 (left, MEA) and fluorocarbon group (right, FDT) used for the molecular functionalization on v-MoS2. High-resolution XPS spectra for v-MoS2, MEA-MoS2, and FDT-MoS2 after the doping process: (c) C 1s; (d) N 1s; (e) F 1s.

(CF3-terminated thiol, FDT) can withdraw electrons from v-MoS2 because of the high electronegativity of fluorine atoms and can provide a p-doping effect.37 MEA and FDT have the same spacer group composed of two carbon atoms, and the length of two carbon atoms as a spacer group is short enough to transfer charge between the functional group and the MoS2 surface, as also reported in previous studies.26,37,44,45 MEA-treated (MEA-MoS2) and FDT-treated (FDT-MoS2) samples were immersed in ethanol and tetrahydrofuran (THF) for 12 h in order to completely remove the residual unbound molecules, which are only weakly physisorbed on v-MoS2 sheets. After the moleculeattachment and washing process, the chemical states of v-MoS2, MEA-MoS2, and FDT-MoS2 were investigated by XPS to confirm the secure binding of thiolterminated molecules on MoS2. The splitting of C 1s spectra indicates the presence of CN bonds on MEA-MoS2 and CF bonds on FDT-MoS2, respectively (Figure 2c). The entire C 1s peak of MEA-MoS2 and FDT-MoS2 shifted slightly toward higher binding energy compared to that of v-MoS2. As shown in Figure 2d, Mo 3p3/2 peaks at 395.1 eV were observed for all samples, while only MEA-MoS2 has clear NH and N 1s peaks. In contrast, the F 1s peak was obtained for only FDT-MoS2 SIM ET AL.

due to the CF2 and CF3 groups in FDT (Figure 2e). These XPS analysis results led to the conclusion that MEA and FDT molecules were tightly bound on the surface of MoS2 after the repeated washing processes. We also characterized the surface topography of MEA-MoS2 and FDT-MoS2 by AFM before (Figure 3a and 3b) and after the doping process (Figure 3c and 3d) to further confirm the binding of thiol molecules. The surface roughness of the silica substrate after MEA and FDT treatments (blue circle) did not change significantly because the weak binding of SiO2 and thiol groups may not be sufficient to prevent detachment of the molecules during the washing process. However, the surface roughness of MEA-MoS2 and FDT-MoS2 (yellow circle) was increased from 0.0865 to 0.2815 nm and from 0.107 to 0.3905 nm, respectively, suggesting the existence of the molecules after the washing process. These AFM scanning results suggest that the thiol-terminated molecules were tightly absorbed only on v-MoS2. In order to investigate the surface charge transfer effect between the functional groups of thiol molecules and v-MoS2, we also performed photoluminescence (PL) measurement because the n-doping and p-doping effects of the molecules can induce red- and VOL. 9

’

NO. 12

’

12115–12123

’

12118

2015 www.acsnano.org

ARTICLE Figure 3. AFM and PL characterization on functionalized MoS2 via thiol chemistry. (a,b) AFM images of 4-layer MoS2 after thermal annealing on a SiO2/Si substrate. (c,d) AFM images of the same MoS2 samples after the doping process using MEA and FDT, respectively. The surface roughness of each sample was calculated using these AFM images collected from the blue circle (substrate) and yellow circle (MoS2 surface). PL spectra of functionalized MoS2 before and after doping process; (e) n-doping phenomenon (MEA), (f) p-doping phenomenon (FDT).

blue-shift of the PL peaks, respectively. To minimize the effect of spatial inhomogeneity, the PL spectra were measured within the channel region after fabricating MoS2 FETs. We confirmed that the PL and Raman spectra obtained from the channel region have consistent peak positions, as shown in Figure S5c and S5d. If MEA donates electrons to v-MoS2, these excess electrons would be bound with photoexcited electron hole pairs to form trions.46 Figure 3e depicts that the measured PL spectra of MEA-MoS2 was red-shifted and noticeably weakened, which is an indication of n-doping. On the other hand, the PL spectra of FDT-MoS2 were blue-shifted and more intensified because of the p-doping effect by F-containing groups and the resultant enhancement of the PL from excitons (Figure 3f). These PL characterization results reveal that various functional groups can be combined with the thiol-termination group to induce surface charge transfer and attached molecules can consequently act as a dopant for modulating the electrical properties of v-MoS2. We now demonstrate how the surface charge transfer doping affected the performance of the FET devices, as schematically illustrated in Figure 4a and 4b. Figure 4c presents the transfer curve of v-MoS2 FET doped with MEA. The measurement was carried out in a glovebox and with a drain-source voltage (VDS) of 0.5 V. The variations of the threshold voltage (Vth) obtained from the transfer curves after doping and thermal annealing are summarized in Table 1. MEAtreatment on v-MoS2 led to a negative shift of Vth and a slight increase of the drain current (IDS), which implies SIM ET AL.

