Chemical and Electronic Repair Mechanism of Sulfur Defects in MoS 2 Monolayers
Anja Förster, Sibylle Gemming, Gotthard Seifert, David Tománek
CChemical and Electronic Repair Mechanism ofDefects in MoS Monolayers
Anja F¨orster, † , ‡ , ¶ Sibylle Gemming, ‡ , § , (cid:107) Gotthard Seifert, ‡ , ¶ , ⊥ and DavidTom´anek ∗ , † † Physics and Astronomy Department, Michigan State University, East Lansing, Michigan48824, USA ‡ Center for Advancing Electronics Dresden (cfaed), 01062 Dresden, Germany ¶ Theoretical Chemistry, Technische Universit¨at Dresden, 01062 Dresden, Germany § Helmholtz-Zentrum Dresden-Rossendorf, Institute of Ion Beam Physics and MaterialsResearch, Bautzner Landstrasse 400, 01328 Dresden, Germany (cid:107)
Institute of Physics, Technische Universit¨at Chemnitz, 09107 Chemnitz, Germany ⊥ National University of Science and Technology, MISIS, Moscow, Russia
E-mail: [email protected]
Abstract
Using ab initio density functional theory cal-culations, we characterize changes in the elec-tronic structure of MoS monolayers introducedby missing or additional adsorbed sulfur atoms.We furthermore identify the chemical and elec-tronic function of substances that have beenreported to reduce the adverse effect of sul-fur vacancies in quenching photoluminescenceand reducing electronic conductance. We findthat thiol-group containing molecules adsorbedat vacancy sites may re-insert missing sulfuratoms. In presence of additional adsorbed sul-fur atoms, thiols may form disulfides on theMoS surface to mitigate the adverse effect ofdefects. Keywords transition metal dichalcogenides, 2D materials, ab initio calculations, electronic structure, de-fectsThere is growing interest in two-dimensional(2D) transition metal dichalcogenide (TMD) semiconductors, both for fundamental reasonsand as potential components in flexible, low-power electronic circuitry and for sensor appli-cations.
Molybdenum disulfide, MoS , is aprominent representative of this class of TMDs.A free-standing, perfect 2D MoS monolayerpossesses a direct band gap of 1.88 eV at the K -point in the Brillouin zone. Most commonlyused production methods for MoS monolay-ers are chemical vapor deposition (CVD) andmechanical exfoliation of the layered bulk ma-terial, as well as sputter growth atomiclayer deposition (ALD) of the precursor MoO and subsequent conversion to the disulfide un-der reducing conditions and at high tempera-tures. A direct ALD process using H S andMoCl or Mo(CO) is another possibility toobtain MoS monolayers. The CVD techniqueis probably best suited for mass production, butthe synthesized MoS layers lack in atomic per-fection. The most common defects in these lay-ers are sulfur and molybdenum vacancies, aswell as additional adsorbed sulfur atoms. Eliminating or at least reducing the adverse ef-fect of such defects is imperative to improvethe optoelectronic and transport properties ofTMDs.1 a r X i v : . [ phy s i c s . c h e m - ph ] O c t igure 1 V A (a) (b) A V Mo S
Figure 1: (Color online) (a) Perspective and (b)top view of the optimized geometry of an MoS monolayer containing a sulfur monovacancy (V)and a sulfur adatom (A).In search of ways to mitigate the adverse ef-fect of defects, different methods have beensuggested, including exposure of MoS to su-peracids or thiols. In the related MoSe system, Se vacancies could be filled by S atomsfrom an adjacent MoS layer. In the presentstudy, we focus on the reactions of thiols withdefective MoS monolayers.First, we characterize changes in the elec-tronic structure of MoS monolayers introducedby missing or additional adsorbed sulfur atomsusing ab initio density functional theory (DFT)calculations. We provide microscopic informa-tion about the chemical and electronic functionof thiols as a theoretical background for theunderstanding of the successful use of thiols,which have been reported to reduce the adverseeffect of sulfur vacancies in quenching photolu-minescence and to improve the electronic con-ductance of defective MoS . We found that ad-sorbed thiols may re-insert missing sulfur atomsat vacancy sites. We also found that in presenceof sulfur adatoms, thiols will form disulfides onthe MoS surface, which both mitigate the ad-verse effect of defects.In Figure 1 we display the structure of adefective MoS monolayer with a sulfur mono-vacancy (V) and an additional sulfur adatom(A), since these defects are known to sig-nificantly affect the electronic properties ofMoS . The formation energy of the sulfur va-cancy is 2.71 eV and that of the sulfur adatomis 1.07 eV. Consequently, the recombination en-ergy of a sulfur vacancy and a sulfur adatom is − .
89 eV. In spite of the large energy gain, nospontaneous healing will occur in a system with both defect types present due to the high acti-vation barrier of ≈ . , and also with a study of vacancy de-fects. Defects affect drastically the electronic struc-ture in the vicinity of the Fermi level. Settingapart the inadequacy of DFT-PBE calculationsfor quantitative predictions of band gaps, weshould note that in our computational approachwith (large) supercells and periodic boundaryconditions, also defects form a periodic array.In spite of their large separation, defect statesevolve into narrow bands that may affect theband structure of a pristine MoS monolayer.The effect of a sulfur monovacancy, as well asthat of a sulfur adatom, on the density of states(DOS) of an MoS monolayer around the bandgap region is shown in Figure 2.As seen in Figure 2c, sulfur monovacanciesintroduce defect states within the band gapand their superlattice shifts the DOS down by0.16 eV with respect to the pristine lattice. Thedefect states are localized around the vacancyas seen in Figure 2a. The effect of a superlat-tice of sulfur adatoms, addressed in Figure 2band 2d, is to reduce the DFT band gap from1 .
88 eV to 1 .
72 eV, in agreement with publishedresults.
Defect sites play an important role as catalyti-cally active centers and as sites for functional-ization reactions of 2D MoS . Sulfur vacanciesin particular are considered to be important nu-cleation sites for a functionalization with thiolmolecules R-SH. The likely possibility of an ad-sorbed thiol group transferring a sulfur atom tothe vacancy and thus repairing the defect is par-ticularly appealing. In this case, the detachedhydrogen atom may reconnect with the remain-ing R to form R-H and fill vacancy site of MoS with sulfur, asR-SH + MoS V → R-H + MoS , (1)where MoS V denotes the MoS layer with a sul-fur vacancy.An alternative reaction has been proposed to2 igure 2 (a) (b) Sulfur Monovacancy Sulfur Adatom D O S (c) D O S (d) Figure 2: (Color online) Ball-and-stick modelsof (a) a sulfur vacancy defect and (b) a sul-fur adatom defect in an MoS monolayer. Den-sity of states (DOS) of MoS with (c) a vacancyand (d) an adatom defect. The DOS and theposition of the Fermi level are shown by solidblue lines in defective lattices and by dottedblue lines in the corresponding pristine latticesin (c) and (d). The DOS has been convolutedby a Gaussian with a full-width at half maxi-mum of 0 . .
003 e/bohr .benefit from the STM tip current in an STMstudy. In the first step of reaction (2), simi-lar to reaction (1), a hydrogen atom is removedfrom the thiol as its sulfur atom fills the pre-vious vacancy, determining the reaction bar-rier for both reactions (1) and (2a). The re-moved hydrogen atom will then form H anddesorb from the MoS surface. The rest R isstill bound to the sulfur atom, adsorbed at thesulfur vacancy site. The final assumption of theproposed mechanism is that the R-groups areremoved with the support of the STM tip, as represented in reaction (2b).R-SH + MoS V →
12 H + R-S-MoS V (2a)R-S-MoS V ST M −−−→ R • + MoS (2b)There is evidence in the literature supportingboth reaction (1) (References [21], [28], [29])and reaction (2) (Reference [22]).The authors of Reference [30] propose yetanother reaction (3a). Instead of the thiolmolecules repairing the sulfur vacancy, theyform an adsorbed R-SS-R disulfide at thesurface of MoS while releasing a hydrogenmolecule. We also considered the possibil-ity that instead of desorbing, the hydrogenmolecule will fill the vacancy defect as describedin reaction (3b),2 R-SH + MoS → R-SS-R + H + MoS (3a)2 R-SH + MoS V → R-SS-R + H -MoS V (3b)Based on a previous study and the obser-vation of H S as well as H C=CH duringthe reaction of C H SH with bulk MoS inReference [28], we also considered a sulfur atomadsorbed on the MoS surface, identified asMoS A , as the driving force for the observeddisulfide formation.In this case, the reaction to form the disul-fide R-SS-R is divided into the following twosteps. In reaction (4a), one thiol reacts withthe adatom to R-S-S-H and, in the follow-upreaction (4b) with a second thiol, to R-S-S-R.An alternative reaction with a sulfur vacancyfollowing reaction (4a) is also possible. Similarto reaction (1), the SH-group of R-S-S-H cancure the vacancy defect, leading to the reduc-tion of R-S-S-H to the thiol R-S-H in Reaction(4c), R-SH + MoS A → R-SS-H + MoS (4a)R-SH + R-S-S-H → H S+R-SS-R (4b)R-SS-H + MoS V → R-SH + MoS (4c)To better understand the above reactionmechanisms (1) – (4), we performed DFT calcu-lations to compare the energy associated withthe pathways of these reactions. For the sake3 igure 3 Mo S H C