n-doping of v-MoS2. Nevertheless the degree of change observed upon doping via attachment of MEA molecules is relatively small, and it may be attributed to the adsorbed H2O and O2 acting as a p-dopant for MoS2.43 In addition to adsorption on the MoS2 surface, the hydrophilic NH2 group can absorb a large quantity of H2O molecules. Thus, to remove physisorbed H2O and O2, the doped MoS2 samples were annealed at 150 °C for 10 min, and additional FET characterization was carried out to tell apart the original doping effect by thiol-terminated molecules from the parasitic effects from H2O and O2. A significant negative shift of Vth and a marked increase of IDS were observed after the annealing process, as shown in Figure 4c. On the other hand, even without additional thermal treatment, a clear positive shift of Vth and a substantial decrease of the overall current were obtained upon binding of FDT, indicating significant p-doping of v-MoS2 (Figure 4d). In order to eliminate the additional contribution caused by the adsorption of H2O and O2 molecules, we performed the same annealing treatment as applied to MEA-MoS2. Despite the decreased p-doping effect caused by the removal of H2O and O2, the IDS of the FDT-MoS2 FET device remained lower than that of v-MoS2, indicating that its n-concentration is still significantly lower than that of v-MoS2. In addition, its Vth was also larger compared to that of the v-MoS2 FET device. Both observations confirm the p-doping effect by the attached FDT molecules. These transfer characteristics of the devices based on MEA and FDT-treated MoS2 are consistent with the PL analysis results shown above. VOL. 9

’

NO. 12

’

12115–12123

’

12119

2015 www.acsnano.org

ARTICLE Figure 4. Electrical characterizations of FET devices based on the functionalized MoS2. (a, b) Schematic illustration presenting charge transfer doping of a back-gated MoS2 FET with MEA and FDT molecules, respectively. (c, d) Transfer characteristics of MEA-MoS2 and FDT-MoS2 obtained with VDS = 0.5 V; before doping (black line), after doping (blue line), after doping and annealing (red line). (e) Changes in carrier density before and after doping process (measured at VGS = 0 V).

Threshold Voltage (Vth)a of MEA-MoS2 and FDT-MoS2 FET Devices TABLE 1. Variation in

MEA-MoS2

Vth [V] a

FDT-MoS2

undoped

doped

annealed

undoped

doped

annealed

25

38.5

< 40

27

4

18.5

Vth was defined at VGS having a drain current of 1 nA (constant current method).

From these transfer curves, we calculated the carrier density (n2D) to quantitatively compare the doping levels of MEA-MoS2 and FDT-MoS2 using the equation n2D = (IDSL)/(qWVDSμ), where q is the electron charge, L and W are the channel length and width, respectively, and μ is the mobility at VGS = 0. The mobility was extracted using the following equation with the gate capacitance (Ci) of 1.15  108 F/cm2 (SiO2 300 nm). μ ¼

DID L DVGS WCi VDS

The estimated n2D values of MEA-MoS2 and FDT-MoS2 are presented in Figure 4e. The carrier density was increased from 9.4  1011 cm2 to 1.4  1012 cm2 (Δn = 4.6  1011 cm2) by the MEA treatment on v-MoS2. After thermal annealing, significant enhancement of the carrier density (Δn = 3.7  1012 cm2) was observed, due to the removal of adsorbed molecules. In the case of FDT-MoS2, the estimated carrier density changes (Δn2D) were 7.0  1011 cm2 after doping with FDT and 1.8  1011 cm2 after postdoping annealing. These doping concentrations are comparable with recently reported values of n-doping with potassium and benzyl viologen (∼1  1013 cm2)31,32 and p-doping with octadecyltrichlorosilane (OTS) (∼1.4  1011 cm2).35 Furthermore, we also confirm the p-doping effect using 2,2,2-trifluoroethanthiol SIM ET AL.