Reaction 2a Reaction 1
Product
Vacancy healing process + ½H educt transition state product-3.0-2.5-2.0-1.5-1.0-0.50.0 E a =0.22eV E R =-3.09eV E ne r g y [ e v ] Reaction coordinate Reaction 1 Reaction 2a E R =-0.90eV Figure 3: (Color online) Reaction scheme forthe sulfur vacancy healing process caused by ex-posure of MoS with vacancies to CH SH. E R denotes the reaction energy and E a the activa-tion barrier. The same initial state can lead totwo different final states via the same transi-tion state. The favorable reaction (1), shown indark blue, leads to a free CH molecule. Theenergetics of reaction (2a) is displayed in lightorange.of easy understanding, we consider the smallmethanethiol molecule CH SH as a representa-tive of thiols.We limit our study of vacancy repair processesto reactions with MoS monolayers that containone sulfur monovacancy per unit cell. We ana-lyze which reactions with thiols are favorable torepair vacancy and adatom defects. Our resultsalso unveil the likely cause of apparent contra- dictions in the interpretation of experimentalresults obtained by different researchers. Results/Discussion
Vacancy Repair
The majority of published results indicate thatthiol molecules interacting with sulfur-deficientMoS may fill in sulfur atoms at the vacancydefect sites. Reaction pathways for the twovacancy-healing reactions (1) and (2a), whichhave been proposed in the literature, aresketched in Figure 3. We note that reaction (1)has been studied in greater detail for a differ-ent thiol and agrees with our findings for themodel compound H C-SH.We find that reactions (1) and (2a) are bothexothermic and require crossing only a low ac-tivation barrier of 0 .
22 eV, since they share thesame transition state shown in Figure 3. Thelarger energy gain E R = − .
09 eV in reaction(1) in comparison to − .
90 eV in reaction (2a)suggests that the former reaction is thermody-namically preferred.Figure 4a shows the DOS and partial densi-ties of states (PDOS), projected on individualatoms, of the product of reaction (1). Figure 4bprovides the corresponding information for re-action (2a), and Figure 4c provides a detailedview of the PDOS for the CH -group and theconnected sulfur atom. In both cases, the de-fect states associated with sulfur monovacancieshave been removed. In the final state of reac-tion (1) the DOS is completely restored to theundamaged state of the semiconductor. For re-action (2a), on the other hand, the Fermi levelis shifted to the lower edge of the conductionband due to the CH -group. Therefore, onlythe preferred repair reaction (1) leads to bothan electronic and a chemical repair of MoS . Disulfide Formation
A different reaction scenario (3a) has been pro-posed in Reference [30], suggesting that disul-fides are formed when thiols interact withMoS . We investigated the MoS surface bothin its pristine state and in presence of sulfur4 igure 4 (a) Product of Reaction 1 (b) Product of
Reaction 2a totalSMoCS(C)H(c) CH adsorbed on MoS in the product of Reaction 2a
Figure 4: (Color online) Electronic structure ofproducts of the vacancy healing process shownin Figure 3. The total density of states (DOS)and partial densities of states (PDOS) of theproducts of reactions (a) (1) and (b) (2a). (c)PDOS of the CH -molecule bonded to a sulfuratom in an MoS monolayer. The PDOS of thesulfur atom connected to C in CH , S(C), isshown by the brown line. All DOS and PDOSfunctions have been convoluted by a Gaussianwith a full-width at half maximum of 0 . and reaction (3b) on a sulfur-deficient MoS substrate. The schematic reaction profile in-dicates that both reactions involve significantactivation barriers. As seen in Figure 5, reac-tion (3a) is endothermic with a reaction energyof E R =0.39 eV and involves a high activationbarrier of 2.91 eV. Reaction (3b) near a sulfurmonovacancy is only slightly endothermic, witha reaction energy E R =0.02 eV, and involves asomewhat lower activation barrier of 2.13 eV.Thus, reaction (3b) is energetically more favor- Figure 5
Product 3a
Disulfide formation process