(TFET) with three fluorine atoms (in Figure S6a) with a much smaller number of fluorine atoms per molecule compared to FDT with 21 fluorine atoms. The overall length of TFET is much shorter than that of FDT and is comaprable with that of MEA. As shown in Figure S6b, TFET also provides a p-doping effect, although TFET-MoS2 demonstrates a relatively smaller decrease of drain current than the case of the FDT-doped FET device. Also, the change of carrier density (Δn = 1.85  1011 cm2) and the shift of threshold voltage (ΔVth = 4.7 V) were smaller than those (Δn = 7.0  1011 cm2 and ΔVth = 8.5 V) of FDT-MoS2, as depicted in Figure S6c, which can be explained by the different number of fluorine atoms in the functional groups. This suggests that the surface charge transfer effect dominates the doping effect and can be modulated quantitatively by choosing appropriate functional groups. The n-doping with MEA and p-doping with FDT were also confirmed by high-resolution XPS spectra of Mo 3d and S 2p, as shown in Figure S7. The Mo 3d and S 2p peaks shifted to higher binding energy, which is caused by a reduction of the work function after the charge transfer.6 In contrast, the FDT treatment would increase the work function of v-MoS2 due to the charge transfer from v-MoS2 to FDT, which was confirmed by the negative shift of the Mo 3d and S 2p peaks. The doping effect by the chemisorbed molecules also influenced the output characteristics of MoS2 FETs. As the n-doping effect increases the channel conductivity, the relative contribution from contact to total resistance rises, and thus more nonohmic behavior is observed in the output characteristics of MEA-MoS2 (Figure S8a). On the other hand, in the case of FDTMoS2, due to the decreased channel conductivity resulting from the p-doping effect, the channel resistance was relatively more dominant in the total VOL. 9

’

NO. 12

’

12115–12123

’

12120

2015 www.acsnano.org

ARTICLE Figure 5. Investigation of doping stability. (a) Transfer characteristics of v-MoS2 measured in glovebox and air (in linear scale on the left side and in log scale on the right side). (b) Transfer characteristics of MEA-MoS2 FET measured in air. (c) Variation of transfer characteristics after exposure to air for 7 days and second annealing. The current level was maintained when the same sample was stored in a N2 filled glovebox after the second annealing. (d) Change of total resistance after each step (storage in air, second annealing, and storage in N2) at VGS = 40 V.

resistance, leading to more ohmic-like characteristics (Figure S8b). To investigate the effect of the MoS2 thickness on doping density, we compared 2-layer, 4-layer, and bulk MoS2 using MEA, as shown in Figure S9. The trend of the n-doping effect upon the molecular MEA-functionalization and thermal annealing is similar to the results obtained with a 4-layer MoS2 FET, as shown in Figure S9a and S9b. The change of carrier density increased with a decreasing number of layers, as presented in Table S1. The double-layer MoS2 FET showed the highest enhancement of current at VGS = 0 V (about a 19.7-fold increase) and carrier density changes (Δn = 7.6  1012 cm2) after thermal annealing in a glovebox, consistent with previous studies.36,47 Furthermore, we measured the transfer curve of MEA-MoS2 in ambient conditions to confirm the air stability of the doping effect. Because our v-MoS2 obtained by annealing at 250 °C has abundant sulfur vacancies, H2O and O2 molecules are easily absorbed on the vacancy sites in ambient conditions, and thus the current level is lower than that measured in a glovebox (Figure 5a, plotted in a semilog scale). After doping with MEA and subsequent annealing at 150 °C for 10 min, the transfer curve measured in ambient conditions (Figure 4c) is similar to the results obtained in a glovebox, as shown in Figure 5b. However, the current level significantly decreased after storage in air for 7 days due to the adsorption of H2O and O2 on MEAMoS2 (Figure 5c). Thus, the total resistance (Rtot) of this sample underwent a 103-fold increase at VGS = 40 V (Rtot of the first annealed MEA-MoS2 sample = 4.4  105 Ω; SIM ET AL.