Educt 3a (pristine MoS ) Product 3b Educt 3b (S monovacancy) + ½H Mo S H C educt transition state product-0.50.00.51.01.52.02.53.0 E a =2.13eV E R =0.39eV E ne r g y [ e V ] Reaction coordinate Reaction 3a Reaction 3b E a =2.91eV E R =0.02eV Figure 5: (Color online) Reaction scheme of thedisulfide formation process involving exposureof an MoS monolayer to two CH SH molecules.E R denotes the reaction energy and E a the ac-tivation barrier. Reaction (3a), shown in darkpurple, occurs on pristine MoS . Reaction (3b),shown in light green, occurs on a sulfur-deficientMoS substrate.able than reaction (3a).The near-neutral reaction energy of reaction(3b) can be explained by the Kubas interac-tion of transition metal η -H -complexes. Itmeans that the excess H molecule in the prod-uct of reaction (3b), which is attached to anMo atom at the sulfur vacancy site and indi-cated by a circle in Figure 5, still retains the H-H bond character. According to Reference [34],the Kubas interaction energy is in the range of0 . − . igure 6 (a) Product of Reaction 3a (b) Product of
Reaction 3b totalSMoCH(c) Atoms of the Kubas complex of
Reaction 3b
S inC H S Figure 6: (Color online) Electronic structureof products of the disulfide formation processshown in Figure 5. The total density of states(DOS) and partial densities of states (PDOS)of the products of reactions (a) (3a) and (b)(3b). (c) PDOS of the 3 Mo atoms and theH molecule attached to the vacancy that formthe Kubas complex in the product of reaction(3b). All DOS and PDOS functions have beenconvoluted by a Gaussian with a full-width athalf maximum of 0 . R = − .
09 eV in thermodynamic equilibrium.The PDOS functions characterizing the prod-ucts of reactions (3a) and (3b), visualized inFigure 5, are shown in Figure 6. The productof reaction (3b) still contains a defect state inthe gap region, indicating that the chemisorbedH molecule is incapable of electronically re-pairing the effect of the sulfur vacancy. Thisis seen in the PDOS of the Mo atoms of theKubas complex surrounding the vacancy defectin Figure 6c. The product of reaction (3a), onthe other hand, shows no indication of a defect Product 4b Product 4c Product 4a + Sulfur vacancy
Adatom healing process Figure 7 + ½H S Mo S H C + educt transitionstate interm.state transitionstate prodcut-4-3-2-1012 E R =-2.81eVE R =-0.18eVE a ~3.00eVE R =-0.86eV E ne r g y [ e V ] Reaction coordinate Reaction 4a Reaction 4b Reaction 4c E a =1.00eV Figure 7: (Color online) Reaction scheme ofthe adatom healing process that starts with re-action (4a) and leads to disulfide formation inpresence of extra sulfur atoms on MoS , shownin dark red. Subsequent ligand exchange reac-tion (4b) is shown in brown. Alternative subse-quent vacancy repair reaction (4c) is shown inlight blue. E R denotes the reaction energy andE a the activation barrier.state, since the vacancy-free MoS monolayer isnot affected much by the physisorbed disulfide,as seen in the PDOS of Figure 6a.We can thus conclude that the disul-fide formation reaction (3a), suggested inReference [30], is endothermic. The alterna-tive reaction (3b) on a sulfur-deficient MoS substrate displays a lower activation barrierand an end-product stabilized by the Kubasinteraction, but is still weakly endothermic andthus unlikely. In the following, we propose analternative pathway towards disulfide forma-tion.6 igure 8 (a) Sulfur adatom on MoS (b) Product of Reaction 4a (c) Product of
Reaction 4c totalSMoCS ad S in C H S S in CH SHH
Figure 8: (Color online) Electronic structure ofproducts of the adatom healing process shownin Figure 7. The total density of states (DOS)and partial densities of states (PDOS) of (a)MoS with a sulfur adatom (without CH SH),(b) the product of reaction (4a), and (c) theproduct of reaction (4c). All DOS and PDOSfunctions have been convoluted by a Gaussianwith a full-width at half maximum of 0 . Adatom Repair
The postulated alternative reaction requires ex-tra sulfur atoms adsorbed on the MoS surface,which act as nucleation sites for the disulfideformation. The reaction leading to the forma-tion of disulfide R-SS-R in presence of sulfuradatoms consists of two steps, described by re-actions (4a) and (4b), as well as the alternativereaction (4c) following reaction (4a), as shownin Figure 7.In reaction (4a), a CH SH molecule in-teracts with the reactive sulfur adatom tomethylhydro-disulfide (CH SSH), releasing − .