Rtot after 2 days in air = 1.23  108 Ω; Rtot after 7 days in air = 5.11  108 Ω, as shown in Figure 5d. However, the second annealing at 150 °C for 10 min restored the current level to that of the first annealed sample, and consequently the total resistance was also recovered. Moreover, this low resistance state was maintained during additional storage (4 days) in a glovebox. This behavior reveals that the thiol molecules bound with MoS2 were not degraded during the long storage in an air environment and thermal annealing at 150 °C, suggesting this doping method lends excellent stability. CONCLUSIONS In summary, we have demonstrated a highly effective and stable surface charge transfer doping mechanism based on thiol-based molecular functionalization. We showed that the vacancy sites of MoS2 can be artificially created by thermal annealing and can act as the binding sites for thiol-terminated molecules. As a demonstration, we showed that amine-containing MEA and F-containing FDT molecules can be selectively chemisorbed and tightly bound on sulfur vacancies via the strong interaction between the thiol and the S-vacancy. Surface charge transfer between MoS2 and functional groups of bound molecules and a consequent doping effect were confirmed by systematic characterizations using XPS, PL spectroscopy, Raman spectroscopy, and electrical measurements on FET devices. Significant enhancement (Δn = 3.7  1012 cm2; MEA-treated and thermal-annealed) or reduction (Δn = 1.8  1011 cm2; FDT-treated and thermal-annealed) of the carrier density was obtained VOL. 9

’

NO. 12

’

12115–12123

’

12121

2015 www.acsnano.org

and recovery properties by thermal annealing. This method also would be applicable to CVD-grown MoS2, which intrinsically contains many vacancies. Furthermore, we anticipate that this simple mechanism based on the tight binding between vacancy sites and thiol molecules can be extensively applied to other 2D semiconductors and may open a new possibility for systematically modulating the doping levels of various 2D TMD materials for more practical device applications.

METHODS

Open Innovation Lab Project from National Nanofab Center (NNFC).

Materials. Molybdenum disulfide (MoS2) was purchased from 2D SEMICONDUCTORS, INC. 2-mercaptoethylamine (NH2-terminated thiol, MEA), 1H,1H,2H,2H-perfluorodecanethiol (CF3-terminated thiol, FDT) and 2,2,2-trifluoroethanthiol (TFET) were purchased from Sigma-Aldrich. AZ5214E photoresist, AZ300MIF developer, and AZ400T lift-off solution used in photolithography were purchased from Microchemical. INC. Preparation and Doping of Vacancy-Containing MoS2. Few-layer MoS2 was mechanically exfoliated from bulk MoS2 using a Scotch tape method and was placed onto 300 nm-thick SiO2 on a heavily p-doped Si substrate. Optical microscopy was used to identify the thickness and location of MoS2 sheets. The samples were placed in the middle of a tube furnace and annealed at 250 °C for 1 h. For the monolayer formation of MEA and FDT molecules on v-MoS2 samples, the samples were soaked in a solution of 1/40 (v/v) MEA/ethanol and FDT/THF for 72 h. They were then rinsed with ethanol and THF, blown dry with N2, and baked at 100 °C for 30 min in a dry glovebox. After being baked, the MEA treated and FDT treated samples were immersed in ethanol and tetrahydrofuran (THF) for 12 h to remove unbound molecules. Characterization of Doped/Undoped MoS2. The thickness and the surface roughness of the v-MoS2 samples were obtained using atomic force microscopy (AFM, XE-70, Parksystems). All AFM images were collected with 10  10 um scan size and a 0.1 Hz scan rate. Raman spectroscopy and photoluminescence (PL) analyses were performed using a LabRAM ARAMIS system (Horiba scientific) with a 514 nm excitation laser under ambient conditions (scan range: 350 to 450 cm1 for Raman analysis and 580 to 720 nm for PL measurement). An X-ray photoelectron spectroscopy (XPS) analysis was carried out using a multipurpose X-ray photoelectron spectrometer (Sigma Probe, Thermo VG Scientific). Fabrication and Characterization of MoS2 FET Devices. Back-gated MoS2 FETs were fabricated using conventional photolithography. To pattern the source and drain electrodes, we used a photoresist (AZ5142E) and a mask aligner (MDA-8000B) followed by a development process using AZ300MIF developer. A Ti/Au (20 nm/80 nm) electrode was deposited by electron beam evaporation followed by a lift-off process with AZ400T. All devices were then annealed at 200 °C in H2/Ar for 1 h to improve the contact of the MoS2/Ti interface. The aforementioned doping process was performed after fabricating the FET devices. Electrical properties (transfer and output characteristics) of the FET devices were measured using a HP4156A semiconductor parameter analyzer in a nitrogen-filled glovebox. To remove the H2O and O2 absorbed on v-MoS2 during the doping process, the doped v-MoS2 FETs were annealed at 150 °C for 10 min in a glovebox before characterization. Conflict of Interest: The authors declare no competing financial interest. Acknowledgment. This work was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF-2013R1A1A2061440). This work was supported by the National Research Foundation (NRF) of Korea funded by the Ministry of Science, ICT and Future Planning (NRF-2014R1A2A1A11052860). This work was also supported by

SIM ET AL.

ARTICLE

compared to v-MoS2 samples. Moreover, we demonstrated that the carrier density change under an air environment as a result of O2 and H2O adsorption can be reliably recovered by simple thermal annealing for desorption of the adsorbed molecules, suggesting the chemisorbed dopant molecules have high stability. We showed that our thiol-based molecular doping strategy provides multiple advantages of charge transferinduced doping capability, excellent binding stability,

Supporting Information Available: The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.5b05173. Supplementary figures. (PDF)

REFERENCES AND NOTES 1. Borysenko, K. M.; Mullen, J. T.; Barry, E. A.; Paul, S.; Semenov, Y. G.; Zavada, J. M.; Nardelli, M. B.; Kim, K. W. First-Principles Analysis of Electron-Phonon Interactions in Graphene. Phys. Rev. B: Condens. Matter Mater. Phys. 2010, 81, 121412. 2. Schwierz, F. Graphene Transistors. Nat. Nanotechnol. 2010, 5, 487–496. 3. Kam, K. K.; Parkinson, B. A. Detailed Photocurrent Spectroscopy of The Semiconducting Group-Vi TransitionMetal Dichalcogenides. J. Phys. Chem. 1982, 86, 463–467. 4. Mak, K. F.; Lee, C.; Hone, J.; Shan, J.; Heinz, T. F. Atomically Thin MoS2: A New Direct-Gap Semiconductor. Phys. Rev. Lett. 2010, 105, 136805. 5. Huang, Y. L.; Chen, Y. F.; Zhang, W. J.; Quek, S. Y.; Chen, C. H.; Li, L. J.; Hsu, W. T.; Chang, W. H.; Zheng, Y. J.; Chen, W.; et al. Bandgap Tunability at Single-Layer Molybdenum Disulphide Grain Boundaries. Nat. Commun. 2015, 6, 6298. 6. Lin, J. D.; Han, C.; Wang, F.; Wang, R.; Xiang, D.; Qin, S. Q.; Zhang, X. A.; Wang, L.; Zhang, H.; Wee, A. T. S.; et al. Electron-Doping-Enhanced Trion Formation in Mono layer Molybdenum Disulfide Functionalized with Cesium Carbonate. ACS Nano 2014, 8, 5323–5329. 7. Wi, S.; Kim, H.; Chen, M. K.; Nam, H.; Guo, L. J.; Meyhofer, E.; Liang, X. G. Enhancement of Photovoltaic Response in Multilayer MoS2 Induced by Plasma Doping. ACS Nano 2014, 8, 5270–5281. 8. Cheng, R.; Jiang, S.; Chen, Y.; Liu, Y.; Weiss, N.; Cheng, H. C.; Wu, H.; Huang, Y.; Duan, X. F. Few-Layer Molybdenum Disulfide Transistors and Circuits for High-Speed Flexible Electronics. Nat. Commun. 2014, 5, 5143. 9. Nam, H.; Wi, S.; Rokni, H.; Chen, M. K.; Priessnitz, G.; Lu, W.; Liang, X. G. MoS2 Transistors Fabricated via PlasmaAssisted Nanoprinting of Few-Layer MoS2 Flakes into Large-Area Arrays. ACS Nano 2013, 7, 5870–5881. 10. Lopez-Sanchez, O.; Lembke, D.; Kayci, M.; Radenovic, A.; Kis, A. Ultrasensitive Photodetectors based on Monolayer MoS2. Nat. Nanotechnol. 2013, 8, 497–501. 11. Novoselov, K. S.; Jiang, D.; Schedin, F.; Booth, T. J.; Khotkevich, V. V.; Morozov, S. V.; Geim, A. K. Two-Dimensional Atomic Crystals. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 10451–10453. 12. Liu, H.; Ye, P. D. D. MoS2 Dual-Gate MOSFET With AtomicLayer-Deposited Al2O3 as Top-Gate Dielectric. IEEE Electron Device Lett. 2012, 33, 546–548. 13. Radisavljevic, B.; Radenovic, A.; Brivio, J.; Giacometti, V.; Kis, A. Single-Layer MoS2 Transistors. Nat. Nanotechnol. 2011, 6, 147–150. 14. Chou, S. S.; De, M.; Kim, J.; Byun, S.; Dykstra, C.; Yu, J.; Huang, J. X.; Dravid, V. P. Ligand Conjugation of Chemically Exfoliated MoS2. J. Am. Chem. Soc. 2013, 135, 4584–4587.