86 eV due to the formation of a stable disulfide bond. The estimated activation bar-rier for this reaction is close to 1 eV, which isconsiderably lower than the values for the cor-responding reactions (3a) and (3b) in absenceof an extra sulfur adatom.Electronic structure changes during theadatom healing process are displayed in Fig-ure 8. The DOS of the product of reaction (4a),shown in Figure 8b, shows no defect-relatedstates in the band gap, indicating chemical andelectronic repair of the sulfur adatom defectthat is seen in Figure 8a.In the subsequent reaction (4b), shown inFigure 7, a second CH SH molecule interactswith the methylhydro-disulfide CH SSH, lead-ing to the exchange of the hydrogen atom with amethyl group and formation of hydrogen sulfide(H S) as a side product. This reaction is mildlyexothermic, with an overall reaction energy of − .
18 eV. Even though the combined reaction(4a) and (4b) for the formation of CH SSCH is strongly exothermic with a net energy gain of − .
05 eV, the activation barrier for the ligandexchange in reaction (4b) is prohibitively highwith E a ≈ + 3 eV, which essentially suppressesthe formation of CH SSCH following reaction(4a).Therefore, we investigated reaction (4c) as analternative follow-up process to reaction (4a).In reaction (4c), the CH SSH molecule inter-acts with a nearby sulfur vacancy defect. Thisreaction is similar to the vacancy healing reac-tion (1) and consequently is strongly exother-mic with a reaction energy of − .
81 eV. Reac-tion (4c) is barrier-free and thus occurs spon-taneously. As seen in in Figure 8c, describingthe product of reaction (4c), the defect-relatedstate above E F has been removed from theDOS. This means that following the adatom re-pair and disulfide formation reaction (4a), reac-tion (4c) will take place in case that also sulfurvacancies are present. The two reactions willthus heal both vacancy and adatom defects.Our above considerations offer an attrac-tive explanation why disulfide formationwas observed in Reference [30], but not inReferences [21], [22], [28] and [29]. Initially,reactions (1) and (4a) plus (4c) have takenplace in all samples that contained vacancies.7acancy healing as primary outcome of re-actions reported in References [21], [22], [28]and [29] could likely be achieved due to anabundance of vacancies in the samples used.We may speculate that the MoS sample ofReference [30] contained more sulfur adatomsthan sulfur vacancies. In that case, all vacancydefects could be repaired, but some adatomdefects were left unrepaired in the sample ofReference [30]. At this point, lack of vacancydefects would block reactions (1) and (4c). Theonly viable reaction was (4a), which repairedadatom defects, leaving a pristine MoS surfacebehind with disulfide as a by-product . Thisspeculative assumption is also consistent withthe observation that the electronic structure ofMoS has remained unaffected by the reactionleading to the formation of disulfide. Conclusions
We studied three different reaction paths ofthiols, represented by methanethiol (CH SH),with a defective 2D MoS monolayer. Weshowed that the repair of sulfur monovacan-cies by adsorbed CH SH is an exothermic re-action releasing up to 3 eV. In another possi-ble reaction between CH SH and MoS , lead-ing to the formation of disulfide, we found thatpresence of sulfur vacancies lowers the reactionbarrier due to the Kubas interaction at the de-fect site. The corresponding reaction involv-ing MoS with sulfur adatoms instead of vacan-cies, on the other hand, leads to disulfide forma-tion and releases about 0 . Methods/Theoretical
To obtain insight into the reaction processes,we performed DFT calculations using the SIESTA code. We used ab initio
Troullier-Martins pseudopotentials and the Perdew-Burke-Ernzerhof (PBE) exchange-correlationfunctional throughout the study. Except forsulfur, all pseudopotentials used were obtainedfrom the on-line resource in Reference [39]. Thepseudopotential of sulfur has been generatedwithout core corrections using the ATM codein the SIESTA suite and the parameters listedin Reference [39]. All pseudopotentials weretested against atomic all-electron calculations.We used a double- ζ basis set including polariza-tion orbitals (DZP) to represent atoms in crys-tal lattices, 140 Ry as the mesh cutoff energyfor the Fourier transform of the charge density,and 0 K for the electronic temperature. Weused periodic boundary conditions with largesupercells spanned by the lattice vectors (cid:126)a =(12 . , . , .
00) ˚A, (cid:126)a = (6 . , . , .
00) ˚A, (cid:126)a = (0 . , . , .
23) ˚A to represent pristineand defective 2D MoS lattices. The unit cellsof defect-free MoS contained 16 molybdenumand 32 sulfur atoms, and were separated by avacuum region of ≈
15 ˚A normal to the layers.The Brillouin zone was sampled by a 4 × × k -point grid and its equivalent in larger su-percells.The above input parameters were found toguarantee convergence. In particular, we foundthat using the larger triple- ζ polarized (TZP)instead of the DZP basis and increasing themesh cutoff energy affected our total energydifferences by typically less than 0 .
01 eV. Wefurthermore validated the ab initio pseudopo-tential approach used in the SIESTA code bycomparing to results of the all-electron SCM-Band code and found that energy differencesobtained using the two approaches differed typ-ically by less than 0 . until noneof the residual Hellmann-Feynman forces ex-ceeded 10 − eV/˚A. In addition to the defaultdensity matrix convergence, we also demandedthat the total energy should reach the toler-ance of (cid:46) − eV. To eliminate possible arti-facts associated with local minima, we verifiedinitial and final state geometries by perform-ing canonical molecular dynamics (MD) sim-8lations using the NVT-Nos´e thermostat with T = 273 .
15 K and 1 fs time steps.Due to the complexity of the reaction energyhypersurface and the large number of rele-vant degrees of freedom, approaches such asthe nudged elastic band, which are commonlyused to determine the reaction path includingtransition states, turned out to be extremelydemanding on computer resources. We fo-cussed on transition states only and initiatedour search by running canonical MD simula-tions starting from a set of educated guessesfor the geometry. Following the atomic tra-jectories, we could identify a saddle point inthe energy hypersurface, where all forces actingon atoms vanished, and postulated this pointin configurational space as a transition state.To confirm this postulate, we ran MD simula-tions starting at a slightly altered geometry ofthe postulated transition state. We concludedthat the postulated transition state is indeedthe real transition state once all trajectoriesreached either the initial (educt) or the final(product) state. The activation barrier was de-termined by the energy difference between theinitial and the transition state.
Author InformationCorresponding Author ∗ E-mail: [email protected]
Notes
The authors declare no competing financial in-terest.
Acknowledgement
We thank Jie Guanand Dan Liu for useful discussions and Gar-rett B. King for carefully checking the bib-liography. This study was supported by theNSF/AFOSR EFRI 2-DARE grant number via the W3 programme.
References
1. Li, H.; Yin, Z.; He, Q.; Li, H.; Huang, X.;Lu, G.; Fam, D. W. H.; Tok, A. I. Y.;Zhang, Q.; Zhang, H. Fabrication of Single-and Multilayer Mos Film-Based Field-Effect Transistors for Sensing NO at RoomTemperature.
Small , , 63–67.2. Castellanos-Gomez, A.; van Leeuwen, R.;Buscema, M.; van der Zant, H. S.;Steele, G. A.; Venstra, W. J. Single-LayerMoS Mechanical Resonators.
Adv. Mater. , , 6719–6723.3. Chang, H.-Y.; Yang, S.; Lee, J.; Tao, L.;Hwang, W.-S.; Jena, D.; Lu, N.; Akin-wande, D. High-Performance, Highly Bend-able MoS Transistors with High-K Di-electrics for Flexible Low-Power Systems.
ACS Nano , , 5446–5452.4. Ellis, J. K.; Lucero, M. J.; Scuseria, G. E.The Indirect to Direct Band Gap Transi-tion in Multilayered MoS as Predicted byScreened Hybrid Density Functional The-ory. Appl. Phys. Lett. , , 261908.5. Splendiani, A.; Sun, L.; Zhang, Y.; Li, T.;Kim, J.; Chim, C.-Y.; Galli, G.; Wang, F.Emerging Photoluminescence in MonolayerMoS . Nano Lett. , , 1271–1275.6. Pachauri, V.; Kern, K.; Balasubrama-nian, K. Chemically Exfoliated Large-AreaTwo-Dimensional Flakes of MolybdenumDisulfide for Device Applications. APLMater. , , 032102.7. Li, H.; Wu, J.; Yin, Z.; Zhang, H. Prepara-tion and Applications of Mechanically Ex-foliated Single-Layer and Multilayer MoS and WSe Nanosheets.