VOL. 9

’

NO. 12

’

12115–12123

’

12122

2015 www.acsnano.org

SIM ET AL.

33. Du, Y. C.; Liu, H.; Neal, A. T.; Si, M. W.; Ye, P. D. Molecular Doping of Multilayer MoS2 Field-Effect Transistors: Reduction in Sheet and Contact Resistances. IEEE Electron Device Lett. 2013, 34, 1328–1330. 34. Yang, L. M.; Majumdar, K.; Liu, H.; Du, Y. C.; Wu, H.; Hatzistergos, M.; Hung, P. Y.; Tieckelmann, R.; Tsai, W.; Hobbs, C.; et al. Chloride Molecular Doping Technique on 2D Materials: WS2 and MoS2. Nano Lett. 2014, 14, 6275– 6280. 35. Kang, D.-H.; Kim, M.-S.; Shim, J.; Jeon, J.; Park, H.-Y.; Jung, W.-S.; Yu, H.-Y.; Pang, C.-H.; Lee, S.; Park, J.-H. High-Performance Transition Metal Dichalcogenide Photodetectors Enhanced by Self-Assembled Monolayer Doping. Adv. Funct. Mater. 2015, 25, 4219. 36. Li, Y.; Xu, C. Y.; Hu, P. A.; Zhen, L. Carrier Control of MoS2 Nanoflakes by Functional Self-Assembled Monolayers. ACS Nano 2013, 7, 7795–7804. 37. Lee, B.; Chen, Y.; Duerr, F.; Mastrogiovanni, D.; Garfunkel, E.; Andrei, E. Y.; Podzorov, V. Modification of Electronic Properties of Graphene with Self-Assembled Monolayers. Nano Lett. 2010, 10, 2427–2432. 38. Lee, C.; Yan, H.; Brus, L. E.; Heinz, T. F.; Hone, J.; Ryu, S. Anomalous Lattice Vibrations of Single- and Few-Layer MoS2. ACS Nano 2010, 4, 2695–2700. 39. Wu, J.; Li, H.; Yin, Z. Y.; Li, H.; Liu, J. Q.; Cao, X. H.; Zhang, Q.; Zhang, H. Layer Thinning and Etching of Mechanically Exfoliated MoS2 Nanosheets by Thermal Annealing in Air. Small 2013, 9, 3314–3319. 40. Yamamoto, M.; Einstein, T. L.; Fuhrer, M. S.; Cullen, W. G. Anisotropic Etching of Atomically Thin MoS2. J. Phys. Chem. C 2013, 117, 25643–25649. 41. Kim, I. S.; Sangwan, V. K.; Jariwala, D.; Wood, J. D.; Park, S.; Chen, K. S.; Shi, F. Y.; Ruiz-Zepeda, F.; Ponce, A.; Jose-Yacaman, M.; et al. Influence of Stoichiometry on the Optical and Electrical Properties of Chemical Vapor Deposition Derived MoS2. ACS Nano 2014, 8, 10551–10558. 42. Nan, H. Y.; Wang, Z. L.; Wang, W. H.; Liang, Z.; Lu, Y.; Chen, Q.; He, D. W.; Tan, P. H.; Miao, F.; Wang, X. R.; et al. Strong Photoluminescence Enhancement of MoS2 through Defect Engineering and Oxygen Bonding. ACS Nano 2014, 8, 5738–5745. 43. Yue, Q.; Shao, Z. Z.; Chang, S. L.; Li, J. B. Adsorption of Gas Molecules on Monolayer MoS2 and Effect of Applied Electric Field. Nanoscale Res. Lett. 2013, 8, 425. 44. Xie, Z. T.; Baldea, I.; Smith, C. E.; Wu, Y. F.; Frisbie, C. D. Experimental and Theoretical Analysis of Nanotransport in Oligophenylene Dithiol Junctions as a Function of Molecular Length and Contact Work Function. ACS Nano 2015, 9, 8022–8036. 45. Hoffmann, M. W. G.; Mayrhofer, L.; Casals, O.; Caccamo, L.; Hernandez-Ramirez, F.; Lilienkamp, G.; Daum, W.; Moseler, M.; Waag, A.; Shen, H.; et al. A Highly Selective and SelfPowered Gas Sensor via Organic Surface Functionalization of p-Si/n-ZnO Diodes. Adv. Mater. 2014, 26, 8017–8022. 46. Mak, K. F.; He, K. L.; Lee, C.; Lee, G. H.; Hone, J.; Heinz, T. F.; Shan, J. Tightly Bound Trions in Monolayer MoS2. Nat. Mater. 2013, 12, 207–211. 47. Chen, M. K.; Wi, S. J.; Nam, H.; Priessnitz, G.; Liang, X. G. Effects of MoS2 Thickness and Air Humidity on Transport Characteristics of Plasma-Doped MoS2 Field-Effect Transistors. J. Vac Sci. Technol. B 2014, 32, 152105.