Acc. Chem. Res. , , 1067–1075.8. Varrla, E.; Backes, C.; Paton, K. R.; Har-vey, A.; Gholamvand, Z.; McCauley, J.;9oleman, J. N. Large-Scale Production ofSize-Controlled MoS Nanosheets by ShearExfoliation.
Chem. Mater. , , 1129–1139.9. Samassekou, H.; Alkabsh, A.; Wasala, M.;Eaton, M.; Walber, A.; Walker, A.;Pitk¨anen, O.; Kordas, K.; Talapatra, S.;Jayasekera, T.; Mazumdar, D. ViableRoute towards Large-Area 2D MoS UsingMagnetron Sputtering.
2D Materials , , 021002.10. Kastl, C.; Chen, C. T.; Kuykendall, T.;Shevitski, B.; Darlington, T. P.; Bo-rys, N. J.; Krayev, A.; Schuck, P. J.;Aloni, S.; Schwartzberg, A. M. The Impor-tant Role of Water in Growth of MonolayerTransition Metal Dichalcogenides.
2D Ma-terials , , 021024.11. Keller, B. D.; Bertuch, A.; Provine, J.; Sun-daram, G.; Ferralis, N.; Grossman, J. C.Process Control of Atomic Layer Deposi-tion Molybdenum Oxide Nucleation andSulfidation to Large-Area MoS Monolay-ers.
Chem. Mater. , , 2024–2032.12. Browning, R.; Padigi, P.; Solanki, R.;Tweet, D. J.; Schuele, P.; Evans, D. AtomicLayer Deposition of MoS Thin Films.
Mater. Res. Express , , 035006.13. Kwon, D. H.; Jin, Z.; Shin, S.; Lee, W.-S.; Min, Y.-S. A Comprehensive Study onAtomic Layer Deposition of MolybdenumSulfide for Electrochemical Hydrogen Evo-lution. Nanoscale , , 7180–7188.14. Vancs´o, P.; Magda, G. Z.; Pet˝o, J.; Noh, J.-Y.; Kim, Y.-S.; Hwang, C.; Bir´o, L. P.;Tapaszt´o, L. The Intrinsic Defect Struc-ture of Exfoliated MoS Single Layers Re-vealed by Scanning Tunneling Microscopy.
Sci. Rep. , .15. Addou, R.; McDonnell, S.; Barrera, D.;Guo, Z.; Azcatl, A.; Wang, J.; Zhu, H.;Hinkle, C. L.; Quevedo-Lopez, M.; Alsha-reef, H. N. Impurities and Electronic Prop-erty Variations of Natural MoS CrystalSurfaces.
ACS Nano , , 9124–9133. 16. Liu, K.-K.; Zhang, W.; Lee, Y.-H.; Lin, Y.-C.; Chang, M.-T.; Su, C.-Y.; Chang, C.-S.; Li, H.; Shi, Y.; Zhang, H. Growth ofLarge-Area and Highly Crystalline MoS Thin Layers on Insulating Substrates.
NanoLett. , , 1538–1544.17. Lee, Y.-H.; Zhang, X.-Q.; Zhang, W.;Chang, M.-T.; Lin, C.-T.; Chang, K.-D.;Yu, Y.-C.; Wang, J. T.-W.; Chang, C.-S.; Li, L.-J. Synthesis of Large-Area MoS Atomic Layers with Chemical Vapor Depo-sition.
Adv. Mater. , , 2320–2325.18. Zhan, Y.; Liu, Z.; Najmaei, S.;Ajayan, P. M.; Lou, J. Large-AreaVapor-Phase Growth and Characteriza-tion of MoS Atomic Layers on a SiO Substrate.
Small , , 966–971.19. 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 inMolybdenum Disulphide Monolayers. Nat.Commun. , , 6293.20. Amani, M.; Lien, D.-H.; Kiriya, D.;Xiao, J.; Azcatl, A.; Noh, J.; Mad-hvapathy, S. R.; Addou, R.; KC, S.;Dubey, M.; Cho, K.; Wallace, R.M.;Lee, S.-C.; He, J.-H.; Ager, J.W.;Zhang, X.; Yablonovitch, E.; Javey, A.Near-Unity Photoluminescence QuantumYield in MoS . Science , , 1065–1068.21. 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. To-wards intrinsic charge transport in mono-layer molybdenum disulfide by defect andinterface engineering. Nat. Commun. , , 5290.22. Makarova, M.; Okawa, Y.; Aono, M. Selec-tive Adsorption of Thiol Molecules at Sul-fur Vacancies on MoS (0001), Followed by10acancy Repair via S–C Dissociation.