VOL. 9

’

NO. 12

’

12115–12123

’

ARTICLE

15. Coleman, J. N.; Lotya, M.; O'Neill, A.; Bergin, S. D.; King, P. J.; Khan, U.; Young, K.; Gaucher, A.; De, S.; Smith, R. J.; et al. Two-Dimensional Nanosheets Produced by Liquid Exfoliation of Layered Materials. Science 2011, 331, 568–571. 16. Feng, Q. L.; Zhu, Y. M.; Hong, J. H.; Zhang, M.; Duan, W. J.; Mao, N. N.; Wu, J. X.; Xu, H.; Dong, F. L.; Lin, F.; et al. Growth of Large-Area 2D MoS2(l‑x,)Se2x, Semiconductor. Adv. Mater. 2014, 26, 2648–2653. 17. Wu, S. F.; Huang, C. M.; Aivazian, G.; Ross, J. S.; Cobden, D. H.; Xu, X. D. Vapor-Solid Growth of High Optical Quality MoS2 Monolayers with Near-Unity Valley Polarization. ACS Nano 2013, 7, 2768–2772. 18. Kang, K.; Xie, S. E.; Huang, L. J.; Han, Y. M.; Huang, P. Y.; Mak, K. F.; Kim, C. J.; Muller, D.; Park, J. High-Mobility ThreeAtom-Thick Semiconducting Films with Wafer-Scale Homogeneity. Nature 2015, 520, 656–660. 19. Najmaei, S.; Liu, Z.; Zhou, W.; Zou, X. L.; Shi, G.; Lei, S. D.; Yakobson, B. I.; Idrobo, J. C.; Ajayan, P. M.; Lou, J. Vapour Phase Growth and Grain Boundary Structure of Molybdenum Disulphide Atomic Layers. Nat. Mater. 2013, 12, 754–759. 20. Schmidt, H.; Wang, S. F.; Chu, L. Q.; Toh, M.; Kumar, R.; Zhao, W. J.; Neto, A. H. C.; Martin, J.; Adam, S.; Ozyilmaz, B.; et al. Transport Properties of Monolayer MoS2 Grown by Chemical Vapor Deposition. Nano Lett. 2014, 14, 1909–1913. 21. Zhu, W. J.; Low, T.; Lee, Y. H.; Wang, H.; Farmer, D. B.; Kong, J.; Xia, F. N.; Avouris, P. Electronic Transport and Device Prospects of Monolayer Molybdenum Disulphide Grown by Chemical Vapour Deposition. Nat. Commun. 2014, 5, 3087. 22. Hong, J. H.; Hu, Z. X.; Probert, M.; Li, K.; Lv, D. H.; Yang, X. N.; Gu, L.; Mao, N. N.; Feng, Q. L.; Xie, L. M.; et al. Exploring Atomic Defects in Molybdenum Disulphide Monolayers. Nat. Commun. 2015, 6, 6293. 23. van der Zande, A. M.; Huang, P. Y.; Chenet, D. A.; Berkelbach, T. C.; You, Y. M.; Lee, G. H.; Heinz, T. F.; Reichman, D. R.; Muller, D. A.; Hone, J. C. Grains and Grain Boundaries in Highly Crystalline Monolayer Molybdenum Disulphide. Nat. Mater. 