J.Phys. Chem. C , , 22411–22416.23. Surrente, A.; Dumcenco, D.; Yang, Z.;Kuc, A.; Jing, Y.; Heine, T.; Kung, Y.-C.;Maude, D. K.; Kis, A.; Plochocka, P. De-fect Healing and Charge Transfer-MediatedValley Polarization in MoS /MoSe /MoS Trilayer van der Waals Heterostructures.
Nano Lett. , , 4130–4136.24. Noh, J.-Y.; Kim, H.; Kim, Y.-S. Stabilityand Electronic Structures of Native Defectsin Single-Layer MoS . Phys. Rev. B , , 205417.25. Santosh, K.; Longo, R. C.; Addou, R.; Wal-lace, R. M.; Cho, K. Impact of IntrinsicAtomic Defects on the Electronic Struc-ture of MoS Monolayers.
Nanotechnology , , 375703.26. Ghorbani-Asl, M.; Enyashin, A. N.;Kuc, A.; Seifert, G.; Heine, T. Defect-Induced Conductivity Anisotropy in MoS Monolayers.
Phys. Rev. B , ,245440.27. Sim, D. M.; Kim, M.; Yim, S.; Choi, M.-J.; Choi, J.; Yoo, S.; Jung, Y. S. ControlledDoping of Vacancy-Containing Few-LayerMoS via Highly Stable Thiol-Based Molec-ular Chemisorption.
ACS Nano , ,12115–12123.28. Peterson, S. L.; Schulz, K. H. EthanethiolDecomposition Pathways on MoS (0001). Langmuir , , 941–945.29. Wiegenstein, C. G.; Schulz, K. H.Methanethiol Adsorption on DefectiveMoS (0001) Surfaces. Phys. Chem. B , , 6913–6918.30. Chen, X.; Berner, N. C.; Backes, C.; Dues-berg, G. S.; McDonald, A. R. Functional-ization of Two-Dimensional MoS : On theReaction Between MoS and Organic Thi-ols. Angew. Chem. Int. Ed. , , 5803–5808. 31. Benziger, J. B.; Preston, R. E. Organosul-fur Chemistry on Tungsten (211) Sur-faces. 1. A Comparison of Methanethiol andMethanol. Phys. Chem. , , 5002–5010.32. Kubas, G. J. Five-Co-Ordinate Molyb-denum and Tungsten Complexes,[M(CO) (PCy ) ], which ReversiblyAdd Dinitrogen, Dihydrogen, and OtherSmall Molecules. J. Chem. Soc., Chem.Commun. , 61–62.33. Gordon, J. C.; Kubas, G. J. Perspec-tives on How Nature Employs the Prin-ciples of Organometallic Chemistry inDihydrogen Activation in Hydrogenases.
Organometallics , , 4682.34. Skipper, C. V. J.; Hamaed, A.; An-tonelli, D. M.; Kaltsoyannis, N. The KubasInteraction in M(II) (M = Ti, V, Cr)Hydrazine-Based Hydrogen Storage Mate-rials: A DFT Study. Dalton Trans. , , 8515–8523.35. Soler, J. M.; Artacho, E.; Gale, J. D.;Garc´ıa, A.; Junquera, J.; Ordej´on, P.;S´anchez-Portal, D. The SIESTA Methodfor Ab Initio
Order-N Materials Simulation.
J. Phys.: Condens. Matter , , 2745.36. Perdew, J. P.; Burke, K.; Ernzerhof, M.Generalized Gradient Approximation MadeSimple. Phys. Rev. Lett. , , 3865.37. Monkhorst, H. J.; Pack, J. D. Spe-cial Points for Brillouin-Zone Integrations. Phys. Rev. B , , 5188.38. Troullier, N.; Martins, J. L. Efficient Pseu-dopotentials for Plane-Wave Calculations. Phys. Rev. B , , 1993–2006.39. http://departments.icmab.es/leem/siesta/Databases/Pseudopotentials/periodictable-gga-abinit.html , ac-cessed on August 1, 2016.40. Hestenes, M. R.; Stiefel, E. Methods ofConjugate Gradients for Solving LinearSystems. J. Res. Natl. Bur. Stand. , , 409–436.111. BAND2017, SCM, Theoretical Chem-istry, Vrije Universiteit, Amsterdam, TheNetherlands, http://scm.comhttp://scm.com