2013, 12, 554–561. 24. Lauritsen, J. V.; Nyberg, M.; Vang, R. T.; Bollinger, M. V.; Clausen, B. S.; Topsoe, H.; Jacobsen, K. W.; Laegsgaard, E.; Norskov, J. K.; Besenbacher, F. Chemistry of One-Dimensional Metallic Edge States in MoS2 Nanoclusters. Nanotechnology 2003, 14, 385–389. 25. Peterson, S. L.; Schulz, K. H. Ethanethiol Decomposition Pathways on MoS2(0001). Langmuir 1996, 12, 941–945. 26. Kim, J.-S.; Yoo, H.-W.; Choi, H. O.; Jung, H.-T. Tunable Volatile Organic Compounds Sensor by Using Thiolated Ligand Conjugation on MoS2. Nano Lett. 2014, 14, 5941– 5947. 27. Yu, Z. H.; Pan, Y. M.; Shen, Y. T.; Wang, Z. L.; Ong, Z. Y.; Xu, T.; Xin, R.; Pan, L. J.; Wang, B. G.; Sun, L. T.; et al. Towards Intrinsic Charge Transport in Monolayer Molybdenum Disulfide by Defect and Interface Engineering. Nat. Commun. 2014, 5, 5290. 28. Kibsgaard, J.; Chen, Z. B.; Reinecke, B. N.; Jaramillo, T. F. Engineering The Surface Structure of MoS2 to Preferentially Expose Active Edge Sites for Electrocatalysis. Nat. Mater. 2012, 11, 963–969. 29. Xie, J. F.; Zhang, H.; Li, S.; Wang, R. X.; Sun, X.; Zhou, M.; Zhou, J. F.; Lou, X. W.; Xie, Y. Defect-Rich MoS2 Ultrathin Nanosheets with Additional Active Edge Sites for Enhanced Electrocatalytic Hydrogen Evolution. Adv. Mater. 2013, 25, 5807–5813. 30. Yu, N. N.; Wang, L.; Li, M.; Sun, X. T.; Hou, T. J.; Li, Y. Y. Molybdenum Disulfide as A Highly Efficient Adsorbent for Non-Polar Gases. Phys. Chem. Chem. Phys. 2015, 17, 11700–11704. 31. Fang, H.; Tosun, M.; Seol, G.; Chang, T. C.; Takei, K.; Guo, J.; Javey, A. Degenerate n-Doping of Few-Layer Transition Metal Dichalcogenides by Potassium. Nano Lett. 2013, 13, 1991–1995. 32. Kiriya, D.; Tosun, M.; Zhao, P. D.; Kang, J. S.; Javey, A. Air-Stable Surface Charge Transfer Doping of MoS2 by Benzyl Viologen. J. Am. Chem. Soc. 2014, 136, 7853–7856.

12123

2015 www.acsnano.org